Note: If you came here from The Red Desert looking for a Unitarian Universalist atheist's essay on religion, you probably want "Is Rational Religion Possible?" This one is on libertarianism.
The Dog Ate My Manifesto:
Reflections on Moderate Libertarianism
Peter A. Taylor
(Last modified June, 2012)
Dedicated to Glen R. Taylor
"What is remembered lives."
This is the long version. Please read the short version (about 21 k) before attempting to wade through this version.
Table of Contents
Part 1: Laissez Faire
2. The Prisoners' Dilemma
3. Objection 1: Judging Criteria
4. Objection 2: Distribution of Benefits
5. Objection 3: Knowledge Problems
6. Objection 4: Insecure Boundaries
7. Objection 5: Imperfect Competition
8. Objection 6: Free Rider Problems
9. Welfare Economics
10. Interpersonal Utility Comparisons
Part 2: Democracy
11. Public Choice Theory
12. Parallelism Between Laissez Faire and Democracy
13. Rational Ignorance
14. Rent Seeking
15. Sensitivity Comparison
16. Problems with Public Choice Theory
17. A Closer Look at Intangibles
18. The Malleability of Values
Part 3: Libertarianism as a Coping Strategy
19. Coping Strategies
20. Flavors of Libertarianism
21. Possible Reasons for Rigidity
Part 4: My Position
22. The Four Directions
23. Indefensible Positions
24. Decision Tree
26. Appendix A: The Ethics of Logrolling
27. Appendix B: Education Vouchers
28. Appendix C: Foreign Policy
29. Appendix D: Rational Irrationality
30. Appendix E: Health Care Reform
31. Appendix F: Minimum-Wage Laws
32. Appendix G: Border Control and Immigration
Here is the really short version. This is the
framework I'm advocating for making political decisions.
Part 1: Laissez Faire
I initially had three objectives in writing this essay. My primary purpose originally was to score some points on my "liberal" (social democrat) friends, to convince them that I'm not crazy and that moderate, "small-l" libertarianism (classical liberalism) makes sense, even if they still don't agree with it. My second purpose was to clarify and thus strengthen the moderate libertarian position, distinguishing it from and defending it against the radical (anarchist) libertarians, whom I regard as false friends. My impression was that the moderate position is generally perceived as something of a slippery slope, undermining the moderates' efforts both within the libertarian movement and outside it, and I hoped to give the moderates some traction. My third purpose was to organize and clarify my own thinking. This last objective has proved much more time consuming than I imagined it would be, and has taken on more importance to me than the other two. I hope my efforts in this regard will prove useful to others like me, who are attracted to libertarian sentiments, but who are aware that there are limits to how far these ideas can be pushed, and who are unpleasantly unclear about just where those limits are. The writing process has certainly helped me to clear up some cognitive dissonance.
In order to make my arguments cogent for a hostile audience, I am focusing on practical arguments. When I do talk about moral theories, I view them mainly as secondary principles that follow from practical first principles. There are several problems with this. One problem is that confirmation bias may be at work. My readers may understandably suspect me of cherry-picking the practical arguments, and want to know more about my biases, as I understand them. (See chapter 21.) Another problem is that my practical arguments may be indeterminate, and my biases come in when I start talking about decision-making under uncertainty. If I have serious doubts about the consequences of government coercion, I would rather risk committing a sin of omission than a sin of commission. This is typical of libertarians. I think a good utilitarian case can be made for preferring sins of omission, but this is also partly a matter of temperament.
My original working title for this essay was "A Fuzzy Logic Libertarian Manifesto." "Fuzzy" in this context means multivalent logic, the mathematics of ambiguity and overlapping set membership. As you might gather from this, I see political decision-making problems as being fraught with ambiguity, and I am not as interested in answering questions as I am in making sure I am asking them correctly. I am not as interested in justifying specific moderate doctrines as I am in setting up the problem right. As I have read and argued about politics, my views have evolved into a decision tree: someone says "There ought to be a law", and I have a mental flow chart I use to see if this is reasonable or not. I will present my decision tree towards the end of this essay, after first presenting the rationale behind it. I expect few people to find my position intuitively obvious.
My decision tree starts off by admitting that the free market does not automatically solve everything. It goes on by trying to balance market failures against government policy failures. It then uses a set of "firewalls" or "tripwires" to limit the risks associated with these policy failures. It ends by giving one of three answers to the challenge, "There ought to be a law:" (1) I may agree; (2) I may think it is unreasonable; or (3) I may disagree with the understanding that the person making the claim is reasonable but has slightly different values or beliefs than I do. What gives this logic its "libertarian" flavor is the set of "tripwires," the checks against the casual use of force for "good" causes.
Defining "libertarian" and using it in a consistent way has been surprisingly difficult for me. I expect that many people will say I'm not a "real" libertarian. I am sorely tempted to abandon the word, and call myself a "liberal," but in the US, unless I qualify "liberal" carefully with words like "classical" or "Jeffersonian," this would get me confused with what in Europe would be called a social democrat. Instead, I qualify "libertarian" with terms like "moderate" and "small-l." I use the word "libertarian" in a weak and admittedly vague sense. The central question is, "When am I justified in using force?" To me, the defining characteristic of a libertarian is a certain reluctance to use force offensively, to "initiate coercion," even in a good cause. Coercion includes force, fraud, theft, and related threats. A strict, radical libertarian may steadfastly refuse to initiate coercion under any circumstances, but I will use the term to describe anyone who distinguishes between offensive and defensive good causes, and who is clearly more reluctant to use force offensively. In contrast, I regard both conservatives and social democrats as being far more concerned with having a "good" cause than a defensive one. I will also require a libertarian to have fairly conventional ideas about what is offensive and defensive, and to accept responsibility for decisions that he hires other people to carry out: a libertarian should be reluctant to vote for a politician who proposes to use force offensively. I sometimes call myself a "postlibertarian," following Jeffrey Friedman, to emphasize that I have considered and rejected the strict anarchist libertarian position.
For sake of clarity, I will describe several different flavors of "libertarians:"
1. Abolitionist anarchists, people who think the free market solves all problems immediately, and government should be done away with right now.
2. Gradualist anarchists, people who think that for the moment, there are some problems for which we need government, but who hope that some day these problems will be solved by other arrangements.
3. Minarchists (ultraminimalists), people who think that the optimum level of government is as little as possible, subject to the constraint that some government is necessary to preserve order and defend against other governments. In industrial engineering jargon, this minimum would be called a "corner solution" to the problem of finding the optimum level of government.
4. "Apologetic" limited government libertarians, people who think an attractive society needs more government than the minimum amount needed for stability but who still regard government as somewhat immoral, and are embarrassed to take positions advocating "immorality." In industrial engineering terms, this optimal amount of government would be called an "interior point."
5. "Unapologetic" limited government libertarians (classical liberals). These are libertarians who are not embarrassed to advocate an amount of government that is determined by practical tradeoffs. Again, this is an interior point.
I regard the unapologetic limited government position as being right, and the other four as being wrong. The distinction between the apologetic and unapologetic limited government libertarians is important. Part of the point in my circulating this essay is to get the limited government advocates to stop apologizing for making practical tradeoffs. But I think it is also important to point out to some readers that "libertarianism" is not monolithic; there are many conflicting versions, and to the consternation of many libertarians who seem to deal with ambiguity very poorly, it is a fuzzily defined set of views with many unresolved doctrinal disputes.
Update: As Will Wilkinson says in Libertarian Democraphobia, "I think this divide [between anarchist and limited-statist libertarians] is far wider than is reflected in the libertarian community, and part of the reason is that limited-government libertarians tend to internalize more of the anarchist framework than they logically should." I also like his term, "neoclassical liberal."
2. The Prisoner's Dilemma
Since many of the practical tradeoffs that voters have to make involve coping with lack of good information, much of my emphasis throughout this essay will be on information. I want to catch people who are being dishonest or naive about their motives. I also want to know why and when people make bad decisions.
One obvious reason why someone may make a bad decision is stupidity. However, this explanation is grossly overused, as is illustrated by a game theory paradox called the Prisoners' Dilemma.
In this game, there are two prisoners, A and B, and a prosecutor who has enough evidence to send them each to prison for two years for a crime they committed together. If the prosecutor could get more evidence, he could send them each to prison for four years, but the only way he can get more evidence is for one prisoner to betray the other in a confession, in exchange for which the confessor will get a one year lighter sentence. The prisoners are kept separate from one another; they have to decide independently whether or not to confess, and they cannot change their minds after finding out what the other prisoner did. Even if they could bargain with each other, they have no way to enforce their bargains or convincingly commit themselves to carry out any threats. They only play this game once. I also assume that each prisoner cares considerably more about the amount of time he spends in prison than he does about the amount of time his accomplice spends there.
If neither prisoner confesses, they each spend two years in prison, for a total of four man-years. If one confesses and the other does not, the confessor gets one year and the other gets four, for a total of five man-years. If they both confess, they each get three years (four minus one), for a total of six man-years. Collectively, they are best off if neither of them confesses, and worst off if they both do. But individually, no matter what A does, B is better off by confessing, and vice-versa, as long as the two decisions are made independently of one another. So if the two prisoners are "rational," in this situation they will both behave in a way that is antisocial in the context of their two-person microcosm of society. They bring suffering on one another not because they don't understand the way their world works, but because they understand it entirely too well.
Richard Dawkins discusses this game and a variant of it called the Iterated Prisoners' Dilemma at length in The Selfish Gene. The difference between these is that in the iterated version, the same two players play the same game repeatedly. This allows them to retaliate in one game for what the other player did in the previous game. In this situation, the strategies that work the best are variants of "tit for tat," doing whatever the other player did in the previous game most of the time. Players of this game are better off most of the time if they don't confess except in retaliation, and if they usually, but not always, retaliate against the other player's confessions.
Unfortunately, in multiplayer versions of this game, it is impossible to retaliate against specific players, and we are back where we started. This is essentially the situation Garrett Hardin described in "The Tragedy of the Commons" (Science, 1968). (The article may be found in its entirety at http://www.dieoff.com/page95.htm .) Suppose you are one of 100 ranchers, each with 10 cows that graze on a common pasture that is an appropriate size for a total of 1000 cows. If you bring in an additional cow, the commons will be slightly overgrazed; not enough to render it unusable, but enough that the cows won't get quite as fat as they used to. Suppose that collectively having 1000 cows on the commons will result in them each bringing $1000 profit to the ranchers, but with 1001 cows, they would each be worth $990. Collectively, the ranchers are better off by $9010 if they limit the herd to 1000. But the rancher who introduces the 1001st cow now has 11, so he personally is better off by $890. The other 99 ranchers are each worse off by $100. Unfortunately, it isn't just one rancher who has perverse incentives--each of the 100 ranchers has similar incentives.
If the number of ranchers is small, it is likely that they can eventually negotiate a voluntary agreement to limit their herds. Mancur Olson's The Logic of Collective Action (Cambridge, Mass.: Harvard University Press, 1965) is a classic that explains how complicated this can be. These negotiations resemble another game theory paradigm, the game of "chicken" (see Dennis Mueller, Public Choice II, ch. 2, Cambridge University Press, 1989, ISBN 0521379520) wherein one player's best move does depend on the other players' moves, unlike the Prisoners' Dilemma. Voluntary agreements become much more difficult with large numbers of ranchers. Each of the ranchers has an incentive to be intransigent and to at least act stupid. Hardin warns that "...natural selection favors the forces of psychological denial."
In the original Prisoners' Dilemma, irrationality is not a factor. The prisoners cannot fairly be accused of irrationally failing to negotiate with one another because negotiations were ruled out from the start. But in the Tragedy of the Commons, negotiation is difficult, not impossible, so even if the ranchers are perfectly rational in their independent decisions about how many cows to graze, they may well be accused of behaving irrationally during any subsequent negotiations. But the negotiations may break down or be prohibitively difficult even if they are rational, especially if the number of ranchers is large. The individual ranchers' behavior is irrational, contrary to their individual long term interests, only to the extent that they have the opportunity to negotiate with one another and they conduct the negotiations foolishly.
In the case of the original 2-player Prisoners' dilemma, altruism could easily make a difference. If prisoner A cared about prisoner B's prison time 51% as intensely as he cared about his own, he would not confess. This is plausible. Two people might be in love. But in the Tragedy of the Commons, this is less plausible. Here the problem is not so much that too many people are too selfish, but that they are set up in a social system that is too sensitive to selfishness. It only takes one rogue to spoil things for everyone. The others are faced with the dilemma of either following suite (adding more cows) or doing futile things that tend to reward the rogue (removing their own cows).
Another way out of the Prisoners' Dilemma is to use force. In Hardin's words, "mutual coercion, mutually agreed upon." This could be the Mafia enforcing a "no squealing" rule among its members. In polite society, it could be political action. But the political medicine we should prescribe depends on the disease we diagnose. Being stupid and having perverse incentives are not the same thing. If we diagnose the tragedy of the commons as being the result of stupidity, we might be tempted to attack the problem by requiring IQ tests for ranchers. If the problem is perverse incentives, having cleverer ranchers could make the problem worse. Responding to perverse incentives is not irrational; irrationality only comes into the picture when a political candidate tries to solve the problem and is told to go to hell by the people he's trying to help.
Some people cling to what I regard as a "white light and fluffy bunny" philosophy, that values are objective, and that if everyone were sufficiently enlightened about values, these sorts of conflicts would disappear. This makes no sense to me. I see a fundamental arbitrariness at the heart of the issue, limited only loosely by our predispositions as biological organisms. (A good discussion of "transcendent" vs. "empirical" ethics from a biologist's perspective is Chapter 11 of Consilience, by Edward O. Wilson, who has little time for Kant. See http://www.theatlantic.com/issues/98apr/biomoral.htm .) If our boat is sinking, and there is only one life preserver on board, which of the two of us should get it? Our apparent interests are simply in conflict, and if you give me the life preserver anyway, I will be grateful, but I still don't see why I should regard your values as objectively right. Even if everyone on the planet agrees on some value (ie. whose life is more valuable), this is still subjective opinion, not objective fact. I see "enlightenment" as a combination of emotional sensitivity and an awareness of long term consequences, not as objectivity. If "right" values are defined in terms of people's happiness, they are subjective and frequently conflicting; if they are not defined in those terms, why would either of us care about them? Even if you convinced me that values were objective, unless we agreed on which ones were the right ones, it still wouldn't solve the problem.
In terms of the Prisoners' Dilemma, what are objectively right or wrong are the conclusions about good and bad strategies that we derive from the logic of the game. The "tit for tat" style of "justice" that is associated with the Iterated Prisoners' Dilemma is objectively a relatively good strategy, or part of "natural rights," in the sense that it is a logical consequence of the payoff matrix (and other assumptions) that is part of the definition of the game. But this conclusion is relevant to the real world only if this payoff matrix happens to correspond to the values of real world people, which may or may not be true, and which is still subjective either way.
The Prisoners' Dilemma is a central part of my world view, but it is not enough. In the next six chapters, I want to make an exhaustive list of all the reasons that make sense to me for why someone might oppose laissez faire. Some of these reasons make obvious sense in terms of the Prisoners' Dilemma, but not all of them.
3. Objection 1: Judging Criteria
The first reason the free market doesn't solve everything has to do with there being 6 billion people on the planet who can't agree on criteria for judging the results. Limiting consideration to the USA, we are still talking about a quarter billion. If the free market produces a utilitarian outcome, there may be a ruling class, or a tyrannical majority, that doesn't like that outcome. I may have to buy off the powers that be before I can argue for utilitarianism. Note that I am using "utilitarian" in the formal sense of the greatest good of the greatest number, material or nonmaterial, not the informal sense of being crude or ignoring intangibles. In the case where population changes may be important, I will need to specify carefully which group of people I am looking at, before or after the population change, but for most purposes this is not important. (See David Friedman, "What Does Optimum Population Mean?" Research in Population Economics, Vol. III (1981), Eds. Simon and Lindert.)
If everyone has equal bargaining power, "reasonable" bargaining positions tend to be clustered around utilitarianism, as we play a 6 billion player implicit bargaining game. I reject metaphysical arguments that the State should follow some higher purpose than the purposes of the people who constitute it. Furthermore, I see the 6 billion people's opinions of what is desirable as being irredeemably subjective. If the State pursues the will of God, it is because people choose to, or because of divine constraints on their behavior, but it should not be because people are confused about the difference between constraints and objectives. Similarly, trees will have rights only if people choose to assign rights to them, either because they like trees or as a way of dealing with ecological constraints, but not because trees have outwitted humans at the bargaining table. God, trees, and social justice are relevant only if, only because, and only to the extent that they affect the subjective happiness and the behavior of the 6 billion participants in the negotiations. Any writer who starts off by insisting that there are certain basic rights is, in my opinion, asserting the conclusion.
Why should bargaining focus on utilitarianism? Why not uniform happiness? Why not uniform distribution of tangible goods? Why not focus on the well-being of the least well-off? I borrow the idea of a "veil of ignorance" from John Rawls: I want to choose a set of rules that will maximize my expected happiness, but I choose without knowing the future in detail. I also expect that a group's success in bargaining situations correlates with the amount of effort they put into them, which in turn I expect to correlate with how sensitive their well-being is to the outcome. Also, in so far as my motives are altruistic, I tend to most want to help those people who would most benefit from my help. But I also don't care about minor departures from utilitarianism. I care that utilitarianism is an approximation of what I can hope for from an implicit bargaining game, and I care that there will be people in positions of relative power who have conflicts of interest involving other people's well being.
This negotiating game will never end conclusively, but I expect more or less utilitarian positions to command the most popular support in democratic elections, especially as we make decisions affecting the distant future, when we can't be sure on which side of the railroad tracks our great-great-grandchildren will be living. But the first objection I will acknowledge to laissez faire is that there may be "powers that be" who have to be bribed in order to allow the rest of us to pursue utilitarianism. Sometimes I may have a conflict of interest, too.
Pop quiz: What does "utilitarian" mean as I am using it? (1) The greatest good of the greatest number. (2) Crude, ignoring intangibles.
4. Objection 2: Distribution of Benefits
The second reason the free market doesn't solve everything is that, even in situations where markets are "efficient," there is no particular reason to expect that the distribution of wealth (or any other advantages or desiderata) that they produce is anywhere close to what would be desirable from a utilitarian standpoint. Even if I were happy with the current distribution, I see no reason why the technology of the future should produce a similar distribution for my grandchildren*. In economic jargon, baking the biggest and best pie is an "allocation" problem, whereas dividing it up is "distribution." Deciding whether to make heating oil or gasoline from a tanker full of crude oil is an allocation problem. Making sure Grandmom in New England has enough money to buy heating oil is a distribution problem. As Thomas Schelling argues in the excellent lead essay in Choice and Consequence, practical problems usually appear as a muddle of both allocation and distribution issues, such as how to respond to an OPEC oil embargo. There are limits to the degree to which these issues can be separated--if every one of n people is guaranteed an equal share of the pie, each individual's incentive to cooperate in baking the pie is only 1/n of what his incentive ought to be for the free market to work ideally. Also, you can't produce an allocation improvement (a surplus) without opening a can of worms about how the surplus should be distributed. National defense produces a big surplus. Nevertheless, I want to separate allocation and distribution as much as practical. These problem classes have very different flavors to them, and procedures such as markets or voting rules that work well for one class of problems typically work very badly for the other. I want outcomes that are attractive both from an allocation and from a distribution standpoint, and to get a good compromise, I will often need different tools for the different jobs. I also don't want a solution to an important but relatively straight-forward and easy problem to have to wait for the solution of a difficult and controversial but relatively unimportant one. I also want to separate allocation and distribution in order to simply understand the issues better, and in particular to recognize when a special interest is promoting a bad solution to an allocation problem in order to hide a distributional agenda. But in general, no matter how "efficient" markets are at solving allocation problems, I have no guarantee that the resulting distribution of wealth will be attractive. So the second kind of objection to the free market is that the distribution of wealth (any benefits) may be unacceptable.
* Ed Leamer asks whether a computer is like a forklift or a microphone? A forklift increases worker productivity, but it is also an equalizer. A 5th percentile Japanese woman and a 95th percentile Scandinavian man can lift the same weight if they are driving the same model of forklift. A microphone, on the other hand, creates a "superstar" effect. One Enrico Pavarati with a microphone can put thousands of second-rate singers out of work. "Every town had its supply of entertainers in the Middle Ages but now a single entertainer can serve the whole world."
While there's no particular reason for expecting laissez faire to produce a satisfactory distribution of wealth in general, there is a simple textbook argument for expecting free trade normally to be better than widespread trade restrictions (protectionism) from a distribution standpoint (as well as allocation). In poor countries, labor is plentiful, and therefore cheap, relative to capital. In rich countries, labor is relatively scarce, and therefore expensive, and capital plentiful. If free trade is allowed, businesses tend to buy labor from poor countries where it is cheap, and borrow capital from rich countries where interest rates tend to be relatively low. This tends to drive wages up in poor countries and down in rich countries. It also tends to lower interest rates in poor countries and raise them in rich ones. These distributive results are good for the poorest people (labor in poor countries) and the richest of the rich (capital in rich countries), and bad for the others (labor in rich countries and capital in poor ones). If we accept the presumption that a dollar is worth more to a poor man than a rich one (declining marginal utility of income), and worth dramatically more to the very poorest, the benefits to the poorest of the poor should outweigh any harm this brings to anyone else. (The claim that wages in poor countries are relatively insensitive to the demand for labor is hard to reconcile with declining marginal utility of income.) Besides, the distributive harm to workers in rich countries is partially if not completely offset by the improvement in allocation, seen as generally lower prices (David Friedman shows in Ch. 19 of Price Theory that both countries clearly benefit in terms of allocation.) There are possible counter-arguments, some of which I discuss in Chapter 9, but if someone tells you that free trade is bad for the poor, he owes you an explanation, and you should be suspicious.
Another textbook example where laissez faire appears to be better for the poor than the obvious alternative is in setting the terms of rental contracts (see pp. 28-9 of Price Theory). At least, Friedman shows that in the long term, having government dictate "favorable" terms for renters (ie. the terms of eviction) in rental contracts results in the average renter being worse off in absolute terms than with free (competitive) markets (which may not necessarily be the same thing as laissez faire). The renters end up being forced to pay for things that cost more than the renters were willing to pay voluntarily. Thomas Schelling discusses the related issue of rent control (in the same context as gasoline rationing) in Choice and Consequences (pp. 4-6). He says that he can usually convince his students that gasoline rationing, for example, is a bad idea, and that the moral issues are not what the students thought they were initially, but that this often takes longer than 50 minutes, and he never tries to do it in just one class period. It is conceivable that "favorable" terms for renters might hurt landlords sufficiently more than they hurt renters to be "good" from a distribution standpoint (makes things more even), but the renters still end up materially worse off.
The remaining reasons why the free market doesn't solve everything have to do with what economists call "market failures," failures to achieve something approaching "Marshall efficiency," which is one of several ways to define solutions to the allocation problems. (This is David Friedman's term, from Price Theory, pp. 440-442.) Basically, if a dollar were worth the same to a rich man and a poor man, Marshall efficiency and overall utility would be the same thing. Marshall efficiency is important because people can exchange dollars more easily than they can transfer happiness from one person to another, and we can observe dollars changing hands but can't measure changes in happiness. We can write theorems about Marshall efficiency. It allows us to separate the parts of political economics that we have a firm grip on from those that we don't. There is no solution to the distribution problem corresponding to Marshall efficiency because there is no consensus on how big a piece of the pie a particular person deserves. Please note that Marshall efficiency is not the same thing as Gross Domestic Product (GDP), or any of the measures of tangible wealth in For the Common Good, by Daly and Cobb. A man's decision to work shorter hours because he realizes he would rather go fishing more and eat out less often probably increases Marshall efficiency, but it decreases GDP.
Part of the reason this is confusing is because there are different kinds of economics. There is the "science" of economics, and then there is what James M. Buchanan calls the "pure logic of choice." A scientist has to have theories that are "falsifiable," logically capable of being disproven by experiment or observation. This in turn requires that the theories describe things that can be measured. Unfortunately, what I am interested in is utility, which can't be measured. The pure logic of choice, on the other hand, is a logical framework for figuring out what questions to ask; it tries to set the problem up correctly, and if it turns out that the results depend on knowing things that are hard to measure, so be it. What I am trying to do here is set the problem up correctly, so I am interested in Marshall efficiency, which is not a measurable quantity, and if I ever forget that, I am in Big Trouble. I will occasionally switch back and forth between making factual, "scientific" claims and describing how to set up the problem, but I need to be very careful when I do, and part of your job as reader is to catch me when I mess this up. What will probably get me in trouble is assuming in a particular situation that people's goals are simple enough to analyze, and then forgetting that I made that assumption.
Pop quiz: Is economics about money and material goods (true or false)?
We live in an age of 'policy wonks' who judge programs by their effect on productivity, or output, or work effort. Wonkian analysis uses the jargon of economics while ignoring its content. Economists view the wonks' fixation on output as a bizarre and unhealthy obsession. Wonks want Americans to die rich; economists want Americans to die happy.
-- Steven E. Landsburg, The Armchair Economist
David Friedman defines economics in Price Theory as "that way of understanding behavior that starts from the assumption that individuals have objectives and tend to choose the correct way to achieve them." He also makes the point that outside of academia, too many "economics" papers have conclusions that were decided in advance by the people paying for them. Economic jargon seems to be the preferred language of people who are playing con games on the voters. One consequence of this is that economics has a bad reputation.
In order to argue that markets are Marshall efficient (the First Fundamental Theorem of Welfare Economics), economists need to make some assumptions, which I have clumped into four categories, following David Friedman, from whose "Should Medicine be a Commodity" article I quote ( http://www.daviddfriedman.com/Academic/Medicine_Commodity/Medicine_Commodity.html ). This gives rise to four additional objections to free markets.
5. Objection 3: Knowledge Problems
One of the necessary assumptions is "Perfect Knowledge: Individual producers know the cost of all alternative ways of producing and the market price for what they produce; individual consumers know the price and the value to them of goods they consume." Part of "perfect knowledge" is "rationality." For this reason, in Free to Choose, Milton and Rose Friedman specifically added taking care of children and the mentally ill to Adam Smith's list of the proper functions of government. The word "value" is a bit of a sleeper here. In order for me to have perfect knowledge, I need to know what I want and how badly I want it. (The existence of the advertising industry shows that this is often unrealistic.) I also need to be able to judge my doctor's competence, I need to not be crazy or stupid, and I need to not make too many random errors about important decisions. Random variables don't make much difference to this argument if one person's knowledge is as good as the next person's, but asymmetric information is more of a problem. The market for used cars is screwy because the seller knows more about what a car is worth than the buyer; the buyer doesn't want to offer an above-average price for an unknown quantity, so people who sell good used cars tend to get penalized for it. People tend to be better informed about things they buy frequently (ie. a good price for a box of raisins) than rarely (ie. a good price for a house).
"Rationality" does not mean that everyone carries a calculator around with him all the time. It just means that people can generally figure out how to get what they want, often simply by imitating their friends. McKenzie and Tullock (The New World of Economics) claim that for economists' purposes, mental patients and laboratory rats are quite rational enough. But if you don't have a pretty good idea what someone's objectives are, it's hard to say if they are pursuing them rationally; uncertainty over people's rationality tends to get lost in the uncertainty over their objectives. Also, in some situations (ie. games like poker that involve bluff and intimidation), it may be in someone's best interests to at least appear to be irrational, and we have "the rationality of irrationality." We saw this to some extent in the Tragedy of the Commons.
6. Objection 4: Insecure Boundaries
Another assumption is "Private Property: Property rights are defined and costlessly enforced and can be costlessly transferred for all scarce goods." Property can be public or private, and many things are not defined as property at all, such as fish that are outside of any nation's territorial waters. These are a "fugitive resource," and fishermen have perverse "use it or lose it" incentives similar to the ranchers' incentives in the tragedy of the commons. From the standpoint of society as a whole, it might be better to avoid catching fish below a certain size, or from populations that are too thin to replenish themselves quickly. But if a particular fisherman doesn't harvest a particular fish when he has the chance, the vast majority of any benefit that society gets from his abstinence will go to someone else. Furthermore, there might not even be any such benefit, because the next fisherman may not be so scrupulous. If property is not transferrable, it may be wasted even though there are other people who want it, which happens often with water rights in the American west. Note that a high crime rate can be a source of market failure, and that for these purposes, a legal system that behaves erratically has the same effect as crime.
Another obvious problem is air or water pollution, something with a negative value that one can't avoid simply by refusing to buy the polluter's goods. If I live downstream of a medieval tannery, I as an individual can't avoid having dead fish in the river by wearing cotton instead of leather. I may be "external" to the transaction between the leathermonger and the leather buyer, but I still suffer because of it. Things like pollution are sometimes called "external costs" or "negative externalities." This again is like the multiplayer Prisoners' Dilemma. If the leathermonger has a thousand customers, my decision to buy leather only increases the harm I suffer from his pollution by a tenth of a percent, even though the costs to society as a whole, if there are a thousand people living downstream of him, may be more than the value to me of wearing leather.
The term "scarce goods" is something of a sleeper. If we want to consider unpleasant intangibles, such as my neighbors' possible distaste for the bumper stickers on my car, we need to expand this "private property" assumption (or our interpretation of it) to include any intangibles that people deem important, including psychological boundaries. Augmenting the term "scarce goods" with words like "services" and "amenities" would help, but it is not enough. David Friedman has explained libertarians' views of rights as all boiling down to "the right to be left alone," but there is a gap between a libertarian's idea of being "left alone" and a solid utilitarian argument. In order for my list of potentially valid objections to laissez faire to be complete, I want my neighbors to be able to say that their objections to my bumper stickers, their loss of utility that is potentially greater than my gain, fits into this fourth category. I think there is a compelling case for free speech, free expression, etc., but I want that to be one of the conclusions of my arguments, not an assumption that I have inadvertently swept under the rug.
I don't want to understate the importance of intangibles. As H. L. Mencken put it,
What avails it for a man to have money in the bank and a Ford in the garage if he knows that his neighbors on both sides are watching him through knotholes, and that the pastor of the tabernacle down the road is planning to have him sent to jail? The thing that makes life charming is not money, but the society of our fellow men, and the thing that draws us toward our fellow men is not admiration for their inner virtues, their hard striving to live according to the light that is in them, but admiration for their outer graces and decencies--in brief, confidence that they will always act generously and understandingly in their intercourse with us. We must trust men before we may enjoy them.
Notes on Democracy, pp. 174-5, Alfred A. Knopf, New York, 1926.
If it seems ridiculous to you to suggest that intangibles be thought of in the same terms as tangible property, consider the copyrighting of "intellectual property." "Intellectual property" is abstract and inconvenient to legislate about, but if the issues are important enough, we generally accept this inconvenience, and whether we accept it or not, we acknowledge the underlying problems.
The logic of the argument for the economic efficiency of laissez faire does not depend on distinguishing between people's material and nonmaterial interests, but I choose to do so for two reasons. One is that if intangibles are not addressed explicitly, there is likely to be confusion about whether they are included. People are likely to either overlook them or to accuse other people of being blind to them. The other reason is that intangibles naturally belong in a class by themselves because they are so much more difficult to reach agreement over. Consequently, I will divide arguments about markets failing due to lack of property rights into two subclasses, tangibles and intangibles (4a and 4b).
The remaining two categories of assumptions can be lumped together as "no transaction costs," but I prefer to keep them separate because they seem less abstract that way. The "costless transfer" assumption regarding property rights (objection 4) is necessary for there to be no transaction costs, as is adequate information (objection 3), but "transaction costs" also include the time and effort people spend searching for what they want and negotiating with one another.
7. Objection 5: Imperfect Competition
The third assumption that goes into the economic efficiency argument is "Perfect Competition: Every producer and consumer is so small a part of the market that the quantity he produces or consumes does not affect the market." It may seem obvious that monopolies are a good argument for government interference with the markets, but since I am interested in figuring out just how good an argument they are, I need to disentangle the different reasons people have for worrying or not worrying about them much. The way I have introduced monopoly is in the context of economic efficiency and transaction costs. Technically, this is correct. Monopolies fall into two idealized cases, single-price monopolies and price-discriminating monopolies, with some different internal logic, but real-world monopolies inevitably involve either complicated negotiations or customers who are willing to pay what something costs to produce, but aren't getting it. I raise this issue in the context of Marshall efficiency, and technically, the loss of efficiency associated with monopolies is often modest. Also, technically, figuring out how a regulatory agency should deal with the situation is often problematical. But there is more to antitrust than this.
One of the reasons I emphasized "fuzzy logic" in my original title for this essay was the obvious one that the assumptions that go into my argument are seldom if ever completely true, but only true to various degrees. For example, imperfect competition includes not just monopolies and monopsonies (situations where there is only one buyer), but any situation where choices of whom to buy from or sell to are limited. Any real-world situation will fall into this category to some extent, and so I remind my readers that we are in the ambiguous realm of "fuzzy" or multivalent logic. Monopolies also illustrate a less obvious point, that real world problems do not necessarily fit neatly into any one specific category, even though the category may be named after that particular problem (Bart Kosko discusses fuzzy logic in terms of overlapping set membership in his book, Fuzzy Thinking). "Company" towns are notoriously objectionable places to live, but the reasons they are objectionable are mainly because of the distribution of wealth and intangible side effects (ie. "social justice"). In general, legitimate arguments to laissez faire, as well as similar objections to political intervention, will not fall simply into one of my classes of objection; their logic will flow in differing degrees down several paths, dividing and recombining as appropriate. It is important to remember that monopolies are objectionable for their distributional effects and their intangible social ills, not merely because they raise transaction costs.
Ironically, the "perfect" idealized price-discriminating monopoly is potentially as economically efficient as perfect competition, yet price-discriminating monopolies, companies that charge different prices to different customers, are probably the more obnoxious kind. From a technical standpoint, the most troublesome are bilateral monopolies, with only one buyer and one seller, which tend to resemble the game of "chicken," complete with standoffs, bluffing, and expensive breakdowns in the negotiations.
Some forms of monopoly are due to a cohesive ethnic group or subculture acting as a single entity towards certain outsiders. The Mormon church reportedly made a lot of enemies this way in its early history.
Earlier I expressed concern about what the technology of the future would do to my arguments. Computers are one of my reasons for concern, for several reasons that involve tendencies for computers to promote monopolies. The first reason for this tendency is that the average cost of copies of a piece of computer software (and similar kinds of increasingly important information) decreases with the number sold. The first copy is horribly expensive to produce, but the second and subsequent copies can be copied for a matter of pennies, far cheaper than a competitor could develop a similar product independently. This is the textbook recipe for a natural monopoly. (Dennis Mueller, on p. 235 of Public Choice II, defines "natural monopoly" in terms of "falling long-run average costs.") A second reason is that the advantages of standardization are particularly acute for computer software, and standardization often means buying from the dominant supplier even when he does have serious competition. A third reason is that with modern computers, the optimum size of corporations is increasing. In the past, the size of corporations was limited partly by the number of subordinates that one manager could supervise and the number of layers of management that could exist before communications breakdowns became prohibitively expensive. Computers have increased the number of employees that one CEO can manage effectively and made it easier to communicate across layers of management (and physical distance). As with the distribution of wealth, even if I were happy now, there is no guarantee that I would take a complacent attitude towards monopolies in the future.
Mancur Olson presents an even more disturbing argument in The Rise and Decline of Nations. He was puzzled by the fact that, after Germany caught up with Britain in GDP after World War II, Germany's economy continued to grow faster. Olson came to the conclusion, with convincing evidence, that cartels and oligarchies tend to develop over time in politically stable countries with fixed borders. Just as a small group of ranchers is likely to eventually negotiate a socially beneficial agreement over grazing, so do other groups of professionals eventually negotiate cartels. Olson draws a sharp distinction between "laissez faire" and "free markets;" such cartelized markets are not "free" in his view. He argues convincingly that the Great Depression was due at least in part to these sorts of cartels, which often develop only over long periods of time. He gives new meaning to John Maynard Keynes' comment that "In the long run we are all dead."
Some problems relating to imperfect competition involve "firm-specific capital," investments of money or time on things that are only useful if you are dealing with a specific company or person. For example, if I become dependent on elaborate spreadsheets that work with one company's spreadsheet program, I may lose a great deal of work if I decide to switch to another company's program. Other examples are non-portable medical insurance and retirement plans. The cost of changing jobs may be prohibitive even if I hate my job. Markets that are characterized by high transaction costs such as these examples illustrate are sometimes said to be "sticky."
8. Objection 6: Free Rider Problems
The last assumption is "Private Goods: Every producer can control the use of the goods he produces." The opposite of a "private good" is a "public good," which Friedman defines in The Machinery of Freedom (2nd ed., Open Court, 1989) as "an economic good which, by its nature, cannot be provided separately to each individual, but must be provided, or not provided, to all the members of a pre-existing group." A flood control dam upstream of a heavily populated valley is a public good. More importantly, so is military protection. Like fugitive resources, public goods can be viewed in terms of the Prisoners' Dilemma, but they are things people produce rather than things that occur naturally. Instead of having an incentive to overconsume something that isn't mine, with a public good I have an incentive to try to weasel out of helping to pay for something that I hope to get anyway. If an overfishing problem is serious enough, the fish population may crash, whereas if a public good problem is serious enough, the public good will never be produced in the first place.
Consider how difficult it would be to defend a nation against an invasion without a government-backed army. It is militarily impossible to defend the houses of people who have contributed to a militia from an invading army without also defending their neighbors. Thus, in the absence of a government that can levy taxes or draft soldiers, my incentives to be reasonable with anyone who tries to negotiate a unanimous or near-unanimous agreement for citizens to support a militia are again like those of the rancher on the commons. My reasonableness costs me money and possibly an increased chance of being killed if there is a war. The benefit, assuming the number of citizens is large, is a tiny reduction in the odds of everyone in my society losing a war. I pay all of the additional cost of my "reasonable" contribution to the militia, and get only a trifling fraction of the additional benefit. I may contribute for intangible reasons, such as patriotism, but I have obvious tangible reasons not to.
How difficult it is to produce a public good without using force depends on the number of people involved and the cost/benefit ratio. It also depends on how uniformly the benefits are spread. If a small subgroup gets the lion's share of the benefits, they may decide to foot the entire bill themselves. Paul Samuelson requires a "pure" public good to be one whose benefits are uniformly distributed (his definition of a public good is more complicated, involving "jointness of supply" and "non-excludability of benefits," but in my understanding, the only difference in effect between Samuelson's and Friedman's definitions is to blur the distinction between monopoly and public good problems). Often, a small public good is produced as a side effect, perhaps inadvertently, of producing a private good. For example, by taking the flu vaccine, I not only protect myself from getting the flu, I also prevent myself from passing it on to my co-workers. Similarly, in the process of doing applied research for the sake of developing a commercial product, I may do some basic research that advances the state of the art of my entire industry. I may publish a scientific paper to promote my career that results in someone in the Third World averting a crop failure. Effects like this are sometimes called "positive externalities." Again, I experience all of the disadvantages of the flu vaccine, but I only get some or most of the benefit, so I tend to not take the vaccine as often as would be consistent with utilitarianism.
9. Welfare Economics
Figuring out how to produce the right amount of a public good, or the right amount of a private good that has positive or negative externalities associated with it, is called "welfare economics." Although I have emphasized that these problems are all similar in having a strong flavor of the Prisoners' Dilemma, and may be lumped together as "spillover effects" or "neighborhood effects," there are a variety of public policy tools that may be appropriate to different cases. National defense is provided directly by the government, and it is hard for me to imagine how else that might be done. Privatization should solve an overgrazing problem if cheap fences are available (ie. barbed wire). By "privatization," I mean converting the commons to so many pieces of private property, not the outsourcing of government services that is often now perhaps disingenuously referred to as "privatization." (Also, beware the possibility that "privatization" will be used as a pretext for giving public property to one's friends, at the expense of the poor.) Class-action lawsuits are sometimes proposed as an approach to dealing with pollution and similar problems. If there are no relevant property rights already defined, lawsuits can sometimes establish them without requiring action by the legislature ("common law" as opposed to "legislation"). If the relevant rights are already defined, these lawsuits can sometimes be used to enforce them independently of public prosecutors. The classical approach to welfare economics, attributed to A. C. Pigou, is to tax things like pollution and subsidize things like inoculations. This works best if the amount of harm done by each additional ton of pollution is constant or decreasing, and can be estimated easily. This way, the government chooses the price of pollution, and the market chooses how much to pollute. (As with disciplining small children, there is enough difference between the problem of encouraging desirable "start" behavior and the problem of discouraging undesirable "stop" behavior that some distinction between them should be maintained; it is probably not a good idea to think of a subsidy as a negative tax.) Another approach, associated with Ronald Coase, is tradable quotas. This works better if there is a fairly specific amount of pollution that is tolerable, before the damage rises sharply. This way, the government chooses the amount of pollution and the market chooses the price. An analogous approach to subsidizing a public good would be to contract by competitive bid for a specific amount to be produced. The stereotypical "command and control" approach to environmental regulation is used for political reasons, not because economists approve of it. The preferred approaches are market-like, even if the markets are artificial.
Military defense takes first prize as the most compelling reason why the most extreme form of libertarianism, anarcho-capitalism, won't work in a world like ours. Second prize goes to environmentalism. As Jeffrey Friedman put it Critical Review v. 6 p. 443, "The environment is the libertarian Waterloo: it reveals the flaws of the doctrine in a way that seems to ensure that no 'answer' is forthcoming." It is hard to imagine the Earth's population of 6 billion negotiating to limit their output of greenhouse gases without government involvement. It is similarly hard to imagine "fencing off" privatized chunks of the atmosphere. Many environmental issues are also very intangible, such as species diversity and beauty.
Consider the effects of population growth on the utility of the people already living on the planet. In an ideal world, my property rights would be defined so perfectly that how many children other people had would be none of my business; it wouldn't affect me or my family unless we chose to allow it to affect us. In the real world, the way property rights are defined is a compromise between protecting people against the worst cases of nonconsensual interference in their lives and allowing people to take action without having to get prior written permission from everyone on the whole planet. Property rights can't be defined perfectly, and higher population density means more and more compromises have to be made.
Population growth is another example of the "fuzzy logic" problem of a real-world objection to laissez faire that doesn't fit neatly into just one category. A high birthrate in another part of the world may not directly affect my happiness, but it exacerbates problems with insecure boundaries, imperfect competition, and free rider situations. My family may choose not to have as many children as our neighbors, but we would still experience the effects of a high population density. For one thing, a high population density increases the opportunities for crime. Negative externalities such as the spread of communicable diseases are also made worse. With one person per square mile, I don't have to worry about whether my neighbor's septic tank is leaking, but with 100,000 people, a good sewage treatment system is essential. If the population is heterogeneous, nonuniform population growth also imposes growing political costs on the slower growing subculture, even with the private police and court systems that some anarchists suggest. This includes not only the risk of being dispossessed outright, but also a certainty that the ownership of publicly owned property like national parks will be spread out over a larger public. This is analogous to "stock dilution," a corporation issuing new shares of its stock and giving them away to outsiders. Immigration poses similar issues.
The imperfect competition effects are more subtle. High population density makes commercial competition more intense, but unfortunately, the way property rights need to be defined is as a compromise to avoid the worst of many problems involving monopolies, public goods, and both positive and negative externalities. If the rights of students in a college dormitory protecting them against noise pollution are defined too narrowly, one party animal can keep everyone else from being able to study or sleep, which is an externality problem. If their rights are defined too broadly, anyone in the building may be able to object to any detectable amount of noise. In this case, anyone who wants to play music will have to treat several hundred people as each having monopoly privileges over their own ears, and will have to deal with "the holdout problem." With a given definition of rights, more people means more monopolists. More people also means that the definitions will probably need to change.
Some property rights compromises involve one person imposing risks on another. If I fly an airplane over your land, there is some small probability that I will crash, and maybe kill one of your cows, or possibly even you. Your property rights may require that I maintain some minimum altitude, but they normally do not forbid overflights. Regardless of whether I have insurance, or am able and willing to compensate you or your next of kin for the damage I do, someone is bearing some risk associated with the fact that we are both using the same land at the same time. In effect, this property rights definition that allows overflights treats the land partly as your private property, a "scarce good," and partly as an unowned resource, plentiful empty space. As the population density increases, the risks become more expensive, and eventually it will become clear that empty space near where I want to fly is a scarce good, and one for which I do not have property rights defined. As Zechariah Chafee wrote, "Your right to swing your arms ends just where the other man's nose begins." High population density means there are more chins nearby. Some of those "chins" are tangible and some involve intangible things including goodwill.
Finally, and probably most important, high population density and the associated human also reduces the amount of public goods that I receive from undeveloped land and healthy ocean. On Spaceship Earth, whose responsibility is it to maintain the CO2 scrubbers? Technically, some of these things may be public goods or positive externalities and some of them may be unowned resources, but the distinction is not clear to me, and the point appears to me to be academic. If I buy a piece of land and refuse to sell it to developers, are the wild bees that live there and pollinate my neighbors' crops a positive externality or an unowned resource? I don't care. Either way, if I sell out to developers, my neighbors are going to have to pay for services, using domesticated bees, that they used to get for free, and something wild and beautiful has been lost. His children have also probably lost a place to play.
Herman Daly has spoken of the scale of human activity within the broader context of the natural world. If I understand him correctly, he applies terms like "externalities" only within the sphere of human activity, and he refers to the effects of human expansion on the rest of nature as a problem of "scale." He likens "externalities" to the "epicycles" in Ptolemaic astronomy. I see some justice in this comparison, but unfortunately, as far as I can see, he offers me no Keplerian ellipses to take their place, just the already familiar welfare economics of Pigou and Coase. I prefer to think of the decisions humans make regarding nature as a subset of the decisions human beings make in general, and I am willing to be sloppy in my use of terms which I do not use on a professional basis. Consequently, I see no need to separate objections to laissez faire involving "scale" from the six objections I have already recognized, but I encourage others to do so if they find that it clarifies the issues involved. I also note that the idea of "scale" goes well with Coasian tradable permit-style welfare economics, in which essentially the scale of human activity within some domain is set by the government, but markets are otherwise allowed to function.
In For the Common Good, Herman Daly and John Cobb present some communitarian arguments against free trade, which also illustrate how real world problems refuse to fit nicely into simple categories. (For my readers who have also read Daly and Cobb, it also illustrates how hard it is to translate arguments from one worldview to another.) One of their arguments is distributional (my objection 2): they consider an inequitable distribution of wealth within a single country to be a more serious social problem than differences between countries. As I argued in Chapter 4, free trade tends to equalize both interest rates (the return on capital) and laborers' wages across countries, whereas trade restrictions tend to maintain high wages in wealthy countries and low wages in poorer countries. Other arguments involve risks associated with trade disruption due to political, military, or natural disasters. The people who buy imported food and the people who buy domestic food share equally in any food shortage that results from a trade cutoff. I call this a problem with insecure intangible boundaries (objection 4b); I may have been keeping some local farmer in business for the last ten years while you bought imported food, but during an emergency you have just as much right to buy from him as I do, and he won't have the capacity to feed everyone. Even if there are very efficient insurance and commodity futures markets for dealing with that risk, import buyers are still inflicting risk on domestic buyers, and it is unlikely that most people will have the good information (objection 2) and the secure property rights in an emergency (objection 4a) for these risk markets to be efficient. I also note that the problems such as insecure property rights that people face in developing countries are likely to be more serious than in the US. It is distinctly possible that the political and social systems in poor countries are often unable to cope with the rapid changes likely to be brought about by a sudden transition to free trade (and I would argue that the local political system is likely to be precisely why the country is poor in the first place).
10. Interpersonal Utility Comparisons
Having made the four sets of assumptions associated with objections 3 through 6, I can then demonstrate that the GIGO law (Garbage In, Garbage Out) is in effect. The perfect information and "rationality" assumptions eliminate any internal restraints that prevent people from negotiating for what they want. The property assumptions eliminate any impediments to the negotiations other than the people involved in them, and eliminate harm that takes place outside of negotiated limits. The perfect competition and private goods (or zero transaction costs) assumptions eliminate interpersonal obstructions to the negotiations. With perfect competition, there are large numbers of buyers and sellers or all goods, making all goods "commodities" with well defined market prices, which eliminates any incentive to haggle. The private goods assumption allows every transaction to be broken down into a simple one-on-one exchange. From these assumptions, it then follows that negotiations will move quickly towards an equilibrium in which all of the possibilities for mutually beneficial agreements have been exhausted. This is "Pareto efficiency," the usual definition of economic (allocation) efficiency.
Note that whether or not a particular outcome is Pareto efficient has to do with a mutually beneficial process that can lead away from it, not with the outcome itself considered in isolation. (If you can suggest a deal that makes someone better off, and leaves no one worse off than the status quo, then your situation is not Pareto efficient; otherwise it is.) The motive for defining economic efficiency this way is largely to avoid having to make interpersonal utility comparisons. A Pareto efficient outcome is said to be "on the Pareto frontier" because there are an infinite number of such possible outcomes, not one unique solution. Not only are there an infinite number of possible starting points for the negotiations, but for any starting point, there are an infinite number of possible outcomes within the "bargaining range" within which people can haggle. A negotiation may lead from a particular point that is not Pareto efficient to one that is, but that doesn't mean that any point on the Pareto frontier is necessarily better in any meaningful sense than any point that isn't. If you pick two possible outcomes at random, typically neither one will be a Pareto improvement over the other. For example, if Bill Gates owned the entire world, that would be Pareto efficient, because there would be no way to improve on the situation without making it worse from his viewpoint, but that doesn't make the situation an improvement from my viewpoint over one in which the world was conquered by Red Hat Linux. While it is hard to argue that a Pareto improvement by itself is a bad thing, it is easy to be unhappy about the point from which the negotiations started, and it is easy to prefer a different Pareto improvement that leads to a different point on the Pareto frontier, say, one in which my side was able to drive a slightly harder bargain.
Although the Pareto criterion sometimes allows me to compare different possible outcomes without requiring me to make interpersonal utility comparisons, this is of little advantage to me because, for my purposes, I have to deal with distribution of wealth issues anyway. I also find that I have to pay too high a price for a Pareto improvement in terms of the way property rights need to be defined. Specifically, I want to relax the private property assumption that eliminated negative externalities. This assumption prevents me from doing things I desperately want to do without getting your permission because they might annoy you, however slightly. For example, I and my date may be extremely horny and want to make love in my college dormitory room, but in order to show this outcome to be Pareto efficient, we would have to negotiate with each and every person within earshot who might possibly be annoyed, however slightly, by the squeaking noises coming from the mattress. Again, the motive for looking for Pareto improvements is largely to avoid interpersonal utility comparisons. If I say that it's okay for me to blow off negotiating with you over my behavior and blow off compensating you for the damage I do to you because this damage is so small compared to how much it benefits me, I am comparing your loss of utility to my gain. In order to be able to defend property rights that are reasonable compromises between their various problems, I need to be able to defend these comparisons.
David Friedman does this by using Alfred Marshall's definition of an improvement rather than Vilfredo Pareto's. Marshall defined a change to be an improvement if the benefits outweighed the costs, where these are measured in terms of what people are willing to exchange in order to get or avoid them (ie. in terms of money). The difference between a Marshall improvement and a Pareto improvement is that a Pareto improvement in effect requires a set of side payments to be made from the beneficiaries to the victims so that no one is left worse off. Marshall does not. Under normal circumstances, the Marshall "frontier" is the same as the Pareto frontier. (The differences involve situations such as alcoholic parents who take advantage of the grandparents' altruism towards the children, in which case Marshall's definition of "improvement" is actually slightly more restrictive. See http://www.daviddfriedman.com/Academic/Marshal_Pareto/Marshal_Pareto.html .) But in general, the processes that Marshall allows are more realistic than Pareto's, and the place one ends up at on the frontier from any given starting point is likely to be different. Marshall allows any two arbitrary situations to be compared. Pareto does not, but the shape of the frontier doesn't depend on a specific destination for a change, only whether one is possible.
Using Marshall efficiency allows me to relax the private property assumption. Instead of worrying about "market failure" every time there is a negative externality, however slight, I now only have to worry if I suspect that the dollar value of the harm done to the victims is greater than the dollar value of the gain to the beneficiaries. The price I have to pay for this relief is that I have to worry more than before about the distribution of wealth. If, on average, the beneficiaries and the victims are equally wealthy, this is probably not a problem, but now, even if we start off with an acceptable distribution of wealth, there is no guarantee that the poor will be at least as well off after the allocation problem has been "solved." I have, in effect, reclassified some of the headaches I described under "objection 4" as "objection 2."
An effect that simply transfers money from one person's pocket to another's is called a "pecuniary externality," and has no effect on Marshall efficiency because the Marshall criterion is based on measuring the value of things by their exchange value (money). But from a distribution standpoint, we still have to pay attention.
My list of objections to laissez faire is more abstract and daunting than I would like, but I believe it is complete. If you have an argument for government intervention that doesn't seem to fit into any of these categories, one of us has made a mistake. It may seem that I present the argument for the attractiveness of laissez faire as a very leaky bucket, but I believe I have now got the leaks lined up in relatively orderly rows, and I am now ready to consider the alternative.
Part 2: Democracy
11. Public Choice Theory
There is no difficulty in proving any ethical standard whatever to work ill if we suppose universal idiocy to be conjoined with it....
--John Stuart Mill, Utilitarianism
I have been discussing "laissez faire" or "markets" as a process, and I now propose to do the same for democratic politics. Many people talk of "capitalism" and "democracy" as being sets of values, perhaps corrupt and healthy ones, respectively, but I regard any discussion of them on that basis as sterile rhetoric. For my purposes, capitalism and democracy are processes, akin to computer programs that run over a large network. In industrial engineering jargon, one may describe these as multi-objective decision-making algorithms for a massively parallel data processing system. Just as it is wrong to confuse the objective with the constraints in an optimization problem, so it is wrong to confuse the objective with the algorithm. Someone who designs anti-aircraft missiles could use Newton's method to minimize the miss distance to the target aircraft, while his next-door neighbor who works on electronic countermeasures uses Newton's method to maximize the miss distance. Perhaps a psychologist can argue that the use of Newton's method tends to predispose people to want to shoot down airplanes in the long run; if so, I believe that that point needs to be stated explicitly and supported by evidence, because I see no reason why a reasonable person would merely assume such a thing. Similarly, I follow "Mr. Dooley" in disbelieving that men can be turned into angels through an election. I encourage people to reflect on their values, but my immediate purpose is not to change people's values, but to see how good a job these two algorithms do at achieving the values people already have. However, since the idea that people's values are conditioned by their economic circumstances is central to much of socialist thought, I will specifically address this issue in Chapter 18.
Just as I have discussed "markets" without insisting that the term only applies in cases in which, for example, water rights are defined in a particular way, I use "the democratic political process" to refer to a variety of rules, from initiative and referendum to multicameral legislatures. A rule, such as needing a 2/3 majority for a Constitutional amendment, may be associated with a particular ideological norm, such as "consensus" as opposed to "majoritarian" democracy, but by voting for a politician who endorses a particular rule, I do not control what my neighbor's ideology will be. There are many ways to define property rights, and many ways to "reinvent government." In some cases the differences are important.
In addition to defining "laissez faire" and "democracy" as processes, in order to make a fair comparison of the ways they work, I have to make compatible assumptions about the people who use them. Specifically, I need to be consistent in my assumptions about how rational people are and how much information they soak up simply because they are curious. I also need to be consistent in assuming how selfish or altruistic people are. If I want to incorporate models of how people's values change over time in various environments, I still need to be consistent about the initial conditions (and the models need to make sense).
Notice that I am more worried about making a fair comparison (consistent assumptions) than about making accurate assumptions. It is as if an economist and a political scientist were trying to decide whether to order a large (16") pizza or two mediums (10") for the same price, and they want to know which is more food. I am afraid that the economist will make the simplifying assumption that pi is 2 in estimating the area of the large pizza, and the political scientist will make the simplifying assumption that pi is 5 in estimating the area of the mediums: they will end up ordering the mediums when the large is the better deal. I don't care that their estimates are bad. I care that they are inconsistent.
It is not essential, but still desirable, if the logical framework I use in discussing government "policy failures" is similar to the logic I use in discussing "market failures." This approach is called "public choice theory," which is a branch of economics, which makes it easy to make sure I am consistent in the assumptions I use in comparing laissez faire with democracy. It is also the view of politics I am most familiar with; part of my attraction to it may be lack of awareness of the alternatives. It is also possible that I am giving economists too much credit for work that is not theirs out of my confusion. In any case, I reserve the right to be eclectic.
I anticipate that some people will reject my arguments as being necessarily wrong simply because they don't like the way I frame the situation. This is a bit like saying that, if you would use rectangular coordinates to set up a math problem, and I do it in polar coordinates, my answer must be wrong. Obviously, this is not the case. A poor choice of coordinate systems may be inconvenient, make a problem harder, and may make certain kinds of errors more likely, but it doesn't make the answer wrong. So bear with me. My answer could still be right, and maybe this framework will grow on you.
One of the simplest models of how democracy works is the Median Voter Model. In its simplest form, there are two candidates for office, 100% voter turnout, and the voters are distributed in an internally consistent way along a single left-right spectrum of opinion (if voter A likes candidate X better than Y because Y is too far to the right, A would like Y even less if Y moved further to the right). This situation is analogous to having two otherwise identical ice cream vendors taking two positions along a long, narrow beach, with customers walking to whichever one is closer. The optimum place to sell ice cream is in the center (where the median customer is). If both vendors sell right next to each other at the center, they each get half the business. If either of them sets up anywhere else, the vendor in the center gets more than half, and the one who moved gets less than half. This may explain the common perception that there isn't a lot of difference between the two major candidates in many American elections. (McKenzie and Tullock have a good discussion of the Median Voter Model in The New World of Economics.)
Another important concept is the Condorcet paradox. Suppose three people, A, B, and C, are trying to agree on a restaurant. A prefers Chinese, then Italian, then Mexican. B prefers Italian, then Mexican, then Chinese. C prefers Mexican, then Chinese, then Italian. What is their collective opinion? Two of the three prefer Chinese over Italian, two of the three prefer Italian over Mexican, and two of the three prefer Mexican over Chinese. No matter what their current plans are, there is an alternative that is preferred by a 2/3 majority. Each individual can rank his preferences in order (their individual preferences are "transitive"), but the group cannot (the group's preferences are "non-transitive"). For this reason, when one speaks of a group decision, it is often necessary to specify the process by which it was made. One consequence of this is that groups that are governed by majority rule are vulnerable to agenda manipulation, and are sensitive to organizational rules such as how soon after a motion has been defeated it may be reconsidered. Without such rules, a legislature could conceivably find itself playing a seemingly endless game of "rock, paper, scissors," a phenomenon known in public choice literature as "cycling." In any case, the agenda setter is a very powerful figure. (See Mueller, pp. 63-4, 87-9.)
Although there may be a stigma in many people's minds associated with "logrolling" and political compromise in general, these are as important to a public choice argument for the benevolent working of democratic government as trade is for economic efficiency in the free market. For a simple example of how this might work, consider a town meeting in a hypothetical town with ten voters. Suppose that four of the ten are gay, and willing to pay $100 per year to avoid being persecuted. They are also in favor of gun control, and willing to pay $10 per year to ban handguns. Another four of the ten are NRA members, who are willing to pay $50 per year to not have a handgun ban, and $5 per year to persecute gays. The remaining two voters are willing to pay $10 per year to ban handguns and $5 per year to persecute gays. Marshall efficiency will be higher by $370 if gays are not persecuted, and $140 higher if there is no handgun ban. If the current laws persecute gays, even though six of the ten would prefer to maintain the status quo, the remaining four have incentives to offer large enough side payments to be able to buy their votes. If the minority can agree on how to divide the cost, and the majority on how to divide the bribe, the result will be Marshall efficient and could potentially have the unanimous support of both groups. The handgun law could be considered separately, with similar results.
Earlier I expressed concern about a cohesive ethnic group or subculture acting as a monopoly or cartel. The political analog of this is bloc voting, and it, too, will screw up my argument. People who vote as a bloc are acting "altruistically" with respect to the other members of their bloc, but not with respect to society as a whole. The US Presidential Electoral College is an example where the legal framework in effect allows each state to force its voters to vote as a bloc, which causes a great deal of mischief. See the discussion of the "unit rule" by Neal Peirce and Lawrence D. Longley in Electoral College Primer, or my review of it in Three-Fifths of a Floridian.
Unfortunately, if we multiply the numbers of people involved in these town meetings by a million, and consider votes that take place independently, by initiative and referendum, it is very unlikely that any such side payments can be negotiated. In such a case, two different "tyrannical" majorities are likely to pass (or refuse to repeal) legislation that is Marshall worsening.
But now consider a legislature that is elected by proportional representation (PR), in a society in which gay rights and gun control are either important enough issues to determine people's political affiliations, or in which they correlate well with other issues that are that important. (Examples of PR include the Irish "Single Transferable Vote" system and the German "mixed-member" hybrid party list system, and I discuss at length in Instant Runoff Voting.) If there are only ten members in the legislature, well informed and loyal to their constituencies, and there is a cheap, effective way to make side payments from one legislator's constituency to another's, we are back in a situation similar to the original town meeting example. Unfortunately, bloc voting by political parties is likely to be a problem.
If direct side payments are effectively prohibited, but the legislative agenda is established long enough in advance of any votes, "logrolling" may suffice to accomplish a similar result. In this case, the four legislators elected by gay-friendly voters should be more than happy to trade their votes on gun control to the NRA-friendly faction in exchange for the latter's votes on gay rights, and vice versa. Enforcing the bargain may depend on these issues having the potential to come up again in the same legislature (ie. within two years for the US House), or on the cooperation of the agenda setter (ie. the Speaker) or relevant committee chairmen. With the two issues combined in a package, the Marshall efficient package (gay rights and gun rights) might pass with 80% of the legislature behind it. More complicated deals involving more issues might receive higher levels of support.
These "more complicated deals" might involve side payments, perhaps taking the form of military spending supported by a general tax, but capable of being spent in any desired combination of Congressional districts. Alternately, a side payment could take the form of a tax break for heterosexuals who don't own guns. This may sound like a very strange idea, but if you are an American, and you take a close look at the instructions for the federal income tax Form 1040 next April 15th, I'll wager that you'll find plenty of other things that are equally strange.
Now consider a legislature elected by American-style single-member plurality (SMP) elections, also known as "first past the post" (FPTP). Two-way races are characteristic of SMP (see Political Parties, by Maurice Duverger). Suppose also that the gays and NRA members are uniformly dispersed throughout the country, so that there are no districts in which they form a local majority, and are thus unable to directly elect representatives of their liking. In this case, instead of explicit bargaining that takes place within the legislature, after a general election, now any bargaining must take place implicitly, before the general election, in the candidates' official and unofficial "platforms." For example, if Smith publicly opposes gay rights and publicly condemns the NRA, he can expect to lose, 80% to 20%, to Jones, who supports both. But now the fun begins. What of Johnson, who opposes gay rights but supports the NRA, and Miller, who supports gay rights and condemns the NRA? In a two-way race, if we assume that voters vote "sincerely," for whichever candidate proposes the package that is the most desirable, either Johnson or Miller can expect to defeat Jones, 60% to 40%; the gays are still assumed to have a slight preference for gun control, and the NRA members for persecuting gays. But for the same reasons, in a two-way race, Smith can expect to defeat either Johnson or Miller 60% to 40%, and we are playing "rock, paper, scissors" again.
Important though this "cycling" may be in theory, in practice it is rare. An easy way to remove cycling from a political model and make the overall model determinate is to use a probabilistic model of voter behavior. Instead of assuming that voters always vote, and always vote for the platform that is the tiniest little bit more attractive to them, assume that their behavior has a random component to it, and that the probability that a voter will vote for a particular candidate increases only gradually with the attractiveness of his platform. Depending on the details, in such a model, Jones' Marshall efficient platform might consistently prevail over that of any of his three potential rivals. It is easy enough to rationalize such a probabilistic model by invoking modest amounts of voter misinformation or insufficient incentives to vote at all (a problem to which we shall return). Another way to get rid of cycling is to observe that the fact that the outcome is indeterminate with "sincere" voting is already enough to put the voters' behavior into the realm of probability theory. (See Schelling, The Strategy of Conflict.) After a few cycles of gays and gun owners sometimes being persecuted and sometimes not, the gays and NRA members who voted respectively for Miller and Johnson are likely to rediscover the old saying that if something seems too good to be true, it probably is. By electing Jones, even when another candidate seems preferable in the short term, the gays and gun owners form a coalition that ends the cycling with a close second-best result for 80% of the voters, to the only slight annoyance of the remaining 20%. The outcome is once again Marshall efficient.
12. Parallelism between Laissez Faire and Democracy
This approach to understanding politics may seem odd, but it conveniently parallels the approach I used in discussing laissez faire. The claim that political action can be expected to produce desirable results is vulnerable to variations on the same six objections to the claim that laissez faire can be. We have (1, "judging criteria") the same controversy over whether utilitarianism is the most appropriate criterion for judging outcomes. Even if someone who proposes a wealth redistributing world government could convince me that my wealth would be used justly and wisely in the Third World, I might still not be willing to part with it. We have (2, "distribution of benefits") an analogous question over whether the political process produces an acceptable distribution of wealth, or benefits in general. How voting rights are defined will clearly affect this. Then comes the allocation part of the problem. We have (3, "knowledge problems") an analogous question over whether the people making decisions are sufficiently rational and well informed. Perfect information would require everyone to know what the "market price" of political support would be (how big a military base has to go in a typical representative's district in order to get his support for $X worth of pork barrel in someone else's district?). We have (4, "insecure boundaries") somewhat analogous concerns that people may be harmed in ways that are not under the control of the relevant authorities, or that some necessary transfers cannot be made efficiently. In the case of laissez faire, all scarce goods were assumed to be private property, with zero enforcement costs, and costlessly transferable. In the case of democracy, the analogous assumption is that all scarce goods are either private or public property, all property may be costlessly transferred, voting rights and public decision making rules have been defined, all laws are costlessly enforced (including private property rights), and the public (ie. the legislature or "the state") may reverse itself at any time on what will be public and what will be private property and how it will be distributed. Specifically, the "costless transfer" clause implies that side payments (bribes, subsidies, etc.) may be targeted at swing votes with no unintended spillover (more on this later). In the case of laissez faire, the "relevant authorities" were the courts and the specific people being harmed, who defended their interests using property rights. In the case of political action, the relevant authorities are the governing majority coalition of lawmakers, and people defending their interests using their voting rights, or providing various kinds of incentives to voters and their representatives. The main point of the analogy to property rights is that all harm takes place within limits determined by the legislature. People's boundaries are secure except where violations are allowed by the legislature. Under this assumption, all scarce goods are fundamentally public property, although private property may exist in a democratic context just as corporations with voting stockholders may exist in the context of private property; in principle, the corporation exists at the pleasure of the stockholders, and private property rights exist at the pleasure of the legislature. We have (5, "imperfect competition") an analogous objection that the way society or its power structures are organized may cause negotiations to go slowly or produce perverse results. The corresponding assumption is that there are no political monopolies, oligarchies, voting cartels, or other price-fixing arrangements that limit vote trading. Like goods in the idealized free market system, there are large numbers of independent buyers and sellers of votes; votes on any given bill become a commodity with a well-defined market price. There is thus little or no haggling, and there must be no permanent, disciplined coalitions such as British political parties unless their behavior is in turn effectively disciplined by voters (implicit bargaining) or they are individually very weak. Finally, we have (6) "free rider problems," or perhaps this should be "non-excludability of benefits." In the case of democracy, the corresponding assumption is that any benefits may be offset by targeted taxes, which can not be avoided or evaded, and which impose no "excess burden" (the taxes have no unintended effects on people's behavior). This assumption enables any coalition that benefits from passing a bill to raise revenue needed for side payments up to the full exchange value (dollar value) of the collective benefits.
Once again, we have the GIGO law. If we make the optimistic assumptions that none of the last four objections apply, we have assumed away all of the reasons why the outcome will not be Marshall efficient. We have (3) assumed away any internal restraints that prevent people from negotiating for what they want or from negotiating in order to avoid what they don't want. We have (4) assumed away any impediments to the negotiations other than the people involved in them, and eliminated harm that takes place outside of negotiated limits. The (5) no oligarchy and (6) excludability or taxability assumptions eliminate interpersonal obstructions to the negotiations. Again, these last two assumptions can be covered by a single "zero transaction costs" assumption. It then follows that negotiations will move quickly towards an equilibrium in which all of the legislation whose benefits have a greater exchange value than their costs (Marshall improvements) has been passed, and any coalition that tries to pass bad legislation (Marshall worsening) will be undermined by a more stable Marshall efficiency-promoting coalition.
From my standpoint, the chief difference between the laissez faire and the democracy efficiency arguments is that under democracy, there is no fundamental difference between "start" behavior and "stop" behavior. The free market is supposed to operate like a ratchet; negotiations enable people to turn the crank forward (progress as defined by Marshall), and the mechanism of property rights prevents the crank from turning the wrong way. A good ratchet design (good property rights definitions) is a compromise between allowing the crank to turn easily forwards (avoiding monopolies and public good problems) and not letting it turn backwards very much (avoiding public bads). Market failures occur because the ratchet is either too tight and sticky to move forward (ie. unable to provide national defense) or too loose to prevent backsliding (ie. unable to protect the environment). Under ideal private property institutions, a Marshall worsening transaction may be blocked by any individual whose property rights are involved, simply and passively by refusing his consent. Negotiations are only needed to move a transaction forward. A farmer who doesn't want to sell his land to developers and is tired of talking about it may safely take his phone off the hook (hopefully!). In contrast, under democracy, using simple majority rule, it may take as much negotiation to block a Marshall worsening as it does to promote a Marshall improvement. Democracy operates more like a tug-of-war in which there is a theoretical bias in favor of Marshall improvement, but no one is allowed to tie their end of the rope to a tree and take a rest. As the saying (attributed to Gideon J. Tucker) goes, "No man's life, liberty, or property are safe while the legislature is in session." The ACLU can never safely take their phones off the hook.
Although there is no fundamental difference between "start" and "stop" behavior under democracy, a good constitution can introduce biases that tend to preserve the status quo. This makes it more difficult to pass legislation in general, but Marshall worsening legislation is more strongly affected by this than Marshall improving legislation because proponents of the latter have more resources at their disposal with which to engage in logrolling: Marshall improvement means they can produce a bigger pie. This is like using chemotherapy against cancer: the medicine makes both healthy and cancerous cells sick, but the healthy cells are better able to cope. Constitutions may be formal (ie. the US) or informal (ie. Britain). Conservative biases sometimes take the form of requiring supermajorities (more that 51% of the legislature being needed to pass certain types of bills). Multicameral legislatures, presidential veto powers, and a Bill of Rights with judicial review also tend to preserve the status quo. Social conventions and moral restraints may be regarded as informal constitutional rules if they are widely respected. One approach toward understanding moderate libertarians is that they are offering a set of moral restraints that they would like to see play this informal constitutional role.
In the ideal case of zero transaction costs, Marshall efficiency can be achieved regardless of how voting rights are defined. The Coase theorem guarantees in this case that votes will be traded to the highest bidder. Votes are valuable, so voting rights definitions will always affect the distribution of wealth, but allocative efficiency is affected only because of departures from this ideal case. As with laissez faire, Marshall's definition of an improvement is more relevant than Pareto's in the real world.
13. Rational Ignorance
Obviously, one important reason why people make bad decisions (decisions that lead to Marshall worsenings as well as unintended perverse distributions of wealth) is because of bad information. In order to get a handle on when bad information is a problem, and what can reasonably be done about it, we need to relax the perfect information or "rationality" assumption. One way to do this is to retreat to a meta-assumption, that people may not start the game with perfect information, but they make rational decisions about when it is in their best interest to look for more information and where to get it. Consider my buying a car. I am not generally very well informed about cars, but I can go to the local library and read Consumer Reports. The benefit of making a good decision on a car that costs many thousands of dollars is worth a few tens of dollars of my time at the library. There are many others in my situation, so the publishers of consumers' magazines take advantage of economies of scale by doing research on cars and selling the information to mass markets. I still need to worry about whether or not to trust my sources of information, but the main cost to me of acquiring the information is the amount of time I have to spend trying to understand it. It might take only a few seconds for you to email me a huge file of automotive data, but it will still take me hours to digest it. In the case of buying a car, the information is worth a few hours of my time.
On the other hand, if I am trying to decide which brand of hair trimmer to buy, and neither of them costs more than, say, $20, it isn't worth my time to make a serious attempt to figure out which one is a better buy. I am better off just guessing, and if the one I buy breaks down soon enough to annoy me, I'll just buy the other one next time. The cost of a mistake is trivial, and I might guess right anyway. So when I buy a hair trimmer, I don't do any research; I rationally choose to be ignorant when I make my purchase.
Now consider a western state with a mixture of urban areas, where the definitions of water rights are relatively unimportant, and rural areas, where they are more important. In the previous examples, rational ignorance was not a major problem; the people who are interested in cars and hair trimmers are the only ones who need to worry about having enough information about cars and hair trimmers, and the market for consumer information works well enough, albeit imperfectly. But if water rights are being determined by initiative and referendum, or by legislators who are elected largely by urban people, whoever brought the issue up for consideration has given the urban voters a dilemma. The urban voters can either spend their time learning about water rights that are basically somebody else's problem, or they can vote ignorantly. The "dutiful" urban voters are bearing an external cost (negative externality) imposed by the initiator of the bill or initiative, and the other urban voters are imposing uninformed decisions on the rural farmers and ranchers. If everyone is "dutiful," and only 1% of the population is directly interested in water rights, the cost of having informed decision makers (voters) is arguably a hundred times higher than it would be if only the interested parties were involved. Is it worth it? If you were a member of an altruistic association of urban preachers with state-wide influence, should you urge your association to urge their congregations to read up on water rights so they can vote wisely? It may not be worth their time. Like an individual buying a hair trimmer, a society that uses elections to make decisions about water rights might be better off guessing more or less randomly than making a serious effort to become well informed. A high level of ignorance may very well be a utilitarian outcome, or in game theory jargon, it may be "group-rational."
Suppose the preachers decided to call on their congregations to become better informed about water rights. If you were a typical member of one of these congregations, what incentive do you have to act on their call? Like the potential contributor to a militia, you might have intangible reasons to get involved (perhaps to have something interesting to talk about over lunch, or perhaps you will learn something as a side effect of having lunch with a romantic interest who grew up on a ranch), but you have no obvious tangible incentives. If your motives are limited to tangible considerations, the rational thing for you to do is not to go out of your way to do anything about being ignorant.
A similar result holds true even if the issue under consideration is a public good that affects everyone more or less equally, such as funding for control of infectious diseases, or basic medical research. Let's consider a hypothetical example: Two scientists are competing for grant money in a nation with 100 million voters. One claims his $100 million XYZ research project (a $1 cost for the average voter) has a 50% chance of bringing the country $2 billion worth of benefits, uniformly distributed, or an average expected benefit of $10 per voter. His competitor claims that XYZ has a 1% chance of bringing $1 billion in benefits, making it worth ten cents to the average voter. I estimate it would take me $100 worth of my time to do a survey of people I trusted who knew something about XYZ that would tell me with 90% confidence who is right. Dennis Mueller (p. 350) estimates that, with 100 million voters, the odds of any one voter changing the outcome of an election are .00006. (I get .00004 using Stirling's formula and a binomial distribution, with the preferred candidate having a 50% chance of winning any tie. P=1/SQRT(2*Pi*n), where n is the greatest even integer less than or equal to the number of voters. I think the difference is that Mueller is deciding whom to vote for, given a committment to voting, where I am deciding whether to vote, given a preferred candidate.) David Friedman, using different assumptions, has estimated the odds of a US Presidential election coming down to one vote as .0000001. What should I do? If I am narrowly self-interested, the value to me of a correct outcome from a national referendum on XYZ is either ninety cents or $9, depending on who is right. The cost to me of getting good information is $100. But the probability that this information will correctly change the way I vote is only effectively 40% (I had a 50% chance of guessing right in the first place, and the information has a 10% chance of being wrong), and the probability that my vote will change the outcome of the referendum is .00006, so the overall value to me of this information is at most .0216 of one cent (.00216 of a cent if XYZ is a boondoggle). Once again, I bear 100% of the cost of my involvement, and I get only a trivially small fraction of the benefit, if there is any. The odds that my behavior will make a noticeable difference in the outcome (ie. the odds of an election coming down to one vote) are very small. I might become well informed as a side effect of reading Scientific American out of curiosity, or out of a professional interest, or because I ate lunch with someone who was interested in XYZ, but I have essentially no tangible incentive to inform myself about XYZ. This remains true even if we increase the potential value of the project by two orders of magnitude, high enough to get our hypothetical panel of preachers to take an interest in it. As Mancur Olson emphasized, information about a public good is itself a public good. We should thus expect that the marketplace for honest political information should have a very strong flavor of the Tragedy of the Commons.
One objection to this line of reasoning is the claim that there are organizations like the League of Women Voters (LWV), that are political equivalents to Consumer Reports or Underwriters' Laboratories (UL). There are other groups, such as the National Taxpayers' Union, the Sierra Club, and various special interest groups, that also provide information that certain groups of voters seem to appreciate having, and which at least appears to provide more direct advice on how to vote. Does this disprove the theory of a tragedy of the commons in political information? From the standpoint of the pure logic of choice, all we can say is that, if people do read the LWV Voters' Guides, their motives for doing so are not obvious. From a scientific standpoint, we would hypothesize that only the obvious motives are important, and then test the theory by looking at the quality of the information and at how widely it is used. In the case of groups like the LWV, my perception is that (1) not enough people take advantage of them, and (2) the information isn't sufficient, consisting mainly of politicians' evasive answers to very limited sets of questions. In the case of the more directly partisan groups, my view is that the advice these different groups give seems to be contradictory and inevitably appears to be either interested or ideologically biased. I also claim, based again on my own limited and unscientific observations, that most of the people who get information from these biased sources do not make any serious attempt to balance it by using a variety of very different sources. I dare say that my mother in law, who gets mail from the Republican National Committee, does not balance it with information from the Democratic, Libertarian, Green, or Reform parties. I conclude that voting based on this sort of information more closely resembles the process of buying a hair trimmer than the process of buying a car. On the demand side, voters appear to have little incentive to get adequate, accurate or balanced information, and this near-total lack of demand for good information seems to be matched by an equally near-total dearth of worthwhile supply.
Another objection to the claim of a tragedy of the commons in political information is that that's why we have legislatures, so we can delegate the job of being well informed to a small group of professional decision makers. In direct "Athenian" democracy, people chose specific policies directly and chose technical experts to execute those decisions, but in modern representative democracies, we choose representatives. Furthermore, we choose representatives who will be in office for several years, six in the case of US Senators. The issues Senators vote on are not predictable in detail six years in advance. Thus, we voters choose general policy positions, but not specific proposals. Legislators may have to become experts on various issues (if they can't trust hired experts), but the voters only need to be able to judge the legislators' trustworthiness, their loyalty to their constituents.
My reaction is that voting for a politician to represent my interests in a series of decisions is like trying to buy a "lot" of aircraft-quality high-strength bolts from a vendor I'm not sure I can trust. If I test all of the bolts to failure, I have nothing left to install in my airplane, and it defeats the point of buying them. If I don't test any, I have no reliable basis for believing that the airplane will be safe. So I should test a sample of the bolts. Depending on my philosophy, risk tolerance, etc., I may test a large sample or a small sample, but I will have to test a sample if I want to know if I can trust my vendor (and build a safe airplane).
Similarly, if 100 bills come up before the legislature, rather than (1) studying all of them in depth and second guessing my representative on all of them, or (2) assuming blindly that he is acting in my interest, I should pick a few bills, become reasonably expert on the issues involved in them, and evaluate him based on whether I liked his votes on those bills. My "test" is whether I approve of the way my representative voted on particular bills, after I have studied them in depth (not knowing how things will actually turn out in twenty years). This needs to be a random sample of the bills. If I only evaluate bills involving issues where I have low information costs, such as matters that relate to what I do for a living, I am using a biased sample, a sample of issues in which I have a special interest, and which thus rewards my representative for favoring "special interests." Similarly, I should look at low-profile issues as well as high-profile ones to see how he behaves when fewer people are paying attention. This sampling is further confounded by the "logrolling" process as discussed in Chapter 26; my Senator may have sold his vote on one bill in order to buy someone else's vote on another bill, in which case I need to know what he bought with his support. In order to be able to compare my Senator to another candidate, I also need to "test" the challenger. This requires me to get straight answers out of him, publicly (to make sure he doesn't give different answers to different interest groups), on how he claims he would have voted had he been in office, and then I have to guess about whether he is telling the truth.
This spot-checking strategy may reduce the overall cost of the political information we need in order to vote intelligently, but does it reduce it enough to matter much? From the standpoint of the pure logic of choice, the low probability of a major election coming down to one vote would seem to dwarf the tangible incentives to do research on most issues. Again, from this standpoint, all we can say is that, if voters do go to the trouble to be very well informed, their motives are not obvious tangible ones. From a scientific standpoint, we would again hypothesize that only the obvious motives are important, and then test the theory by looking at the participants' behavior. If large numbers of voters actually engage in meaningful spot-checking, and base their behavior at the polls on the results of their research, we would expect that the successful politicians would be ones who give straight answers, whose answers are consistent across different audiences, and whose elected behavior is consistent with their campaign promises. We would also expect to see some evidence of random sampling and in-depth study of issues by large numbers of voters, who would remember being lied to, and punish such behavior at the polls. On the other hand, if the rational ignorance paradigm is realistic, the marketplace for political information would be characterized by evasion, "issueless campaigns" focused on name recognition, and voters who are indifferent towards dishonesty and politicians' biases towards special interests.
Addendum (8-15-2007): The above argument confuses ignorance with gullibility. Bryan Caplan argues in The Myth of the Rational Voter that rational voters could and would make allowances for their ignorance, but that real voters do not. Also, unbiased random errors tend to cancel in the aggregate, but errors resulting from cognitive biases do not. The problem is lack of mental discipline rather than lack of information. He calls this "rational irrationality." See my book review (aka Appendix D): caplan.htm .
Addendum (11-5-2010): Or as Jamie Whyte put it,
Modern politics is just as you should expect it to be when votes are cast by ignorant people taking advantage of a low-cost source of emotional gratification.
Addendum (9-8-2009): I'll give the last word here to Megan McArdle:
People don't know what's in various bills, because bills are very complicated, so they just project whatever they think would be neat onto the ones authored by politicians they like....
14. Rent Seeking
Another important reason why people make perverse decisions is related to insecure boundaries. In my previous example of wild fish as a "fugitive resource," I discussed fishermen having perverse "use it or lose it" incentives; they have an incentive to overconsume scarce resources. But it's actually worse than that. Not only do they have an incentive to overconsume, but they also have an incentive to waste additional resources by competing with one another to overconsume. Ordinarily, when two buyers compete with one another for a limited supply of something, the high bidder's loss from having to pay a higher price is a gain to the seller, and the low bidder pays nothing. The extra money that changes hands is a "pecuniary externality," as discussed in Chapter 10, and only matters from a utilitarian standpoint if it is a problem in terms of distributive justice. If the competition is in the form of investing in something that the seller likes enough to trade for, again the loss to the buyer is balanced by at least as great a gain to the seller. But in the case of competing for a fugitive resource, there is no seller; rather than producing an advantage for some seller, the consumer's investment goes into taking resources away from some other consumer without compensating him. Mueller (p. 244) makes the general point that while economists tend to regard competition as a good thing, this is true only if it is price competition; otherwise competition is usually bad. From society's perspective, building technologically advanced fishing boats that are ever more effective at reducing a wild fish population that is already too small is effort doubly wasted, but from the perspective of an individual fisherman, the only choice he has the authority to make is between spending more to get less and finding another line of work.
Besides making socially harmful investments in the pursuit of fugitive resources, people also make harmful investments in pursuit of the status of being a monopolist. This might be a natural monopoly pursued under laissez faire, or a government-created monopoly pursued under democratic politics. The above-normal profits a monopolist receives are sometimes called "monopoly rents," and the socially harmful competition for them has come to be called "rent seeking." Mueller's chapter on rent seeking in Public Choice II seems focused on monopolies. Gordon Tullock defines rent seeking more broadly (p. 43 of Government Failure) as "the use of resources for the purpose of obtaining rents for people where the rents themselves come from some activity that has negative social value." David Friedman, in The Machinery of Freedom, avoids the term "rent seeking" entirely, and instead talks about "the economics of theft." I apologize if my use of "rent seeking" is technically incorrect, but for my purposes, I am content to use it as a metaphor for any investments people make in the process of playing negative-sum games. These negative-sum games will typically involve prizes such as monopoly status that are not legally considered "property," but as Friedman's "theft" analogy suggests, the logic is similar in situations where property rights are defined, but not easily enforced, or even in situations where property rights are enforceable, except that the owners have had the wool pulled over their eyes.
Friedman offers burglary as an example of a negative-sum game involving investments: the burglars and fences are spending valuable time transferring things from people who are willing to pay retail for them to people who aren't, and the victims and potential victims are investing in locks, police, and prisons. Another example is lawsuit abuse. Yet another example is "cost shifting" in the health care industry. Still another example is the elderly couple who get divorced in order for one of them to be eligible for Medicaid benefits even though the other is wealthy.
Rent seeking, as I use (or misuse) the term, can occur under private property institutions for many reasons. Some things, like fugitive resources, are not property. Some property rights are hard to enforce (crime). A privileged position as a monopolist (David Friedman, Price Theory, p. 475) is not property, and can be thought of as violating either the perfect competition assumption or the private property ("secure boundaries") assumption, and perhaps similar comments apply to certain types of "impure" public goods. Rent seeking also occurs because of knowledge problems. A business owner may not know that an employee is taking bribes or has a conflict of interest (the "principal-agent problem"). A more serious problem is lapses of rationality. The main point of the advertising industry seems to be to create and exploit many of these lapses. As Robert Costanza suggests, people may really want self esteem, but they can often be confused into buying sports cars instead. I find it hard to relate rent seeking specifically to public goods, but I think of rent seeking in general as a parasitic disease that infects any institutions that (1) can't defend their valuables, (2) are indecisive or poorly informed, or (3) have high decision making costs.
To the extent that private property exists within democratic political institutions, democracy is plagued by the same sources of rent seeking, but in addition we have some new sources. One is that, as with a conventional monopoly, an investment may have to be made in order to acquire or defend a privileged position as an oligarch. Some side payments may not be easily targeted to swing voters, or the size of the payments may not be easily controlled. In addition to exploiting individual irrationality and the conflicts of interest within the private sector, under democratic politics, rent seekers may also exploit any quirks there may be in the group decision making process. As the Condorcet Paradox illustrates, some of these quirks are unavoidable; group preferences may be non-transitive. These quirks include agenda manipulation, rational ignorance, and any flaw or asymmetry in the ease of negotiation that tends to make it easier for one faction to secure votes than another, such as asymmetric information. Generally high transaction costs tend to bring about results that resemble the initiative and referendum example, or "tyranny of the majority." Asymmetric information tends to enable "the exploitation of the [numerically] great by the [numerically] small." Concentrated interests also have the latter effect; a small group of people with an intense interest in something are easier to organize than a large (dispersed) group who have the same total amount of money riding on some issue, and it is likely to be necessary to get organized in order to collect the resources necessary for lobbying and in some cases, side payments. A large group of people who are already organized (ie. a corporation or a labor union) tend to act politically as a concentrated interest.
As I suggested earlier with the analogies to a ratchet and a game of tug-of-war, rent seeking is a bigger problem under voting institutions than under property institutions. As Dennis Mueller puts it (p. 243), "The entire federal budget can be viewed as a gigantic rent up for grabs by those who can exert the most political muscle." And on p. 241 he says, "The iron law of rent seeking is that wherever a rent is to be found, a rent seeker will be there trying to get it." However, the most famous examples of rent seeking involve things that don't show up on the budget, such as monopoly privileges. On p. 236, Mueller presents an argument by Peltzman that a vote-maximizing politician has an incentive to compromise between the interests of producers and consumers, so that regulation usually produces oligopoly-like results that are intermediate between perfect competition and pure monopoly. There may be political incentives to regulate a pure monopoly, but less incentive to regulate oligopolies. Regulating competitive businesses such as agriculture is likely to be more politically expedient than regulating oligopolies. It is important to remember that creating artificial monopolies is not merely a transfer of wealth, but results in a net loss. The Civil Aeronautics Board (Mueller, p. 238), which formerly regulated airline fares, prevented the airlines from engaging in price competition, but did not prevent them from competing on the basis of frills and schedules (often resulting in flying half-empty planes). Passengers paid higher prices than they would have otherwise, but instead of this money being a transfer to the airline stockholders, most of it was wasted in providing the passengers things that were worth something, but not as much as they cost. Mueller (p. 242) reports Krueger's study that rent seeking involving import-export licenses in India and Turkey wasted 7.3 and 15% of the respective nations' GNP. If rent seeking were merely bribery, the value of the resources spent on it would be a transfer, but efforts that go into things like obtaining training that isn't really necessary to get an impressive degree or certification (credentialism), lobbying, and advertising are waste.
The monopoly examples illustrate that a $1 loss to the consumer is not a $1 gain to the monopolist, but even when a transfer can be made efficiently (ie. transferring an import-export permit from one firm to another), some fraction of the value is likely to be dissipated in various forms of rent seeking, which could be anything from hiring a lobbyist to failing to pursue a job opportunity in order to maintain Medicaid eligibility. How large is the domain within which rent seeking takes place, and what fraction of the value of the resources being fought over is wasted in the competition? We are again in the realm of fuzzy logic. Everything is stealable to some degree, but some things can't be transferred efficiently, or are worth much more to one person than another, or are subject to constitutional or moral restraints. The waste due to rent seeking may be a highly nonlinear function of how efficiently a prize can be transferred from one faction to another. In some cases, the losses are likely to be comparable in size to the value of the goods being transferred, but different models of how political institutions work give different results.
How does lobbying work? Some lobbying costs, such as contributions to a politician's election campaign, look like bribery (initiated by lobbyists in order to get politicians to act favorably) or extortion (initiated by politicians threatening to act unfavorably). But unless the particular politician is spending lots of his own money on his campaign, a dollar in campaign contributions may not translate to a dollar of personal income. These contributions are clearly not the same thing as the "side payments" that would be made to a legislator's constituents, that would enable him to support a bill while honestly acting in his constituents' best interests. Other types of lobbying cost might be attempts to provide the legislators or voters with honest information, which might be viewed as a "transaction cost" similar to a salesman providing training on how to use a new machine he is trying to sell, providing that the goods or services being exchanged actually belong to the people whose representatives are being lobbied, and the representative is acting in good faith. But some lobbying looks very much like rent seeking, investments made in order to win negative-sum games.
For purposes of illustration, consider a number of simple models of how political lobbying works:
1. The faction that spends the most money gets the prize. Factions guess at how much other factions will spend, based perhaps on what they spent last year, and either give up or spend more, in a style somewhat similar to a sealed-bid auction. They will often guess wrong. This model leads to the entire value of the prize being wasted, on average.
2. The probability of a faction's winning is equal to the money spent by that faction divided by the total spent by all factions. This game could be played "sealed bid" style or by taking turns, "poker" style, with the players each knowing how much the others have already committed, and gradually increasing their own lobbying commitments. This model gives the same conclusion, roughly 100% waste.
3. In this poker-style nonlinear probabilistic model, the probability that a faction will win the prize is the square of its spending divided by the sum of squares for all factions. Lobbyists take turns making commitments. The optimal strategy is for the first player to spend 50% of the value of the prize and for the other players to fold. This results in 50% waste.
4. This model is similar to the second, except that the contributions, no matter how large, are only partially effective in influencing decisions. If lobbying only has a 10% chance of influencing the decision, firms only have incentives to waste an average of 10% of the value of the prize.
5. Lobbying expenditures have nonlinear, diminishing effects on influencing decisions. The first $1M gives you a 10% chance of winning, the second $1M an additional 5% chance, the third an additional 2%, etc. This might make sense if people get "saturated" with advertising, or if too much lobbying attracts the attention of the press, and legislators act differently when they are being watched more carefully.
6. The competition among lobbying factions may be monopolistic. For example, a trade protectionism issue may pit a consumer advocacy coalition against a trade association and labor union coalition. Banning certain Japanese cars might be worth $1 billion to the carmakers, but if there are only two bidders, if the highest spender wins, and if the consumer groups are too dispersed and unorganized to be able to spend more than $10 million, there may be little reason for the carmakers to spend much more than $10 million. Similar results will occur if the prize is worth much more to one faction than to any of its competitors, if the goods involved cannot be transferred efficiently, or if one faction has a decided advantage in the effectiveness of its lobbying (such as a bias in the legislature to preserve the status quo, or a larger supply of photogenic widows and orphans).
One response by the consumer advocates would be to use the government to help them organize themselves, analogous to the role the National Labor Relations Board (NLRB) plays in supporting labor unions. Legal barriers to using the government's money to lobby the government could be sidestepped by framing the consumer advocacy work as an "information campaign," as NASA does with its propaganda on the benefits of having a "space program." Setting aside any questions regarding the sincerity and effectiveness of the consumer advocates, a law that taxes consumers to pay for services that benefit consumers might find widespread support. Under the perfect information and zero transaction cost assumptions in the simple "town meeting" model of government, such a law might easily pass. But under perfect information and zero transaction cost assumptions, the lobbying we are concerned with would be completely ineffective, and the problem we are proposing to solve would not exist in the first place. With poor information and high transaction costs, we are back to the need for advocacy work that would not be funded appropriately, if ever, until after it was already successful.
Can lobbyists work on the basis of contingency fees, paid from the public treasury? I suppose this is possible, but it would be politically unpalatable, hard to camouflage, and would reopen questions about the advocates' sincerity. If you were a venture capitalist, would you loan money for such a scheme? I wouldn't unless I were sure the "consumer advocates" were not sincere. Such things are more likely if decisions are made through the courts than through the legislature (ie. state governments suing tobacco companies in federal court?). In any case, the use of the government's money to lobby the government will certainly promote rent seeking waste.
7. The lobbying factions may collude with one another and form a cartel. This could limit lobbying expenditures to the minimum necessary to influence the legislature. A successful lobbying cartel could become a de facto government, in the "Praetorian guard" or "power behind the throne" fashion. David Friedman has argued (The Machinery of Freedom, p. 155) that a rational and cohesive ruling class would be much more efficient and single-minded than the hodge-podge of conflicting rip-offs that seem to characterize American politics. Regarding the Civil Aeronautics Board, after explaining how he estimates the numbers, he states, "...we still have the curious spectacle of a ruling class that steals a billion dollars from itself and spends eight hundred million for the privilege." But a partially effective cartel could limit the waste associated with rent seeking considerably.
8. Results similar to the seventh model could be obtained by implicit bargaining. Lobbying might be likened to a poker game with very high betting limits, in which players can see how deep the other players' pockets are, and what many of their cards are, before any chips are committed to the game. If it is reasonably obvious who will win the game, the other players will quit the game before it begins. In this case, the lobbying budget of the most advantaged faction works by threat more than by actual expenditure. A similar model might be a tug-of-war in which the teams size each other up before the game, and the obviously weaker team walks away without either team expending much effort.
These models could be extended ad nauseum. The simple models (1 and 2) tend to result in 100% waste. This leads to David Friedman's Second Law: "The government can't give anything away." (At least not efficiently.) Combinations of the factors seen in models 4, 5, 6, and 8 could be very effective at limiting the waste. The results are sensitive to factors that can't be established convincingly without serious case-by-case research.
As I argued earlier, in comparing the ratchet and tug-of-war analogies, the problem of rent seeking occurs even under strict laissez faire capitalism, but tends to be much more severe under democratic government. It is not clear how the problem can be mitigated without radically curtailing either the legislature's discretionary power or such things as people's rights to free speech and petition for redress of grievances. As Dennis Mueller concludes (p. 246), "We are back to the need for fairly fundamental constitutional reforms to attack rent seeking seriously."
Texas politicians aren't crooks; it's just that they tend to have an overdeveloped sense of the extenuatin' circumstance. As they say around the Legislature, if you can't drink their whiskey, screw their women, take their money, and vote against 'em anyway, you don't belong in office.
--Molly Ivins, The Progressive,
15. Sensitivity Comparison
Having constructed comparable arguments for why either laissez faire capitalism or democratic politics might yield attractive results under ideal circumstances, we can now consider how sensitive these arguments are to real-world departures from the ideal. While I think David Friedman's anarcho-capitalism is plausible enough to be a suitable basis for a science fiction novel, I don't think any reasonable person could maintain that in today's world, military defense in unnecessary, or that governments are unnecessary to provide it. I also agree with Jeffrey Friedman's view of the environment as another libertarian "Waterloo." Certain kinds of governments have compelling advantages in providing certain public goods and in protecting against certain public bads. Other problems, such as the distribution of wealth, monopolies, and imperfect information (including irrationality), provide strong arguments that government intervention is, if not absolutely essential, then at least highly desirable, if it can be done in a suitably disciplined way. As Garrett Hardin put it, "We must admit that our legal system of private property plus inheritance is unjust--but we put up with it because we are not convinced, at the moment, that anyone has invented a better system. The alternative of the commons is too horrifying to contemplate. Injustice is preferable to total ruin." Or as John Stuart Mill put it (Utilitarianism, p. 59, Bobbs-Merrill Co., Indianapolis, Oskar Piest, ed.), "We should be glad to see just conduct enforced, and injustice repressed, even in the minutest details, if we were not, with reason, afraid of trusting the magistrate with so unlimited an amount of power over individuals." Rational decisions about political intervention require that we have some idea of how far our "magistrate" can be trusted.
The question of when we can realistically expect democratic politics to be an improvement over laissez faire is central to David Friedman's article, "Should Medicine be a Commodity?" Friedman uses Marshall efficiency as his criterion for judging improvements, and derives rough generalizations about the types and severity of market failures that need to exist in order for political solutions to be likely to be superior. I am trying to relate the behavior of governments more directly to utilitarianism, and do so in a more general way than would be appropriate to a discussion of health care. Instead of trying to show that political intervention is a bad idea in some particular business, I am trying to argue that intervention is risky, as a prelude to selling my version of libertarianism as a system for controlling those risks.
I begin again with the assumptions that utilitarianism is an acceptable criterion for judging the results of either system. Whether relaxing these assumptions weakens the laissez faire or the democracy argument more depends on who the relevant "powers that be" are and the context of their decision. Ordinary voters in a wealthy state who are voting on whether to ratify a new constitution, if they are part of a federation that includes poor states, may want to enshrine laissez faire principles in their constitution in order to preclude any forcible redistribution of wealth. On the other hand, beneficiaries of "corporate welfare" may like the way the legislature redistributes wealth just fine. For my purpose of appealing to ordinary voters in ordinary times, I am content that the objections that I will allow to utilitarianism will carry little enough weight that I may ignore them.
The second pair of assumptions that I need to relax is that either system, laissez faire or democracy, will produce an acceptable distribution of wealth, or benefits in general. In one respect, democratic political action should have a major advantage in being able to distribute wealth in a utilitarian way, relatively evenly: if the right to vote is distributed equally throughout society, increasing the importance of voting institutions would tend to equalize the distribution of wealth by making this right more valuable. On the other hand, many of the factors that hinder political institutions in making Marshall improvements also have distributional consequences. The poor who receive direct welfare benefits and the not-poor who are pleased to live in a more equitable society are dispersed interests who compete at a disadvantage against many concentrated interests. Rich people in general are a less dispersed interest than either the poor or the middle class. The rich and the middle class tend to be better informed than the poor. They certainly have more money and probably more leisure time to spend on lobbying and rent seeking. The cause of social justice also suffers a competitive disadvantage for other reasons. Its ambiguity and subjectivity lead to difficult negotiations among the people working towards it (high transaction costs). It is also information intensive, potentially involving large numbers of decisions about giving small amounts of aid to strangers, each in their own particular circumstances, some through no fault of their own, but some through behavior that altruistic people might not want to reward or "enable." It appears to me to be very difficult to reduce distributive justice to a formula, yet the voters may with reason be reluctant to give much discretionary power to an organization that has access to the public treasury. The proper administration of a welfare program is a public good (difficult to produce), and the improper administration is a private good (easy to produce). Some of the good done by any such programs will be negated by the "displacement" of voluntary charities, especially if the accomplishments of those programs are exaggerated in the public's perceptions. Finally, distributive justice is undermined by another form of "irrationality:" self-deception on the part of the middle class and wealthy people on whose political support the social justice activists depend. "Corporate welfare," for instance, often involves things like sports fans who can afford tickets to professional games telling themselves that they are doing the poor a favor by spending poor people's tax money on sports arenas. Social security is a pyramid scheme that depends for much of its support on elderly people, who are often relatively well off, thinking that they, rather than the poor, are worthy objects of charity. Medicare and Medicaid are a matched pair, one designed to transfer wealth to the elderly (who, statistically speaking, are not poor) and the other to make the whole arrangement sound like charity for the poor.
Again, the point of this is not that political efforts towards distributive justice are impossible or unnecessary, but that they are difficult and risky. Unfortunately, public choice theory does little to guide the judgment calls that are needed here, even in conjunction with other approaches. While I find statistics like the Gini measure of income inequality disturbing and somewhat interesting, they don't tell me what outcomes are acceptable and what processes will achieve them. (See Friedman, Price Theory, p. 402. A Gini coefficient of zero means wealth is uniformly distributed. A coefficient of one means one person owns everything.)
The third pair of assumptions to relax are that neither system suffers badly (from an allocation standpoint) due to knowledge problems. As was argued in the discussion of rational ignorance, considerations of information costs in general, and specifically the rational ignorance argument, suggest that democracy is much more sensitive to information problems than capitalism is, and that we should tend to limit democratic political intervention to matters of common knowledge. We should be particularly wary of situations involving asymmetric information. Knowledge problems provide one set of arguments for why political solutions may be more attractive in small homogeneous societies or political units than in large diverse ones. The quality of the news media, the "fourth estate," is an important consideration, as is the general level and quality of education. Although this may be outside the realm of public choice theory, one should also be concerned with how well the representatives know their constituents' wishes (more on this later). Finally, one should be concerned that whenever a government agency has a unique role, such as NASA in the manned space program, that agency will function in some ways as a monopoly, creating monopoly-like difficulties for anyone trying to find out what is possible at what price. Generally speaking, information problems tend to be arguments for laissez faire.
The fourth pair of assumptions to relax is that neither system suffers too badly (again, from an allocation standpoint) from problems involving insecure boundaries. I have discussed the implications of insecure boundaries for democratic politics mostly in terms of rent seeking. As I argued in the rent seeking discussion, democratic institutions tend to have more kinds of problems with rent seeking than market institutions, but it isn't clear to me how the seriousness of the problems associated with, for example, fugitive resources in the fishing industry compare with the corresponding problems associated with decisions that are made in legislatures and regulatory agencies. Obviously, if fugitive resources are the main argument for political interference in some market, the fugitive resource problems need to be fairly serious for there to be a reasonable expectation that political institutions will do better, but in many cases we are concerned with the more difficult decision of comparing the seriousness of a public good problem in the free market with the seriousness of a rent seeking problem in government.
Again, the imperfect competition and free rider assumptions may be lumped into a single "zero transaction cost" assumption, but I prefer to separate them, so the fifth pair of assumptions to relax is that neither laissez faire nor democracy suffers too badly from imperfect competition. Democracy suffers disproportionately from having crowded agendas--one can have competing private firms, but one wouldn't want multiple legislatures making decisions in the same jurisdiction. One often does have competing agencies, such as the US Coast Guard effectively competing against the Navy for Congressional funding to operate patrol boats on the rivers in Viet Nam during the war, or NASA and the Air Force competing to develop reusable launch vehicle technology. British socialized medicine allows private medicine to practice in parallel with it. The Soviet Union had competing "design bureaus" in its military industrial complex; Tupolev, Mikoyan-Gurevich, Sukhoi, etc. Theoretically, this kind of competition should force the agencies closer to producing Marshall efficient outcomes, but in practice, in the US such "duplication of effort" is more often condemned and restricted than promoted. McKenzie and Tullock (The New World of Economics) have gone so far as to suggest that the first thing one should do to promote more efficient government is to abolish all of the commissions that go around making recommendations on how to do so, because these commissions seem to invariably suggest that "duplication of effort" (read "competition") be eliminated (ie. the Grace commission). Certain key positions, such as President or Speaker of the House, are inherently monopolistic. Whether elections in general limit the number of competing "firms" (political parties) very severely depends on the election system used, a point which I will discuss later. The particular election system used most often in the US effectively limits the number of serious parties to two. The Single Member Plurality (SMP) system is almost synonymous with the two party system. As with insecure boundaries, if imperfect competition is the main argument for political interference in some market, the imperfect competition problem needs to be fairly serious for there to be a reasonable expectation that political institutions will do better, but we will often need to compare one combination of problems with laissez faire (including monopoly) against a different combination of problems with government (including crowded agendas).
Finally we come to the sixth pair of assumptions, that neither laissez faire nor democratic politics suffer too badly from "free rider" problems. At first glance, it is not obvious why legislatures should ever suffer seriously from public good problems, at least in the presence of perfect information. The costs of using the government to produce a public good are presumably borne by the taxpayers, and can in principle be distributed the same way as the benefits, given perfect information. Representatives get paid the same salaries whether they work for or against a particular bill, and are presumably rewarded or punished at the polls accordingly by voters, given perfect information. Riders can presumably be added to a bill that reward the bill's sponsors and many of its supporters (or their constituents) for whatever efforts were needed to get the bill passed. As long as the problem is purely one of non-excludability of benefits, and the number of representatives is small compared to the number of people interested in the public good, democratic government might be expected to have a major advantage over laissez faire.
There are a number of replies to this view. One reply is simply that it is hard to say anything intelligent about the way governments work without relaxing the perfect information assumption. A second reply is that there is synergy among the various classes of market failures we have been considering. As we saw in the discussion of "rational ignorance," the cost of information about some issue like overgrazing may be relatively small, but a public good problem may prevent the information from being disseminated and absorbed. As Mancur Olson emphasized, information about a public good is itself a public good. A third reply is that the "perfect information" assumption was meant to be limited to technical matters, such as how much would it cost to build cars that are slightly more fuel-efficient, and was never meant to suggest that people go into negotiations knowing in advance what the outcome is going to be. If everyone has the same amount of disposable wealth and would benefit the same amount from a public good, then there is an obvious likely outcome to the negotiations (a "Schelling point"), but otherwise, the "transaction costs" due to trouble reaching agreement on how to divide the benefits can be significant enough to prevent a Marshall-improving bill from passing. That's why the "zero transaction costs" assumptions were needed. Fourth, legislators may get "paid the same" whether or not they promote the public good, but if my time and political resources (and society's resources) are limited, any attempt on my part to promote the public good may conflict with my attempts to promote my private good.
Although I suggested that side payments, in the form of riders attached to a bill that arranges to produce a public good, could compensate the bill's promoters for any trouble or expense they might have borne in promoting it, there are problems with this claim, too. There is a justifiable stigma against using the government's money to lobby the government. Also, does a legislator bring home more pork barrel by working hard to pass a beneficial piece of legislation or by playing hard to get? The people who work hard to get a good bill passed can try to reward themselves by putting pork barrel riders in it, but they could also try to get riders put in without working hard (and perhaps by playing hard to get). Whether or not the riders pass depends on their bargaining strength and skills, which may have little to do with their hard work and public spiritedness, and may even be negatively correlated with those things. We still have the basic problem of the ranchers with the tragedy of the commons, that they have an incentive to be unreasonable. But it still appears that the problem isn't as bad because there are fewer people actively involved in the negotiations and because a substantial minority of unreasonable negotiators can simply be outvoted and ignored.
Consider the costs of immunization against infectious diseases. Immunizations are partly a private good, enormously decreasing the odds of the recipient getting the disease, but also partly a public good, because the rest of the population enjoys slightly lower odds due to there being fewer potential carriers. Under laissez faire, the public good associated with immunizations is likely to be produced to some extent as a side effect of the private good associated with them. If an immunization costs me $10, and it's worth $20 to me and another $15 to the rest of society, I will want to get one despite my not being able to capture all of the benefit, and I can make this decision for myself without having to hold an election. If some transaction cost raises the effective cost of the immunization by $2, my logic is not greatly affected: this simply brings my $10 cost up to $12. Under democracy, the calculation is different. If the bill to provide for immunizations passes, my share of the total costs and benefits may still be roughly $10 and $20, respectively, but the odds that my effort will make the difference between whether or not the bill passes are much smaller. If it now costs me $2 worth of time and stamps to try to get my Congressional Representative to vote for the immunization bill, and this has one chance in ten thousand of altering the outcome of the vote, my cost is $2 and my expected benefit is ($20-$10)/10000, or 1/10 of a cent. If the "transaction cost" of trying to influence my representative were lower, and the good in question were a more nearly pure public good, laissez faire may not compare so favorably. If one thinks of "fixed" transaction costs and "variable" operating costs, the public good problem in democracy comes from the "variable" costs being discounted by the odds that one person's efforts will affect the outcome, whereas the "fixed" costs are not. The champions of a good bill also have a disadvantage in that much of their work is a sunk cost by the time the logrolling takes place.
The central paradox here is that, while a public good (national defense) is the granddaddy of arguments for depending on government rather than markets, any attempt to use government to produce public goods is partially frustrated by the fact that good government is itself a public good. Government has the advantage that the "variable" costs of actually producing the good (ie. syringes needed for immunizations) may be paid for from the public treasury, but the "fixed" decision-making costs (ie. time and stamps) probably cannot be. Where public goods are concerned, government tends to be a "Jack of all trades, master of none," able to solve harder public good problems than laissez faire can solve, but not able to compete well at solving easy problems. This is particularly true where there are synergistic effects involving both information problems and public goods. In so far as lobbying can be explained in terms of public choice theory, it probably needs to be understood in terms of this synergy. One must also observe again that real-world decisions involve comparing inefficient markets that may be dominated by public good problems against inefficient political institutions that may be dominated by information problems, and in both cases there may be synergistic effects.
In so far as I take public choice theory seriously, I draw three general conclusions from it. One is that, while government intervention is essential for such things as national defense, we shouldn't expect much from it. It tends to excel at producing relatively pure public goods for large, homogeneous, well informed publics. (It isn't clear to me how the tradeoffs between democracy and laissez faire change with the size of the society.) Governments tend to cause trouble where there are high and asymmetric "transaction costs," conflicting interest groups with different levels of organization (concentrated vs. dispersed), some of whom have to do substantial amounts of lobbying or "educating" the public in order to obtain a sympathetic hearing (exert influence). I also conclude that consensus-oriented political institutions are more trustworthy than simple majoritarian ones: a law passed by a 2/3 majority is more likely to be a Marshall improvement and less likely to be an artifact of disorganization on the part of those injured by it than a law passed by a 51% majority, so to some extent a law can be judged by its "pedigree." Public choice theory predicts that legislatures tend to pass laws whose beneficiaries are well informed, obvious, or concentrated and whose victims are poorly informed, hidden, or dispersed. (The effects of how much leisure time people have are similar, but are more in the realm of political science than public choice theory.) Money is an obvious factor, capable of bribery, hiring lobbyists, and purchasing information and organization, even if it isn't entirely clear how lobbying should be modeled. The theory also predicts that the cost of information will tend to set thresholds: the benefits these laws bring should be large enough to be significant to beneficiaries, but small enough to not attract other people's attention unless the legislation is popular. Costs should be distributed widely enough that they are not significant to any one victim. High transaction costs, such that logrolling is effectively prohibited, tend to produce referendum-like results. Also, high profile symbolic issues like abortion are more in the realm of psychology than public choice theory.
A second conclusion is that the level of uncertainty in what we can expect governments to do is huge, and the stakes are very high. Sometimes a democratic government will protect its people from a Hitler, and sometimes, out of 250 million people, it will throw 2 million of them in prison for victimless crimes, such as using a recreational drug that is currently out of favor with the dominant religious authorities (usually in combination with having the wrong skin color).
My third conclusion is that we need some sort of "Safety, Reliability and Quality Assurance" (SRQA) system for controlling these risks. A handgun needs a safety. To mix metaphors, government is a loose cannon that needs to be tied down, but not jettisoned. I prefer to think in terms of using a strong poison like chlordane to stop termites from destroying a house: I believe it is often necessary, but we need to keep it away from the children, and we need to use carefully thought-out procedures in handling it. Every new application for chlordane should receive careful scrutiny. We mustn't be cavalier in our use of it.
16. Problems with Public Choice Theory
There are a number of possible objections to the arguments above. If democracy and laissez faire capitalism worked the way that public choice theory assumes that they do, I sometimes find it hard to imagine either of them working at all, let alone working as well as they do. In point of fact, the economic and public choice theory efficiency proofs presented above showed that our six idealized assumptions were sufficient to produce attractive outcomes, but they did not show that these assumptions were necessary. There seems to me to be a great deal of psychological and sociological "glue" that holds civilization together and prevents either our property or our voting institutions from unraveling as quickly as they otherwise would in the face of violations of these six rather hokie assumptions. One might plausible suppose that this "glue" works better in the case of democracy than capitalism (or vice versa). If this "glue" that works outside of the theoretical models is more important than anything in the models, the theory is in big trouble.
While the economics and public choice efficiency proofs are technically only sufficiency proofs, the "sensitivity analysis" depends for its relevance on a belief that assumptions 3 through 6 are also more or less necessary for either system to produce a Marshall efficient outcome. Its relevance also depends on believing that the degree of perversity of real world outcomes can be reasonably estimated from considering how seriously these assumptions are being violated. If the sociological "glue" is too important relative to the incentives assumed by economic or public choice theory, the "sensitivity analysis" is mostly irrelevant.
One may well ask whether the public choice models of how governments work are realistic enough to be useful. Even if this is the only body of theory we have that can be used to predict how a novel new institution will behave, if the state of the art in economics and public choice theory is as bad as the state of the art of medicine in the days of bloodletting, we might be better off just using our intuition, while maintaining a strong bias in favor of tradition.
In the case of economics, to me, the most obvious embarrassment is the advertising industry. Although some small portion of advertising can be rationalized in terms of mnemonic devices, in terms of disseminating actual information, or in terms of "commitment strategies" (the frontier banker won't skip town with your money because he's invested too much of his own money in his opulent building), the vast majority of it seems oriented to insulting people's intelligence, to putting them emotionally off balance so they can be manipulated, and to just generally flying in the face of the assumption that people are rational. But economic theory does pretty well in making sense of commodities markets.
The situation in public choice theory is quite a bit worse. Here the most conspicuous embarrassment seems to be "the paradox of voting," to which Dennis Mueller devoted an entire chapter of Public Choice II (chapter 18). B.F. Skinner is reported to have calculated that the odds of a major election coming down to one vote were comparable to the odds of being struck by lightning on the way to the voting booth. Others have compared the odds to those of being hit by a car on the way there. From a "scientific" standpoint, if I hypothesize that people are "selfish" and "rational," that is, they behave in such ways as to pursue obvious, tangible objectives, it is very hard to reconcile this hypothesis with the fact that large numbers of people actually vote at all. In US elections, there are frequent complaints that the policy positions of the major party candidates are nearly indistinguishable from one another. It's hard to argue that, for the average voter, a chance to decide between Tweedledum and Tweedledee is worth dying for. Furthermore, political scientists tell us that people with more education are more, rather than less, likely to vote, so we can't easily explain the phenomenon by suggesting that people are misinformed about the odds. It would appear that some relatively subtle, intangible motives are at work.
Addendum (8-15-2007): Bryan Caplan specifically condemns both the Self-Interested Voter Hypothesis (SIVH) and the assumption of voter rationality in The Myth of the Rational Voter. See my review (Appendix D): caplan.htm .
In some situations, it may not be important to know exactly what someone's motives are. For example, if I know that you recently retired, and am wondering whether this resulted in a Marshall improvement, it may not matter much to me whether you now spend your time fishing or playing chess. If I have enough information to have some assurance that you are mentally competent and that you captured the lion's share of the benefits and bore the lion's share of the costs of what you were doing both before and after your retirement, I don't need any further details to conclude tentatively that no great Marshall worsening has occurred. But in the case of voting, your life is strongly affected by the way other people vote. If you vote foolishly, you may only suffer one part in 250 million of the ill consequences, if there are any. If I want to draw any conclusions about whether democracy is Marshall efficient, I need to know the details.
Dennis Mueller's chapter on the paradox of voting reminds me of David Friedman's explanation of why someone who has just bought homeowner's insurance (paying to reduce risk) might turn around and buy lottery tickets (paying to increase risk, and at very bad odds). From an investment standpoint or from a risk aversion standpoint, this makes very little sense, but it does make sense as a consumer good, as entertainment. Basically, some people enjoy imagining themselves as rich, and a lottery ticket is a sort of visualization aid that makes the fantasy seem richer and more believable. It's a little like a brief game of "Dungeons and Dragons" (a fantasy role-playing game, or FRPG). Mueller suggests that voting is a similar exercise in fantasy: the odds of an election coming down to one vote being essentially nil, the voter has no real incentive to be ruthlessly honest and calculating, and may instead safely indulge his imagination (another example of "the rationality of irrationality").
For example, if I am a middle class person who is actually very selfish, but I prefer to think of myself as altruistic, I can vote in favor of a ballot proposition to transfer wealth from the upper and middle classes to the poor, and pat myself on the back for my generosity, secure in the knowledge that the odds are astronomical against my vote making any actual difference in the amount of taxes I pay. Self-deception or posturing may take many forms. I may pretend to be more altruistic than I really am, or perhaps more deserving: if I am wealthy and elderly, I may congratulate myself on how righteous and deserving I am, worthy of transfer payments from the young, whereas a clear-eyed reading of the moral teachings I subscribe to might instead indicate that alms should go to the poor. I am also likely to pretend to be more influential than I really am. When I talk to people about the lack of incentives voters have to vote at all, let alone to vote thoughtfully, a typical response is "What if everyone thought that way?" There seems to be a natural human desire to believe that other people will do what I do. And the XYZ Bureau that will be created by the referendum that I am now voting for will be staffed by people who think the same way I do (the Right Way (tm)!) and will want to do what I want them to do, even though I will not actually be following XYZ in the news. This last point also brings up the fact that I am likely to think that I'm smarter, wiser, and better informed than I really am, which also will influence what policy positions I support, and the number and diversity of public servants I consider myself as a voter competent to supervise. It is also possible that I will cherish a view of my society as being more cohesive than it really is, which may encourage me to vote to create institutions that won't work the way I expect them to.
The above explanations seem to belong in the realm of psychology. For me, peer pressure seems to be an important reason to vote, but why should any of my peers find it in their interest to exert such pressure on me? This reminds me of the discussion by Mancur Olson of the need for "selective incentives" in order for a large ("latent") group to produce public goods. Participation in the political process, even if it is largely symbolic, seems to confer social status. If this is true, we are in the realm of sociology rather than psychology, but either way we are outside the realm of economics. We may find some statistical correlations between voting and other behavior, but it seems unlikely that we will be writing theorems about social status or incorporating it into public choice models any time soon.
Mueller closes his chapter with a paradox, that public choice theory starts by assuming that people are selfish, but that it is quite likely that when people vote, they selfishly choose to pretend to be altruistic. So there is some support for viewing people as behaving more altruistically in the voting booth than they do in the marketplace. (Again, see Appendix D.)
A different, and more ominous view of political behavior is presented in the popular psychology book, Families and How to Survive Them, by psychologist Robin Skynner and Monty Python's John Cleese. These authors describe much of political behavior in terms of "paranoia," people's tendency to adopt a posture of trying to appear wholly good (along with their friends) and depicting their opponents as wholly bad. In this view, much of political activism bears a strong resemblance to soccer hooliganism. Although Skynner and Cleese say that some amount of "paranoid" thinking is normal and harmless, they focus mostly either on extreme forms of behavior by political activists on the fringes, or on the behavior of members of Parliament who engage in serious bargaining, and appear to shift in and out of "paranoid" thinking patterns as suits their purposes. A better discussion of the similarity between the behavior of ordinary voters and sports fandom is found in David Friedman's Hidden Order, in chapter 19, "Law and Sausage" (as in Bismarck's quote that it is "better not to see them being made"). The point here is that being a Democrat or a Republican is a lot like being a Dallas Cowboys or a Philadelphia Eagles fan. Asking why people vote is like asking why people go to football games. Besides seeing displays of some unusual skills, you perhaps get a feeling of camaraderie with the other fans and you get to have a lot of fun cheering for "your" side, and perhaps expressing mock hostility towards the "other" side. The odds that your behavior will affect the outcome of the game may be about the same as the odds that your vote will determine the outcome of an election. Again, public choice models seem to be a long way from being able to incorporate this level of psychological and sociological sophistication, and it is not clear what sort of biases this introduces into the comparison between markets and legislatures.
If irrationality in advertising is an embarrassment to economic theory, it is also an embarrassment in political ads, and brings us back to the question of the importance of money in political campaigns. Political scientists Nelson Polsby and Aaron Wildavsky (Presidential Elections 1988) have argued that, in presidential elections, once a candidate gets the nomination of a major political party, there is enough money available for advertising that voters are effectively saturated, and at that point, differences in political spending are not very important. As they write on p. 48, "The problem, then, is not to explain why money is crucial, but, on the contrary, to explain why it is not." They also argue that voters in a general election are very heavily influenced by party affiliation (ie. Cowboys fans are pretty consistent in cheering for the Cowboys rather than their opponents), which also tends to make "late money" relatively unimportant. Money plays a much more important role in winning party nominations, when no candidate has a unique claim to his party label. Public choice models based on analogies to competitive markets suffer doubly from the anti-competitive, "two party" nature of the plurality voting systems commonly used in the US: obviously, when there are only two organizations that are capable of fielding viable candidates in the general election, the market is disturbingly monopolistic, but the similar voting systems used internally by the major parties for selecting nominees also cause serious problems in those contests with self-fulfilling prophesies. Primary voters are only allowed to indicate support for one candidate on their ballots, so they are put in a position of either "throwing their votes away" or biasing their decisions towards one of the two candidates they think are most likely to win: either the "front runner" or his leading rival. "Early money," particularly early money that is collected in such a way as to make one eligible for matching federal campaign funds, is the principal means of establishing oneself as the front runner. The front runner becomes a "Schelling point," and his nomination a self-fulfilling prophecy. I commend the Reform Party for adopting the Instant Runoff Vote (IRV, aka the Alternative Vote or the "Australian ballot") instead of plurality voting for their presidential nominations. I discuss voting systems in more detail in Chapters 22 and 25, with links in Chapter 25 to some additional articles.
"Late money" is probably a great deal more important in local races, where name recognition is a problem. It is probably particularly important in "non-partisan" elections where party affiliation is not printed on the ballots. I once asked a Pasadena, Texas city councilman how important "single-issue" voters were, and his response was that the most important class of voters were the "zero-issue" voters, who vote based on nothing but name recognition. Consequently, political advertisements tend to be bland, vaguely patriotic looking, "issueless" ads, focused on improving name recognition while avoiding mentioning any position that might alienate anyone. Having the candidate's name in large letters is obviously important for name recognition. I suspect that some voters get a feeling of power or improved social status when the candidate they vote for wins, and might tend to vote for a candidate whose name they see a lot because they find that they feel better after the election when they do so. But whether we are discussing "early" or "late" campaign money, it would seem that advertising is more important in understanding politics than it is in understanding conventional markets.
We have already discussed the importance of "agenda setters" in legislatures. Analogies to conventional markets would lead us to expect that these positions of power would be abused, but here again, sociological considerations might make the problem more or less serious, and it is not immediately obvious which, and to what extent.
The number of parties in a legislature is typically small, in the quirky realm of game theory rather than the predictable realm of competitive markets, where commodities, or in this case, votes, have well-defined prices, and no one person's behavior has much effect on anyone else. This also tends to throw a wrench into the public choice machinery. Another factor is the legal unenforcability of political promises, and the relative lack of effective commitment strategies. Again, the analogy with conventional markets is strained. I find it disturbing that Mueller claims on p. 94 of Public Choice II that political scientists tend to regard logrolling as a bad thing. While Buchanan and Tullock, in The Calculus of Consent (free online), present logrolling as a good thing, they acknowledge that it could be a bad thing, depending on the nature of transaction costs in the political marketplace, as I have discussed in terms of concentrated interests having an advantage over dispersed ones. As logrolling is central to the public choice understanding of how democracy is supposed to work, the lack of confidence that logrolling is even usually a good thing is a bad sign. It certainly appears that politicians are chosen more for their psychological skills than for their platforms. (I defend logrolling as a good thing in Appendix A.)
If the behavior of typical voters is "irrational" from the perspective of public choice theory, the behavior of typical political activists is only slightly less so. In Political Parties, political scientist Maurice Duverger describes at length the conflict between activists and candidates with a "fight to win" attitude and those with the attitude of "fight to feel good about losing." Polsby and Wildavsky cover much the same territory in discussing the difference between those who are willing to make great compromises in order for their candidate to wield power and those who regard supporting their candidate as a means of self-expression, and have very little reason to compromise. Public choice models typically assume that the candidates and their backers are motivated by something relatively tangible, like money or power that is contingent on winning elections, but just as the odds of an election coming down to one vote are astronomically small, the odds that the actions of a single volunteer or minor contributor will push his candidate over the top are only slightly less remote. In so far as candidates are dependent on volunteers and voluntary contributions, these volunteers put constraints on the candidates' pre-election behavior that cannot reasonably be ignored. But all the volunteers can reasonably expect out of their efforts is a warm feeling or some other side effect of the efforts they put in, such as perhaps meeting interesting members of the opposite sex, because the odds of making a difference in the election itself are too low to justify serious efforts on that account. Activists will get a warm feeling from doing "the right thing," but some will perceive "the right thing" as helping a moderate candidate win, and some will perceive "the right thing" as setting an example of moral righteousness, win or lose. Judging by the religious right in the US, the latter activists are very important, but they would seem to defy the ability of public choice economists to incorporate them into simple, logical models. One may still suppose that the behavior of these activists is goal-oriented in some sense, but as with the paradox of voting, we are left with a dilemma: we don't have a plausible "scientific" public choice model of either voting or unpaid political activism because intangibles are too important, and we can't puzzle out any "philosophical" story about utility because there is too much coupling between one person's behavior and another person's happiness. A simple model would predict that hardly anyone would normally vote or do volunteer work for a political party. We know that the system works better than that, but we can't say how much better.
There is also typically significant conflict between political candidates and party leaders, both of whom have enough influence to often determine the outcome of an election, but who have different incentives. While there is a "theory of the firm" in economics that tries to understand analogous conflicts that are internal to individual businesses, again the motives of the people involved are less clear under democracy. Professional politicians presumably want to get reelected under normal circumstances, and are willing to make almost any necessary compromises in order to do so. Party leaders who are making decisions about someone else's career tend to be much more willing to risk alienating substantial numbers of voters in order to achieve their policy objectives. Where the candidates appeal in their arguments to the need to win elections, the party leaders' mantras tend to involve "principle, intransigence, purity," as Duverger puts it. In the US, the fact that important nominations are determined by "primary" elections tends to severely undercut the authority of party leaders and results in amorphous, undisciplined parties. The European parties Duverger mainly writes about are relatively "disciplined:" any member of the British Parliament who votes against the wishes of his party leaders risks that he will not be nominated in the next election, regardless of how popular he is with the voters. But as Duverger writes, all political parties are oligarchies. Again, this makes them hard to reconcile with competitive public choice models. Where a competitive firm appeals to its customers, the emphasis on "principles, intransigence, purity" suggests that the workings of political parties are much worse than public choice theory would predict, but we can't say how much worse.
As an aside, I am familiar with the "principle, intransigence, purity" flavor of internal party politics from my exposure to the Libertarian Party. I was a bit horrified in reading Duverger to discover that this pattern is relatively normal.
The fact that political activism tends to be irrational is reinforced by the fact that the organizers are largely self-selected, at least at the grass-roots level. Volunteer organizations in general tend to be dominated by the people who are most likely to show up at organizational meetings. Often there is a shortage of candidates for officer positions. The people most likely to come to these meetings are the ones who are the most excited about the organizations' goals and prospects. The people who are the most excited are in many cases the people with the most unrealistic expectations. In other words, volunteer organizations are often dominated by crazy people.
Most of what I have had to say above is related to the fact that good government, even to the level of good political parties, is a public good. Because of the prohibitive "transaction costs," the complexity of the negotiations that would be needed to get everyone to agree to do his fair share of the work, this public good tends to be produced only as a side effect of some rather inscrutable private goods. Let us now consider the effects of imperfect information in the organization of political parties, including any "rational ignorance" that may be a synergistic effect of imperfect information and public good problems.
The best example of this of which I am aware is the way US presidential nominations are made, as described by Polsby and Wildavsky. One may think of the distribution of information within a political party as having a pyramid shape, with a small number of very well-informed "insiders" at the top, a moderate number of serious activists in the middle, and a large number of loosely affiliated people at the bottom, who vote in that party's primary elections but have little or no other involvement. Similar comments apply to the distribution of information about any specific topic within society as a whole, with "special interests" at the top, the "attentive public" in the middle, and the "mass public" at the bottom.
If two Senators are seeking the Democratic Party nomination for President, and the decision between them is made largely by other Democratic Senators, there is the obvious problem that the decision makers are not representative of the common people. If the decision makers came from poor or middle class families, they may be "sprung from the people," as Duverger puts it, but as successful national level politicians, they are no longer "of the people." But at least they know the candidates, having worked with them for at least a few years, and are in a position to evaluate them based on personal observations. Voters in primary elections, by comparison, are almost completely dependent on campaign propaganda and the news media for their information. Polsby and Wildavsky are particularly disturbed by the practice of holding many primaries at the same time (ie. on "Super Tuesday") since it eliminates the possibility of using the primaries as "trial heats" in which candidates may be evaluated before the big race.
I would like to drive home the point that voters in general are more dependent on the news media than are people in conventional markets. As stated before, information about a public good (such as good government) is itself a public good, and people have perversely little incentive to acquire and absorb it. For the most part, then, the political information most people have they get as a side effect of what for them is a private good: entertainment. For all their pious attitudes, political columnists are basically in the same business as sports columnists, the entertainment industry, and they tend to cover both political "races" and sporting events with an emphasis on generating excitement and on reporting who's winning, rather than why anyone should care who wins. Again, when I'm buying a car, I have a much greater incentive to read Consumer Reports than I do to read the League of Women Voters' Guide when I'm voting, and consequently, I am less dependent for my information on the news media. The political system's dependency on the entertainment industry makes it hard to analyze.
There is also the matter that incentives to acquire specialized information are not uniformly distributed throughout the population. For example, in the film, "The best little whorehouse in Texas," under democratic politics, the whorehouse was closed down by a television journalist who appealed to voters who overwhelmingly were not patrons of that establishment. In Nevada, under the free market, a journalist who wanted to close a similar establishment would need to appeal to its customers, whose knowledge and opinions may conceivably be very different from those of typical voters, and are likely to be held more strongly because those people are more directly affected by it. This also makes the free market less sensitive to the news media.
Democratic political institutions also tend to be harder to analyze than conventional markets because they are more sensitive to altruism. Some forms of altruism consist of the altruist going out of his way in order to "do good." For example, I may feel morally obligated to give some fraction of my income to charity, regardless of whether my contribution makes a noticeable difference in how the world looks to me. But there are other forms of altruism that are comparatively goal-oriented. H. L. Mencken described this in terms of my not being able to enjoy my dinner when there is someone "cutting a sour face across the table." There is nothing in the free market that prevents people from giving money to charity, but typically, the "good" that is being done is a public good from the standpoint of the people who are paying for it. Consider a charity that provides food and shelter to the homeless, and suppose for the moment that the people carrying signs asking for money under the freeway overpasses in Houston are all genuinely homeless, and not professional panhandlers. One normally thinks of the "beneficiaries" of this charity as being the people who directly receive aid, but the people who are paying for it (or who are able to pay and choosing not to) presumably suffer some psychological discomfort every time they drive past a beggar. A charity that reduces the number of beggars on the streets is a public good for its patrons, even though it is a private good for its "excludable" direct beneficiaries. Again, markets don't prohibit people from giving money to charities, but they don't reward it much, either, and a business that is run with an eye towards helping the poor is likely to be at a disadvantage to its competitors that don't have that restriction. Consequently, the charity that businesses do engage in can usually be treated as an afterthought, something that the business' owners choose to pay for, but that has little effect on how the businesses are run. In contrast, as the Paradox of Voting illustrates, it is very hard to explain the basics of how democracy works without having some theories about altruism.
Because altruism is important, especially in political institutions, a good comparison of the the relative merits of laissez faire and democracy needs to consider what effect the respective institutions have on how altruistic people are. Richard McKenzie views altruism, people's willingness to act in a public-spirited way, as a scarce resource that is badly needed for certain social functions, and should not be wasted by asking people to use their altruistic urges to produce things in non-market ways that can just as easily be produced by people acting selfishly in markets. In this view, having people altruistically use the government do things that markets can do just as well is like giving scarce type O blood to a patient with type A+ blood, when there is a plentiful supply of type A+. Herman Daly, on the other hand, regards markets as something that consumes or diminishes this scarce resource, and that must therefore be carefully held in check. Frankly, my sympathies here run with McKenzie. My point here is that (1) it is important to know who is right, and that (2) this question is in the realm of psychology, not economics. I will offer my opinions on this issue in Chapter 18.
Mancur Olson used the term "selective incentives" in describing what I have called sociological "glue." These incentives reward people for producing what would otherwise be a public good, at least for some segment of the population (it may be a bad thing for everyone else). This can be tangible; a co-op may have a natural monopoly that enables it to finance a public good whose beneficiaries are predominantly members of the co-op. In other cases, selective incentives can be intangible things like goodwill or social status. (I elaborate on this slightly in my review (lewy.htm#criteria) of Guenter Lewy's book, Why America Needs Religion.) In either case, it is very difficult to predict whether "selective incentives" will be found in any particular case. It seems to me that "selective incentives" is almost a code word for economists that means, "For my next book, I will have to collaborate with a sociologist."
Finally, we must not forget that, like the rest of economics, public choice theory has very little to say about the distribution of wealth, which is an area in which government intervention appears to many to be very desirable.
Astronomers quail at the three-body problem and throw up their hands in surrender before the four-body problem. Any given moment in history is a problem of at least two billion bodies.
--"F. W. Taylor", from C. M. Kornbluth's novel, Syndic
17. A Closer Look at Intangibles
Although I hope the previous chapter was helpful in understanding the way democratic political institutions work, its main purpose to clarify some of the limitations of public choice theory. Having done this, I will now take some brief glances at politics from some other perspectives. I have no particular system here, but I consider these topics to be either important or interesting.
One of the topics discussed in Polsby's book on Congressional politics (New Perspectives on the House of Representatives, 4th ed. Nelson W. Polsby and Robert L. Peabody, ed., Johns Hopkins U. Press, Baltimore and London, 1992) is the problematical importance of constituents' mail. On one hand, Congress members are professionals, and appear to have strong incentives to find out what their constituents want. On the other hand, for reasons which are similar to the reasons for the rational ignorance problem, the average constituent has little incentive to keep his Representatives well informed, and there are numerous concentrated interests that have strong incentives to feed Congress members distorted information.
The Congressmen Polsby interviews say they take their constituents' mail seriously, but after reading Polsby it's hard to see how they can. Much of this is obviously "stimulated mail," where thousands of people have been asked to write letters by specific lobbying groups. I know of one political activist in particular who bragged of being "a one-man storm of protest." These letters are easy to identify because they are worded very similarly to one another and because almost inevitably, at least one of the writers includes a copy of the lobbyist's original request. Congressional staffers claim to be able to identify the particular lobbyist by the writing style of the stimulated letters. Some of the mail is from non-constituents, especially if the Congress member is head of or on an important committee. Some of the mail is ambivalent about whether a particular bill is good or bad. Also, some of the mail is unrelated to any particular bill at all, and discusses vague philosophical issues, things Congress has no authority over, past business of previous Congresses, or issues that might come up in the future but on which no legislation is currently pending.
Opinion polls are also problematical as sources of practical information for Congress members. From the standpoint of any particular Congress member, a particular poll is overwhelmingly likely to cover the wrong people (not a representative sample of his constituents), or have been taken at the wrong time, becoming obsolete before it becomes useful. Many polls that claim to be "scientific" are in fact deliberate attempts at disinformation. The ones that are sincere can often give meaningless or misleading results because of a slightly wrong wording, cover the wrong topic, or are not specific enough. But the main problem with public opinion polls is probably that there are not enough of them.
As we saw in the discussion of the paradox of voting, the behavior of voters and amateur political activists tends to resemble soccer hooliganism more than it resembles shopping for groceries, and activists are likely to "fight to feel good about losing" rather than to win. Besides sports, one can also look at this behavior in terms of theater or religion.
One of the relevant features of religion in this context is Sam Keen's observation that it is often perverted by being used to make "neurotic claims of moral superiority." This is essentially the same thing Cleese and Skynner describe in terms of paranoid thinking. Another religion-like feature of politics is that words such as "externalities" and "fascist," that were intended to convey some specific technical meaning, come to be used as "shibboleths" instead, words that are used to identify the people who use them as friend or foe, to divide people and put them in narrow little boxes. (The word "externality" seems to be particularly offensive to a leader of the local Green Party, judging by a recent email, but there was no discussion of what it meant.) Similar things happen with certain issues, such as minimum wage laws. Such an issue often becomes a focus of acrimonious partisan conflict, when the technical merits of the arguments on either side of the debate are quite murky. Major political figures also take on obvious ceremonial functions, some of which border on the religious. Dr. Hutchison, an instructor at the local C. G. Jung Center, observed that US presidents are voted out of office when the economy is bad, regardless of whether the current administration can reasonably be held responsible for this. This seems to be a continuation of an ancient practice in pre-Roman northern Europe of ritually executing kings following some disaster such as a crop failure. In some cases, the passage of legislation itself seems to take on a ceremonial function. In the wake an event such as the shootings of students at Columbine High School in Colorado, Congress hastily passes legislation that seems to be a substitute for a prayer vigil in a nation that is forbidden from having a state religion. Some Libertarian Party literature refers to "the Cult of the Omnipotent State." From a propaganda standpoint, I suspect that this sort of language is highly counterproductive, but I agree with the point it makes. Attributes such as omnipotence, which in earlier times would only have been ascribed to a deity, seem nowadays likely to be unconsciously ascribed to some rather unlikely political institutions and the mortals who in some hypothetical future will be running them.
We all deserve to be distrusted in politics. And it is the particular lot of the administrator to be more distrustable than anyone. In quite simple human terms, if a man cannot stand this, he should change his job, not create a mystique of technical indispensability.
--Bernard Crick discussing "scientism" in In Defence of Politics
The same paranoid thinking patterns that can be described in terms of sports fandom or the abuse of religion can also be described in terms of theater. Although the theater analogy generally seems to be more limited than the religion analogy, the theater analogy does a better job of reflecting the fact that political behavior is public behavior, whereas religion is often private. The current controversy over capital punishment seems to me to have a particularly theatrical flavor.
One may well ask how this unconscious and often neurotic behavior affects the legislation that is passed under democratic institutions. I don't have a coherent answer to this, but it's hard to imagine that the effects are positive. Certainly people also do irrational things for similar reasons under property rights institutions. The relevant difference from this perspective is that under property rights institutions, people usually have stronger material incentives that make such posturing more expensive. Greed gets in the way of paranoia.
Another topic that does not seem to be satisfactorily addressed by public choice theory is the role of campaign money. A partial explanation discussed earlier is that it is used to establish "front runner" status in primary elections and in general elections where minor parties may be a factor. A sizable early campaign purse is likely to deter many potential opponents from even entering the race. This is an important point if you are a "power broker," someone who advises wealthy contributors on the worthiness of various candidates, and who is interested in making sure that the voters' choices on election day are tightly constrained. A wealthy contributor or faction may contribute to primary candidates within both parties in hopes that both major party nominees will be on the same side of center.
But how is this money actually spent, and why? Chandler Davidson writes in Race and Class in Texas Politics that much of this money goes for traveling expenses. But from a public choice perspective, why are face-to-face meetings important? If one were simply transmitting data, a phone call should be enough. Why the baby-kissing and flesh-pressing? In the case of the "power brokers" that a candidate meets, part of the motivation for having the candidate travel is to enable the people who are evaluating him to control the meeting so that it functions as something closer to a lie-detector test than to watching a political advertisement. Does the candidate think the way "my kind of people" think, or does he just know what sort of promises we like to hear? Another reason to meet important people in person, behind closed doors, is secrecy. This is more consistent with the view of campaign contributions as bribes. The promises extracted in exchange for financial support may be explicit and illegal, or more likely, vague, implicit, and merely politically embarrassing if made public. Political promises made to different groups may also be mutually inconsistent (to the degree that they are not hopelessly vague), so speaking to different groups in person helps a candidate limit the degree to which one group will hear promises made to other groups with conflicting agendas. But perhaps the most important reason for meeting people in person is simply emotional. Most people who vote or engage in low-level political activism do so because they get a warm feeling from it, and that's basically all they get (as long as the odds of one activist tipping the outcome of the election are small). In economics terms, the "selective incentives" of which Mancur Olson writes in The Logic of Collective Action are largely emotional, not rational, and personal contact is important in building emotional connections, however one-sided these may be in some cases.
Besides travel, money is also used to spread various kinds of information and propaganda. Much of this is false or misleading information. This includes information (true or false) about how popular a bill is with the public, such as the "stimulated mail" described by Polsby. Some propaganda may be seen as deliberate "noise" or distraction from serious issues. Much of it is intended purely to boost name recognition, which helps a candidate for the "irrational" reasons discussed earlier.
Unfortunately, in many cases political propaganda tries to promote and take advantage of the "paranoid thinking patterns" described by Skynner and Cleese by resorting to demagoguery and hatemongering. Although this may include spreading unflattering or false information about various rivals or scapegoats, this sort of propaganda does not so much spread information as it offers people emotional support (encouragement) for indulging their paranoid tendencies. Blatant examples of this are provided by Naziism, the Ku Klux Klan, McCarthyism, and the Serb nationalism of Slobodan Milosevich. Although some of the War on [some] Drugs propaganda may be a sincere attempt to give emotional support to young people for avoiding illegal drugs, much of it is "pharmacological McCarthyism;" instead of addressing drug abuse as "spiritual bankruptcy" among "people like us," we are encouraged to view it as an outside threat from criminals from the ghetto or from Columbia. Our attention is focused on the supply of drugs to the virtual exclusion of any consideration of the demand. Even the propaganda that appears to be sincerely aimed at discouraging drug use does so mainly by inviting young people to view drug users as subhumans.
On a similar, but non-pathological note, propaganda is used to promote the cohesion of groups of people with dispersed interests. Mancur Olson writes in The Logic of Collective Action of the difficulty of getting members of large groups of people to act cohesively, and the need for "selective incentives" to reward individuals for producing "public goods" within those groups. These selective incentives can take the form of tangible goods in somewhat monopolistic markets: the American Medical Association became powerful partly because of its Journal, and State Farm got into the insurance business originally as a side effect of trying to organize farmers. However, these incentives also include what I described as "sociological glue," which includes emotional support, social status, and public recognition that is contingent on an individual's supporting the group's leaders in their political endeavors. Some political advertising can be seen in this light. It can be as simple as a "Union Proud" bumper sticker. The allegiances that result or are strengthened through political action may be rational or not from an economist's perspective. The rational ignorance phenomenon discussed earlier makes it unlikely that true information will have a very strong advantage over false information in being more readily propagated. The leaders of a political faction may or may not be sincere, and the members may or may not be well-informed, sophisticated, or enlightened.
While we're on the subject of campaign money, how is it collected? Some of it may be thought of as bribes, specific contributions from specific people with explicit or implicit strings attached for support of specific pieces legislation that benefit those specific contributors. This is how public choice theory tends to treat campaign contributions. There are also numerous small contributions from people who expect nothing in return but a warm feeling, the most intangible form of "selective incentives." But there are also large contributors who contribute to politicians in order to promote what the contributors regard as public goods, either genuine public goods for society as a whole or public goods for the upper class, such as (arguably) reductions in capital gains taxes. One explanation for this behavior might be that the wealthiest, who gain the most from these tax reductions, are a "privileged" group in Olson's sense of being small enough to be able to negotiate with one another effectively. Perhaps they spend much of their time together in their country clubs hammering out deals about what is a fair share of the burden of supporting agreed-upon political candidates. Perhaps facilitating these negotiations is part of the job of the "power brokers." The alternatives to Olson's small "privileged" groups are "latent" groups which need "selective incentives" because they are too large to be able to negotiate effectively, and "intermediate" groups, which are only occasionally likely to hold successful negotiations. Chandler Davidson, in Race and Class in Texas Politics, writes of "the ninety-nine," the wealthiest stratum of the upper class in Texas as of 1980. For Olson's purposes, 99 is a relatively large number, and Texas is just one of the 50 United States. Davidson makes a point of distinguishing between a "ruling class" and an "upper class," and like David Friedman, denies that the upper class is cohesive enough to be properly described as a "ruling" class. The upper class is too complex and divided to be regarded as a single unit for purposes of understanding American politics (to the frustration of many Marxists). But it does have some significant degree of cohesion, which tends to make it resistant to public choice models (based on "atomistic" and "rational" individuals), which seem unlikely to become sophisticated enough to include realistic treatments of "selective incentives" anytime soon.
The middle class and especially the poor are not as cohesive as the upper class. Partly this is because of numbers, larger groups being harder to organize and it being harder for one person to become personally acquainted with as large a proportion of the other members of his class. The poor in particular have less leisure time, less education, and are less well informed (or at least less influenced by the media and the press). Thus "strong," competitive political parties tend to be more important to the poor than to the rich (see Davidson, Race and Class in Texas Politics). Strong political parties transfer information independently of commercial media and mobilize (motivate as well as organize) volunteers who are not "mercenaries" hired by candidates with large campaign purses. Parties also motivate higher turnout of poor voters and help register them (reduce participation "costs"), altering the demographics of voter and activist participation as well as transmitting information (ie. which candidate to vote for).
Part of what I mean by a "strong" political party is one that is relatively insensitive to money. This requires a political party "information pyramid" with a relatively large and influential middle, consisting of long-term party activists. In the non-competitive (one-party) environment that was Texas in the 1950s, the decisions were made largely by the wealthy insiders at the top of the information pyramid. In the "Super Tuesday" primary elections that characterize modern presidential elections, Nelson Polsby and Aaron Wildavsky (Presidential Elections, 1988) lament that the candidates are selected by large numbers of people, at the bottom of the information pyramid, who know very little besides what they hear from the commercial media and paid political advertising managed by "mercenaries" paid by the candidates rather than party loyalists. (See also Linda Greenstein's article on the Supreme Court's consideration of "blanket primaries," NY Times, 4-25-2000.) Polsby and Wildavsky seem to want to balance oligarchy and ignorance; too much of either makes the nomination process too sensitive to wealth. They want to see more influence by long-term party activists, the middle of the pyramid, with "strong" parties. It is ironic that one remedy for oligarchy is strong political parties, which are themselves intrinsically somewhat oligarchic.
The difference between loyal party activists and "mercenaries" illustrates the point that good government (and the good political parties needed to help produce it) is usually a public good, whereas bad government is more often a private good. Public goods are difficult to produce and often require subtle "selective incentives" or "sociological glue," whereas private goods are relatively easy to produce, often with labor that can simply be hired. From an economics standpoint, political parties are important because they help provide these selective incentives, which are needed to promote the group's shared vision of the good. As David Friedman has suggested, amateur political activists may become well-informed because they enjoy winning arguments at cocktail parties, and may volunteer to help with partisan mailings in hopes of meeting like-minded members of the opposite sex, rather than because either of these things are good for society.
The question of how to make democratic politics less oligarchic raises the question of what other issues besides "corporate welfare" are distorted by the uneven way power is distributed in Congress and the various local and State legislatures. One apparent example of a disconnect between the average voter and the average legislature is abortion. The legislatures seem to be considerably more socially conservative than the median voter. The decision to make abortion legal was settled in the Supreme Court (Roe v. Wade), which despite being deliberately isolated from direct popular influence by the Constitution, decided the issue in a way that was more popular than what appeared to be a likely outcome in Congress or in most of the State legislatures. (See http://www.kff.org/content/2001/3131/PolicyandPolitics.pdf .) This example appears to fly in the face of my utilitarian argument that I expect the amount of political action that people engage in to be closely related to how strongly their personal utilities are affected by the outcome--the pregnant women are strongly affected, and the unborn "victims" are not themselves actual participants in the political process.
Another apparent example of the effects of uneven power distribution is the War on [some] Drugs, particularly the issue of medical marijuana. Jerry Epstein (http://www.dpft.org/mmjbooklet.pdf) reports, "More than 60 polls in over 30 states since 1996 show only minor variations from state to state with average support for medical marijuana use growing until it is now in the 75% range." The 1997 book, Marijuana Myths Marijuana Facts, on p. 163, cites earlier polls showing roughly 2/3 popular support (Lynn Zimmer, Ph.D., and John P. Morgan, M.D., The Lindesmith Center, New York, 1997, ISBN 0-9641568-4-9). Medical marijuana has even won popular referenda, but is resisted by most state legislatures, and Congress won't touch the issue with a 20 foot pole. One possible explanation of this is the argument I mentioned in introducing public choice theory, that referenda don't allow logrolling, and that the reason why medical marijuana supporters appear underrepresented in Congress is that they have traded their votes away for support on other issues that are more important to them. With our SMP elections, this would be implicit bargaining, which takes place in the candidates' platforms before the elections, more or less invisibly. This logrolling explanation suggests that the failure of drug policy reform efforts may actually be a sign that the system is working well, and that the referenda are a bad idea. While this theory is logically possible, I don't believe it--I think it's an excuse for politicians and activists to save face about a decision not to engage in a battle they don't think they can win. My impression is that social conservatives are simply overrepresented in these legislatures, partly for temporary "accidental" reasons and partly because of the irrational nature of political activism. Some of the "accidental" reasons involve the fact that power is not uniformly distributed in legislatures, and that many key positions are presently held by the most senior (and conservative) members of the dominant party in Congress, which is currently the Republicans. A second set of reasons involves less temporary facts. Drug users are convenient victims for people wanting to indulge their paranoid tendencies (Nazis had their Jews, we have our pot users). Also, there are powerful government agencies (ie. BATF) and officers, as well as the private interests in the prison-industrial complex, who owe their status and funding levels to pharmacological McCarthyism. I have already alluded to the fact that the officers in many volunteer organizations are largely self-selected, which magnifies some of the irrational tendencies in political parties. Like Roe v. Wade, I believe the War on [some] Drugs contradicts the argument that political pressure is exerted in proportion to the strength of the voters' interests.
Another possible explanation of the medical marijuana issue is that it is a "wedge" issue, that the people opposing it know that their position is absurd, but are vigorously defending it anyway because they are afraid of medical marijuana being "the thin end of the wedge." In other words, if they back off of an absurdly extreme position, they are afraid that they will expose the arguments for their more moderate claims to closer public scrutiny. The social conservatives may be convinced of their arguments, but may expect that if the voters were to actually think about the War on [Some] Drugs, a majority of them would not be. From this perspective, the purpose of opposing medical marijuana is to suppress serious public debate, which in my mind makes the death of Peter McWilliams and the time he spent in jail doubly tragic. While this explanation defends the rationality of the social conservatives in taking such an absurdly extreme position (throwing hundreds of people in jail), it instead accuses them of throwing millions of people in jail on the basis of arguments which they themselves recognize to be desperately weak.
The National Rifle Association (NRA) also seems to be nervous about "wedge" issues, but I am a little more forgiving of them because I see their behavior as an attempt to cling to a stable bargaining position (a Schelling point that is attractive from a parochial viewpoint) more than as an attempt to cover up a fraudulent utilitarian argument (purporting to be good for nearly everyone). This is a matter of degree.
A couple more of the intangibles that relate to how dangerous voting institutions are can be described as cultural. Does society punish people for engaging in self-deception by lowering their social status, or are large degrees of self-deception accepted as normal? Consider the wealthy elderly person in my previous example, whose religion supposedly teaches that alms should go from the rich to the poor, but who is morally opposed to alms going in any direction but from the young to the old. (I have a particular person in mind as I write this.) Do this person's peers pat him on the back for espousing this view, or scold him? If American society scolded people more for this sort of behavior (moral narcissism?), the prospects for using the government to promote distributive justice might be quite a bit brighter. As it stands, political debate in the US focuses much more on increases in Medicare, alms for the elderly, who vote in larger proportions than the poor. In comparison, Medicaid, alms for the poor, is neglected.
Similarly, cultures may be compared in terms of how tolerant they are of people who are unable to leave their "paranoid" thinking patterns behind them when they exit the soccer stadium. Weimar Germany didn't scold people enough for projecting their evils onto the Jews, and the history of slavery, Jim Crow, and racial profiling in the US is nothing to brag about, either.
What would be different if people were scolded more for indulging in "paranoid" thinking? Davidson's Race and Class in Texas Politics suggests that racism has provided compelling arguments for federal intervention in state politics. High levels of paranoid thinking might be an argument for a stronger federal government, but it is definitely an argument for weaker state and local governments, subject to more checks and balances. High levels of paranoid thinking make me far less inclined to trust public officials with autonomy and discretionary power, especially involving the police and privacy issues, but also to some extent involving complex questions of welfare administration. This also affects my degree of concern over government secrecy and over the use of high property taxes (which I discuss in Chapter 23).
A good religion should teach people to resist these natural paranoid tendencies (ie. love your enemy), and almost all of them seem to do this in theory, but in practice, many religious traditions seem to promote cohesion within their own groups by issuing unwritten standing invitations to their members to indulge their paranoid tendencies towards outsiders. I suspect that all religions are not equally vulnerable to this sort of misuse, but I find it hard to argue this point because the informal teachings (ethos) appear to me to be much more important in this regard than the formal ones.
The fact that intangibles are vitally important even in the running of a business that exists for the expressed purpose of producing very tangible profits is illustrated by the accounting practice of putting a dollar value on "goodwill." In a political context, people sometimes refer to an analogous "moral fabric of society," and I would like to discuss some of what this entails. It could be argued that legal precedents are relatively tangible, and represent a redefinition of property rights, as do "regulatory takings." However, since our legal system depends critically on the "goodwill" (beliefs and attitudes) of jurors, I'm am inclined to regard these legal precedents as intangible. I argued earlier that a legal system that works erratically has a similar effect as a high crime rate. Other forms of goodwill involve social conventions, social precedents and non-binding agreements, often implicit ones. Thomas C. Schelling (pp. 128-131, Micromotives and Macrobehavior, W. W. Norton and Company, Inc., New York, 1978, ISBN 0-393-09009-4) gives an example of people deciding whether or not to water their lawns after having been asked not to by the mayor during a drought. Someone who likes his lawn green may decide to give up watering, thinking there is an implicit "I won't if you won't" bargain, but change his mind after going outside and discovering that too many of his neighbors are "cheating" on it. This decision increases the number of people "cheating," which in turn makes it more likely that the implicit bargain will generally break down or "unravel." (Schelling, an economist, is very good at keeping one foot in the domain of psychology.)
This may be a good place to remind readers that, while the "scientific" side of economics has little to say about things that can't be measured, intangible psychic benefits and costs do motivate and matter to people, and the "pure logic of choice" side of economics considers them as important in principle as tangible goods and bads. As noted earlier, a man's decision to work shorter hours, so he can go fishing more, decreases the measurable Gross National Product, but presumably results in a Marshall improvement, which we can't measure, but which is central to the utilitarian case for libertarianism. Psychic costs include intangible "externalities" like my neighbors not liking my bumper stickers or my taste in house paint. Psychic costs and benefits also include "social justice." This is confusing, because "social justice" is largely "distributive justice," and I made a big point of emphasizing that (1) a Marshall-efficient outcome is defensible as a good solution to an "allocation" problem, and that (2) allocation and distribution were two different things. In terms of the pie-baking metaphor, the size of the pie in this case is a function of how the pie is cut. Once again we see that allocation and distribution cannot be separated entirely, and that describing complex real-world problems adequately in economic terms involves fuzzy logic.
Consider again a private charity that tries to get homeless people off the street (with some success), and a proposal for government to subsidize this (setting aside for the moment any fears of perverse side effects). As David Friedman put it, "It is possible in our society to argue for a government program to help the poor. But the argument is not that the poor, being part of the winning coalition, should benefit at the expense of others. The argument is that by helping the poor we can make everyone better off, that helping the poor is not merely a means to make the poor happier, but a means to reduce crime, make us all feel less guilty, make the cities livable, or whatever." ("Many, Few, One -- Social Harmony and the Shrunken Choice Set," American Economic Review, March 1980). Why might I, who am comfortably well-off, want to help the homeless, and why might I want the government to be involved? There are many possible answers to this, but three seem particularly important to me. One reason we have already considered is the Rawlsian "veil of ignorance," long-term uncertainty about my own or my family's future. (In the shorter term, I may want unemployment or underemployment insurance, and be unable to buy it because of "adverse selection," the tendency for buyers of insurance to be self-selected for being bad risks.) Another reason I might want to help the homeless is that it gives me a warm feeling to use my time or money directly to help them, perhaps for religious reasons, independently of whether this results in a noticeable decrease in their numbers, and independently of how anyone else's money might be involved. However, if this is my only motive, it doesn't explain why I would want the government to force my neighbors to also contribute. (It's possible that I might also enjoy using my time and money to lobby the government to force my neighbors to help the poor, without actually caring whether this makes a noticeable difference in the number of homeless I see, but that's another story.) A third and, I think, the most likely reason I might want to help the homeless, is that it pains me to see them or otherwise to know about their plight, regardless of how much effort I make on their behalf. To the extent that my motive (and that of 999 other well-off people who feel this way) is to get the homeless off the street, helping the homeless is a public good for this "public" of 1000 well-off people. (Presumably it also benefits the homeless, but their interest in getting the government involved is another story.) My $100 contribution might get a homeless person off the street for a month, and this may only be worth $1 per month to me, but there may be 999 other well-off people to whom it is also worth $1. I belabor this point partly to emphasize that distribution and allocation cannot be totally separated, and partly to emphasize the weakness of the stereotypical hard-line libertarian position against welfare programs. From the standpoint of most voters, welfare programs are attempts to provide public goods. It is possible to argue that government aid to the poor is so inept as to be counterproductive, but that's a different argument. A non-anarchist libertarian who claims that voluntary charities will provide an appropriate amount of aid to the poor needs to review the question of why we use governments rather than insurance companies to provide national defense (public goods, Chapter 6).
Recall also that one of the arguments I used to justify utilitarianism as a criterion for judging government actions was that this was a reflection of the strengths of different people's bargaining positions. So despite what David Friedman says, part of what "distributional justice" means in practice is that, openly or covertly, people will fight over other people's property, and sometimes a coalition of the less well off will win one of these fights. So while a plea for distributional justice may be primarily an attempt to get closer to the Pareto frontier, focused on intangible benefits to society as a whole, there is still an element of this that reflects an attempt to move from one point on the frontier to another.
Earlier I suggested that using the logic of economics as a framework for thinking about politics was like describing a rectangular city street layout in polar coordinates. I want to elaborate on this a little. I have a technical background in engineering. Engineering is full of precisely defined quantities, natural laws, crisp logic, and mathematical theorems. Economics is a little fuzzier, but there are still theorems. The logic is built on a solid foundation, like a high-rise building built on rock, even if there are questions about its suitability and relevance for my current purposes. There are no comparable theorems or natural laws in sociology or political science (or psychology, as far as I know). Their logic is more analogous to a ship floating in water.
Another analogy I use in comparing economics and psychology is that they are like two different kinds of river. Economics works best with large numbers of people, so that no one person's behavior is too important to the overall outcome, but it can only consider motives that are relatively simple and obvious. Economists excel at recognizing and avoiding "fallacies of composition and division," errors involving confusion between the properties of a group and the properties of the group's members (this confusion is why the "rock, paper, scissors" scenario in Chapter 11 is called Condorcet's paradox). I see economics largely as a model-building game, like much of engineering, where one starts off with very simple mathematical models, and complicates them only as needed to understand specific problems. On the other hand, depth psychology focuses on one person, or perhaps one person and a handful of his closest associates, but considers a person's motives in great depth. By analogy, economics is a mile wide and an inch deep, whereas depth psychology is a mile deep and an inch wide. Sociology and political science are somewhere in between. From the standpoint of scientific "parsimony" (not making unnecessary assumptions), the breadth-first approach of economics makes sense to me. We generally know how many members a society has better than we know their motives. The popular psychology I have read (ie. Sam Keen, in Hymns to an Unknown God, discussing the sale of land in Central America) has too often unfairly blamed groups of people for being foolish or having bad values when they have problems that can be explained simply by the fact that they are in large or "latent" groups (Mancur Olson's term).
Another difference between economics and the other social sciences I mentioned is that economics is much more focused on tangibles. When people like me try to stretch it to deal with intangibles, and (characteristically) lose track of the assumptions we make in the "abstraction" process, we get slammed by people like Herman Daly, himself an economist, for committing the "fallacy of misplaced concreteness." Conventional political science is more of a descriptive than a predictive, model-building science, and focuses, along with sociology, with trying to get a handle on intangibles, with tools like opinion polls. But the ideas political scientists work with still seem quite abstract to me, and there seem to be few things over which they have consensus. David Friedman suggests in Price Theory that public choice economists may be like the proverbial drunk who is looking for his car keys under a street light, where the light is good, rather than near his front door, where he dropped them. On one hand, I take this analogy more seriously than Friedman seems to, but on the other hand, no one else seems to have found the keys, either.
One thing economists and political scientists do seem to agree on is that it is a real problem that the voters generally don't know enough about what's going on. There is another little paradox here, too. Economics usually starts by assuming people are "rational" in the sense that they know where their interests lie to a large degree, but the main reason for studying economics is because we don't know where our interests lie, and we're trying to figure it out. Also, what little political science I have read seems quick to assume that leaders at least know where the interests of their supporters or their class lie, even if the average person doesn't. For example, it may be argued that without special efforts to inform them, voters either won't know whether a particular candidate supports increasing minimum wage laws, or won't know whether this is a good thing, but it is taken for granted that everyone who is at all informed "knows" that raising the minimum wage is a good thing. This is despite the fact that there is no consensus among economists on this point, and it is, after all, an economics question.
Another set of intangibles are social norms. Robert C. Ellickson (Order Without Law, Harvard University Press, Cambridge 1991) investigated the way ranchers and farmers behaved differently in Shasta Co., California, depending on whether the local ranch land was legally "open range" or "closed range," a matter of some local political importance. (David Friedman wrote a review of this book, "Less Law Than Meets the Eye," for the University of Michigan Law Review, which may be found on his web site, http://www.daviddfriedman.com/Academic/Less_Law/Less_Law.html) The land classification determines legal rules regarding fencing and liability for stray cattle. Friedman writes, "What Ellickson discovered about Shasta County was that interactions between neighbors, with regard to straying cattle and many other things, were controlled not by law but by a system of norms, a private law code having no connection to courts, legislatures, or any other agency of state power." The ranchers believed, or claimed to believe, incorrectly, that the land classification was crucial in determining liability for road accidents involving cattle (Friedman discusses possible reasons for posturing on the part of the ranchers at some length). He continues, "The one form of behavior affected by the legal distinction between closed and open range is one to which it is almost irrelevant." The point of my bringing this topic up is that, as Friedman puts it, "It is hard to see how legal rules can serve the roles assigned to them by doctrinal scholars, economists, feminists, or practically anyone else, if nobody pays any attention to them."
I have a personal anecdote that illustrates a similar point regarding voting systems. Ireland uses a Proportional Representation (PR) election system called Single Transferable Vote (STV), which involves ranking candidates in order of preference on the ballots (1, 2, 3...). An Irish citizen told me during a casual office conversation that he had been taught that the best voting strategy for this system was to rank one's favorites first, their most popular rivals last, and unpopular candidates in the middle. This is incorrect. There are systems such as the Borda Count, for which this strategy makes sense, but not STV. But if the mythology about the system determines people's behavior, mythology may be as important as fact. (The unwillingness of Americans to entertain suggestions to abolish the presidential electoral college provides another example where mythology trumps reality, making it impossible to carry on an intelligent conversation about one of the rarities of political science, a problem that could actually be fixed. Again, see my review of Electoral College Primer at Three-Fifths of a Floridian.)
It is clear that if we are dependent on public choice theory to predict how proposed changes to our political institutions will work, the theory leaves a lot to be desired. But we do know enough to say that democracy is dangerous and our understanding of it very imperfect. We also know enough not to hope for very much in terms of democracy producing dramatically more attractive results for most purposes than property institutions. I look at McCarthyism, politicized racism, the War on [some] Drugs, gay-bashing, and religious intolerance, and my reaction is that many of the stereotypical libertarians' worries are fairly well justified. Weimar Germany may not have been perfectly democratic, but it's hard to explain how Hitler came to power without acknowledging that a lot of Germans voted for him. If libertarianism offers us a means of managing the risks associated with living in a democracy, we have good reason to be interested in it.
18. The Malleability of Values
One of the risks associated with making poor choices between property and voting institutions is that people's values may change for the worse as a result. In describing the "perfect information" (rationality) assumption that goes into the arguments for both markets and voting institutions, I mentioned that people need to know what they want: what outcomes ultimately lead to their greatest happiness; their "values." In fact, people's values do often change over time and due to their experiences, sometimes in fairly predictable ways, which is arguably part of the purpose of military boot camp. As illustrated by Bernard Crick's In Defence of Politics, much of the argument given by socialists for their programs consists of claims that these programs or the resulting distribution of wealth will alter people's values for the better. Charles Murray's Losing Ground makes the opposite case, that the welfare state as we know it in the US is altering people's values for the worse. These arguments are distinctly different from the arguments I have presented so far, which have been arguments that the distribution of wealth under laissez faire or democratic politics may be inconsistent with people's current values. It's possible that a group of people may want to change their own values, and that this may require producing a public good of some sort, but normally one thinks of someone else's values as being the ones that need to be changed using coercion. For example, it may be rational, at least on some level, to decide to see a psychotherapist. Perhaps there is even a significant degree of public good involved, such as prior psychological research that makes therapy more effective, or the clients becoming better neighbors afterwards. On the other hand, the reason why paroled drug users are often required to go to 12-step groups is because their behavior is thought to be irrational.
As an example of how changes in the economics system might affect people's values, consider a medieval shoemaker who helps a neighbor, a competing shoemaker, learn the trade a little better. The student becomes a bit more proficient, and lives a bit better or perhaps has a bit more leisure. The master may notice a slight reduction in the price of shoes, but nothing dramatic. A small payoff in terms of goodwill may make teaching worthwhile. On the other hand, consider a situation where there is a potential or near monopoly, and teaching someone the tricks of the trade is likely to result in the teacher becoming unemployed. The example that comes to my mind is software salesmen working for competing companies in a specialized market with a small number of potential buyers. Here sharing information or even just the failure to compete fiercely means not just having to work slightly longer hours or a slightly lower standard of living, but possibly being laid off by a bankrupt firm. Here there is a strong economic incentive to compete more fiercely, because the participants are playing for larger stakes. If this situation is fairly common, this economic incentive may be reinforced by the psychological effect that the individual salesmen associate lack of competitive fierceness with sleeping under bridges. If the situation is very common, these effects are reinforced by sociological effects, that people who aren't fiercely competitive are seen as "losers" by others and ostracized. You may not want your daughter dating someone who's likely to end up sleeping under a bridge.
An analogous argument against current welfare laws is that, by making it marginally possible to raise a child on welfare, these programs interfere with the teaching of healthy values by obscuring the long term consequences of irresponsible behavior (having children one can't afford). Raising a child alone without a good source of income isn't as obviously stupid as it was in the 1950's, so there isn't as much stigma associated with it.
It would be easy to get lost in a longwinded digression on the differences between ultimate values (ie. self-esteem), instrumental values (ie. trade goods), psychological associations, social norms, and subtle incentives, but for my purposes, I am content to lump these all together as "values." (For a brief digression, I discuss several distinctly different purposes of moral education in my review (lewy.htm#criteria) of Guenter Lewy's Why America Needs Religion.)
The political or economic factors that might influence these values are also hard to enumerate. One factor is inequality of wealth or income. Poverty may drive people to take unpleasant, antisocial occupations (ones with substantial external costs) such as telephone solicitors, which wealthier people would disdain. Wealth disparities may isolate people (I can't afford to play horse polo with you or eat in the same restaurants). Wealth disparities may also bring people together in socially awkward, asymmetric ways, as employer and employee. Allowing greedy people to amass great wealth tends to make them more influential, presumably more likely to propagate their values, than if they were prevented from having much more than average wealth. (Similarly, if political power attracts power freaks, making political offices more important may make power freaks more likely to propagate their values.) Another factor is distributional justice, which is related, but far more ambiguous. If my neighbor is richer than I am, I may "envy" him or resent the distributional "injustice," perhaps depending on the degree to which I think he earned it. If I'm richer, I may indulge my paranoid tendencies and say that his poverty is God's punishment for his sinfulness, and I shouldn't let my children play with his. Fear or insecurity also seem likely to affect one's values, and it is probably necessary to distinguish between fear of random events (ie. stock market crashes and "acts of God") and fear of deliberate behavior by other people (ie. crime, lawsuits, and various kinds of persecution). I am particularly reminded of the earlier Mencken quote about having to be able to trust my neighbor before I can enjoy his company, which seems important if promoting Crick's value of "fraternity" is the point of this whole exercise. It also matters how high the stakes are, whether we are playing for physical survival, comfort, social status, or being able to pursue meaningful avocations without interference.
Since people in the real world may seriously violate the "rationality" assumption, I have to take these arguments about altering people's values seriously, at least in principle. However, I want to distinguish between different types of claims. Critics on the extreme left may make "strong" claims, such as that property institutions are always worse than voting institutions in promoting greed rather than altruism. This strong claim asserts that "capitalism" (the market process itself, the use of property institutions) is psychologically malignant, not just that the distributional outcome has gotten out of hand. This is in contrast to the view Robert Frost described, that "good fences make good neighbors" (increase fraternity, or at least protect it from strife). "Weaker" claims involve arguing that free markets discourage altruism (or even rationality) in some specific cases, but are not worse than the alternatives in other cases. A "very weak" claim would be that laissez faire might not cause bad social norms to develop, but it is inconsistent with the intrusive political action that may be needed in some cases to change these norms. This argument is the basis for civil rights legislation in the US that prohibits racial discrimination in many circumstances, such as buying a house or being served in a bar or restaurant.
I have several reactions to the "strong" claim. One reaction is that, while is it fairly clear to me that the alternative to good fences is bad fences or no fences, it isn't as clear to me to what "capitalism" is being compared. Are we comparing the US to the old USSR, dominated by public property, or to the highland culture of Papua New Guinea described in Jared Diamond's Guns, Germs, and Steel, away from the government's influence, dominated by gathering of wild (unowned) food, where any encounter between members of different clans is likely to be violent? Are we comparing a social system that works in a large industrialized nation with one that only works in small tight-knit groups (50 people or less), incapable of supporting modern medical care? The US itself is a mixed economy--are we comparing it to Sweden (a different mix), to a Hobbesian state of nature, or to a utopian fantasy in which airliners grow on trees and people do not have the same genetic predispositions?
A second reaction is that I regret not being familiar enough with the history of private charity in the US or elsewhere to be able to say much about the "bad old days" before the modern welfare state. Were people generally meaner to one another back then, and if so, to what causes can this be attributed? My impression is that physical mobility (the automobile, etc.) has made it easier for people to escape the consequences of antisocial behavior, and that high population density in cities is an important psychological factor. I suspect that these effects, and the effects of religion and political ideology, are more psychologically important in the evolution of culture than the distribution of wealth, but that is only my impression. The "voluntary associations" de Toqueville wrote about in Democracy in America make the bad old days not sound all that bad, but again, that is only my impression.
A third reaction is that we know enough about how people treat each other in the absence of enforceable property rights to reject the strong claim against capitalism. The claims that rumrunners made to monopoly privileges in various areas during alcohol Prohibition were obviously not enforceable in courts of law. The result was such things as the St. Valentine's Day massacre. Such events were correctly attributed to the nature of Prohibition, and so Prohibition was repealed. Nowadays, we have the exact same phenomenon, but these sorts of shootings are described as "community service homicides," blamed on the physiological effects of various drugs or the race or ethnicity of the people involved. In short, good fences may not be sufficient to make good neighbors, but we know enough to say that bad fences make bad neighbors. Modern liquor store owners do not routinely shoot one another.
A fourth reaction is that we also know too much about childrearing. Why is my 5 year old (at the time of this writing) so possessive of his toys? Is this because of his exposure to the free market? Have my wife and I unconsciously made our expressions of love for him contingent on him not letting his sister play with his toys? Is it because he is insecure, and thinks he might go hungry without his toy dinosaurs? Do parents in Sweden have an easier time teaching their children to share? Do publicly financed day-care centers work differently than private ones and nurseries in churches?
My conclusion is that many human vices (and virtues) are archetypal, including greed and status consciousness. Thomas Sowell described this as the "constrained" view of human nature in A Conflict of Visions. Property rights are a tool for mitigating these archetypal conflicts (limiting them and redirecting them in relatively benign ways), and the (often supervised) one-on-one negotiations between children over toys are part of how children learn to get along. A second possibility or "tool," besides property rights, is for the toys to be unowned: the big kid beats up the little kid and takes the toy whenever he wants. Another possibility is tyranny: an adult separates the children and toys and decides who gets to play with what. A fourth possibility is public property with democracy: make the children compete for the favor of other children who will then vote on who gets the toy. This fourth possibility strikes me as vicious and insecurity-provoking; it's better than the second, especially in small groups, but far worse than the first.
It is fashionable in some circles to describe human property as an arbitrary social convention, but this is dead wrong. Anybody who has ever owned a dog who barked when strangers came near its owner's property has experienced the essential continuity between animal territoriality and human property. Our domesticated cousins of the wolf know, instinctively, that property is no mere social convention or game, but a critically important evolved mechanism for the avoidance of violence. (This makes them smarter than a good many human political theorists.)
--Eric S. Raymond, Homesteading the Noosphere
This is surely oversimplifying, but I tend to think about these sorts of psychological issues (archetypal vices) largely in terms of social status. Money is partly important in influencing social status (not to sneer at the value of high quality health care, though!). How much of self-esteem is composed of introjected ideas about social status? How much of having a sense of "community" is a matter of being in the in-group? How much of a soccer hooligan's "paranoia" is pretending to have high social status by virtue of being part of an artificial in-group?
Suppressing the competition for social status is not easy. Architects have designed "egalitarian" office buildings, with no corner offices, only to see new hierarchies of status built up based on distance to the elevators (Robert Frisbee, How to Peel a Sour Grape). Sam Keen talks of Pacific islanders whose social status is determined by their success at breeding pigs with curly tails (audio tape, "Living the Questions," from Sounds True). I am reminded of Bernard Crick's criticism of the "student politics" that proposes to ban the bomb in order to deal with the threat of nuclear war: it's like trying to eliminate capital punishment by banning hemp and steel. People are going to form cliques no matter how wealth is distributed.
How far can we go in finding causal relationships between institutionalized sets of rules and the values they teach? Can we say that high school sports are good or bad? Do they teach winning at all costs, or do they teach teamwork and good sportsmanship? My reaction is that surely it depends on the coach. Likewise, I know people who have had good and bad experiences with Catholicism. It depends at least partly on the local priest.
Religion is to me a more plausible tool for engineering other people's values in a positive way than coerced wealth redistribution, and it has a bad track record. If the essence of the teachings of Jesus and Siddhartha was not to set store by one's social status, their success as teachers was very limited. Likewise, it has been said that there are values associated with science, such as honesty (your career is ruined if you get caught falsifying data) and not prejudging ideas by their sources, but this seems to be of little use in discouraging people from setting store by their social status or position in a management hierarchy.
So, while I can't dismiss moderate, "weaker" claims that certain aspects of our current economic system have adverse effects on people's values or social norms, I do dismiss the "strong" claims that voting institutions are intrinsically better than property institutions in this regard. In contrast, I am much more concerned that people will come to hate each other for being actively forced to spend too much of their lives playing negative-sum "rent-seeking" games with lawyers and lobbyists, usually "won" by the rich.
A number of factors might drive me to take more interest in the "weaker" arguments for political intervention in markets in order to change values and social norms. One factor is the effectiveness of our "social safety net." Another factor is upward and downward economic mobility: I am less concerned as long as the children of boat people can enter the middle class, and former governors of Texas can go bankrupt (ie. Bill Clements).
Whether the socialists can improve on laissez faire depends on the side effects of their policies. If they aren't careful, (1) they will increase rather than decrease injustice (inequality and injustice are not necessarily the same thing), (2) they will pay too high a price for it in terms of freedom (making it dangerous to be unpopular), or (3) they may actually promote strife rather than reduce it. Am I more likely to hate my neighbor because he has more money than I do or because he's been forcing me to hire lawyers?
If wealth transfers are to be used in an attempt to influence social norms, I expect that the process by which these transfers are made is more important than the amount. If the money that supports programs to help the poor is collected through simple, relatively efficient taxes, it may generate relatively little hostility, but if the transfers are complicated, unpredictable, and labor intensive on the part of the payers, the hostility will be greater. I find the complexity of my federal income taxes infuriating, more so than the amount of money involved (not that much of this money finds its way into the hands of the poor). I also associate much of the erratic behavior of juries with attempts to use civil law as a distributional tool. I expect hostility will be greater if large amounts of the money are wasted in "rent seeking." I also expect that hostility will be greater if large amounts of the money that is taken in the name of the poor end up in other people's hands, although this is arguable because much of the money that goes to the poor is associated to some degree with transfers to the not-poor (ie. Medicaid vs. Medicare). Outright fraud is largely a symbolic issue, but if we're serious about looking at the effects of the distribution of wealth from a psychological standpoint, the symbolic issues may be more important than the "real" issues.
My perspective on goodwill as a function of the distribution of wealth is focused largely on avoiding rent-seeking. Does our political-economic process put us in an environment in which conflicts are limited and resolved, and people interact for mutual gain, or in which conflicts are relatively unbounded and open-ended, and in which any interaction with a stranger is likely to produce at least one loser? If attempts to improve distributional justice are poorly thought out, they are likely to end up, as Bill McNichols put it, "promoting strife in the name of building community." I am inclined to support a publicly funded social safety net not because of the effects it is likely to have on people's values, but in spite of these effects.
However, the "very weak" claim, that political action can strongly affect social norms, and that laissez faire is inconsistent with this sort of political action, appears to be very sound. One aspect of the civil rights movement in the US was that a national majority opposed to white supremacist social norms was able to overturn those social norms where they were supported by local Southern majorities. (In terms of my six objections to laissez faire, the original problem could be described in terms either of irrationality or of lack of competition.) While subtle racism remains troublesome in the US, overt racism is very much less common than it was in the 50s, and progress in this regard has followed the predictions of sociologists nicely (see Pettigrew, "Normative Theory in Intergroup Relations: Explaining Both Harmony and Conflict," PSYCHOLOGY AND DEVELOPING SOCIETIES, vol. 3 (1), 1991, 3-16).
Thomas F. Pettigrew examines four aspects of the process of bringing about social change from the top down ("Advancing Racial Justice: Past Lessons for Future Use," Ch. 9, pp. 165-178, Knopke, Norrell and Rogers (eds), Opening Doors: Perspectives on Race Relations in Contemporary America. U of Alabama Press, 1991):
(1) The perception of inevitability of institutional change is critical.
(2) Behavioral change typically leads to attitude change more than the other way around.
(3) Intergroup contact under optimal conditions can reduce intergroup conflict and improve intergroup attitudes.
(4) Even in the absence of intergroup interracial contact, the disconfirmation of intergroup fears can reduce both prejudice and discrimination.
It is possible to regard race problems in the United States as presenting the federal government with an exceptional opportunity for bringing about top-down social change--exceptional in the sense of being a "man bites dog" story. One could argue that this success story is analogous to the US men's Olympic ice hockey team beating the USSR for a gold medal. It happened once, but it was never a routine occurrence. However, it did happen. Since deliberate federal interference with Southern white supremacist social norms was largely successful, it is not possible for a reasonable person to dismiss arguments for similar political interventions (ie. against violent homophobia or the remaining, more subtle forms of racism) out of hand. These arguments must be considered in detail.
Update, 3-29-2014: Dierdre McCloskey argues in The Bourgeois Virtues that the socialists have gotten it completely backwards, and that the virtues that make modern civilization possible are the virtues of merchants. Here's a sample: http://www.deirdremccloskey.com/docs/pdf/Article_249.pdf .
Part 3: Libertarianism as a Coping Strategy
19. Coping Strategies
If I have done nothing else is the preceding parts of this essay, I hope I have convinced any reasonable skeptics that the voters in a large representative democracy will often have a hard time figuring out which of the functions of society are best left in the private sector and which should be actively controlled by government. It is seldom in the narrowly defined self-interest of most voters to devote the time necessary to studying a political issue in enough depth to really understand it in detail. In many cases, even if the majority of voters did choose to spend their time this way, the cost of a correct decision in terms of the voters' time may be greater than the overall cost to society of a wrong decision.
One approach to this problem is simply to accept that the voters will make a lot of wrong decisions, and not worry about it. If these errors were purely random, they would often cancel each other out, but unfortunately, there seem to be consistent biases in favor of concentrated interests and positions that lend themselves to "paranoid" thinking patterns.
A second approach, often favored by political scientists, is for voters to look to political parties and other trusted leaders for information on how to vote. As I argued earlier, this still leaves me wondering how to figure out whom to trust, both for their sincerity and for their technical competence, and I suggested spot checks with randomly selected issues. The spot check approach reduces the burden on the average voter, but does not eliminate it. Another problem with political parties is that the number of viable, serious ones is severely limited by the election system. In the case of US federal elections, our Single Member Plurality (SMP), "first past the post" system effectively limits the number of serious political parties to two. Whether the suggested sources of information be parties, politicians who have to work with parties, or journalists, all of them have conflicts of interest: it is in their professional interest to be popular (at least within a chosen niche), rather than to be right. Furthermore, any source of political information that achieves a significant following among the voters will quickly acquire a following of lobbyists and conflicts of interest. The political scientists hope that typical party activists are a good compromise, semi-professional in their awareness of the issues, but amateur in their lack of conflicts of interest. We are still left with voters and amateur party activists, with little or no tangible incentives, being asked to sift through vast amounts of biased information about alternative policies.
A third approach, useful in conjunction with the second, is to look for "rules of thumb," what a computer programmer might call "heuristics." Heuristics are used when we don't fully understand the elements or relationships in a problem, or when the problem is so large and complicated that we can't put all the pieces together accurately. (See James P. Ignizio and Tom M. Cavalier, Linear Programming, p. 459, Prentice Hall, 1994, ISBN 0-13-183757-5.) Again we must be concerned with the possibility of consistently making errors in the same direction, and with the need to estimate the size of our errors, but as long as the errors due to our heuristic are in a different direction than our other biases, a heuristic is likely to be illuminating and helpful, even if it is not the final word.
I will now consider some possible heuristics, and how to use them (ie. how to cope with not having the final word). Note that I am discussing "public" rather than "private" morality, arguing that it is in the best interest of society for voters to have good heuristics, not that it is in the narrowly defined self-interest of a typical voter to look at this essay as more than entertainment.
Mill, Kohlberg, and me
John Stuart Mill discussed a similar topic in Utilitarianism, although in different language. He described utilitarianism as his "first principle of morality," and justice as a second principle. Confusion results, according to Mill, because this second principle takes on a life of its own and is confused with a first principle, rather than being subordinate to it. While principles of justice are derivable from utilitarianism, he says (p. 34), "...the supposed corollaries seem to have a more binding force than the original theorem...." But in every case where "justice" and utility conflict, Mill claims that utility will prove to be the higher form of morality.
Lawrence Kohlberg took a similar view in defining his six stages of moral development. (See John P. Dworetzky, pp. 544-5, Psychology, 2nd ed., West Publishing, St. Paul, MN, 1982, ISBN 0-314-85231-X.) In Kohlberg's view, as in Mill's, one does not abandon "morality" when one finds that one's conception of it has perverse consequences. Rather one takes increasingly sophisticated views of morality (that seem to me to be increasingly similar to utilitarianism). As Mill put it, regarding "justice," (p. 78), "By this useful accommodation of language, the character of indefeasibility attributed to justice is kept up, and we are saved from the necessity of maintaining that there can be laudable injustice."
I differ from these views in several respects. For one thing, I am considering moral theories as heuristics, as proposed components of a broader system for making decisions. Second, "morality" seems to me an inherently emotionally loaded word, and so I also tend not to think of a principle as being "moral" until, like Mill's "justice," it has taken on "a life of its own." Third, my commitment to utilitarianism is very limited. My first principle of decision making is enlightened self-interest, and I can only equate this with utilitarianism under special circumstances. Utilitarianism is interesting to me as a bargaining position or as a consequence of John Rawls' "veil of ignorance," not because I expect all sufficiently reflective people to have totally compatible values. For me, utilitarianism is already a secondary principle, and I am inclined to reserve the word "morality" for tertiary principles which, more than utilitarianism, have emotionally laden lives of their own.
In one particular respect, I turn Mill and Kohlberg on their heads: for me, morality is a heuristic, and the whole point of using heuristics is that they are simple, not that they are consistently right. As one computer "fortune cookie" put it, "Heuristics are bug ridden by definition. If they didn't have bugs, then they'd be algorithms." Consequently, when I find a contradiction between morality and utility, I do not redefine morality to eliminate the conflict. Instead, my view is that if you have a moral code that does not occasionally give you wrong answers about what you should do, you had better abandon it and find a simpler one, because the one you have is too complicated to be of any practical use. Instead of worrying about whether my moral reasoning is sufficiently sophisticated, I worry about when to abandon a simple line of reasoning in favor of a more complicated system that goes by another name. This is analogous to flying an airplane that has a primitive autopilot: the autopilot reduces the pilot's workload, but the pilot needs to know when to turn the autopilot on and off, and he still needs to know how to take off and land. Morality is like this autopilot. To me, a moral person is not someone who clings rigidly to a particular moral theory, but a person who pays attention to whether or not a particular theory is appropriate. A moral person breaks his own rules only when he has a convincing "story" about why it is necessary. Perhaps this "morality as storytelling" can be seen as an extension of Mill or Kohlberg rather than a departure from them, but for my purpose of criticizing strict anarchist libertarians, I want to emphasize that I am openly breaking the rules rather than engaging in double-talk about what the rules really are or should be.
Justice vs. Liberty
Mill focused on "justice" as a proxy for utility. Jeffrey Friedman has argued in his various editorials in Critical Review that libertarianism and liberalism both had utilitarian roots. According to Friedman's argument, both groups recognized that direct appeals to utility were prohibitively hard to judge, and both groups saw the need for a proxy, or what I am calling a heuristic. The two groups differ on whether "justice" or "liberty" (no initiation of coercion) is a better proxy for utility. Friedman goes on to argue that neither proxy is adequate, and that both groups need to go back to their utilitarian roots. I was told by Bill Nelson, a philosophy professor, that this is not historically accurate, but Friedman's story strikes me as one of those that "if it isn't true, it ought to be." The two groups seem to me to behave just as Mill described, with these second principles having lives of their own to such an extent as to make it nearly impossible for the two groups to talk to one another. In deciding whether a particular use of force is appropriate, liberals seem to ask, "Is the cause just?" where libertarians would ask, "Is it defensive?" (ie. who threw the first punch?). While I lean towards the latter approach for reasons which I hope to make clear, I agree with Jeffrey Friedman that neither question by itself is adequate, and utility makes better sense as a first principle.
What else might "morality" mean?
One of the problems with presenting morality as a heuristic is that the word "morality" can mean so many different things in practice, most of them with problematical emotional content. I mention some of these meanings here so that I can distance myself from some of them and use the others to illuminate my position better.
Some forms of "morality" appear to be for the consumption of others. For instance, I might try to use "morality" to manipulate the gullible into acting in my interest rather than theirs. (Garrett Hardin had some choice words about this in his "Tragedy of the Commons" article mentioned earlier.) Stigmas against intermarriage between different social classes prevent some members of the upper class from diluting the advantages enjoyed by their peers. Another purpose for morality might be to educate people to act out of enlightened rather than unenlightened self-interest, which is likely to be good for both student and teacher. Another purpose might be to change people's values, or at least to change social norms. This is hard for me to separate from "enlightened" self-interest. Again it is likely to be in everyone's best interest.
Other forms of "morality" are clearly for the consumption of their promoters. This could just be an ego trip. Sam Keen warns in Hymns to an Unknown God that religion is frequently misused as a tool for making "neurotic claims of moral superiority." Rigid adherence to a moral code could also be a way for someone who has trouble coping with uncertainty or ambiguity to avoid getting stressed out.
Many of the moral teachings that seem to benefit their students look to me like heuristics, simple approximations to enlightened self-interest which are correct most of the time. For example, teachings about honesty and keeping one's word can be understood in terms of reputation and commitment strategies that make one more attractive to others. Renunciation of alcohol could be a response to a problem mentioned earlier, avoiding the unstable intra-personal bargaining position of "I'll just have one more drink." Acting in good faith, not cheating on an agreement with others (ie. a peace treaty), helps keep one out of unnecessary future conflicts. Stigmas against sexual promiscuity help protect people from venereal diseases.
On the other hand, saying that "casual sex is an empty experience" is largely matter of taste. This doesn't mean that casual sex is a good idea, but it does remind us that a rule's desirability is subjective. Some "moral intuitions" are simply aesthetic.
Another reason for "moral" rules is to enforce implicit bargains, as in Thomas Schelling's example of voluntary water rationing. The idea is that I won't water my lawn if enough other people don't. There is nothing but peer pressure and conscience to enforce this deal.
Although I brought up the topic of alcohol in order to illustrate an unstable bargaining position, alcohol also serves as an example of another problem, that of desensitization, or progressively impaired judgment. This is often referred to as the "frog boil" problem. The story has it that a frog placed suddenly in a pot of hot water will jump out, but if placed in cool water that is slowly heated, will not. (I haven't tried this.) From a utilitarian standpoint, it makes sense that there should be a penalty associated with any proposal that causes people to make subsequent decisions in a state in which their judgment is likely to be impaired. The moral "heuristic" is simpler: don't go there.
One final "rule of thumb" that deserves mention is to simply use "common sense." This at least has the attractive feature of avoiding much of the emotional baggage associated with "morality." Unfortunately, what common sense means in practice to most people is to weigh the obvious consequences of a decision, and ignore or discount the secondary or more subtle consequences. If it turns out that the "secondary" consequences are equal and opposite to the obvious beneficial "primary" effects, and there are other losses to boot, then we have Frederic Bastiat's "broken window fallacy." This fallacy was the subject of Henry Hazlitt's book, Economics in One Lesson, and seems to me to be at the heart of at least 90% of all the political economics journalism I see in the mainstream television and print media. Bastiat's example was a bakery window broken by a vandal. The "obvious" effect is an infinite stream of new economic activity, starting with the glassmaker, who will now have more money to spend elsewhere, etc. The secondary effect is an equal and opposite stream of aborted activity, starting wherever it was that the baker would have spent his money if he had not been forced to buy a new window. In Bastiat's example, the baker had intended to buy a new suit from the tailor, but generally, this is unknowable. These two streams of activity cancel, and the net effect of the broken window is that the baker is poorer by one suit of clothes.
A recent example of a broken window fallacy in Houston was the successful effort by the promoters of "The Arena" to get the taxpayers of Houston to buy them an expensive building. The sports enthusiasts apparently believed they were getting cheaper ticket prices, and the cost to the taxpayers was hidden by a shell game in which they were promised that any new taxes would be attributed not to the arena, but rather to some unidentified other government service whose budget would have to be raided in order to pay for the arena. (Tax money is "fungible," capable of being reallocated easily.)
Another "common sense" way to evaluate proposed political programs is to test them. Take small steps. If the programs seem to be helping, proceed with the next step. If the programs seem to be causing too many problems, abolish them. There are two problems with this. One problem is that the creation of a government program creates as a side effect an entrenched class of people who depend on it, whether as beneficiaries or as employees. The existence of this new special interest group makes it much harder to repeal a bad law than it is to block its passage in the first place, before people become invested in it. (An existing group of people is easier to organize and motivate than a hypothetical group.) These government employees are also precisely the people on whom the voters are most likely to depend for information on how well the programs are working.
The other problem is that if the negative side effects associated with a particular program are at all subtle or have multiple causes, it may not be obvious to the casual observer how to attribute any new problems that develop. Although I am not a fan of Ayn Rand, I thought she illustrated this point very well in the novel, Atlas Shrugged. Each negative side effect of self-interested political interference in the marketplace was used by the powers that be as an argument and justification for yet greater and more perverse acts of interference. I am reminded of the lemming's comment in James Thurber's "Interview with a Lemming," "...you cut down elm trees to put up institutions for people driven insane by the cutting down of elm trees...." Consider the labor market, which may not work well because it isn't competitive enough in terms of employers. It may be too difficult for workers to change jobs or there may not be enough different firms hiring people in any particular field. However, if government intervention on behalf of the workers (ie. making it hard for them to be fired) has side effects that make it even harder to change jobs (ie. employers think twice before hiring anyone they can't fire easily, especially if there's a lot of paperwork and potential legal problems involved), this can easily make the problem worse by making the labor market even more "sticky" (high transaction costs) than it was to begin with. But the average person is unlikely to know to what extent the difficulty of changing jobs is natural, and to what extent it is due to laws that supposedly protect the worker. This attribution problem is exacerbated by the fact that, given many small steps taken over many years, it's hard for people to have a clear sense of what the baseline for comparison should be. It's hard for inattentive people to do controlled long-term social experiments in a complex and rapidly changing world. The informal "try it and see" approach purports to have an effective feedback mechanism that it does not in practice have.
In short, I argue that "common sense" does not give us an option of avoiding the use of heuristics. Common sense is a heuristic. The choice is between good heuristics and bad ones.
20. Flavors of Libertarianism
My emphasis on utilitarianism in an essay on libertarianism may seem strange. Libertarians often seem to regard an opportunity to make a utilitarian argument as a lucky coincidence rather than a central part of their thinking. But lacking any sort of of a "pope," libertarians can define themselves in many different ways, for many different reasons. I would like to explore some of these possible reasons now. My definition of libertarianism is a very weak one, "a certain reluctance to use force offensively, even in a good cause." In practice, this comes pretty close to the saying, "That government governs best that governs least." Why might this be an attractive doctrine?
To start with, I will acknowledge some obviously bad reasons:
One view of libertarians, offered to me by a Green, is that they are a bunch of rich people looking for tax relief at the expense of the poor. There may be some people who fit this description, but I think this view is largely a result of rich conservatives who like to use libertarian rhetoric, but are quite selective and disingenuous about it. There is also an unfortunate tendency for radical libertarians to become overly focused on tax collections as the most tangible measure of the degree of politicization in society, which to me is an infuriatingly simplistic attitude. In any case, I don't think this is a realistic view of libertarians, and it is certainly not version of "libertarianism" that I am interested in defending.
A second view of libertarianism is to dismiss it as merely a symptom of psychological flaws. Libertarians are sometimes dismissed as antisocial people who have problems with too-rigid boundaries, either calmly but mistakenly defending boundaries that aren't worth defending, or getting so upset with authority figures that they can't think straight. Libertarianism is thus equated with bad social skills. Another version of this view is that the strict anarchist version of "morality" is an outlet for people to indulge in the "paranoid" thinking patterns described by Cleese and Skynner, or the "neurotic claims of moral superiority" described by Keen. There does seem to me to be quite a lot of this going on, particularly within the Libertarian Party, where it tends to be highly visible. Other people may be attracted to strict libertarianism because they are uncomfortable with ambiguity or complexity. These sorts of motives, ie. paranoia and discomfort with ambiguity, may be hard to tell apart, and are characteristic of religious fundamentalism, so perhaps libertarianism should be viewed as a fundamentalist form of classical liberalism. Another, similar attraction that appeals more to moderates is that libertarianism's uncommonly negative view of government allows its adherents to take a posture of superior wisdom relative to the naive masses. The "professional cynics" who take this posture tend to have much lower profiles than the "fundamentalists," but I don't have much feel for their relative numbers. Again, there is nothing here that I am interested in trying to defend.
A third attraction for many people is simply that libertarianism dovetails nicely with their own personal moral intuitions (ie. "live and let live" or "don't throw the first punch"). This is basically a matter of aesthetics. My first reaction to this idea is to wonder how people's moral intuitions got formed in the first place. I expect the process to have utilitarian implications. But different people have had different experiences, resulting in differently developed intuitions, and I want my arguments to be cogent whether my readers' moral intuitions are similar to mine or not. I am also not interested in trying to promote aesthetically pleasing ideas that don't work.
Atlas Shrugged is full of negative consequences of behavior that Ayn Rand considered immoral. My impression is that she would have loved to have had a solid utilitarian argument for libertarianism, but that she despaired of ever producing one, and so she decided to "cheat" by turning her intuitions into axioms.
Another bad reason for being attracted to libertarianism is bad judgment in interpreting questionable evidence. (See the discussion of "induction" on p. 62 of The Engineering of Knowledge-Based Systems by Gonzalez and Dankel, Prentice-Hall, 1993, ISBN 0-13-276940-9. Inductive reasoning is the process by which people with finite amounts of information draw inferences about general principles.) "Moderate" libertarians may find themselves in the position of having a theory about government that seems to work tolerably well 95% of the time, and not to work the other 5%, but the "evidence" is subjective, subject to all kinds of error and fraud, and subject to interpretation. When you see a limited amount of questionable evidence against a theory that generally works, do you conclude that the evidence is in error, or that the theory is? This is a judgment call. But as Peter Wright put it in Spycatcher, "If you look for a pattern hard enough, you'll find it whether it's there or not." A friend whom I considered a "moderate" libertarian seemed in one conversation to be wrestling with this question. I was left with the impression that perhaps one reason why many "moderates" seem wishy-washy about anarchism is because they are not entirely convinced that it's wrong. I am personally convinced that the anarchists are wrong, and so I am not interested in trying to defend their "bad judgment," but I am afraid that the genuinely moderate libertarian audience I had originally hoped to reach with this essay is much smaller than I had thought.
Finally, we come to what to me are interesting reasons for someone to be attracted to libertarianism. These reasons are all more or less utilitarian arguments for moderate libertarianism, but they fall into two categories. Some of them present libertarianism as a heuristic, a way of coping with ignorance about which of several possible outcomes is the most utilitarian. The others present libertarianism as a bargaining position, a strategy for coping with conflicts of interest that might otherwise prevent society from achieving a utilitarian outcome even after one has been identified.
1. One argument for libertarianism as a heuristic is the one David Friedman presented in "Should Medicine be a Commodity?" which I discussed earlier in my "sensitivity analysis" section. This argument states that either laissez faire capitalism or democratic politics can theoretically produce results that are attractive in the sense of being Marshall efficient, but that political institutions are generally more sensitive to real-world violations of the assumptions that go into these efficiency arguments. This supports the belief that "laissez faire" is a good "standalone" heuristic, useful regardless of what other people are thinking, and regardless of whether we are considering the short or the long term.
2. Another argument is that libertarianism is necessary as an antidote for naive "common sense" approaches to political economics that are biased in favor of democratic politics. I discussed "common sense" earlier in terms of letting concentrated interests run wild, people indulging their paranoid tendencies, the "broken window" fallacy, and Thurber's elm tree problem. It often seems to me that common sense is simply a euphemism for not thinking. For example, please indulge me in a short rant about the "War on [Some] Drugs:"
Part of human beings having "paranoid" tendencies is that common sense tells us to divide people into good and bad, and the bad ones' utility doesn't count. Common sense tells many of us that it's perfectly reasonable out of a population of 250 million to have 2 million of them in prison, largely for non-violent drug offenses. Not only do we not need to think about this, but if a high government official (Joycelyn Elders) suggests that we should think about it, she should be fired. The prison statistics suggest that common sense is hard to distinguish from racism. Convicted felons also lose their voting rights--common sense tells us to disenfranchise the pot smokers and then hold a referendum on whether pot should be legal. Common sense tells us that alcohol, which is strongly associated with domestic violence, should be legal, but marijuana, which is associated with lethargy and appetite stimulation, is so frightfully dangerous that it should be listed with the most dangerous "Schedule I" illegal drugs. Common sense also tells many of us that the logic of alcohol Prohibition is completely different from the logic of prohibition of other drugs.
As the saying goes, common sense is what tells you that the world is flat. This arguably creates a need for democracy to have some sort of "quality control" system. In this view, libertarianism may be offered as a sort of intellectual counterweight, even if it tends to underestimate the optimum size of government. Even if libertarianism isn't a particularly good standalone heuristic, it still may be useful if used in conjunction with other common approaches, which have opposing biases. Again, this attractiveness would not depend on whether we are considering the short or the long term.
Update, 1-2-2010: Regarding the pitfalls of "common sense," note these articles on evolutionary psychology by David D. Friedman and Will Wilkinson.
3. A third view that interests me is that libertarianism is not only a means of identifying what's in the public interest, but also a "peace treaty" intended to bring this about. Rather than playing a variety of negative sum games, with each faction trying to win, we collectively agree not to play these games. In this view, libertarianism is a bargaining position, a version of "live and let live" which has not yet won adequate popular support.
In this "bargaining position" view, the attractiveness of libertarianism to some people may depend on looking at the big picture, looking past some immediate short term advantage in whatever negative sum game is being played this week, but it does not depend on having a particularly long term view. One's planning horizon needs only extend far enough to consider the next time one's faction will be out of power. The advantage to society of avoiding negative sum games is relatively obvious and immediate. However, the attractiveness to a particular voter of, say, a particular libertarian Congressional candidate may depend on how popular libertarianism is elsewhere in the country. If I see libertarianism as an existing "peace treaty," I may want my Representative to help uphold the treaty, but my willingness to abstain from playing negative sum games that I think I could win may depend on other people's willingness to do the same.
This view emphasizes some points which I have otherwise largely been ignoring. Early in this essay, I argued that any bargaining position that was much different from utilitarianism was unlikely to be supported by very many voters, and I have devoted most of this essay to trying to figure out what rules were likely to produce utilitarian outcomes. Now my emphasis is on the difference between knowing what policies are utilitarian and having incentives to support them. In the beginning of this essay, I asked my readers to more or less agree to utilitarianism in principle, but there are two problems with this agreement. One is that it was contingent on there not being "powers that be" who know where their bread is buttered and who hope to do better for themselves than is consistent with utilitarianism. One obvious situation where the powers that be may have little interest in utilitarianism is if we have a "tyranny of the majority." In this case, my arguments will not be cogent, and there is little I can do about it. The other, more serious problem with an implicit agreement on utilitarianism, is being able to enforce the agreement. Each of us is a member of various factions that will sometimes be part of a majority coalition and sometimes not. What is to stop a succession of transient majority coalitions from each abusing their temporary power in turn? One possible answer which I discussed in introducing public choice theory is that the voters will be so well-informed and political negotiations will be conducted so efficiently that society as a whole will consistently win these games of political "tug-of-war." This answer appears to be thoroughly unrealistic in light of the subsequent discussion of rational ignorance and negotiation costs. We are thus left with a need for some sort of "peace treaty," a threat that whoever is in power today had better not abuse their power, or the general utilitarian agreement will break down, and today's abusers will be tomorrow's abused. In short, what goes around comes around.
Is it possible for an out-of-power faction that believes the "treaty" has been violated to return to power and retaliate selectively against the perpetrators? This may be tricky--side payments to one faction within the abusive coalition to support legislation to punish the other members--but not impossible. There are then two "comes around" scenarios: selective retaliation when the political winds shift and a general breakdown of the "moral fabric of society."
4. A fourth interesting (to me) view of libertarianism is a variation on the previous one. Again, libertarianism is seen as a bargaining position, but here the argument is that libertarianism is attractive as a "Schelling point," a stable, relatively unambiguous bargaining position that, if widely adopted, would help society stay off of "slippery slopes." Other bargaining positions may look better in the short term because, while strict libertarianism forbids playing a lot of negative sum games, it also forbids a lot of positive sum games, but it arguably makes up for this in terms of stability.
On one level, I expect most people to more or less agree in principle to utilitarianism. But then I argue for rule-utilitarianism: it's too hard to tell if a particular act is utilitarian, so in practice we have to use proxies or heuristics instead of direct appeals to utility. I further argue that a good political doctrine is both a heuristic and a bargaining position intended to influence other people's behavior. Now I argue that the best doctrine is a compromise between being an accurate heuristic and having good properties in terms of bargaining theory. In this context, "good properties" means mainly that the terms of the "peace treaty" are unambiguous enough that interested people can agree on whether or not they are being violated. Another form of instability (the form that Schelling was mainly concerned with in The Strategy of Conflict) is if the bargaining process goes on interminably. A stable bargaining position needs to be perceived as being special in some sense; seen not just as one position among innumerable others, but seen as "the" alternative to further negotiations (and time is valuable). Offering to split a prize 50-50 is a position that has symmetry that is likely to make it seem special, whereas offering to split it 53.8% to 46.2% seems more like an invitation to make a counteroffer. For my purposes, I am interested in whether I can offer a version of libertarianism as a good solution to the grand political bargaining game, a solution that is likely to lead to relatively few demands for a "new deal" with every election, and that is relatively unlikely to collapse in the face of sincere disagreements over whether someone is cheating.
Earlier I claimed that liberals tended to focus on justice where libertarians focus on coercion. One advantage of the latter approach is that, while it is still somewhat subjective, it isn't nearly so as deciding if the cause is just. So, I claim, a "peace treaty" between opposing factions within a democracy is less likely to break down if it is based on libertarianism than liberalism, even if "justice" is arguably a better ("act-utilitarian") proxy for utility. This makes libertarianism more attractive in the long term even if liberalism is more attractive in the short term. This view of libertarianism differs from the previous one in that respect. They are similar in that the usefulness to me of libertarianism in both of these views is dependent on other people having similar views.
A variation on this argument (that libertarianism is nice because it's relatively unambiguous) is that any number of political doctrines might be reasonable if we can get people to pick one that they will be able to work together on as a team. I am reminded of Wiley Larson's image of trying to run a potato sack race with five people having their legs in the sack, each trying to run in a different direction. The details of what people are being organized to do may be less important than getting everyone in the potato sack to run in the same direction.
5. My final "interesting" view of libertarianism is the "frog boil" argument. This view is based on the argument that "common sense" is flawed because people become desensitized to the flaws and risks associated with systems that are either fashionable or familiar. As society progresses down a certain path, the voters' ability to make critical judgments regarding the problems they encounter may be progressively impaired. A proposed example of this process is that abuses that would never have been tolerated in the glory days of the Roman Republic gradually came to be accepted as normal and taken for granted under the Roman Empire, until finally, Rome fell because it was no longer seen by enough of its citizens as worth defending.
One thing that would help with such a desensitization problem is to use a consistent set of standards for judging our institutions. To borrow an expression from Omar Bradley, "Steer by the stars, not by the lights of each passing ship." (This is in contrast to the way executives' salaries seem to be set: "How much should we pay our CEO?" "I don't know. How much do our competitors pay theirs?") According to this view, libertarianism may be attractive in the long term because of its consistency, even if other political philosophies look better in the short term. This is especially likely if libertarianism is seen as a complement to other heuristics rather than as a standalone. In practice, it is often hard to distinguish between a "frog boil" argument and a "slippery slope" argument.
Although I have tried to articulate many different reasons for someone being attracted to "libertarianism," the internal debate within the libertarian movement (and the Libertarian Party) tends to focus more on policy than on rationale. There is a moderate, limited government wing, a strict "abolitionist" anarchist wing, and a "gradualist" anarchist wing. The gradualist position is that, while we can't do without government right now, perhaps we will be able to do without it sometime in the distant future (and that this would be a good idea). While I think the gradualist position is mistaken, the gradualists and moderates should have similar views on policy: we recognize that for the foreseeable future, "laissez faire" is only an approximation of a reasonable government policy. We agree, at least intellectually, on the need to know what the alternatives to laissez faire are and when to use them. We are in the same boat in having to deal with the problem that our "heuristic" is not the final word. Unfortunately, we are also in the same boat in that too many of us still tend to regard "compromise" as a dirty word.
21. Possible Reasons for Rigidity
Which of the above positions make sense? Which of them reflect arguments that might be cogent for reasonable people? And do any of the arguably reasonable arguments explain the libertarians' tendency towards rigidity?
First let me try to separate my own biases. Part of my attraction to libertarianism is a visceral reaction against victimless crime laws. I don't enjoy coercing other people, and I especially don't enjoy being coerced. While I think any reasonable person should conclude that victimless crime laws are generally a bad thing, I can not expect everyone to share my feelings about them and my degree of opposition towards them. A second bias I have is that I am an introvert, particularly one with weak social skills and a taste for the unconventional. I identify with Janis Ian's line in "At Seventeen" about "those whose names are never called when choosing sides for basketball." In other words, if there is a competition involving social or political skills, I expect to be on the losing side more often than not. This means that, if there is some question as to whether we are playing a positive or negative sum game, I am likely to be one of the first ones to suspect it to be negative sum. People with better social skills, or more conventional views and interests, may win more of these games than I do, and may therefore be slower to suspect the games of being negative sum. In a different context, Sam Keen has referred to a tendency such as this as "the epistemological privilege of the underprivileged." I think my experience is typical of many libertarians, so while libertarianism does nothing I know of to cause people to develop bad social skills, the population of people who are likely to question how well democracy works is a biased sample of the population in this and several other respects (also having above average education, and leisure time to spend thinking about politics). A third factor that biases me towards libertarianism is my relatively high discomfort with ambiguity, which also seems to me to be typical of libertarians in general. In the opening paragraph of this essay, I commented that I thought that the ambiguity of moderate libertarianism was a great handicap for the moderate position both within the libertarian movement and outside of it. Much of my motive for describing different flavors of libertarianism in the previous section was to see to what extent libertarians' distaste for ambiguity makes sense as part of some kind of doctrine, and how much of it is merely an artifact of the idiosyncrasies of the biased sample of the general population who are attracted to libertarianism. In order to clarify my own position, I need to pursue this distaste for ambiguity a bit further.
1. My favorite of the flavors of libertarianism is the one David Friedman presents in his "Medicine" chapter, a direct appeal to economic "efficiency." I think this is a good argument as far as it goes, and within the limitations discussed earlier. I am dissatisfied with it for not telling me how to retreat from a strict laissez faire position in an orderly manner, a point which I will try to address below. But by itself, this argument gives me very little basis for an apologia for rigidity.
In this view, crisp logic ("chunky" morality) may lower the cost of the voters' time and trouble in evaluating political proposals and monitoring politicians' performance. People seem to find "true or false" issues easier to deal with than "shades of gray." But the simplicity of crisp logic doesn't justify rigidity in negotiations as anything but a convenience for the negotiators. The simplicity of crisp logic may explain why someone might oppose an expansion of the role of government in a situation where only a tentative utilitarian case has been made for it, but it doesn't justify him clinging to his "principles" even when he knows they're screwed up. He's not doing that as a favor to society.
The above criticism of rigidity in negotiations seems self-evident to me, but it is not self-evident for everyone. I have heard someone argue that taking an extreme, rigid position is good for bargaining purposes because the eventual compromise will be closer to the outcome one actually wants. In other words, this argument claims that large, multiplayer political bargaining games are characterized by the "rationality of irrationality," where the apparently craziest players tend to get their way more than the "reasonable" players do. In fact, we saw this to be the case in the Tragedy of the Commons, where, in the absence of coercion, the unreasonable ranchers got their way. We also see the "rationality of irrationality" in cases where terrorists are able to exert influence far out of proportion to their numbers, again in the absence of effective coercion by a numerically stronger victim population. But democracy is another story: where the majority rules, and even large minorities can be coerced effectively, the shots tend to be called by the median voter, not the most extreme. Under majority rule, if a small faction tries to sell their votes for an unreasonably high a price, the buyer can usually just go to someone else for a better deal. In cases where fringe parties exert a disproportionate influence in Proportional Representation legislatures (ie. Israel), this influence is despite, not because of their extremism. (See the discussion of "the logic of coalition forming" in Instant Runoff Voting.) Where single seat electoral districts are used (ie. the USA), uncooperative fringe parties act as "spoilers," effectively removing their supporters from relevant political involvement, and extremism is even more counterproductive.
In bilateral negotiations where coercion is not allowed (ie. buying a horse in a small medieval town with only one buyer and one seller), the result of intransigence is not clearly so bad, but still the unreasonable player risks blowing the deal. Even in this ideal situation, clinging rigidly to a position that is known to be unworkable guarantees blowing the deal.
Bernard Crick describes this sort of "principled" behavior as "students' politics" on p. 135 of In Defence of Politics. He observes, "There is almost nothing that can do less harm or good to man or beast, or which has less political power, than students' politics."
2. My next argument, the "antidote for common sense," is even less satisfactory as a justification for rigidity. While I agree that "common sense" has unfortunate biases and that a better heuristic is in order, there is nothing in libertarianism that tells me how to conduct negotiations with people of other views, or that the views I espouse should vary in response to the views of others. It seems to me that believing that my heuristic is biased should encourage me to be more flexible in my negotiations with others, rather than less flexible.
3. The "peace treaty" argument might be relevant to explaining libertarians' rigidity for several reasons.
One argument, favored by gradualist anarchists such as David Friedman (p. 147 of The Machinery of Freedom) and Eric Raymond (http://www.catb.org/~esr/writings/anarchist.html), is Friedman's argument that "the logic of limited governments is to grow," and that someday in the distant future, by eliminating government entirely, the problem of enforcing an agreement to keep government limited could be sidestepped. The problem of organizing political activists to resist ill-conceived or crooked expansions of government authority may be likened to the problem of maintaining an army to defend a piece of land. In this view, accepting a "utilitarian" argument to create a government where one does not already exist to solve a minor social problem would be like a small island nation selling land to a hostile country, allowing it to establish a beachhead. Unless the selling price for this land is high enough to compensate the island nation for the cost of fighting a war against an entrenched invader, it is a bad bargain. Similarly, unless the "utilitarian" argument takes future political risks and obligations into account, it is a bad argument.
In the abstract, this gradualist argument seems plausible to me, but it also appears to be irrelevant to the current situation in which we have a government, need it (at least for the time being), and are arguing over how big it should be and how it should operate. In terms of the island beachhead analogy, the problem is that it isn't our island yet. The beachhead argument might make sense as a reason why an anarchist would want to be rigid about preserving anarchy in the future, once it is established, but it doesn't make sense as a reason for (1) making hopelessly premature attempts to establish it, or (2) admiring rigidity in adhering to any particular plan for establishing it in the longer term.
A second "peace treaty"-style defense of rigidity might be that rigidity is an attempt to form an alliance to establish or defend social norms. Libertarian "morality" may be seen as a sort of "sociological glue" that is needed in order to organize political activists to resist special interest groups (to get the people in the potato sack race running in the same direction). It may be argued that morality has to be defined in discrete, unambiguous ways ("chunky" rather than "smooth") in order to take on the emotional "life of its own" that seems necessary to motivate activists and voters. I think this view has some merit. I want there to be a stigma associated with things like voting to spend taxpayers' money on some rich person's privately owned sports arena. I want there to be stigmas associated with lots of things that have bad or unknown side effects, such as opening the cans of worms that strict libertarians try to keep sealed. I just want these stigmas to be modest, measured ones rather than overwhelming ones, because in many cases these cans need to be opened.
A third way that the "peace treaty" view might explain libertarians' rigidity is simply that people need to be consistent about what the terms of the treaty are if they are serious about enforcing it. (This might be a good argument for strict constructionism in Constitutional law, if the voters recognize their role in upholding it. In addition, like a legal contract, the terms of a peace treaty need to be clear and objective enough to be "verifiable" before they can be enforceable.) This argument has at least one obvious problem: libertarianism isn't popular enough to be anything like a central part of an implicit agreement. A second possible problem is that it isn't clear how a political faction that is out of power would go about enforcing such an agreement, or why a faction that was in power would want to honor it. It may be possible for one faction in a transient majority coalition to retaliate selectively against a faction in a previous such coalition, but it is not clear how practical this would be. (If it isn't possible to retaliate selectively, we have a Tragedy of the Commons scenario.) A third problem is that we have a dilemma: strict libertarianism is too impractical to use as a basis for an enforceable agreement (ie. it can't handle national defense problems), and moderate libertarianism isn't defined clearly enough to allow people to treat it that way.
4. The bargaining game logic in the "Schelling point" or "avoid slippery slopes" argument is, again, so closely related to the bargaining game logic in the "peace treaty" argument that it is hard to separate them. I described these above as differing in that one needs to look at the longer term in order to appreciate the slippery slope argument, whereas the benefits claimed by the "peace treaty" argument were more immediate. Another possible difference is that some versions of the "peace treaty" argument refers to an understanding that was supposedly reached in the past, where the slippery slope argument tries to filter out bad potential future agreements. These arguments also differ in that they are focused on attempts to organize or inhibit the organization of slightly different groups of people. The "peace treaty" tries to include all of society. The gradualist anarchist argument is focused on inhibiting the organization of special interest groups that are otherwise hard to hold at bay. Now let's look at what effect rigidity has on the problem of organizing the people who will be trying to resist special interests in a democracy (bearing in mind that we are all members of some special interests but not of others).
We have already considered the case of an island nation, where there is an enormous military advantage in defending the entire island rather than allowing an enemy to establish a beachhead and only defending parts of the island. I used the attacker's military disadvantage of having to attack over deep water as a metaphor for the difficulty of organizing a predatory lobbying effort from scratch, as opposed to using an existing organization. Thomas Schelling, in The Strategy of Conflict, examines a more complicated situation, in which a retreating army ends its retreat at a landmark.
Some landmarks, such as rivers, are attractive defensive positions for immediate physical reasons (just as a bargaining position may have obvious utilitarian merit). But there are other factors. Organizing the defenders may be a factor: a conspicuous landmark may be a good rallying point for poorly disciplined volunteer troops or fractious allies (the potato sack problem). Having a deal that is popular may not be nearly as important as being able to agree on what the deal is. (This is why I expressed some sympathy for the NRA's hard line on the 2nd Amendment.) Similarly, it makes sense that military orders in wartime should be obeyed even if they are poorly conceived -- coordination is that important. This need to agree on what the deal is is particularly true if there is also widespread disagreement among the defenders as to who is really under attack, which is especially likely if their attackers are also divided into quarreling factions. However, a more subtle consideration is the one Schelling had in mind: the implicit bargaining game between the attacking and defending generals.
Although war is often presented as a classic example of a constant-sum or "zero-sum" game (what's good for one side is bad for the other), it typically has both constant-sum and variable-sum aspects. The amount of land being fought over is constant, but the killing and destruction is not. As long as the attacking general is interested in other things besides killing enemy soldiers, he will care how determined the defending troops are to hold their position. If he thinks the defenders can be routed easily, he may attack, but if he thinks the defense will be fierce, he may be more inclined to sign an armistice. Even if he is confident of winning, the price of "victory" may be too high. From this standpoint, stopping a retreat in a mediocre defensive position is a doubly bad idea--the attacker won't believe that you are serious about holding the line there. But if the defensive position is a reasonably good one, and there are no other obvious landmarks to retreat to behind it, both sides are obliged to take the defense seriously. Both sides are likely to be asking themselves, "If the retreat doesn't stop here, where will it stop?" Such a position is not to be abandoned lightly.
Although the analogy between warfare and political lobbying is an imperfect one, these considerations, rallying defenders and deterring predatory behavior, are relevant to both kinds of conflict. Allies may draw defensive lines with an eye towards preserving their alliance, and defend them rigidly, despite a modest tactical advantage in retreating (or some other prima facie utilitarian argument). Rigidity may also be part of a "commitment strategy," a way of convincing a potential attacker that the defense is part of a package deal, and will be fiercer than is justified by the value of the property that is the immediate subject of the attack. Rigidity may help maintain cohesion, and cohesion changes what might be a series of low-profile conflicts into a single high-profile conflict. This is the "politics of mobilization" rather than the "politics of persuasion." This presumably reduces the number of predatory interest groups that may be able to mount an effective attack, or forces them to form their own awkward coalition. (Unfortunately, the distinction between "defensive" and "predatory" interest groups may be largely wishful thinking.) It also attracts greater public scrutiny, which one would expect to benefit the side with the better utilitarian claim on the property in question (unless one also accepts the "antidote for common sense" argument).
My conclusion, then, is that the "slippery slope" argument provides a potential justification for some degree of rigidity on the part of moderate libertarians. But there are some serious limitations to this argument. The most serious one is that the positions the libertarians are trying to defend have to be pretty good ones to start with: close to being utilitarian in the short term, popular, and clearly articulated. Otherwise there is no coalition to preserve. Unfortunately, moderate libertarians have not been very successful at articulating a position that is distinctly different from the abolitionist anarchists' position, which itself is unpopular and indefensible by sane people. They tend to present themselves as half-assed anarchists. In addition, in order to argue convincingly that libertarians' rigidity makes sense in these terms, one would also have to consider the details of the hypothetical coalition they are trying to assemble or hold together, and what the alternative positions are (ie. How far do you have to retreat in order to get to the next "landmark," and how much more readily defensible is it?). But the thrust of the "slippery slope" argument is that obvious landmarks are important simply by virtue of being obvious landmarks, and that in itself implies another limit to how much rigidity makes sense: the psychological importance of a landmark is subjective, and is affected by the amount of effort people put into promoting it. Landmarks don't have to be natural. They can be erected, and often are, like cities, or watch towers that can be seen from far away. So if someone uses a "slippery slope" argument in order to defend a silly position, he invites the question of whether his effort wouldn't be better spent constructing a watch tower in a more sensible spot.
An aggressor may try to exploit differences within a defensive alliance with a "divide and conquer" strategy, also referred to as "bologna tactics" (cutting something large into thin slices before devouring it), as illustrated by Pastor Niemöller's famous poem:
First They Came for the Jews
First they came for the Jews
and I did not speak out
because I was not a Jew.
Then they came for the Communists
and I did not speak out
because I was not a Communist.
Then they came for the trade unionists
and I did not speak out
because I was not a trade unionist.
Then they came for me
and there was no one left
to speak out for me.
-- Pastor Martin Niemöller
5. The Schelling point or "slippery slope" argument is again hard to distinguish from the desensitization or "frog boil" argument. Although an artificial Schelling point "landmark" can be erected, and a conspicuous "tripwire" may become a landmark over time, a "tripwire" is usually easy to set up, and it has a different purpose. A landmark helps organize people, whereas a tripwire prevents people from becoming desensitized, by raising their awareness. Also, a landmark has to be special in some (subjective) way, whereas a tripwire can be located more or less anywhere.
One example of how a "tripwire" works in an unrelated field is the emphasis in the aerospace industry on having "positive margins of safety." Engineers are notorious for making "micromatic calculations from vague assumptions and questionable data," often guessing at the parameters that go into an equation and explicitly multiplying the result by an arbitrary "uncertainty factor" (often several of these under various guises) before using it to calculate a "margin of safety" (MS) which we then present as if it were accurate to three decimal places. We then make an enormous fuss over whether the MS is equal or just a hair greater than zero (safe), or just a hair less than zero (unsafe). In "experimental" aerospace work, if the margin of safety is slightly negative, in many cases, the end result is that a project manager just signs a waiver and ignores the "problem." Superficially, this may seem silly. Why carry on as if there is a clear right and wrong, when everyone involved knows that the results are uncertain? The reason we go through this process is in order to make sure the project manager is aware of the level of risk he is embracing. We can't calculate the probability of a structural failure exactly, and it is never absolutely zero, but a negative MS serves as a trip wire. It tells managers that they are facing a higher level of risk than "normal." One reason we use "crisp" logic here (no shades of gray) is in order to establish clear lines of responsibility. If a failure occurs, is it the engineer's fault, a freak accident, or reckless management? But crisp logic also helps prevent management from becoming desensitized to higher than normal levels of risk. This kind of desensitization seems to have been one of the factors that led up to the Challenger accident of 1986, in which 7 astronauts were killed because of a known problem with the seals in the Solid Rocket Boosters.
David Friedman illustrates desensitization in the political arena with a comment on p. 146 of The Machinery of Freedom, "It took about 150 years, starting with a Bill of Rights that reserved to the states and the people all powers not explicitly delegated to the federal government, to produce a Supreme Court willing to rule that growing corn to feed to your own hogs is interstate commerce and can therefore be regulated by Congress." (This could also be interpreted as an example of a failed Schelling point.)
I think the tripwire argument, that some degree of rigidity helps voters avoid desensitization, has some merit. It provides a better explanation for libertarians' tendency towards rigidity than the argument that utilitarian arguments are unreliable and that simple moral rules are easier to evaluate. But it still doesn't justify a reasonable person in continuing to argue against a position that has already won its utilitarian case.
That pretty much sums up my position regarding Libertarians' tendency towards "purity, principle, and intransigence." Libertarianism makes better sense as a heuristic than as a bargaining position that is distinct from utilitarianism. Viewing libertarianism as a heuristic, the "tripwire" argument does make sense as an reason for why libertarian moral views might have a somewhat discretized, "chunky" texture. But it only makes sense up to a very limited point. Intransigence is still an embarrassment, not something to be proud of. Intransigence may be a sign of immaturity, or it may be a sign of laziness, of not being willing to think about what to do when one's assumptions have been shown to be substantially false. Intransigence could also be part of a sort of fantasy role-playing game in which one pretends to have all the answers and avoids allowing oneself to become aware of any evidence to the contrary. (More spleen-venting along these lines may be found in my rant, The Art of Stupidity.) Perhaps much of the motivation for libertarian radicalism is to avoid having to face difficult problems realistically, as Bernard Crick's remarks about "student politics" suggest. In any case, intransigence is at best an embarrassment. I have argued that having "tripwires" makes sense--there is some rational basis for a very modest degree of moral rigidity on the part of libertarians--but I see no rational basis for intransigence.
1. Do not be idolatrous about or bound to any doctrine, theory, or ideology, even Buddhist ones. All systems of thought are guiding means; they are not absolute truth.
2. Do not think the knowledge you presently possess is changeless, absolute truth. Avoid being narrow-minded and bound to present views. Learn and practice non-attachment from views in order to be open to receive others' viewpoints. Truth is found in life and not merely in conceptual knowledge. Be ready to learn throughout your entire life and to observe reality in yourself and in the world at all times.
-- Buddhist Precepts of the Order of Interbeing (2 of 14)
Update, 2-25-2010: Eugene Volokh (Mechanisms of the Slippery Slope. Harvard Law Review, Vol. 116, February 2003. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=343640) has a 90-page paper where he discusses slippery slope arguments. He classifies my arguments 3, 4, and 5 as all being types of slippery slopes, and provides his own, more detailed classification system. It's too long to summarize, but very well thought out (note also the Condorcet paradox, aka "cycling"). To some extent, I have to eat my words. Many of these slippery slope arguments are stronger than I have given them credit for. But still, there are limits to the persuasiveness of these arguments, and when a person who is not an abolitionist anarchist makes a slippery slope argument that leads to a conclusion that only makes sense if you are an abolitionist anarchist, he has gone too far.
Part 4: My Position
22. The Four Directions
In the previous section I argued that libertarianism was useful and reasonable up to a point, but that the level of rigidity that has come to be commonly associated with it goes well beyond this point, and moderates do their position a disservice by condoning this rigidity. I wish to pursue further the question of which stereotypical libertarian positions are reasonable and which are not. My primary purpose in this is to distance myself, and moderates in general, from the ideologues. At the same time, I also want to shed some light on how stereotypical libertarian positions compare with the alternatives. I don't expect libertarians to acquire a reputation as champions of the cause of "reinventing government," but if we are going to argue convincingly in favor of our stereotypical positions, or even know when to back off, we need to be able to be realistic and fair about what the alternatives are. And perhaps we can offer constructive criticism.
The stereotypical libertarian approach to social problems is to move them more fully into the private sector, through privatization, deregulation, and outsourcing. Much of this essay has been an attempt to make intelligent comparisons between this option and its obvious alternative, of moving social problems more fully into the public sector through regulation and nationalization. A third direction in which reform efforts can go is to address market failures not by moving the problem into the public sector but by changing the way the markets work, perhaps by changing the way certain property rights are defined. This can obviously be done by legislation, but can often be done privately, perhaps by selling easements. This third approach figures prominently in the book, Free Market Environmentalism, by Terry L. Anderson and Donald R. Leal (Pacific Research Institute for Public Policy, 1991, Boulder, CO, ISBN 0-936488-33-6). But if it is realistic to suggest that markets can be reformed, it is also realistic to suggest a fourth direction for reform efforts, that of changing the way public institutions work. Libertarians who want to be taken seriously when they propose redefining property rights cannot simply dismiss "reinventing government" as being inherent nonsense--to do so is obviously inconsistent. In fact, the tradable pollution quotas that Anderson and Leal champion as "free market environmentalism" are created by legislation. Is this an example of "reinventing property rights" or "reinventing government?" It seems to me to be both. The purpose of tradable permits may be to create efficient markets for pollution controls, but it is being done through regulation--unconventional, market friendly, smart regulation, but regulation none the less.
Tradable pollution quotas are "smart" in that they limit pollution to some legally specified, reasonable level at minimum cost. With tradable quotas, everyone who is in a position to reduce their pollution cheaply has incentives to do so, regardless of how politically well connected they are or how much insight the regulatory agency has into their technology. Unfortunately, this doesn't make for very good public theater, so it tends to be unpopular. (It may also fail to advance hidden redistributionist agendas.) But self-righteous theatrics doesn't look any better coming from libertarians than it does from Greens or drug warriors. Reasonable, potentially cogent arguments need to be grounded in something close to utilitarianism, and utilitarianism demands that we consider reforms that operate in all four of these directions, rather than just asserting the conclusion that privatization is the answer.
One proposal for "reinventing government" that I mentioned briefly already was to embrace "duplication of services" as a means of making government agencies or departments compete with one another in a more market-like way, rather than regarding it as something to be eliminated as "waste." The information and incentive problems in government that can be mitigated by interdepartmental competition seem to me typically to be much more serious than the problem of having too much overhead. (See McKenzie and Tullock.)
Another proposal that should be of particular interest to moderate Libertarians and Greens, and anyone who is interested in making it possible for minor parties to run serious political candidates in the US, is fundamental electoral reform. Most elections in the US use Single Member Plurality (SMP, or "first past the post") voting (one representative per district, with no runoffs). As Maurice Duverger wrote in Political Parties, plurality voting is practically synonymous with "the two-party system." Third party candidates act as self-defeating "spoilers" under this system. With only two effective political parties, US politics tends to be unnecessarily oligarchic. The internal workings of the major parties exacerbate this problem.
While SMP is good in that it rewards centrist candidates, thus producing governments that are relatively insensitive to quirks in the way legislatures work (ie. how unevenly power is shared), there are other systems that do a better job of rewarding centrists but that allow more than two parties to field serious candidates without causing any of them to become "spoilers." One such superior system is the Australian "Instant Runoff Voting" (IRV) system. IRV, like Approval Voting and the Condorcet "pairwise runoff" systems, effectively beats the plurality system at its own game. This is as opposed to the various Proportional Representation (PR) systems, which follow a substantially different model of how legislatures are supposed to work. For a lengthy discussion of Instant Runoff Voting, and voting systems in general, see Instant Runoff Voting.
It should be readily apparent that making government work better is not necessarily the same thing as trying to eliminate it. A person who is honestly trying to promote a utilitarian outcome will consider both approaches. A hard-core anarchist, on the other hand, may want government to work badly, so that it will be less popular, and thus easier to eliminate, just as some anti-nuclear activists don't really seem to want the nuclear waste problem to be solved. This can sometimes lead anarchists to advocate policies that would make government less effective, efficient, just, or well-disciplined, rather than more, such as supporting bills that "cut" taxes for some people by complicating the tax codes and creating loopholes or making "tax expenditures," that benefit the politically powerful, without any corresponding cuts in "corporate welfare" or other spending (all of which ultimately has to be paid for by someone). One also sometimes hears arguments for replacing the income tax with sales taxes on the grounds that sales taxes supposedly become prohibitively inefficient at lower tax rates, limiting the government's taxing power. But if the voters can determine the form that taxes will take, why can't they determine the tax rate?
This sort of sophistry presents a number of challenges to moderate libertarians. One problem is that moderates who aren't careful to keep fresh batteries in their bullshit detectors are occasionally going to fall for some of it. A second problem is that moderate libertarians have a great deal of trouble adequately distancing themselves from these nutcase positions. Part of this problem is that most audiences don't want to be bothered with learning twenty different conflicting definitions of "libertarianism," but part of the problem is also a reluctance on the part of the moderates to condemn people whom they perceive as allies. Part of the problem also, I believe, is that even if moderates don't want to defend the radicals' concepts of libertarianism, they don't see themselves as having clear alternatives. Finally, even when moderates do manage to distance themselves from specific nutcase positions, the moderates still too often embrace attitudes that influence their thinking about other issues, acting as ideological blinders.
Specifically, most of the moderate libertarians I have met seem to have been infected with a general attitude of disdain for the task of making government work, even though they acknowledge it as necessary. This makes it very difficult, for example, for them to address fundamental issues involving electoral reform. Libertarians will readily participate in or contribute to cover the legal expenses associated with specific ballot access lawsuits, but they have been slow to take an interest in issues like Instant Runoff Voting or Proportional Representation, despite the fact that these reforms strike at the heart of the two-party system and would enable serious minor party candidates to compete on an even footing with those of the "Demopublican" duopoly that libertarians like to complain about so much. This results in the irony for me that for several years the best way I could find to promote libertarian ideas was by supporting the Green Party, which at times seemed almost diametrically opposed, but which was more aware of and serious about fundamental electoral reform. (This is not to be confused with campaign finance reform, which I regard as being necessarily superficial due to the limitations imposed by the First Amendment's commitment to free speech. See Nelson Polsby's essay on this topic in New Federalist Papers.)
My point here is that if you're a moderate, and someone you know does not maintain an open mind on reforms in any of these four directions, then that person is not your ally. Radicals (people who make fools of themselves while claiming to speak for "libertarians") make serious debate impossible. In fact, much of the motivation for radicalism may be to avoid dealing with difficult issues. You can't reform a bureaucracy if you won't admit that you need one.
23. Indefensible Positions
Before I proceed to suggest a set of doctrines for moderate libertarians, there are some basic facts of political life that I want to lay out and some often heard stereotypical libertarian doctrines that I want to criticize.
The first fact of life that I wish to establish is that freedom is context dependent, not an absolute. As Bernard Crick put it (In Defence of Politics), "Freedom depends both on some distinction and on some interplay between private and public actions...." Almost every action will have some effect on someone else, and is thus public to some extent. If I'm forbidden from doing any activity that has any effect, however slight, on other people without their permission, I am not meaningfully free. But if other people cannot stop me from poisoning them, they are not free. Where do you draw the line? Am I allowed to smoke a cigarette in my backyard? How many photons from my light source have to hit your retina before it becomes a tort? There is no useful definition of "freedom" that doesn't require balancing public bads and transaction costs (monopoly problems). Similarly, there is no absolute distinction between free speech and harassment (or even a "denial of service" attack). Crick writes, "Politics are the public actions of free men. Freedom is the privacy of men from public actions." Being in a free society means that there is an appropriate balance between being able to act publicly and being able to object to public actions. At some point, legislators and judges simply have to use good judgment, and there is no way to avoid this. (Jeffrey Friedman is very good on this topic.)
Another fact of life, as discussed earlier, is that distribution cannot be entirely separated from allocation. Again, you can't produce benefits without opening a can of worms about how the benefits should be distributed, and national defense is a sizable benefit. Similar remarks apply regarding paying for public expenses. The problem is compounded by the fact that many of the goods and ills we are concerned with are intangible, and we can only guess at their relative importance to different people.
A related point is that, even though people may recognize property rights as an important bulwark against tyrannical government, and while it is possible to put up barriers that delay political action and give people time to reconsider, it is not possible to put property "above politics" in any permanent sense. As Crick says, "Politics cannot embrace everything; but nothing can be exempted from politics entirely." Thomas Schelling makes essentially the same point in a more general way in The Strategy of Conflict; in any negotiation, if one of the participants thinks that some fact is relevant to the negotiation, this fact will affect his thinking and behavior, and thus it is relevant. (A hypothetical libertarian anarchist society may not have anything that they call a "government", but will still have "protection agencies" that make decisions about settlements, and practice "politics" as Crick defines it: the sharing of power. In a free society, nothing can be completely exempted from politics.)
Along with these "facts of life", there are a couple of arguments for anarchism that illustrate the aforementioned need for moderate libertarians to keep fresh batteries in their bullshit detectors. One of these is David Friedman's argument in The Machinery of Freedom that government should be eliminated because "the logic of limited government is to grow." Friedman is a gradualist, someone who hopes to shrink government slowly into oblivion rather than suddenly and disruptively eradicate it. But if you can shrink government, why can't you keep it shrunk? Similarly, Eric Raymond (http://www.catb.org/~esr/writings/anarchist.html) points to the slide from Weimar to Nazi Germany to illustrate how wise it would be to not have governments. But Nazi Germany was defeated by armies of other governments. Voters in countries other than Weimar Germany had no choice over whether to disband the German government, and it's a good thing for the entire world that they didn't disband the governments they did have control over. As Friedman put it, even if there is no government to defend, there will still be governments to defend against.
An argument that anarchy is desirable is not complete without an argument that it is possible. Friedman is reduced to hoping that the fatal problems with anarchy that seem insoluble at present may be solved at some distant point in the future. But new social problems are caused by scientific, technological, and economic developments that may not be possible to direct even with governments, and certainly can not be directed without them. There is no basis for believing that new problems will not spring up at least as fast as old ones are solved.
Having laid out a few political facts of life, I now turn to discussion of particular doctrines.
It is sometimes argued (e.g. David Friedman's discussion of Social Security in Machinery) that governments tend to take from the poor and give to the rich more often than they take from the rich and give to the poor. This is often used as a rebuttal to criticisms that the poor would be worse off under libertarian doctrines. In David Friedman's case, this makes a certain kind of sense: if voluntary means could be found to address the more serious market failures, so that government could be eliminated entirely, the poor might very well be no worse off in terms of distributional justice than they are now. But once we admit, as Friedman does, that we have not yet figured out how to make anarchy work, then we are left with a government whose actions will have distributional effects, whether they are intentional or not. The problem then becomes not how to eliminate these effects, but how to limit them.
There are a number of different, purportedly utilitarian claims that libertarians sometimes make regarding the welfare state:
1. Our current welfare programs are flawed and should be reformed.
2. Our current welfare programs are counterproductive and should be eliminated until we figure out how to fix them.
3. Massive confiscations of property should be avoided.
4. Welfare programs, programs obviously intended to transfer wealth to the poor, do in fact help the poor, but should be abolished anyway in exchange for also abolishing all programs that transfer wealth to the not poor. In other words, all transfer payments should be lumped together as a "package deal," and the package should be eliminated.
5. The government should change the way it operates in order to make transfers of wealth to the not poor harder to hide (simplify the tax laws, etc.). In other words, honest accounting would help.
My point here is to distinguish between "honest accounting" and the "package deal," and to criticize the "package deal" argument. Unless you have already figured out how to make anarchy work, you can't eliminate all government programs that help the not-poor, for example by producing public goods, and you can't stop dishonest people from advocating programs that line their pockets by claiming that they are producing public goods, such as using tax money to build sports arenas. If we had honest accounting, we wouldn't have net transfers of wealth to the not-poor in societies that claimed to be caring for the poor. In the absence of honest accounting, the package deal removes public support for the poor in exchange for nothing but continued and perhaps further obfuscation. In short, the "package deal" is not useful as a utilitarian guide for running a government, however limited or temporary.
Is there a consistent moderate libertarian argument against the welfare state? Military defense and helping the poor are both public goods, at least for the not-poor. We're just relatively inept at producing welfare programs that clearly work well. Part of the reason for this is that the voters have mixed views about what they are willing to support. There are also inherent problems setting up institutions that help people without giving them unacceptably perverse incentives, especially when we do not trust our institutions with much discretionary authority. But no one as far as I know has written a proof that it can't be done. In short, the utilitarian case against identifiable corporate welfare is quite clear, but the case against the welfare state is muddy. When evaluating welfare programs, we simply have to read the fine print and think about it.
Another bogus doctrine is that all or most roads should be privatized. David Friedman discusses this in The Machinery of Freedom, 2nd ed. (p. 72), with a passing remark about the difficulty of negotiating rights to the use of access roads. The monopoly problems are obvious, and no basis is given for expecting that workable solutions will be forthcoming. In other words, it would be nice if we knew how to avoid the public provision of roads without horrible monopoly and transaction cost problems, but we don't. This doctrine essentially calls for implementing a solution that we don't have.
A more plausible, but still highly questionable doctrine is that (private) deed restrictions are better than (government) zoning. Friedman discusses this in his "Capitalist Trucks" essay from Liberty magazine (http://www.daviddfriedman.com/Libertarian/Capitalist_Trucks.html). The basic argument is that deed restrictions are accepted only as the result of consensual sales contracts, but that zoning is the result of majority or representative majority voting (initiating coercion of some minority), and that therefore deed restrictions can be expected to be generally better than zoning, or at least safer in terms of avoiding negative sum games.
I think he has a point here: political and economic arrangements can to some extent be judged by their pedigrees, the trustworthiness of the processes by which they came about; and due to transaction costs, democracy (zoning) tends to be untrustworthy. But I think this is a very weak argument for two reasons: (1) the "free market" process that created the deed restrictions was seriously flawed (highly monopolistic and subject to a variety of information problems), and (2) Friedman's distinction between government ("an agency of legitimized coercion" as defined on p. 112 of Machinery) and free market process is basically a historical curiosity with little relevance to the actual functioning of the resulting institutions. (Deed restrictions seem to be more "sticky" than zoning, tending to be frozen images of what some corporation thought its customers would want 40 years ago, but this seems a minor point.)
In the case of my subdivision, the supposedly more trustworthy "free market" process created an institution called a "Community Association" that seems to have almost all the characteristics of a local government except for checks and balances. The primary difference is that, according to Friedman's definition, I don't get to call it a "government."
I had far from perfect information when I bought this house, and even if I had known better, if I wanted to buy a house, my choices would still have been limited to either living under deed restrictions or living much farther away from my job and driving in heavy traffic. My practical alternatives are exactly the same whether this situation came about through "government" or not: go along with it or move somewhere else, much farther away.
Dennis Mueller discusses "exit" and "voice" systems in Chapter 9 of Public Choice II. An exit system has low exit costs: participants can leave easily, or credibly threaten to leave. In a voice system, participants can complain, but the exit costs are high enough that they can be abused fairly badly before their threats to leave need to be taken seriously. It seems to me that what makes governments and government-like institutions obnoxious is not a historical point about whether or when coercion was first initiated, and how property rights were defined 40 years ago, but the fact that they are "voice" and not "exit" systems. Rather than defining "government" this way and being anti-government almost to the exclusion of all else, I suggest that libertarians would be better off if they argued that voice systems with large numbers of people in them are obnoxious. Whether something is technically a "government" or merely government-like is not really all that important, and many of the institutions that radical libertarians have advocated as alternatives to "government" are very government-like.
Another example where being pro-liberty is not the same as being anti-government is free speech. The US Government is constitutionally prohibited from abridging freedom of speech, but corporations are free to have very restrictive codes of conduct regarding speech. This is further complicated by the fact that part of the reason for these codes of conduct is fear of lawsuits in government courts. If the transaction costs associated with changing jobs is low enough, this may not be a problem, but in practice, finding a similar job with a company that doesn't have a similar speech code may be impossible. Isn't there more to being pro-liberty than being anti-government?
A number of stereotypical libertarian doctrines illustrate problems with the use of crisp logic in a fuzzy logic world. Two in particular are abortion rights and children's rights. In an ideal, crisp logic world, a sperm and ovum would come together, and at some point within the next 35 years and 9 months, they would suddenly and completely be transformed from basically medical waste into something worthy of running for high public office. In reality, the transition from medical waste to full humanity is imperceptibly slow and gradual (as mental deterioration and death can also sometimes be). Libertarians need to be on guard against false dilemmas (the fallacy of the excluded middle), platform planks that either call for children to be treated as full adults or treated as their parent's livestock.
False dilemmas also figure prominently in discussions of gun control and the 2nd Amendment to the US Constitution. No sane person would seriously argue that mentally ill people should be allowed to own their own personal nuclear weapons, or that the police and the army should be prevented from carrying firearms. We are all compromisers in this regard, so we shouldn't be too pious about standing on our principles.
Another issue on which I think libertarians (at least the LP Texas platform) are too quick to take a "principled" stand is "hate crimes." The argument against hate crime legislation is that the same number of people are murdered whether the motivation is racism or, for example, sexual jealousy, and that any additional punishment is an attempt to punish a "thought crime" (or to assign a higher legal value to some people's lives than others) rather than any actual criminal behavior. In so far as this is the actual intent of hate crime legislation, I am in sympathy with the LPT platform. For a full discussion of why punishing "thought crimes" is a bad idea, I refer the reader to J. S. Mill's essay, "On Liberty." (The principle of equality before the law follows from the defense of utilitarianism I offered in Chapter 3.) However, there is a consensus among libertarians that not only non-defensive violence, but also the threat of such violence is wrong, and should be illegal. I argue that the harm done by a murder is not just the value of the life taken, but the harm done by threatening innocent people. If I murder (or threaten to murder) my wife's hypothetical lover, I am threatening one person, or perhaps a handful of people at most. On the other hand, if I threaten to murder anyone with dark skin in a city with a million dark-skinned inhabitants (or potential inhabitants), I am threatening a million people (perhaps in a deliberate attempt to force them to move away). In so far as the intent of hate crime legislation is to punish more egregious threats more severely, libertarians should have no objection to it.
Update: As a practical matter, I agree with Andrew Sullivan's view of hate crime legislation:
The real reason for hate crime laws is not the defense of human beings from crime. There are already laws against that - and Matthew Shepard's murderers were successfully prosecuted to the fullest extent of the law in a state with no hate crimes law at the time. The real reason for the invention of hate crimes was a hard-left critique of conventional liberal justice and the emergence of special interest groups which need boutique legislation to raise funds for their large staffs and luxurious buildings. Just imagine how many direct mail pieces have gone out explaining that without more money for HRC, more gay human beings will be crucified on fences. It's very, very powerful as a money-making tool - which may explain why the largely symbolic federal bill still hasn't passed (if it passes, however, I'll keep a close eye on whether it is ever used).
There are other stereotypical libertarian positions that seem to me to be reflections of political activism as a form of self-expression rather than as serious policy recommendations. (See Maurice Duverger's Political Parties for a good discussion of politics as self-expression, and the related attitude of "fight to feel good about losing" as opposed to "fight to win.")
One of these apparent "self expression" positions is support for Congressional term limits. While this would upset career politicians who are focal points for libertarian hostility, political scientists such as Nelson Polsby (New Federalist Papers) believe that the effect term limits would have on Congress would be to increase the relative strength of lobbyists and large campaign fund donors.
Similarly, in order to support the abolition of limited liability for corporations on utilitarian grounds, one would have to have a higher opinion of the reasonableness of the US court system than appears to be common either within or outside of the libertarian movement.
Another apparent form of self-expression is opposition to the income tax, as manifest by proposals to switch from a national income tax to a national sales tax. Sales taxes tend to be regressive since the poor spend a larger fraction of their income on consumer goods than the rich. There is no particular reason for thinking that income taxes are more vulnerable than sales taxes to pressure by lobbyists for complicated breaks for their special interests. The main supposed advantage of sales taxes is that they can be avoided relatively easily by changing consumption patterns, allegedly putting a lower practical limit on how high taxes can get than an income tax does. Is this true? Why should my behavior be different depending on whether my income is taxed on the way in or on the way out? Even if the claim is true that sales taxes become impractical before income taxes do, if the would-be tax reformers are powerful enough to determine the type of tax, why can't they determine the tax level directly?
The case for using property taxes instead of income taxes is even worse. It seems plausible to me that the value of the property one owns is a reasonable proxy for the relative benefits one receives from such government services as military defense, and possibly their costs ("gold without iron is an invitation to plunder" according to Machiavelli). But if property taxes are mostly based on government officials' hypothetical estimates of what a house might sell for when it isn't for sale and hasn't been for years, there is tremendous opportunity for favoritism. One trick is to make the tax rate relatively high and generally estimate low on property values. Estimate very low on your friends' houses, but estimate accurately on your enemies' houses. You can practice plenty of favoritism, but no one will be in a good position to complain. Or just make your enemies' estimates high, "reassess" often, and force them to go to court every year. This is alleged to be a common practice in Arkansas.
We have already discussed some of the potentially objectionable effects of population increases. The stereotypical libertarian support for unlimited immigration is vulnerable to similar objections. I might rationally object to my government allowing unlimited immigration just as a stockholder might rationally object to a corporation's board of directors diluting his share of the corporation's commonly-owned wealth and his degree of control over what remains by unnecessarily giving away additional shares of common stock. This is on top of the effects of higher population density. To make matters worse, there is no "limited liability" to protect me against the effects of bad decisions made by my fellow voters, as there would be with a corporation. This is not to say that a sound argument for unlimited immigration cannot be made, but a sound argument would require much more than a simple appeal to "principle."
Another "self-expression" issue for many libertarians is wanting to abolish the Environmental Protection Agency (EPA). If anything, this is sillier than wanting to abolish the regular police department, because the environmental damages the EPA is intended to combat are overwhelmingly public in nature. Pollution is a public bad and species diversity is a public good. Many of the crimes that the regular police are intended to combat are private in nature (specific crimes against specific victims), and could relatively easily be investigated and prosecuted by rent-a-cops and insurance companies. Not so with air pollution. So while there may be many reasons for demanding reform of the EPA, unless you think people have a natural right to poison their neighbors, it makes no sense to oppose the existence of the EPA.
Last on my list of stereotypical frivolous libertarian self-expression issues is tax expenditures. It is sometimes argued in this regard that the amount of harm done by a tax is proportional to the amount of money collected, and "tax expenditures" reduce the amount of money collected, so therefore libertarians should support tax breaks for the rich, off-budget spending, and various other tax gimmicks and scams. I find this argument curiously inconsistent with other libertarian positions in several respects.
For one thing, libertarians are adamant that threats count as coercion just as surely as actual violence does. So if taxation is theft, and if threatening to rob someone if he doesn't invest some of his money in your brother-in-law's company is coercion, why isn't it coercion to force him to invest in real estate under penalty of higher taxes? [I need to find a good example of a tax expenditure, preferably one endorsed by a prominent LP figure.] There is little fundamental difference between a tax and a fine. If you must have an overly simplistic rule of thumb about how bad taxes are, it might be better to look at the peak tax rates, not just the amounts collected.
A radical may try to rebut this by arguing that it is coercion, just not as bad, the proof being that the victim prefers to buy the real estate rather than pay the extra tax. But this overlooks a number of other points, which the radicals seem to understand perfectly well when it suits them.
Taxes have an "excess burden" associated with the fact that the amount of something sold goes down when it is taxed. Some activity that would have been marginally worthwhile without the tax does not take place with the tax, and the loss of utility due to this reduction is not balanced by money collected in taxes. This is a non-linear function of the tax rate. Thus, the more non-uniform the taxes (that is, the more games being played with the tax code), the higher the excess burden. The famous "Laffer curve" is a consequence of excess burden.
There are higher transaction costs associated with more complicated taxes, as is painfully clear to anyone who has dug through the instructions for filling out a US income tax Form 1040.
There is also the problem of "rent seeking," discussed in Chapter 14. The amount of effort people have an incentive to spend on lobbying the government and bribing politicians is a strong function of how effective those efforts are expected to be in winning special treatment. As David Friedman argued in Machinery, the value of the amount of effort and resources wasted in trying to win special favors is hard to estimate, but a reasonable guess is that it likely to be close to the total expected value of the favors. So a tradition of government granting tax favors to special interests is a real economic and political curse, and from a radical libertarian's point of view, the "savings" in terms of money not taken in taxes is essentially all likely to be wasted.
Finally we get around to considering the poor suckers who are left holding the bag when the politically powerful succeed in dodging their share on the tax burden. When the government spends money, someone has to pay for it sooner or later; if the government defaults on bonds or debauches the currency, that, too is a rip-off. What the radicals would like to happen would be for government spending to go down to match tax revenue. The other possibility is that taxes (or inflation or defaulting on loans) go up. It is quite possible for some combination of both to take place. On the other hand, it is also possible that the powerful, seeing that their share of the tax burden has gone down, find it in their best interest to demand even more government spending: the powerful see to it that they get a disproportionate share of the benefits, and they have already seen to it, with the help of the radical libertarians, that the weak pay a disproportionate share of the cost. I think the latter outcome is more likely, and judging by their rhetoric, one would think that the radical libertarians would also expect government to do the bidding of the politically powerful. I am at a loss to explain why they are inconsistent in this, except that it may stroke some people's egos to tell one another that most people's intuition is wrong.
Before I move on, I have a few more words on political facts of life:
One wild goose that I don't intend to chase is the "self-enforcing constitution." This is a hypothetical constitution that creates limitations and incentives for the people operating under it that preserve it in the face of bad faith or extreme ignorance on the part of the voters. As Crick argued (p. 176), "...Israel fulfills every classical condition for 'liberty having no relevance to a city under siege', Ghana few; yet one is a political system, the other is not - it seems hard to avoid the conclusion that deliberate intention enters into it." He mentions an old Whig saying, "No constitution is better than the character of the men who work it."
Other important facts of life, discussed earlier, are that (1) allocation and distribution can't be entirely separated, (2) we don't have generally agreed upon criteria for judging distributional justice, and that (3) we have nothing along the lines of an "efficiency proof" for expecting either markets or legislatures to be good at promoting distributional justice.
We also have no way to construct a legal system that will defend people's rights in the face of juries that are hostile to those rights. Nor can we define property rights in ways that are "timeless," rights definitions that cannot be or do not need to be modified over time. Finally, we do not know how to create bureaucracies that do not require good judgment and integrity on the part of the bureaucrats and supervision on the part of the voters (Death of Common Sense).
No one value, be it liberty, equality, fraternity, love, truth, reason, even life itself, can at all times override all the others or be sure never to contradict them.
-- Bernard Crick, In Defence of Politics
24. Decision Tree
What, then, is left of libertarianism after we discard the anarchism? Is it merely a vague attitude? For me, moderate libertarianism is a decision tree, a way to put together a "story" about what's the best way to solve a problem, and a "moral" person is someone whose "story" makes sense. The specific details of my decision tree may differ from other moderate libertarians, but the assumptions that go into it and the general conclusions should be familiar to them.
My decision tree starts with someone saying "There ought to be a law...." and tries to evaluate this claim in a three step process. The first step is to categorize the arguments against laissez faire as well as those against democratic politics in terms of the six objections I discussed in Part 2 of this essay, using fuzzy logic if necessary. The second step is to weigh the seriousness of these arguments in the light of a number of economic, game theory, and political science paradigms that have been discussed throughout this essay. The third step is to verify that a number of "firewalls" have not been breached without compelling reasons (aka "morality as storytelling"). I may conclude that the proposed law is really a good idea, that it is clearly a bad idea (perhaps because there is an obviously better way of writing it), or that the conclusion is sensitive to one's values and beliefs.
Fuzzy Categorization System
The often overlapping categories of objections to laissez faire or to democratic politics I recognize are:
1. Utilitarianism may not be the most appropriate criterion for judging outcomes.
2. The distribution of wealth or other benefits may be objectionable.
3. The outcome may be "inefficient" due to knowledge problems (ignorance or error).
4. The outcome may be "inefficient" due to insecure boundaries ("property rights" ill-defined or hard to defend).
5. The outcome may be "inefficient" due to imperfect competition (monopoly).
6. The outcome may be "inefficient" due to free rider problems ("public goods").
Any or all of these objections may apply to either laissez faire or to democratic politics.
Weighing Arguments (Paradigm Check)
Weigh the seriousness of the various objections. If several of them are important, they should be ranked in importance. A number of paradigms should be reviewed as they provide helpful reality checks on many of the objections:
1. The Prisoners' Dilemma. The incentives facing a group of people collectively are not necessarily the same as the incentives facing the individuals who make up the group. To assume that these incentives are the same is to commit either a fallacy of composition or a fallacy of division.
2. The Broken Window fallacy. Most of the economics arguments one hears on the evening news in favor of government programs seem to involve weighing a visible stream of economic activity promoted by the government against an invisible stream of activity that would take place in the absence of such a program. Typically, the invisible stream is erroneously ignored.
3. Rent Seeking. Often much of the value of a special advantage or favor (ie. a monopoly position or some sort of "free lunch") is dissipated in wasteful competition to acquire the advantage or favor.
4. Rational Ignorance. Acquiring the information needed to make a correct decision is often more costly in various ways than the likely consequences for a particular individual of an incorrect decision.
5. Crowded Agendas. Legislatures often take on more responsibilities than they can attend to in a timely and thorough way.
6. The Attentive Public. Ask a randomly selected voter about a randomly selected act of the legislature, and the odds are he doesn't know much about it. If he does know much about it, it is likely to be because he has a personal stake in it. So as a general rule, the only voters who know what's going on are people who have conflicts of interest. Democracy works as well as it does largely because of exceptions to this rule, the "attentive public," who through some stroke of altruism, curiosity, or fate are well informed, unbiased, numerous, and influential enough to prod the legislature to resist the influences of the special interests and the ignorant. These people tend to be badly overworked.
Now make sure that important "firewalls" have not been breached without compelling reasons. Despite the reality checks provided by the paradigms examined in the previous step, there are several reasons why this is still necessary. First, while the above paradigms help explain why many "policy failures" occur, I don't think they shed enough light on the severity of the failures or the potential for mischief that various kinds of political action entail. Second, because many, if not most, of the people advocating various kinds of political action (or inaction) have conflicts of interest, I need a "bullshit detector," a way to separate what are probably sincere differences of opinion from what are probably lies. Third, weighing these risks is difficult, and I want to catch some of what seem to me to be obvious errors, the taking of unnecessary risks.
Since I don't have quantitative measures of political risk, but only qualitative ones, my attempts to control risk take the form of discrete "firewalls" rather than an index of risk that varies over a continuous range. I have tried to group these "thou shalt not" rules into groups of comparable importance, and arrange the groups in increasing order of importance.
However, before listing these, I would like to draw your attention to a few things that are specifically not included. First, there is nothing in this concept of libertarianism that forbids anyone acting altruistically, as the word is commonly understood. Some amount of altruism is essential to making any society work; I am merely trying to make the altruists' work easier. Second, I decline to endorse the communitarians' principle of "subsidiarity," that government functions should be performed at the lowest level possible. Getting a job done at the right level of government involves many tradeoffs, not the least of which is the relative quality of local news coverage. Third, I decline to endorse the initiative and referendum movement. I don't see any reasonable basis for expecting this sort of "direct democracy" to work better than representative democracy. Fourth, I regard term limits, at least for Congress, as making Congress more sensitive rather than less sensitive to professional lobbyists and power brokers.
1. My first "firewall" is stereotypical libertarianism. Don't initiate the use of force, deception, theft, or threats of the same. Obviously, this one will be breached every time taxes are levied to pay for government services. One should still try to minimize the harm done by the breach. As discussed earlier, one can't simply measure harm by the amount of money collected in taxes, any more than one can measure the harm done in a mugging by the amount stolen, ignoring the victim's injuries. For the most part, my prejudices on the relative seriousness of different types of coercion are weak. Although I would rather see people lied to than killed, I tend to take lying on the part of a democratic government more seriously than low levels of force, because it interferes with the workings of democracy itself. The voters need to know what's going on so they can supervise it.
2. My second firewall has two parts. One is to sincerely try to create win-win situations. The other part is to keep government programs small and (especially) simple, with clear lines of responsibility.
One of the ways we try to create win-win situations in the US is by honoring the US Constitution's 5th Amendment "takings" clause: "...nor shall private property be taken for public use without just compensation." Whether the framers of Constitution intended this to apply to "regulatory takings" or not, the idea of avoiding "uncompensated regulatory takings" where practical makes sense in principle, despite its vulnerability to misuse. My commitment to the takings clause is not absolute (it is only a part of my second firewall), and the framers clearly did not intend it to apply to taxation, but it is worth thinking for a moment about why not. If a broad-based tax is levied in order to produce a public good that benefits everyone (ie. national defense), then to a first approximation, this tax is self-compensating. However, it will usually be impractical to figure out who really does or does not benefit from a typical government activity (and to what degree), how much a particular person's taxes are going for that purpose, and how to compensate him.
One reviewer of Richard Epstein's book, Takings, suggested adding a complementary "reverse takings clause," that public property (property owned collectively by the voters) also should not generally be given away for private use without just compensation. Again, this is not an absolute rule for me, as it will almost certainly be violated by attempts to address distributional issues, but most "reverse takings" seem to be gifts to the powerful at the expense of the general public (ie. sales of water or grazing rights at below-market rates, or extensions of copyright from 14 years to the author's death plus 70 years so that Disney won't lose its monopoly on Mickey Mouse).
The example of below-market rates brings up the issue of full-cost pricing. Users of publicly owned resources or publicly provided services should be required to pay at least their marginal costs. Otherwise, my second firewall has been breached, and the provision of those goods needs to be justified in terms of public charity or some other specific reason.
Since a piece of legislation will usually affect different people in different ways, even the most beneficial legislation may be bad for some people. This means that creating win-win situations often requires us to make side payments to people who would otherwise be hurt. People who consider legislative "logrolling" to be immoral often explain their objections in terms of these side payments being immoral "bribes." But are these side payments "bribes" that are used to induce people to violate their moral obligations to look out for the best interests of others, or are they "compensation" used to ensure that more nearly everyone's best interests are being preserved? My reservations about logrolling are not because some people are getting side payments, but because some people who ought to be getting them are not. The fact that we are discussing my second rather than my first firewall means that we have already established a need to coerce people rather than allow them to be "holdouts" who try to strike an unreasonably hard bargain; but in order to avoid breaching the second firewall, we must convince ourselves that even the "holdouts" are actually getting just compensation, whether they admit it or not.
Some legislation fails to create win-win situations in ways that seem deliberate, by dividing people into adversarial "righteous" and "unrighteous" classes. I described some of this legislation earlier as "theatrical." One example of this is the War on [Some] Drugs. Anyone, knowing that alcohol is legal, who is looking for a rational explanation for why marijuana is a "Schedule 1" drug (the same legal category as heroin and cocaine) is in for a frustrating time. A more ambiguous example of adversarial government behavior is the way job discrimination claims are handled. The stockholders of a corporation presumably want to hire managers who will evaluate employees and potential employees in an unbiased way, so as to maximize profits. Even the biased managers themselves will often want to do better in the interests of their careers. Yet rather than helping stockholders and managers do better at what is already in their best interest, the stockholders are faced with lawsuits awarding punitive damages. Some of this I can rationalize as necessary prosecution of fraud, and perhaps a sociologist can explain why such an adversarial relationship is necessary, but I suspect much of the government's behavior here is theatrical.
The other part of my second firewall (not directly related to the win-win criterion, but I think roughly equally important) is to keep government activities small and simple. The most obvious example of what I don't want is the complexity of the US income tax. The main reason for wanting this to be simple is to limit opportunities for hidden shifting of the tax burden from the rich to the poor. More generally, I want all laws to be relatively simple in order to limit any hidden redistributions of wealth. Other reasons for wanting simplicity are "transparency" (so that the "attentive voters" can follow what's going on), the need to keep legislative agendas from getting too crowded, and the need to limit the opportunities for "rent seeking" (wasteful competition, in this case for government favors). "Command and control" style regulations violate this second tripwire; effluent taxes and tradeable permits do not.
I mentioned "clear lines of responsibility" along with simplicity. Apart from wanting "transparency" in the workings of government itself, I also want citizens' legal responsibilities to be defined clearly. Phillip K. Howard writes of "positive" (sword-like) and "negative" (shield-like) definitions of rights in The Death of Common Sense. (This language is reportedly from an essay by Sir Isaiah Berlin, "Two Concepts of Liberty.") The US Bill of Rights contains "negative" definitions, descriptions of things that are forbidden. "Positive" rights are things that are required, such as a particular New York antidiscrimination law whose effects Howard describes on pp. 113-6, that requires handicapped access to any "public accommodation." The catch here is that handicapped access is only required to public accommodations that actually exist, so if it turns out to be impractical to provide handicapped accessible toilet kiosks, the effect of this law is to ban badly needed toilet kiosks for the non-handicapped without bringing any benefit at all to the handicapped. Much of the problem with the "positive" rights Howard describes is that they often specify what must be done, but not who is responsible for doing them. Other problems Howard describes are the result of "rights," as powerful legal claims, being defined too casually, perhaps as a result of political theatrics (or perhaps in order to arrange a redistribution of wealth that doesn't show up on the budget).
3. My third firewall also has two parts: avoid creating monopolies and work by consensus. (Although I am trying to maintain qualitative distinctions between my firewalls, in reality voters are going to have to make tradeoffs between major violations of less important firewalls and minor violations of more important ones. The 2nd, 3rd, and 4th firewalls seem particularly prone to blur together, and the increasing importance mostly arbitrary. The anti-monopoly portion of the third firewall might just as easily go with the second firewall.)
The anti-monopoly rule has several corollaries: Occupational licensing should be replaced by certification. Regulatory bodies such as the Interstate Commerce Commission that are supposed to control monopolies may be justified in setting maximum prices, but should be forbidden from setting minimum prices. Government agencies should be encouraged to compete with one another (this flies directly in the face of many attempts to save the government money by eliminating duplication). People should be able to emigrate easily (force governments to compete with one another). Political parties should also be encouraged to compete with one another vigorously, which means that we need good election systems. Again, these prohibitions are not absolute; monopolies are routinely created by copyright laws, but monopolies, including copyright, should not be created without good reasons, such as being specifically needed to create incentives for expensive investment in creative work, and should not be routinely extended (from 14 years originally to now the death of the author plus 70 years, and software patents seem to have no utilitarian basis at all).
We make a grave mistake when we choose to engage in discussions of copyright in terms of "property." Copyright is not about "property" as commonly understood. It is a specific state-granted monopoly issued for particular policy reasons.
-- Siva Vaidhyanathan, http://chronicle.com/free/v48/i47/47b00701.htm
Working by consensus is a bit vague. What "consensus" means to me is that enough discussion has taken place to ensure that opposing views have been considered, and that there is supermajority support for the majority view. This supermajority needs to agree to the reasonableness of the various judgment calls that go into the argument that any new law is beneficial, and agrees either that just compensation is being paid or that there is a compelling reason why not. Libertarians should oppose legislation that is controversial but not urgent unless they are very sure of their arguments.
4. My fourth firewall is to honor "common sense" principles of justice, such as proportionality, impartiality, not punishing involuntary behavior, not condemning without a hearing, and following precedent.
Punishments that are out of proportion to the seriousness of the crime are a strong indication that political theatrics have gotten the better of rational thought, and the term "disproportionate" (in the draconian sense) seems to imply almost by definition that law enforcement has become a negative sum game in the case at hand. Some laws are also disproportionate in another sense, not that the punishment is particularly out of line with the seriousness of the crime, but disproportionate in the "nonlinear" sense that one crime that is only slightly more serious than another is treated as much more serious. For example, a driver whose blood alcohol level is just a hair below the legal limit is let go, so it doesn't make sense for another driver whose alcohol level is just a hair over to be treated the same way as someone whose alcohol level is twice the legal limit (assuming the latter is punished appropriately). The punishment needs to be proportionate to the crime (and jurors may be reluctant to convict if they know that this is not the case).
I also have a problem with punishing the wrong people, which seems to me to be the whole point of curfews.
It is also important that precedents be taken seriously. We can't have a government of laws, not of men, if the interpretation of the law is completely up to the whim of the local judge. This is true even of bad laws. A bad law can be repealed more easily than a nation's legal system can live down a reputation for chaos and arbitrariness, and a bad law is also easier to make allowances and workarounds for.
Although libertarians are not the only people who need to be reminded of these common sense principles of justice, libertarians may need reminding a little more than most people do, precisely because there is nothing peculiarly libertarian about them, and libertarians may tend to take a "not invented here" attitude towards these principles. A second reason to remind libertarians of them is that libertarians tend to have "black and white" views of the law, focussing more on what should be legal or illegal rather than nuanced views of what is an appropriate degree of illegality, or what is an appropriate process for changing the laws. Finally, libertarians need to be reminded of common sense principles of justice because many of them are too attached to their system of thought. Violations of these principles are a sign of desperation, a sign that something isn't working, and libertarians need to be quicker to admit it when this turns out to be the case.
5. My fifth firewall is that, above all, there must be a commitment that power will be shared, a commitment to conciliation and compromise. This language comes from Crick, but Duverger makes similar observations, specifically about Communist or Fascist political parties that are willing to practice politics only long enough to achieve enough power to overthrow the system. Libertarians have tendencies to be contemptuous of democracy and compromise that remind me disturbingly of Duverger's description of Communists.
Bad laws under a democracy can be repealed much more easily than a tyranny can be overthrown. People who want to avoid tyranny need to be careful only to support leaders who are committed to the sharing of power, not matter how strongly they feel about particular issues. It is better to have a crook for a leader than one whose morality is so "black and white" that he feels himself to be above compromise.
Laws that deliberately abuse or insult political minorities are a clear sign that this firewall has been breached. Unfortunately, these sort of laws also appear to me to be quite common. The most glaring examples of leaders who are unwilling to share power may be foreign, such as Hitler, Stalin, or the more modern example of Milosevich, all of whom went out of their way to deliberately harm various minority groups. However, the US has its own share of zero-tolerance drug warriors (the War on [Some] Drugs began largely as an American version of "petty apartheid") and would-be theocrats whose pronouncements are well nigh indistinguishable from those of Osama bin Laden.
The kind of abuse I describe as "petty apartheid" may also be described as "tribal" conflict, culture wars, theatrics, and in some cases "fetishes" and "soccer hooliganism." John McCarthy has a good discussion of American "tribal" conflict at
and there is also a good article on nationalism by George Orwell at
I include some "fetishes" such as certain kinds of support for "gun control" that are reported by people who admit that they do not actually believe that the specific measures in question would be effective in protecting innocent people, and instead appear to be supported merely as a means of annoying gun owners, who these supporters see as political opponents. I put opposition to these sorts of policies as my final firewall because they are obvious negative sum games--that is precisely what they appear to be intended to be.
Much of the support for gun control is not utilitarian or instrumentalist in character: that is, many people support gun control even though they do not believe it is an effective tool for reducing violence. Instead, positions on gun control seem symptomatic of culture conflict, with gun law used as a way of declaring gun ownership and gun owners to be morally inferior, parallel to the way alcohol prohibition was used as a way for older Anglo-Saxon Protestants to condemn the culture of supposedly free-drinking Catholics from Irish or Southern and Eastern European backgrounds.
-- Gary Kleck, Guns and Violence: A Summary of the Field
( See http://www.catb.org/~esr/guns/point-blank-summary.html ) (Kleck also describes gun control as a "fetish," along with capital punishment.)
Free speech is an integral part of the sharing of power, and even if it weren't, free speech would still be important enough to be part of my last firewall. John Stuart Mill argued in On Liberty that free speech is essential because public debate is the best means we have of determining what is true. The best way to deal with a fool is to rent him a hall. (Consensus is no substitute for free speech, especially where religion is involved.) Bernard Crick argued that free speech is also essential for providing the government with the information it needs to govern competently. Tyrants have a notoriously difficult time getting accurate information. Free speech is also essential if the voters in a democracy are going to supervise their government well, and this includes "consciousness raising" and other expression that is more emotional than factual. Finally, I also believe that the pleasure or advantage that some people get from being able to suppress the speech or expression of others is almost always dwarfed by the pain and disadvantage to the people being restrained. The right to free speech does not include fraud, phone calls at 3:00 AM, or yelling "fire" in a crowded theater, or revealing information such as legitimate military secrets, but the utilitarian case for free speech as commonly understood seems to me to be unanswerable (and this includes flag burning).
This last firewall, the commitment that power will be shared, is almost an absolutely inviolable principle for me, but not quite. If civilized society is sufficiently threatened by an invasion of barbarians, I would be willing to set civilization aside temporarily rather than see it pushed aside permanently, despite the risk of a tyrant rising from within. As Crick put it, "If sovereignty is the father of politics, then once we are grown up enough to look after ourselves, we should only fly to him when in very great distress."
For any given proposed government action, this decision tree can lead to any of three possible results: (1) it is a good idea for any "average Joe" with reasonably normal values, (2) it is a bad idea for any such average Joe, or (3) it is too close to call, and depends on the details of Joe's values, even though they may be within the normal range (in other words, your milage may vary). A typical conclusion that a moderate libertarian would draw for a typical proposed piece of new legislation is that it either isn't worth the cost (perhaps it introduces more opportunities for "rent seeking" than it's worth) or else there is a better way of doing it (perhaps it uses "command and control" regulation to control pollution where effluent taxes or tradable permits would be better, or perhaps an economically inefficient law is being advanced for reasons of distributive justice that would be better advanced by making the income tax more progressive). Of course, many programs are simply rip-offs, such as sports arenas.
Unfortunately, much legislation falls into the "your milage may vary" category. I have described this as depending on different people's values, such as (perhaps) my disagreements with my local "community association" over aesthetics and goodwill. However, much more serious problems, such as how to trade off between distributive justice and economic efficiency, involve not only differences in values, but also differences in judgments about how well various government programs achieve their ends, which are further complicated by conflicts of interest. If a doctor opposes a particular obligation to provide below-cost medical care for the poor, is may be very hard to say whether this is because he doesn't care about the poor, because he thinks the law is ineffective, or because he thinks the burden is distributed unfairly. I have little to say about how altruistic people should be or what is fair, but as David Friedman argued in "Many, Few, One -- Social Harmony and the Shrunken Choice Set" (American Economic Review, March 1980), most people differ more in their beliefs about the practical workings of markets and governments than they do over the desirability of conflicting utopian visions. Friedman writes, "...the capitalist society that the socialist attacks barely has a family resemblance to the capitalist society that the libertarian defends. And I have yet to meet the socialist who is willing to defend the society that Hayek believes socialism would produce." I am mainly interested in disagreements over the practical workings. This decision tree is the best consistent framework I have been able to find for discussing honest disagreements, and it serves as a bullshit detector for most of the dishonest ones.
My original purposes in writing this essay were to (1) defend myself against my social democrat friends, (2) rally the moderate libertarians against the radicals within the LP, and (3) organize my own thoughts. My results have been mixed. I'm reasonably satisfied with the case I've presented against the social democrats. I particularly hope my social democrat readers have become familiar enough with the economic efficiency arguments and counter-arguments to be able to use them. I hope they find that libertarians are not necessarily their enemies, and in those situations when they are, maybe we can more quickly pinpoint the conflicts and in some cases resolve them through empirical investigations. I've also made enormous progress in clarifying my own thinking. However, the approach I had intended to use in rallying the moderates within the LP depended on some arguments that I no longer accept, now that I've tried to present them fully. Specifically, when I started this, I took my Schelling point arguments about "landmarks" and stable bargaining positions being needed to hold a coalition together (the potato sack) very seriously. I also thought the "frog boil" argument (and perhaps the "peace treaty" argument) was another way of looking at the same thing, whereas I now see it as a different argument. Consequently, I thought there was a legitimate practical reason for a certain reluctance to compromise on the part of moderate libertarians. I further thought that their apologetic attitude towards supporting limited government and their accommodating attitude towards the radical wing of the LP were reflections of some inarticulate, intuitive grasp of this need for some measure of rigidity. But I have now come to see this reluctance to compromise as a character flaw and a sign of bad negotiating skills. In fact, much of this essay may be regarded as a painfully anal retentive argument that being anal retentive is not helpful in either understanding or participating in politics. What little "crisp logic" I have in my list of tripwires is a symptom of ignorance, something to be apologized for rather than bragged about.
I also thought that more of David Friedman's anarchist arguments could be modified and salvaged for use by limited government advocates. Specifically, I was interested in the "package deal" argument as a partial solution to the current welfare mess. Instead, I have found lots of radical positions I want to oppose, but little in the way of attractive "landmark" positions for moderates to rally around.
I still consider myself a "libertarian" rather than a "liberal" because of my reluctance to initiate the use of force even for a good cause. Like H. L. Mencken, I am less concerned that human relations will be poisoned by inequity than that they will be poisoned by distrust. After reading Crick, I even find some specific areas where I have to distance myself from the "classical liberalism" of Thomas Jefferson and John Locke. I am attracted to Jefferson's rhetoric of natural rights and the consent of the governed, but I fear it does not stand up well to scrutiny. I see no "natural" rights except in the very limited sense that certain game theory problems do have objectively correct solutions. Thus "natural rights" are not generally a very useful concept except for preaching to the choir. The "consent of the governed" is not a very useful concept either. As Crick puts it, "...if there is absolute consent, there could be no government." The legacy of Jefferson and Locke also looms large in discussions of the right to bear arms, a discussion of which I'm afraid is beyond the scope of this essay.
But while I haven't defined myself outside of libertarianism, I have come to see it as suffering greatly from its incompleteness. I now regard my list of tripwires not as a set of underemphasized rallying points that were important to moderate libertarianism all along, but as an extension of libertarianism. I now see libertarianism merely as a heuristic ("laissez faire"), basically silent about what to do when the heuristic doesn't work. Earlier I used an analogy between libertarianism and a primitive autopilot on an airplane. A simple autopilot is useful, but a competent pilot still needs to know when and how to turn it off, and to be able to fly the airplane without it. Libertarianism seems to provide us with only the autopilot. I commend moderate libertarians for having selected a good autopilot, for recognizing that it doesn't always work, and even (with often excessive reluctance) for their willingness to turn it off, but that isn't enough. We need more ideas than that.
So here I present moderate libertarians with a number of challenges:
1. If I have been unfair or have missed some important points, email me and show me where.
2. If the behavior of political parties and legislatures are at all as I have argued (ie. unnecessarily oligarchic), then work for radical electoral reform. As Steven Hill put it, (Fixing Elections: The Failure of America's Winner Take All Politics, Routledge, New York, 2002, ISBN 0-415-93193-2) "Voting systems are to democracy what the 'operating system' is to a computer...." If you think election systems don't matter, reflect for a moment on how the Elian Gonzales case was affected by the fact that Florida has 25 seats in the Presidential electoral college, the largest of the "swing" states. I don't think any thoughtful person can seriously maintain that the case would have played out the same way if US Presidents were elected by direct popular vote. While voting systems may only be of secondary importance compared to the voters' education and culture, problems with voting systems can actually be solved by legislation. I develop my voting system criticisms and recommendations in more detail in some other articles that may be of interest:
|irv.htm||Instant Runoff Voting advocacy|
|spoilers.htm||Models for minor parties|
|florida.htm||Critique of the Electoral College|
I heard that the Libertarian Party (LP) recently endorsed Instant Runoff Voting (IRV) in the National Platform. This pleases me not only because of the extra publicity this is likely to bring to the election reform cause, but also because it is a sign that the LP is more willing to engage in the practical details of making the political system work than I had expected.
3. In several areas, our attitudes need to change. Libertarians need to embrace compromise, and in fact to develop a reputation for good-natured compromise. A good way to get started on this effort would be to read Bernard Crick's In Defence of Politics, something I wish I had done before reading David Friedman's The Machinery of Freedom. We need to think of compromise not as a necessary evil, but, to paraphrase a common saying at my church before the offertory, as a sacrament of a free society.
4. Another attitude that needs to change is we need to not be cavalier about distributive justice. As Crick puts it, "Indifference to human suffering discredits free regimes...." No other aspect of the utilitarian arguments I have presented in this essay is as subjective in its importance or as uncertain in its implementation as distributive justice. We need to be appropriately open-minded and humble.
5. In addition to defending libertarianism responsibly from its enemies on the political left (and of course the religious right), we also need to figure out how to protect it from its false friends. Obviously, its false friends include conservatives who use libertarian rhetoric insincerely, but I am more concerned with protecting moderate libertarianism from its false friends within the LP. These latter false friends often present moderates and radicals as working together at least until some day in the distant future when the moderates' objectives have been achieved and the relationship will have to be reevaluated. In reality, however, the moderates find themselves engaged in trying to solve a set of problems in partnership with radicals who are engaged in trying to deny the problems' existence.
In Chapter 1, I broke libertarians into categories of (1) abolitionist anarchists, (2) gradualist anarchists, and limited government advocates of the (3) "corner solution" (ultraminimalist), (4) apologetic and (5) unapologetic types. Although it may seem that only the abolitionists are necessarily in denial of the problems the unapologetic moderates are trying to solve, the other three types are little more cooperative. The apologetic type are too ambivalent about the utilitarian criterion to be very helpful, the ultraminimalists are overtly hostile to it (at least in its practical application), and the gradualists are only interested in hypothetical scenarios in the impossibly distant future.
Consequently, I am ambivalent towards the LP, which seems to be dominated by radicals. I am hopeful that the LP will be helpful in educating the public about electoral reform. If a reform such as IRV does pass, I am confident that the LP will either respond to the new incentives of a situation in which reasonable minor party candidates can win, or else be replaced by a new party, probably spun off from the Republicans, which will respond to them. However, if no real progress is being made towards electoral reform, the role of the LP becomes much less clear. It has not been notorious as a spoiler party, and it does do some good as a publicity device for economic and "classical liberal" arguments for limited government, but it also does some harm. Some of the harm is due to the tendency of the radicals' foolishness to discredit the moderates, who are too closely associated with them. More harm comes from the wasting of the moderates' resources through unsuccessful attempts to cooperate with radicals. Still more harm comes from the LP's tendency to provide emotional reinforcement to people for clinging to and propagating irrational beliefs.
Barring passage of something like IRV, the moderates are left with a problem in figuring out how to deal with the radicals in the LP. Can the moderates take over the LP? Can the moderates distance themselves from the radicals while being in an inferior position within the LP? Should moderates abandon the LP entirely? Or can the radicals be brought around, shown the light of reason? Maybe some of the radicals are trainable, and are just waiting for the moderates to show them a better mousetrap. My inclination at the moment is to support the LP reluctantly, as a platform for working on electoral reforms that will ultimately give the moderates better options. But I am not happy with the choices, and I am increasingly tempted to describe myself as a "classical liberal" or "postlibertarian" just to distance myself from the radicals.
6. My final challenge to moderate libertarians is for them to prove to the world that they've thought seriously about how to run a government. We need more ideas. In terms of my autopilot analogy, I have talked about the need to learn to fly. This may be impossible to do while in a partnership with people who don't want us to touch the controls, but even without them, it is still a daunting proposition, and it is essential if we're going to be politically relevant. We need to know what to do when our favorite heuristic doesn't work.
Update: Here are Five Reasons Why Libertarians Shouldn't Hate Government (from Reason.com). A good start toward thinking seriously about running a government would be for the Libertarian Party to form a British-style shadow government. I would also like to see an "illuminated" version of the federal government's budget, with all the off-budget shananigans (e.g. tax expenditures) called out with a dollar amount, perhaps with a 1.5 penalty factor for the obfuscation.
Good luck, and if you do build a better mousetrap, send me some drawings.
There is even an important socialist school of market economics, pioneered by Abba Lerner and Oscar Lange in the 1930s, that asserts that pricing in a socialist economy should mimic the pricing of a perfectly competitive free market, that such an economy would be least wasteful of resources, and that extramarket income transfers should compensate for any results that one does not like.
And there is a large body of professional opinion among economists, perhaps more among older than among younger ones, to the effect that markets left to themselves may turn in a pretty poor performance, but not nearly as poor when left alone as when tinkered with, especially when the tinkering is simplistically done or done cleverly to disguise the size and distribution of the costs or losses associated with some 'innocuous' favoritism.
-- Thomas C. Schelling, Choice and Consequence
26. Appendix A: The Ethics of Logrolling
Earlier I expressed uncertainty over whether logrolling is good or bad. It is sometimes argued that logrolling is immoral and should be illegal. Perhaps only certain kinds of logrolling are bad, and are merely immoral, or perhaps all logrolling is so destructive that all kinds should be illegal. In order to respond to these claims, I need to know what the alternatives to logrolling are. One view is that logrolling is inherent in the sharing of power, the "compromise and conciliation" that Bernard Crick calls "politics," and that the alternative to politics is tyranny. In other words, logrolling is good because the alternative is a high level of violence. This view is analogous to David Friedman's argument in chapter 3 of The Machinery of Freedom, that private property (selfish cooperation) is good because the alternative is not altruism, but force. In markets (or legislatures), people are already fully free to behave altruistically; if altruism isn't enough to make the system work all by itself, the only remaining alternatives are selfish cooperation or force. Communism has been tried, and it was discovered that if people are put in a system that doesn't work unless people are extremely altruistic, the result is not extreme altruism, but a system that doesn't work.
Perhaps this is too pessimistic. Instead of saying that the alternative to logrolling is tyranny, it might be more accurate to say that the alternatives are either tyranny or governmental paralysis (gridlock). This paralysis need not be total, but without logrolling it might be enormously difficult for government to fund anything with a geographically concentrated constituency. Presumably national defense would still be funded, as it is a public good with a very large, thoroughly dispersed "public." A flood control dam would be much harder to fund except by a local government whose borders corresponded closely with the area of land that was prone to flooding. To a lot of libertarians, this might not sound bad at all. So even if we can't figure out an effective way to make logrolling illegal, one could still argue that the stigma associated with logrolling is justified. Even if I can't argue that my congressional district would be better off if we tended to vote out representatives who engage in logrolling, I might still be able to argue that the country as a whole would be better off.
There are several reasons why I think this argument is wrong. One reason is that blocking (and certainly overturning) certain kinds of bad legislation often requires logrolling. An example might be "Jim Crow" legislation that is preferred slightly by some majority, but opposed strongly by a large minority. Logrolling can help here, provided that the majority is not so cohesive that some subgroup can't be found that can be bribed with a suitable side payment into switching sides. Logrolling can thus mitigate the "tyranny of the majority," if it isn't too cohesive. Another reason why a stigma against logrolling is bad is because of transparency, as discussed by Gordon Tullock in Chapter 3 of Government Failure: A Primer In Public Choice. The main effect of a stigma against logrolling (or for that matter, a law against it) is not to reduce the amount of logrolling that takes place so much as to force legislators to be sneaky about it. As Tullock explains, logrolling comes in implicit and explicit flavors. Implicit logrolling involves package deals, where, for example, funding for several different projects is bundled into one bill, which many legislators will support, without any need for them to make explicit promises to one another. Explicit logrolling involves the exchange of promises for support for several different pieces of legislation. In both cases, it is typically a matter of public record that a particular representative voted for a particular bill, but it is not necessarily obvious why he voted for it. If there is a stigma against logrolling, all this does is (1) force the representative to lie about his motives (to say, for example, that he thought the XYZ bill would produce a worthwhile public good, rather than that he sold his vote in exchange for Joe's vote on the ABC bill) and (2) force the people who craft legislation to design side payments that are subtle enough to maintain plausible deniability. In other words, the only effect of a stigma against logrolling is to make it more difficult for voters to figure out what's going on and whether their representatives are doing a good job. This is particularly true in the case of explicit logrolling, where legislators are motivated to lie about their reasons for voting for a bill. In the case of implicit logrolling, the lies are more likely to involve the reasons why the members of a committee voted to word the law a certain way or why certain programs were bundled together (ie. why they attached a particular rider). From this standpoint, explicit logrolling is probably worse.
making a futile attempt to discourage all logrolling
trying to discourage logrolling in general [sorry for the false dilemma],
what I want is to make logrolling work better. One thing that I
think would help would be to reform the seniority and committee
structures of Congress, so that power is distributed more evenly, but
I'm not sure what a "reformed" Congress would look like. I
would also like to remove the stigma against logrolling, at least to
the extent of allowing representatives to tell the truth about why
they support or oppose various bills. There is also the possibility
(discussed at length by Buchanan and Tullock in The Calculus of
Consent) of requiring a reinforced majority (ie. 60%) to pass
legislation. Apart from requiring a constitutional amendment, this
is dangerous because it could lead to excessive gridlock,
particularly in a "separation of powers" system such as the
USA. A complete removal of the stigma against logrolling would make
it possible to use more efficient forms of side payments, such as
block grants (ie. cash) targeted to particular legislative districts,
rather than inefficient payments such as "earmarks" in the
NASA budget or funding for projects such as the Lawrence Welk Museum,
where the desired transfer is a disguised side effect of an insincere
attempt to produce a public good. Allowing these relatively
"efficient" side payments might lower the amount of money
Congress wastes, but I hesitate to endorse them because I'm afraid
that various kinds of "inefficiency" are mostly what
prevent our current problems with "rent seeking" from being
Much as I resent my tax money being wasted on the Lawrence Welk Museum, I would not want there to be too powerful a stigma against "inefficient" transfers because I often don't trust the voters to be able to tell whether a program that produces a public good for someone in a different state is worth what it costs. I'm depending at least in part on the outcome of the logrolling process to tell me that. But I might want "efficient" cash-like transfers to be reserved for situations where compensation for a particular negative side effect is clearly in order, such as putting a nuclear waste disposal facility in someone's backyard.
In any event, I reject the claim that logrolling should be made illegal, and I claim that the stigma against it is at least partially misplaced. The stigma should not be against logrolling per se, but against (1) lying about it (including trying to disguise the transfers as public goods), (2) abusing a position of disproportionate power within Congress (committee chairs), and (3) trying to screw somebody (failure to make reasonable efforts to ensure that the game being played remains a positive sum game). In an uncertain world, logrolling is a useful if imperfect tool for avoiding gridlock while still getting a reality check (especially when combined with a reinforced majority) that a particular piece of legislation is likely not doing very much more harm than good.
Update, 7-30-2012: I am increasingly uncomfortable with the above reasoning. Consider the Dennis Mueller quotation, "We are back to the need for fairly fundamental constitutional reforms to attack rent seeking seriously." Then consider the Bernard Crick quotation, "No constitution is better than the character of the men who work it." When I combine these two ideas, the result I get is that the only thing that stops the US from degenerating into an orgy of rent-seeking is a sense of honor among voters. The Buchanan-Tullock model of amoral, self-interested vote trading doesn't work. There have to be strong moral restraints against political abuse. We only have any basis for believing that logrolling (or democracy in general) is beneficial if it is held within narrow, legitimate limits.
Consider two examples. A big payoff to Nevada as compensation for storing nuclear waste at Yucca Mountain would have been legitimate. The Cornhusker Kickback (for supporting Obamacare) was an illegitimate exploitation of disproportionate power. The question is how do we discriminate between legitimate and illigitimate?
With property institutions, it makes sense to have a refutable presumption that voluntary transactions make the world a better place. But with legislatures, the presumption should go the other way. Legislation, unless it is by unanimous consent of the voters, is involuntary. Furthermore, the Fifth Amendment "takings" clause is ignored ("regulatory takings") more than it is honored, and there is no complementary restriction in the Constitution against public property being given for private use without compensation (no "reverse takings clause"). Logrolling should therefore be presumed guilty until proven innocent. My answer to the question of legitimacy is unchanged: look for lying, abuse of disproportionate power, and negative sum games. But I think it is important that we have the presumption of illegitimacy.
27. Appendix B: Education Vouchers
Public schooling is an issue where I think my decision tree turns in a pretty respectable performance. What are the arguments for government involvement in education, and what form should this involvement take? I am particularly interested in an argument that will explain why so many of the people around me are insistent that government should provide schooling directly rather than by issuing vouchers.
I will skip over the possibility that utilitarianism is a bad criterion for judging education policy. I've never heard anyone argue that position.
Distributional justice provides a plausible argument for educating poor children. A child's education can't readily be sold for wine money (an in-kind rather than cash transfer), and it provides the basic human capital children of poor people need to compete with wealthier children later on in life. It also explains opposition to tuition tax credits, which are useful only to people in relatively high tax brackets. However, this doesn't explain opposition to vouchers. To some extent, we already do have "school choice" for people who are rich enough to buy a house based on the local school's reputation and commute to work. This is part of the reason the current system works as well as it does.
I am going to focus on "knowledge" problems as the best explanation for opposition to vouchers, but let me get the other possibilities out of the way first.
At least one argument for government involvement in education involves problems with property rights definitions (what I call "insecure boundaries"). This is the argument Milton and Rose Friedman presented in Free to Choose. Since a banker can't repossess someone's education (no costless transfers here!), bankers won't loan money for education at "efficient" levels. This argument also fails to explain opposition to vouchers.
Arguments involving imperfect competition might help a little in explaining why a small town might want to maintain at least one small public school. If the town is too small to support more than one school of an efficient size, vouchers would not provide meaningful "school choice." But if the argument is true that a second school would not be efficient, offering vouchers once the first public school is established should have no effect, positive or negative, and if the argument is false, wanting more competition is a powerful reason why vouchers would be desirable. The optimum size for an elementary school might be 200 or so pupils, but my impression is that a school a fraction of that size would be viable. Regardless of what might be true in a small town, imperfect competition is no reason for opposing vouchers in a city the size of Houston.
The most common arguments for government provision of education are public good arguments (the free rider problem). It is claimed that if I pay to educate my child, this benefits the rest of society to such an extent that my interest in my own child's well being won't cause me to spend nearly what would be an appropriate amount of money from society's standpoint. I find this a curious argument partly because it again fails to explain my friends' opposition to vouchers, and partly because it strikes me as far too weak an argument to justify the frequency with which it is brought forth.
If my son plays hooky, will I be upset at him because he is letting down the rest of society, or because he is letting himself down? There is probably some degree of public good in his education, but I expect most of the value of it to be a private good. I expect the vast majority of people to capture the vast majority of the benefits of their education through the higher salaries their education is likely to bring them. (David Friedman develops this point about being able to capture most of the benefits of one's work in Chapter 2 of The Machinery of Freedom.) The belief that education is mostly a private good is also illustrated by people's attitudes toward making education compulsory. The good of the unwilling child being compelled to attend is traded off against the good of the children whose education is being disrupted by the unwilling child's behavior, but I have never heard anyone argue that the disruptive child's attendance is necessary for the good of anyone but himself.
Another public good argument would be that there is a distinction between education and training, where training is what is necessary to get a good job, and education is a mixture of what is necessary to enjoy life more fully and what is necessary to be a good citizen. What is needed to get a good job and to enjoy life more fully are private goods, but what is needed to be a good citizen is a public good. For example, I have read books by Crick and Schelling in order to enjoy life more fully, because I find them stimulating, but in a well-functioning education system, they might be required reading. This argument may provide at least a weak justification for the opposition to vouchers, because parents choosing private schools might regard a "good citizen" curriculum as an expensive frill that won't help their child land a good job. This seems a weak argument to me because it doesn't seem all that difficult for the state to test whether a child knows, for example, the term of office for a US Senator, and to withhold funding for schools that don't teach it. This argument also doesn't explain why primary schools are public, when (it seems to me) little meaningful civics teaching takes place until much later, when the "three R's" have been mastered. While I personally find this "good citizen" argument attractive, it doesn't seem to be important in the minds of most public school advocates I've talked to. Most of them seem focused on trying to provide a safe environment that can deliver the private goods, teaching the "three R's." Civics classes seem to be perceived by almost everyone concerned with schooling as "icing on the cake."
One reason why civics classes may be perceived as "icing on the cake" is that much of the "good citizen" education that is a public good comes as a free side effect of the private goods of being able to read a newspaper, getting a good job that allows some leisure time, and being familiar with the humanities. Most of what I know (or think I know) about politics is the result of side effects. For example, looking for books on aerospace technology in the school library, I stumbled across Herman Kahn's book about nuclear deterrence, Thinking About the Unthinkable. Is there really any appreciable difference between children who are educated for the good of society and children who are educated for their own good? Do home schooled and Catholic schooled children know how long a Senator's term is? Even if the state doesn't require private schools to teach civics, much of the information may be soaked up anyway as a form of entertainment, as I argued in Chapter 13 is true of most political information anyway.
John R. Lott, in Freedomnomics, pp. 188-192, points out another flaw in the civics class argument: having the government in general, and especially having members of government employee unions, teaching future voters what to think about the government is an obvious case of self-dealing. (I tend to share Arnold Kling's view: "Public sector unions are institutionalized corruption, pure and simple.") I find Lott's argument especially persuasive in light of Antonio Gramsci's "long march through the institutions," as illustrated by the role played by President Obama's mentor, the leftist terrorist, Bill Ayers, in the Chicago educational system (see Chicago Annenberg Challenge). It is very easy for a government-controlled system for educating voters about the proper role of government to provide negative value. Lott observes on p. 192 that totalitarian countries spend disproportionately on education (e.g. at the expense of health care), and that "...when Communism collapsed in Eastern Europe and the Soviet Union, education spending fell sharply in the newly free coutries while health care spending soared." He concludes,
Governments have always used public education as a form of indoctrination, and the more authoritarian a regime is, the more constraints we find against private education. It's interesting to note that countries with socialized medical care usually allow much more competition between government hospitals or doctors than they allow between schools.... The reason for this is clear in light of the above discussion: while competition between hospitals or doctors produces better health care, competition between government schools reduces the effectiveness of government indoctrination efforts.
I would add to this that it isn't even necessary from Gramsci's standpoint to actively pervert the teaching of civics or morals. All that is necessary is for them to displace competing institutions and then allow natural human cognitive biases to reduce the next generation of voters to the level of political and economic sophistication of our ancestors of 100,000 years ago.
David Friedman has a rather harsh webbed article on public schooling that also discusses public good arguments:
Alex Tabarrok also questions the external benefits of non-STEM majors.
I have several explanations to suggest for why these weak public good arguments are presented so often. One reason is that a lot of people haven't really thought about it very much, and have a vague impression that the term, "public good," is some sort of magical incantation that justifies whatever sort of government intervention they think is desirable. Another reason for presenting a public good argument is because it is politically correct, and perhaps a lot of people's real reasons for wanting the government running the schools are not. An obviously politically incorrect reason for wanting the government running the schools would be a conflict of interest, which I would have if I were a politician looking for a position running the bureaucracy. Other people with conflicts of interest would be leaders in the public employees' unions. I don't know where the interests of the "rank and file" teachers lie, and I don't think they really know, either, despite often strong opinions; they might be paid even less in the private sector, but might also have less frustrating working conditions. Another politically incorrect argument is an ad hominem: if Pat Buchanan is in favor of privatization, I'm against it, QED. Some of the opposition to school choice may seem like knee-jerk resistance to change, but I think much of it is not opposition to change or to competitive markets in the abstract, but opposition to specific flaws that people expect to be written into the fine print of the enabling legislation. When school choice is being championed by the likes of Jerry Falwell, it may be hard to imagine that the resulting legislation will not be badly flawed.
I return now to "knowledge" problems. There are two obvious arguments that fall into this category. One is that the parents who would be making the decisions about how much education to buy under laissez faire are not themselves educated well enough to appreciate the importance of education. David Friedman argues in his webbed article that this doesn't explain why more than one generation of citizens needs to be educated by the state. I think it's a pretty lame argument, too, because it's no secret that there is a correlation between education and income. The other obvious "knowledge" argument is that the problem isn't that parents are ignorant, but that many of them are irrational. This argument does explain why I might want the state to provide education directly rather than subsidizing it using vouchers and allowing parents to choose schools for themselves. That is, I may feel perfectly capable of shopping for a school for my child, but those people on the other side of the railroad tracks are so irrational that they'll send their children to some fundamentalist school that teaches them that the Earth is only 6000 years old and that humans aren't really mammals.
Historically, in the USA, "irrationality" may have more typically been defined as "being Catholic." Again quoting Lott (p. 190):
In the 1820s, in New York and in other states, legislators became concerned that many students were receiving the wrong type of education. It was not that children were going uneducated—in 1821, about 93 percent of New York's school age youths were already attending private schools. As expressed in legislative debates, the fear was that students educated in private Catholic schools would learn the wrong values and end up becoming criminals.
This led the state to subsidize Protestant schools using vouchers, but (p. 191)
To compete, they began teaching more of what Catholic parents and students wanted—reading, writing, and math—and less of what they didn't want—Protestant religious training.
Eventually, the state solved this problem by nationalizing the schools.
The modern controversy about teaching evolution or "creation science" has supplanted concerns over Catholicism leading to wine-drinking and criminality, but the pattern is the same. From the voters' standpoint, what is really going on is mainly an attempt to establish social dominance, not an attempt to teach technical knowledge that the average child will ever remember or use. As Robin Hanson explains,
Politics isn't about Policy....
School isn't about Learning....
Similarly, Brian Micklethwait asks,
Does it matter that a lot of people are wrong about evolution? ...
In general, I surmise that being a creationist might make you generally more suspicious of leftist/statist policies, on account of those favouring such policies tending to be so scornful of creationism. If so, then as far as I'm concerned: good for creationism.
I hasten to point out that this argument, that the state has to nationalize education because too many parents are irrational, is politically incorrect. It accuses substantial numbers of voters of being irrational. These are voters on whose wisdom democracy depends, with whom politicians curry favor and solicit campaign contributions, and with whom all right-thinking people claim to be, at least theoretically, equal. It makes sense to me that people would be reluctant to use this argument openly. Furthermore, consider Frank's paradox: People can't be honest about their motives in so far as these motives involve prestige, because the admission that one is concerned about one's prestige makes one seem insecure, which lowers one's prestige.
In addition to being politically incorrect, this argument may also be exactly backwards in many parts of the USA. Here in the Bible Belt, fundamentalism is such a powerful force that keeping "creation science" out of the public schools requires constant threats of legal action, and the Texas Republican Party was reported recently to have issued a statement denouncing the "myth" of separation of church and state. Theoretically, the disestablishmentarians should be able to win these battles, but in many cases, it is not clear who is indoctrinating whom. Irrationality will in some cases be an argument for school choice rather than against it.
But assuming that the disestablishmentarians can generally win in the public schools, the question remains whether school choice can be restricted in such a way that the children of fundamentalists are given adequate exposure to modern biology. The public schools don't and can't coerce children to have particular beliefs, but merely coerce them into being familiar with certain beliefs. If children in private schools can be tested to ensure that they are familiar with the same ideas as children in public schools, such as the theory of evolution, then the irrationality argument against vouchers (the argument that only the government can teach biology and geology correctly) fails. The government could also try to restrict private schools that accept vouchers to maintaining a level of separation from religion that is comparable to the level achieved in the public schools, but I expect this to be hard to do because the customers of the private schools would tend to be self-selected for getting along well with a particular school's biases (although withholding funds would be easier for the government than it is with public schools).
Another possible irrationality argument for government involvement in education is that an important part of education is the teaching of values (social norms), and the government, presumably with the support of an overall majority of the voters, may wish to teach a different set of values than many parents wish to have taught. As I explained in Chapter 18, I take this argument seriously in the context of the civil rights movement, but am otherwise very skeptical of it. Private schools that accept vouchers can be made subject to laws such as anti-discrimination laws that are similar to those followed by public schools, so I wouldn't see this as a strong argument against vouchers even if I did take it seriously as a reason for government regulation. I include the argument for completeness rather than because anyone takes it seriously. I certainly don't want the government telling my children what religious beliefs they should have.
Of the arguments I have presented for wanting the government to subsidize children's education, the distributive justice argument seems by far the most convincing to me. The two arguments that make sense when used to oppose vouchers, the "good citizenship" public good and the irrationality argument, seem quite dubious to me, and definitely not strong enough to stand up in the face of demonstrably underperforming public schools. Vouchers (ie. competitive outsourcing) make much better sense to me than the status quo. But when I talk to non-libertarians about vouchers, I typically get three reactions. One reaction is that vouchers would "destroy the public school system." I find this odd, because the public school system supposedly exists in order to benefit the children, and the whole point of school choice is to do just that. Perhaps the people who make this statement are just not articulating their objection clearly enough for me to understand it, but this reaction makes it sound like the children exist for the benefit of the National Education Association (NEA) rather than the other way around.
A second, more thoughtful reaction is to observe that some children are more expensive to educate than others, and to predict that the public schools would end up with only children who are unusually expensive to educate. In other words, public schools might be more (appropriately) flexible in distributing resources than a voucher program is likely to be. (One the other hand, US public schools are internationally notorious for going too far in trying to promote equality, as even Bernard Crick, a socialist, comments in In Defence of Politics, "mainstreaming" retarded children and even trying to "educate" children who belong in hospitals, to the considerable detriment of all the rest.) I have two rejoinders to this prediction. My first is that this may be a valid criticism of particular voucher proposals, depending on what the numbers are and what is written in the fine print, but it is not a generic argument against vouchers. Vouchers could be offered on a sliding scale, depending on a child's known difficulties and the parents' economic means. My other rejoinder is that a voucher program doesn't have to be as flexible as the public schools are in order to do a better job on average, especially if public schools continue to exist, at per-child funding levels that are somewhat higher than the voucher payments. The purpose of the schools is to serve the best interests of the children, and I don't care if the work of the public sector teachers is more difficult than their private sector counterparts as a result. The already existing "alternative education" public schools will probably continue to be needed in either case (vocational schools, too).
A third reaction to vouchers is that they would violate separation of church and state (ie. the "establishment" clause of the 1st Amendment of the US Constitution, although I may want stricter separation of church and state than the Constitution technically requires). Part of this reaction is the observation that parochial schools have a transient advantage over secular schools in that the former already exist in substantial numbers; parents who want to send their children to secular private schools will mostly have to wait for them to be founded. One answer to this objection would be to enact that only secular schools will be eligible for vouchers. It would be unfortunate if this is necessary, because it would delay the widespread availability of private schools, and possibly raise their cost, but it is not fatal to the case for vouchers.
However, strange though this may sound coming from me, I don't think this restricted eligibility would be necessary, at least in urban areas. The reason this may sound strange is that I have a hair trigger for reacting to violations of separation of church and state in the public schools. Why would I not have a similar hair trigger for "outsourced" schools, schools that are privately run, but subsidized by the state? Let me turn the question around: why do I have this hair trigger in the first place? It's OK for the Army to have chaplains, and it's OK for completely private schools to push religion, but it's not OK for public school teachers to push religion. (While I'm at it, do I trust public school teachers to teach classes in comparative religion? It would be desirable, if they could do it evenhandedly, but is this realistic?) Is it OK for outsourced (government funded) education contractors to (A) have openly displayed religious preferences and (B) to push their religion? There are two considerations here that I think are important. One consideration is that children are involved. If a government employee (ie. an Army chaplain) encourages an adult to attend a religious event, I trust an adult to be able to say, "No, thanks." I don't trust a child to be able to resist an adult authority figure (or peer pressure) as easily, and so the threshold for what I consider coercion in the public schools is a lot lower, and is age-dependent. (I also expect an Army chaplain to act in a professional manner.) The other consideration is that government schools have strong if not absolute monopoly power. Even if school attendance were not compulsory, the public and private schools are competing on an extremely uneven playing field. In a highly competitive market, if a parent doesn't like the religious orientation of a school, he can take his child elsewhere, but currently, if the public school is objectionable, the decision to go elsewhere is quite expensive. It seems to me that it is this government-backed monopoly consideration that makes religion in the public schools objectionable, rather than government financing. Few people would object if a government employee on official travel were to spend some of his per diem in a vegetarian restaurant run by a Buddhist temple, or if a construction contract were won by a small company with ties to a Masonic lodge, or if a government employee were discovered to be a tithing member of a church. By this reasoning, whether or not vouchers should be restricted to secular schools would depend on there being at least one reasonable secular school in a particular area. As long as secular public schools continue to exist, I don't see a problem (and if the public schools are not reasonable, that only strengthens the case for outsourcing).
It seems a little strange to me that it is considered "establishing a state religion" to treat schools the same way regardless of whether or not they have religious affiliations. No one seems to have objected on religious grounds to having a level playing field between secular and parochial private schools before the government became involved, and the hard core disestablishmentarians don't seem to object to having the government's thumb on the scales in favor of secular schools, but changing our minds and deciding that we want the playing field to be level after all apparently is a violation.
Suppose you were hired as an advisor by a former communist country with a substantial Jewish minority, with government-owned grocery stores that are subsidized to the point of almost giving food away, but set up by militant atheists who refused to sell kosher food. The Jews responded by setting up an independent kosher food industry at great expense to themselves (the only significant alternative to the government-run industry). The new government wants to privatize the food industry, but now has strict rules about separation of church and state, and the Minister of Agriculture objects that privatizing the grocery stores would create a huge bias in favor of Judaism, and that it is therefore a matter of the strictest principle that the food industry remain a de facto government monopoly for the rest of eternity. How impressed would you be with this argument?
It seems to me that the private parochial schools have competed with the state under an enormous disadvantage for a long time, a disadvantage which secular private schools have generally not been able to cope with. It seems disingenuous to argue that we have to maintain the status quo for fear that parochial schools would have a temporary advantage until secular private schools become better established. To paraphrase Lenny Bruce, the public school system cripples secular private education, and then arrests it for limping.
I don't see any inherent, insuperable problems with voucher programs. How well they work will depend on how fast this "outsourcing" is phased in, and at what funding level, among other factors. Voucher programs could easily be sabotaged by allowing the enabling legislation to be written by (1) the NEA, (2) Jerry Falwell, or (3) wealthy interests trying to frustrate the redistributive agenda. But they can't be dismissed out of hand. In order to argue against a particular voucher proposal, you're going to have to crack the cover and read the fine print.
Finally, I note that Michael Strong has a thought experiment in Be the Solution in which policies regarding education and housing are reversed. He imagines being busted for "truancy" for failure to live in government housing. The only reason I can see why coercing people into public education makes more sense than coercing them into public housing is John Lott's argument: it's all about indoctrination.
28. Appendix C: Foreign Policy
After reading Eric Raymond's Anti-Idiotarian Manifesto (AIM), I decided not to write my own article on foreign policy, but simply link to the AIM: http://www.catb.org/~esr/aim/index.html . Mr. Raymond's short article on the background to the AIM (why it was necessary to coin a new word) is also well worth reading: http://www.catb.org/~esr/aim/background.html .
Steven den Beste expressed a similar, "Jacksonian," view of foreign policy:
You leave me alone and I'll leave you alone. You play fair with me and I'll play fair with you. But if you f*** with me, I'll kill you.
I also like this quotation from Admiral Fisher:
If you rub it in, both at home and abroad, that you are ready for instant war, with every unit of your strength in the first line and waiting to be first in, and hit your enemy in the belly and kick him when he is down, and boil your prisoners in oil (if you take any), and torture his women and children, then people will keep clear of you.
I share the libertarian tendency towards isolationism in foreign policy. The polite version of isolationism was perhaps best articulated by John Quincy Adams (speech to the U.S. House of Representatives on July 4, 1821):
Wherever the standard of freedom and independence has been or shall be unfurled, there will her heart, her benedictions and her prayers be. But she [America] goes not abroad, in search of monsters to destroy. She is the well-wisher to the freedom and independence of all. She is the champion and vindicator only of her own.
The impolite version was articulated by Thomas Carlyle's fictional character, Flimnap:
Tumble and rage along, ye rotten waifs and wrecks; clash and collide as seems fittest to you; and smite each other into annihilation at your own good pleasure. In that huge conflict, dismal but unavoidable, we, thanks to our heroic ancestors, having got so far ahead of you, have now no interest at all. Our decided notion is, the dead ought to bury their dead in such a case: and so we have the honor to be, with distinguished consideration, your entirely devoted,--FLIMNAP, SEC. FOREIGN DEPARTMENT.
29. Appendix D: Rational Irrationality
Bryan Caplan argues in The Myth of the Rational Voter: Why Democracies Choose Bad Policies that voters' lack of incentives to exercise mental discipline ("rational irrationality") is distinct from, and far more serious, than is their lack of incentives to become well-informed (rational ignorance). This argument strengthens the overall case for choosing property institutions over voting institutions, but it also undermines much of classical public choice theory, including much of my chapters 13 and 16.
Caplan's findings lead inexorably to the conclusion that democratic governance can be improved only through reforms based on realistic assumptions about human cognition.
-- Timur Kuran
I have a review of this book posted at caplan.htm .
30. Appendix E: Health Care Reform
As I see it, there are basically three problems we are trying to solve with regard to health care reform:
1. I want efficient health care for people who can afford it when they need it. This is "routine" medical care, at least from a financial standpoint.
2. I want efficient insurance markets (risk pooling) for people who can afford their statistically expected future medical costs, but not their actual costs. This is financially "catastrophic."
3. I want efficient charity for the poor, or those who are some combination of relatively poor and relatively unlucky.
Routine medical care is mostly a job for laissez faire capitalism. There are some problems I want government to address (How do you set prices for emergency services? How do you pay for research?), but these are relatively minor. In theory, the medical insurance industry should also mostly be a job for laissez faire capitalism, restoring insurance companies to their proper function of risk pooling, getting them out of all the other functions they've been dragged into, and making them smaller and less important. In practice, the insurance problem has been made into a hard problem because it has become conflated with both of the other problems. Charity is an intrinsically hard problem. This is a job for a program like Medicaid.
The reasons why I expect laissez faire capitalism to mostly work for routine medical care are standard economics textbook arguments, such as David Friedman presents in "Should Medicine be a Commodity?" (http://www.daviddfriedman.com/Academic/Medicine_Commodity/Medicine_Commodity.html). There are some idealizing assumptions involving property rights, rationality, competition, and neighborhood effects. These assumptions lead to "economically efficient" outcomes. That is to say, in so far as these assumptions are true, we can construct Bryan Caplan's "rickety bridge" that connects private interests with the public interest (I like this better than Adam Smith's "invisible hand" metaphor). An intelligent debate about this would cover the degree to which the various assumptions are being violated, the sensitivity of markets to the violations, and a realistic comparison of the alternatives. For example, in the case of emergency medicine, competition is severely limited or non-existent. But mostly, getting people who have money to spend it on themselves is pretty easy, so routine care is an easy problem. Self-interest is a plentiful resource. We mainly just need to stop screwing up this easy problem by combining it with harder problems.
Insurance companies are bureaucratic and arguably need more regulation than most other businesses because of information problems that violate the "rationality" assumptions. But the point of the needed regulation is to prevent fraud. Otherwise, the usual arguments for laissez faire mostly apply, so having efficient insurance markets is still a relatively easy problem. Again, my main concern is to stop screwing up an easy problem, in this case by using insurance companies as clandestine government policy tools. Their proper function is risk pooling, and they are a huge waste of money and opportunity when they get involved in routine medical care, charity, job changes, or tax law.
Charity is a hard problem. This is partly because altruism is a scarce resource, but also because of a lack of general agreement on who is a worthy recipient and of how much. I'm going to focus on how to pay for it, and assume that something like Medicaid will disburse the money.
We are not debating whether or not to provide some level of charity. We are also not debating whether, for the forseeable future, the government will be funding much of this charity.
We're also not really debating whether health care is a "right." What if I'm 85 years old, and there is an operation that costs $100M, and would extend my life for two weeks? The language of rights is simply not helpful in making the sort of tradeoffs we need to make here. Similar comments apply with respect to "universal" health care. We are arguing over how to pay for medical charity and how to ration it, not over whether to provide it.
There are basically three ways to pay for health care for the poor. The first method is voluntary charity. This depends on large numbers of people being altruistic enough to spend their own money on the poor, unconditionally on what other people do ("strong" altruism). The second method is on-budget government spending supported by overt taxes. This depends on a majority of voters being altruistic enough to spend either other people's money on the poor, or their own on condition that other taxpayers do the same ("weak" altruism). The third method is off-budget spending in the form of manipulation of the tax laws and poorly understood, unfunded mandates. This method depends on some combination of voters being more altruistic than they're given credit for; or voters being stupid, and powerful elites making unpopular, altruistic decisions away from the public spotlight.
Strictly voluntary charity is not currently politically feasible because it depends on institutions and behavior patterns that mostly do not currently exist, and will not spring into existence until some time after we achieve much better clarity over what things the government will and will not pay for. In the longer term, there is little basis for hoping that voluntary charity would ever come close to providing the "economically efficient" level of funding, since there is a significant neighborhood effect. And even an "efficient" level of funding doesn't entirely satisfy questions about distributional justice. Radical libertarians can still argue that government programs tend to be even worse, but even if they are ultimately successful with this argument, until their new institutions are up and running, I still want the government involved in paying for medical care for the poor.
The question then becomes how to pay for the government's involvement in medical charity. I have described the choices as on-budget and off-budget, but the off-budget funding can be further divided into "tax expenditures" and unfunded or underfunded mandates. I contend that broad based, overt taxes and on-budget spending are the best way to do this, both from a transparency and an efficiency standpoint.
The transparency argument is obvious. The amount of tax money that isn't collected because of "fringe benefits" not being taxable doesn't show up on the budget, much of the cost of an employee's compensation package doesn't show up on his pay stub, and the average person has no way of knowing what sorts of "cost shifting" went into the payments his insurance company made. The average voter doesn't know how much the current system is costing him or anyone else, where the money is going, or whom to hold responsible, but the voter is ultimately responsible for supervising the government's actions. Even with Medicaid, some of the funding comes from the federal government and some from the states, which tends to confuse matters. But this is the height of transparency compared to unfunded mandates.
The efficiency argument is less obvious. It is tempting to think of taxes as a pure transfer from a citizen's pocket to the government's, but this is not so. Taxes affect not only the effective prices of whatever is being taxed, but also the quantity produced. The reduced production causes a loss of utility known as "excess burden." The size of the loss depends on the shapes of the supply and demand curves. This is a nonlinear effect, as illustrated by the Laffer curve, but the effect is usually small if taxes are low and broadly based.
The fact that health insurance "fringe benefits" are not taxed is partly a historical accident. It began in WWII as a way for the government to allow critical industries to get out from under the Rooseveldt administration's wage and price controls without having to admit that they were counterproductive. (See Free to Choose by Milton and Rose Friedman. See also http://www.theobjectivestandard.com/issues/2007-winter/moral-vs-universal-health-care.asp) for some early history of tax shenanigans involving Blue Cross/Blue Shield.) But the decision that "fringe benefits" didn't count as "income" transformed the medical insurance business into a massive tax avoidance scam. But participation in this tax avoidance scam is so widespread that, instead of a small class of taxpayers ripping off the other taxpayers, mostly what we have is the middle class ripping itself off, and paying for the privilege in terms of having their employers and their insurance companies dragged in as middlemen every time they see a doctor.
The nice thing about simple, conventional taxes is that there is a bright line between what the government is taking and what it is not taking. Conventional taxation is a minor departure from David Friedman's assumptions about property rights, and markets are not usually very sensitive to it. Open-ended, unfunded mandates are another story. One of Friedman's assumptions is that the cost of enforcing one's property rights is negligible. When this is not true, expensive goods and services are up for grabs, and economists worry about what they call "rent seeking." An example would be a civil lawsuit. A plaintiff may be willing to spend up to $49 in legal fees for a 50-50 chance at $100 in damages, and the defendant may be willing to spend up to $49 for a 50-50 chance to keep his $100. It's easy to come up with models in which the amount of resources wasted in fighting over hard-to-protect wealth is roughly equal to the value of the wealth in question. (See David Friedman's discussion of the Civil Aeronautics Board in The Machinery of Freedom, ch. 38.)
I did a Google search on "cost shifting" and came up with a wikipedia entry on the Emergency Medical Treatment and Active Labor Act (EMTALA), an unfunded mandate for emergency medical care regardless of ability to pay. (http://en.wikipedia.org/wiki/Emergency_Medical_Treatment_and_Active_Labor_Act) The article reports that "55% of U.S. emergency care now goes uncompensated." This legislation is intended to help the poor, but by taking the form of an unfunded mandate, it functions as a punitive tax on the industry that it should be supporting, and falls most heavily on hospitals that disproportionately serve the poor. The fact that this "tax" is concentrated on one industry rather than being broad-based sharply increases the excess burden of the tax, but the fact that it is open-ended results in the wasteful avoidance behavior associated with "rent seeking."
In the health care industry, rent seeking may take the form of hospitals shutting down badly needed emergency rooms, or trying to make them as unattractive and inconvenient for indigent patients as possible. Rent seeking may also include some amount of illegal immigration. The costs that the hospitals do incur largely get shifted to paying customers and their insurance companies, who similarly waste money trying to avoid picking up someone else's tab.
There is a saying from the Watergate scandal: "It's not the crime, it's the coverup." This saying appears to me to apply to government care for the poor. The problem isn't that the government spends too much money on the poor, but that it tries to hide the costs through contortions that end up costing more than the amount being hidden.
I suggest that advocates for the poor have greatly underestimated the voters' willingness to pay for health care for the poor. What I called "strong" altruism may indeed be scarce, but I claim that "weak" altruism is plentiful enough to be able to fund programs like Medicaid enough to provide much better care than the poor are getting now. But we need to stop trying to be so clever about getting people to pay for more than they are collectively willing, because through such cleverness we end up like Wile E. Coyote in the Roadrunner cartoons.
Nothing is more hateful to wisdom than excessive cleverness.
-- attributed to Seneca the younger by Edgar Allen Poe
The first step towards transparency would be to dope-slap any "reformer" who tries to blur the distinction between insurance and charity. Asking for "coverage" for a pre-existing condition is not asking for risk pooling, it is asking for charity.
Medicaid should be reformed, probably expanded, and funded appropriately out of general funds. (Ideally, either at the state level or the federal level, but not both.) This should not be an entitlement, but should be something that the legislature is forced to re-examine every year. It will always be a controversial subject, involving difficult tradeoffs, and the controversy needs to be allowed to take place openly and continually. Medicaid reform is tricky in part because it needs to be luck-tested and stupidity-tested as much as means-tested. Also, the dropoff in benefits with income has to be very gradual or the working poor will get stiffed.
All unfunded or underfunded mandates ("uncompensated regulatory takings") should either be repealed or given just compensation. Take the 5th Amendment seriously.
Make all fringe benefits (employer contributions to health insurance) taxable. This will get insurance out of the routine medicine business and go a long way towards solving the portability problem. No one in his right mind wants to pay for car insurance that has to get involved every time he gets his oil changed. There may also be some regulatory changes needed to clarify insurance companies' financial responsibility when a customer first becomes sick. Homeowners' insurance companies can't weasel out of paying a claim just because the customer changed jobs before his house repairs were completed.
Medicare is complicated. The short answer is that Medicare should be abolished, as welfare for the not-poor. The long version is that Medicare is part of Social Security, and suffers from the same case of multiple personality disorder as the rest of Social Security. Alternately, one can look at Social Security as a matryoshka doll or as three pots of money that have been comingled. One component of Social Security payments can be rationalized as a reasonable return on past forced savings by the recipients, very conservatively invested, paid in the form of an annuity. In so far as the recipients are among the deserving poor, a second component can be rationalized as a peculiar kind of charity, limited to the elderly, and paid for using a regressive tax on younger workers. This leaves a third component that is an intergenerational transfer. The first component should be paid exclusively as a regular Social Security payment. The second component should be redirected to Medicaid. The third component should be abolished, or failing that, should be part of the regular Social Security payment.
Governments should stop artificially restricting the supply of medical care. I like Milton Friedman's proposal (see Free to Choose) for a constitutional amendment abolishing occupational licensing. Licensing should be replaced with certification.
EMTALA also illustrates legislation that partially suppresses the making of tradeoffs involving quality of care. Tradeoffs like this are necessary, and more ability to make them openly would be one of the benefits of replacing licensing laws with certification. One benefit of getting insurance companies out of routine care is that there are tradeoffs between the cost of expensive diagnostic tools and the health of the patient. I may feel that a marginally desirable MRI may be worth what it costs if my insurance company is paying for it, but not if I'm paying for it. Getting insurance companies out of the loop as much as possible reduces these conflicts of interest.
The AMA should be treated as a cartel and subject to anti-trust laws (see Free to Choose). Any laws against price competition or advertising prices should be abolished.
There need to be courts with the authority to set prices for emergency services, analogous to admiralty courts that determine salvage fees. This function is currently being performed largely by insurance companies. (David Friedman discusses this function of insurance companies for non-emergency sevices in a 5-18-2009 blog post.)
Tort reform would be nice, but if the underlying problem is with the juries, I don't know what a solution would look like. A different voir dire process? Tort reform would help alleviate the problem of "defensive medicine," medicine oriented more towards avoiding potential lawsuits than towards serving the actual interests of the patient.
Drug approval policies should generally be more liberal (more like "caveat emptor"). But this is not to disparage the government's role in regulating antibiotics and the treatment of infectious diseases.
The way we currently pay for drug research is through awarding patents (monopoly privileges); if you don't want individuals to pay monopoly prices, the best alternative is probably to have the government offer to buy the patents rather than give the government the responsibility for drug research and development. (But patent law in general needs to be reformed.)
Some closing remarks on distributional justice:
My support for a relatively free market solution, as opposed to something more like the British system, depends on a tradeoff between distributional justice and economic efficiency considerations. The British and Australian systems, and especially the Canadian system, are clearly more egalitarian than what I propose. They are presumably also more distributionally just, although this is a more murky question. I can also imagine how someone could look at the US system as it actually exists today, and especially in the way it is described by political activists, and decide that the British system doesn't sound dramatically more inefficient than the current US system. Robin Hanson claims, with some statistical support, that health care spending in the US could be cut in half with little effect on outcomes (see http://www.overcomingbias.com/2007/05/rand_health_ins.html and http://www.cato-unbound.org/2007/09/10/robin-hanson/cut-medicine-in-half/ ). But I am not defending the current US system, I am advocating something with minimal political interference and legerdemain.
One reason I think that efficiency is important is because I think people's well-being is more sensitive to the quality of their medical care in absolute terms than in relative terms. I am unwilling to decrease the life expectancy of a lucky person by two years in order to increase the life expectancy of an unlucky person by one year. I also have grave doubts about both the justice and the practical consequences of policies that tend to promote equal outcomes for all people without regard to their behavior (see Jodie T. Allen's discussion of the negative income tax, http://www.econlib.org/library/Enc1/NegativeIncomeTax.html ). I have even graver doubts about the government's ability to administer a non-egalitarian politicized health care system justly. I do think that dramatic improvements are possible with a relatively free market in medicine, relative to either the current US system or the British system. A "single payer plan" of any sort, is by definition a government monopoly, and most medical care is overwhelmingly a private good (putting a cast on my broken leg benefits me far more than it affects my neighbors). I find it hard to imagine why anyone would expect a government monopoly to be relatively efficient at producing a private good.
But as redistribution of wealth schemes go, subsidized health care for the working poor seems fairly sane. Similar arguments can be made to those I have used in defense of education vouchers: both a basic education and basic medical attention are worthy uses of money, they benefit rich and poor alike, and they aren't something that the recipient or his parents can sell to buy wine. On the other hand, health care costs are generally less predictable than education costs, and are thus more vulnerable to fraud and more demanding of close supervision by a third-party bill payer.
We should avoid having insurance companies or government acting as middlemen involved in routine health care for the middle class. To me, that makes no sense from either a distribution or an efficiency standpoint. It may be tempting to think that the tax code will be progressive enough that we are helping the middle class at the expense of the rich, but I think this is unlikely. There are limits on the degree to which distribution and efficiency can be separated. Attempts to make the tax code too progressive will either kill the goose that lays the golden eggs (the wrong part of the Laffer curve), or result in much of the tax being passed on to ordinary consumers in the form of higher prices (again, something that is determined by the shapes of the supply and demand curves, not just by government policy). Trying to subsidize the middle class this way brings us back to the current health care situation, where we have the middle class paying for the privilege of ripping itself off.
I don't think the British or Australian systems are insane, I just think that they are unnecessarily inefficient, a poor tradeoff between distribution and efficiency. But the Canadian system goes further, by banning the augmentation of socialized health care with private professional care. This strikes me as pursuing ideological purity to an extreme, and it has to be a mistake for anyone with normal values.
Still, I give the Canadians credit for making an honest mistake. The current US health care system appears to me to be the result of a series of dishonest mistakes.
31. Appendix F: Minimum-Wage Laws
I was having lunch the other day with a mathematician friend of mine, and we were kind of commiserating with each other how, when we said what we did--he's a mathematician, I'm an economist--we often had to stand alone at parties after we admitted what we did. He said that he thought the explanation for why people didn't like our two disciplines was different. "Mathematics is incomprehensible," he said, "but economics just doesn't make sense." Now it's true that economics doesn't make sense to a lot of people. The problem here is, the only way to guarantee low prices was to allow sellers to charge high prices.
— Mike Munger on price gouging (selling ice after a hurricane in Raleigh, North Carolina)
One of the bones of contention that keep coming up between libertarians and progressives is minimum-wage laws. Are these good or bad? Relative to what alternatives? How do you know? And how do you deal with any uncertainty?
Most people seem to approach minimum-wage laws intuitively. They judge the laws based on the sentiments that appear to motivate the laws, and have neither the patience for, nor the trust in economic theory or statistics to go beyond intuition in thinking about the practical consequences. I can sympathize with people for being skeptical about both theory and statistics, especially when the consensus among the professionals seems weak (although my impression is that this has firmed up since Free to Choose was published). But if expert opinion is divided, and most people's intuition is opposed to most professional opinion, any reasonable person would have to acknowledge that the uncertainty level on the trustworthines of popular opinion in this situation is high.
The simplest argument against minimum-wage laws is that there is a better alternative: the Earned Income Credit. But first, I want to illuminate the uncertainties I'm talking about by comparing the minimum wage against the alternative of doing nothing.
In order to make this comparison, we have to know how prices are set. I am assuming that there are large numbers of independent buyers and sellers of unskilled labor, which puts us in the realm of "price theory," with competitive markets and well-defined supply and demand curves. On the other extreme, we could have a situation with only one seller and one buyer (a bilateral monopoly), which puts us in the realm of "bargaining theory." In between, we could have situations such as many sellers with only one cartel acting as a buyer (a monoposony). I think that in a country such as the US, with 310 million people and almost as many automobiles, it's a no-brainer that we are in the realm of competitive markets. Unfortunately, if evolutionary psychologists are correct, the human brain evolved in a very different environment, the "Environment of Evolutionary Adaptation" or EEA, which existed some 100,000 years ago, with people living in tribes of about 50 nomadic hunter-gatherers. In the EEA, bilateral monopolies were the rule, not the exception, so most people intuitively believe in monopolies rather than competitive markets. If you don't think about it, I am afraid you will end up being vaguely dissatisfied with price theory-based arguments, without being able to articulate why. In order for what follows to be cogent, you need to be satisfied that the price of unskilled labor is set by a competitive market, not by a cartel. (See Bryan Caplan's "anti-market bias." Paul Bloom also argues that the human brain is predisposed to believe that any unexplained phenomenon is a deliberate act by some conscious being.)
If we do nothing, the price of unskilled labor is the "market price," set by the intersection of the supply and demand curves. The "producers' surplus," the benefit to the workers of selling their labor, is the shaded area on the accompanying figure between the supply curve and the horizontal line at the market price. On the other hand, if there is a minimum wage, the producers' surplus is the other shaded area. The price goes up, but the quantity of labor purchased goes down. Which area is bigger? It depends on the shapes of the supply and demand curves. The way I drew this figure, it looks like the minimum wage helps, but by changing the shapes of the curves, I could just as easily have made it come out the other way. If the supply curve is nearly flat and the demand curve is steep, minimum-wage laws tend to look good. If the supply curve is steep and the demand curve is nearly flat, minimum-wage laws tend to look bad. In other words, I don't know, and I strongly suspect that you don't, either.
There are other effects of minimum-wage laws. Some of the goods that poor people buy are made by other poor people, so they experience some price increases. They also lose some entry-level opportunities. Part of the problem is that minimum-wage laws often make employer-sponsored training programs prohibitively expensive. Milton Friedman described the effects of the minimum wage as "sawing the bottom rungs off the economic ladder."
It's a no-brainer that if the minimum wage is sufficiently low, it has no effect. It's also a no-brainer that if the minimum wage is sufficiently high, if will cause a prohibitive level of unemployment among the people it is supposed to help, and clearly will be a bad thing. For example, Thomas Sowell says that minimum wage laws encourage unemployed youths to deal drugs. Is there a happy medium? And if so, how do we figure out what it is? What happens if we guess wrong? It should be obvious from the above chart that if the demand curve is fairly flat, and the minimum wage is set too high, it will cause massive unemployment among the people who are supposedly being helped.
One shortcoming of this chart is that it does not distinguish between different groups of unskilled workers. If one group of workers, such as unskilled black teenagers, is regarded as less productive than the average unskilled worker, under a minimum-wage law, they can no longer compete by accepting lower wages than their competitors. The unemployment caused by minimum-wage laws will therefore fall heaviest on the least-preferred classes of workers. (Similarly, Walter Williams discusses the effects of government price-fixing on behalf of white labor unions on black employment in the railroad industry in The State Against Blacks. Laws that purport to be anti-discriminatory often have the exact opposite effect.)
The best way to sabotage chances for upward mobility of a youngster from a single-parent household, who resides in a violent slum and has attended poor-quality schools is to make it unprofitable for any employer to hire him.— Walter Williams
One concept that often comes up in discussions of the minimum wage is the so-called "living wage." As near as I can tell, the idea here is that there is some level of income where human happiness falls off a cliff and hits bottom, and it is desperately important that certain people's incomes be a hair more than this rather than a hair less, even if other people who are even worse off get totally screwed. I illustrate this with a graph of a pair of hypothetical utility vs. income curves. Normally one would expect a utility vs. income curve to be concave down, reflecting the decreasing marginal utility of income. The "living wage" theory assumes that it is concave up near the bottom. This would make sense if, below the "living wage," people either starve to death or are so miserable that they commit suicide, conveniently enabling the proponents of the minimum wage to ignore any further harm that they do these people. There may in fact be some people who would rather slit their wrists than move back in with their parents, but I venture to believe that there are many more who would scrape by with multiple roommates, or even if they did have to live with their parents, would appreciate the ability to earn at least a little bit on the side.
What I think is actually happening is that the proponants of minimum-wage laws like to visualize the good that they are trying to do, and prefer not to think about the harm this entails. The purpose of the utility vs. income graph is to force people to think about the harm.
Given the uncertainties in the slopes of the supply and demand curves for unskilled labor (aka "elasticity"), and in the shape of a typical person's utility vs. income curve, the central question about minimum-wage laws becomes, "How do we make decisions under uncertainty?" I want to separate the tangible consequences from the intangible ones. Roughly, these are the path-independent or "physical" vs. the path-dependent or "moral" ones. The tangible consequences are the ones that show up on the utility vs. income curve. The intangible consequences are more subjective and (legitimately) controversial. If the tangible stakes are high enough, my approach to decision-making under uncertainty tends to be to guess at the probabilities of the various tangible consequences and treat the problem probabilistically (as in decision-making under "risk" rather than under "uncertainty"), perhaps with some bias against risk ("risk-aversion"), ignoring the intangible consequences. But if the practical uncertainty is high and the intangible consequences seem relatively important, my bias would be to try to commit "sins of omission" rather than "sins of commission." Again, I think I can justify this partly in utilitarian terms, which any reasonable person should accept, but it is also partly a matter of my "libertarian" temperament.
Sins of commission include a number of intangible costs. These include decision-making costs. Congress has a crowded agenda, and it costs something in terms of the "attentive voters'" capacity to provide Congress with adult supervision. But there is also an intangible cost associated with coercion. It costs something in terms of "the moral fabric of society" (ie. "goodwill") to stick a gun up someone's nose. Reasonable people will usually understand and forgive if there is a compelling reason for the use of coercion, but if the rationale for coercion is thin, and uncertainties are not honestly acknowledged, understanding and forgiveness will be harder to come by, and the cost in terms of goodwill will be higher. Given the uncertainties associated with minimum-wage laws, even if the only available alternative is to do nothing, I still think a reasonable person should oppose them.
The Earned Income Credit (EIC) makes a lot better sense to me. Instead of trying to transfer money to low-wage workers, simultaneously harming the people who employ them, the EIC subsidizes low-wage workers out of general funds, distributing the cost over taxpayers in general. This causes the market to err on the side of employing more than the "efficient" amount of unskilled labor rather than less. This is clearly better for the workers in question.
Other welfare programs could conceivably be combined with minimum-wage laws in order to mitigate the ill effects, and this might convince me that the net effects were beneficial, but I think the EIC makes better sense.
But the EIC shows up in the government's budget. I think this is good. This is "transparency." But as Robert Frank suggested, transparency may be one of the reasons why the political process works against it. Some voters may be willing to help the poor as long as they believe someone else is paying for it, but if they are forced to think about paying for charity out of their own pockets, they get cold feet. This is especially true if the motivation is wanting to "soak the rich" as much as wanting to help the poor.
Another reason why the EIC tends to lack political support is Bruce Yandle's Bootleggers and Baptists phenomenon. As with alcohol prohibition, some of the political support for the minimum wage is high-minded ("Baptists") and some of it is nefarious ("bootleggers"). High-minded supporters of the minimum wage support it despite its tendency to suppress businesses that depend on cheap labor, but special interests will often support it precisely because it suppresses their competitors. As then-Senator John F. Kennedy put it in 1957,
Of course, having on the market a rather large source of cheap labor depresses wages outside of that group, too—the wages of the white worker who has to compete. And when an employer can substitute a colored worker at a lower wage—and there are, as you pointed out, these hundreds of thousands looking for decent work—it affects the whole wage structure of an area, doesn't it?
This point was not lost on progressives of the early 20th century. Jonah Goldberg, in Liberal Fascism (p. 264, in his chapter on eugenics), quotes Sidney Webb:
Of all ways of dealing with these unfortunate parasites, the most ruinous to the community is to allow them unrestrainedly to compete as wage earners.
and Royal Meeker (advisor to Woodrow Wilson):
Better that the state should support the inefficient wholly and prevent the multiplication of the breed than subsidize incompetence and unthrift, enabling them to bring forth more of their kind.
As Thomas C. Schelling commented in the lead essay in Choice and Consequence,
Minimum-wage laws are thought to have ethical content. But if their main effect, or even their purpose, is to keep the young and the old and the otherwise least valuable employees from working at all, the ethical issues may not be what the proponents thought they were.If you are a high-minded supporter of the minimum wage, consider the possibility that you may have been pwned.
Other articles on minimum-wage laws that you may find worth reading are some by Megan McArdle at The Atlantic
But over the short term, price controls can look like they are working (for some values of the word "working"), because price elasticity is almost always much greater over the long run than short term. The real reason that minimum wages haven't been particularly worrisome in the United States is that there hasn't been much of a long run over the past few decades; inflation and GDP growth have eroded its value before it got too onerous. Of course, supporters of the minimum wage are working very hard to change that.
Whether or not you think that the government ought to be in the business of transferring wealth from one segment of society to another, I hope we can all agree that at least the transfers oughtn't to go upwards.
Most noneconomists believe that minimum wage laws protect workers from exploitation by employers and reduce poverty. Most economists believe that minimum wage laws cause unnecessary hardship for the very people they are supposed to help.
32. Appendix G: Border Control and Immigration
Another bone of contention that has come up a lot lately (6-25-2012) is immigration. This topic has libertarians siding with progressives against conservatives on what are presented as economic issues. Roughly speaking, it puts libertarians in opposition to classical liberals. It is also a topic that I think the libertarians mostly get wrong.
As I suggested in chapter 1, when making decisions under uncertainty, it can be hard to separate practical arguments from moral ones. I believe that the limitations in libertarian moral theory, combined with the difficulty of separating moral arguments from practical ones is a particular problem for libertarians in discussing immigration. Consequently, I will spend more time here discussing libertarian moral theory than I otherwise would.
Suppose that you have land along the Canadian border, and would like to invite your neighbor across the border over for a little pond hockey. Do I, as your fellow citizen, have a moral right to interfere? If he were a US citizen, surely the answer would be no. But what about your Canadian friend? From a moral standpoint, why should we treat foreigners differently from our fellow citizens?
Radical libertarians typically argue that we should not treat them differently. Examples include David Friedman, Bryan Caplan, Alex Tabarrok, and Don Boudreaux.
The obvious libertarian moral view is that, if I forcibly prevent the arrival of the visiting pond hockey squad, I am initiating coercion, which is immoral. I might make a number of counterclaims:
Any of these claims are plausible to me, but I find the third claim the most interesting, because it not only allows the government to take necessary action, but promises to shed light on what actions are morally permissible. If libertarian moral theories are to make any sense in discussing immigration or national defense, at some point we are going to have to acknowledge the government as having some sort of moral standing beyond what libertarians are normally willing to allow. But let's set these arguments aside for now. Suppose that I decide that your visitor is honest and harmless. What comes next?
If your hypothetical Canadian neighbor also wants to leave your private property and drive on a public road, there are secondary moral arguments involving "crowding" or "rivalrous" public goods and services that I and my neighbors provide or collectively own. Even if I think he is thoroughly trustworthy, I may not want to share the road with him during rush hour without compensation. Regardless of whether or not I approve of the welfare state or various social safety nets, I may begrudge him for imposing on them (making him an accessory, in so far as I may find these programs morally objectionable). For sake of argument, I am making a huge, implausible leap of faith here in assuming that issues related to our social safety net can be resolved politically and that the resulting laws will be enforced honestly. Even so, in so far as a nation is a sort of club, with some property owned privately and some owned collectively, the members of the club are within their rights to refuse to share the club's collective resources with non-members, or to charge access fees.
Once a foreign visitor has given us satisfactory compensation for any public resources he is using, the next question is whether he has a right to look for work. Perhaps he has valuable and scarce skills, or perhaps he is unskilled and desperately poor. There is usually more controversy over the latter case, so let's focus on that. If he wants to apply for a job flipping burgers, do I have to let him? This is one of Bryan Caplan's pet issues. Caplan's answer is a vehement "Yes!"
Western moral teachings involve a certain tension between "universalist" sentiments (the Good Samaritan) and "particularist" ones (my tribe are God's chosen people and your tribe aren't). Caplan is really making two arguments here, a libertarian argument and a universalist argument. They tend to go together, but they are logically distinct. The libertarian argument is that I may not initiate coercion. A libertarian may boycott, but may not coerce. The universalist argument is that I shouldn't boycott the poor foreigner, I should rejoice on his behalf, even if it's at the expense of my relatively well-to-do fellow citizen.
As long as the law is equally binding on everyone, and enforced effectively, I have to agree with Caplan that forbidding visitors from working would be initiating coercion, and hence immoral by libertarian lights. In some cases, part of the attraction of illegal alien workers may be that they give employers a means of evading labor laws that are binding on legal guest workers and citizens; if these laws do not bind equally, someone is being treated unjustly. But in the present context of asking what the law should be, we are considering people who are here legally, and so I don't have this excuse for interfering with guest workers.
Caplan's universalism is another story, though. Do I care equally about all people everywhere, or at least all strangers, citizens and non-citizens alike, or do I have concentric circles of moral concern, with people in the inner circles having stronger claims on me than those in the outer circles? I tend to see the world more in terms of the concentric circles and Burkean "little platoons." Extreme universalism is too romantic for my tastes. Libertarianism certainly doesn't imply universalism. My defense of modified utilitarianism (chapter 3) is more likely to be cogent for you if you feel generous toward your fellow citizens, but it doesn't require even that level of generosity. As David Friedman pointed out in "What Does Optimal Population Mean?" when considering utilitarianism, we have to be careful in specifying the set of people whose interests we are comparing across different policies. The relevant population in my bargaining theory argument are US citizens.
Still, the principle of non-initiation of coercion gives us a moral argument for nearly open borders. What rationale might I offer for objecting to an open borders policy for guest workers?
I still don't like the idea of a deus ex machina "easement." If I'm going to back off of libertarian moral theories, I want to do it openly.
I could say that I care about my tribe (particularism) so much that I am willing to sacrifice libertarianism in order to promote "us." This means either dropping the claim that I am acting morally or rejecting the individual as my moral unit of analysis in favor of tribe, culture, nation, religion, or some other grouping. Perhaps I might decide to observe libertarian principles within my group but not across groups.
Another approach would be to plead necessity. Why might I do this? Steve Sailer explains,
We live in a world where violence — perpetrating it and preventing it — is the fundamental fact that social and political organization must deal with.
Thus, all property rights come out of the barrel of a gun.
Once you realize that, the reason why we prefer the welfare of our fellow citizens to that of non-citizens is (to get all reductionist):
They are the ones who would fight on your side.
At least I hope they would fight on my side, although much of modern US politics seems to resemble the Battle of Manzikurt.
As I put it in my third counterclaim, above, libertarianism makes sense "only within the context of some basic level of trust in both the sovereignty of the government and in the government's commitment to protect some reasonably-defined set of its citizens' rights." In a democracy, the government's commitment to defend my property rights depends on the goodwill of my fellow voters. In a society which has been steeped in Progressive ideology for a hundred years, that trust is hanging by a thread. I have strong libertarian leanings, but most of my fellow voters do not. I don't trust my fellow citizens to bear very much temptation. I'm ambivalent here. I am not yet persuaded that a more generous policy toward guest workers would be a bad thing, but I am open to persuasion, and the numbers very much do matter. Necessity is the same reason why limited government libertarians accept taxation in a world in which we need to provide for national defense. If I have to offend my moral sensibilities in order to preserve civilization, so be it.
The next question after whether to allow visitors to become guest workers is whether to allow them and/or their children to become citizens, complete with the right to vote. I might take a universalist attitude towards a potential guest worker population, but in the case of a potential immigrant population, I have a much greater interest in whether this newcomer population has the same attitude towards me. Also, due to an unintended side effect of the Civil War, at least as the 14th Amendment is usually interpreted, it is easy for children of foreigners to acquire US citizenship: all they need to do is for the mother to be here, legally or not, when the baby is born (birthright citizenship). Open immigration proponents may argue that, even if first generation immigrants are spiritual denizens of the third world, their children or grandchildren will assimilate to the political culture of James Madison. This is where the primary controversy lies between the pro- and con- immigration debaters.
Carter Van Carter writes,
I notice (via Mangan) libertarian economist Don Boudreaux thinks the Janjaweed have the right to move en masse to your hometown. And not just the Janjaweed, but also the Tamil Tigers, the Lord's Resistance Army, Hezbollah, the Interahamwe - and many more, as the saying goes.
Many, many, more, if Boudreaux had his way, because he believes everyone from everywhere has an inalienable right to move en masse to America - this in a world where more than 3 billion people live on less than $2 a day. I've described libertarianism as applied autism, I'm now beginning to realize what a terrible aspersion on autistics that was.
How fast will the Janjaweed assimilate? Do the numbers matter? Carter Van Carter and Victor Davis Hanson are clearly concerned with street crime. This is not a trivial problem. Recall Robert C. Ellickson's study of the role of formal law in Shasta County from chapter 17. The first line of defense in maintaining order is the court of public opinion. The police serve as a backstop to culture. A culture that tolerates crime becomes extremely difficult to police. Immigrants such as the ones Victor Davis Hanson describes don't have to vote in order to be a problem. But I am even more concerned with what happens when immigrants from the third world and their children get the right to vote.
What does libertarian moral theory tell us about the right to vote? John Locke wrote that my natural rights included "life, liberty and property," and Gideon J. Tucker observed that "No man's life, liberty, or property are safe while the legislature is in session." David Friedman defines governments as "agencies of legitimized coercion," and takes the view that there is no such thing as legitimate coercion. Regardless of Constitutional theory, in practice, Tucker is indisputably correct. In practice, your having the right to vote means that you and your peers have the power to arbitrarily violate my rights, with only temporary impediments in your way. Don Boudreaux makes a point of not voting (perhaps to draw attention to the paradox of voting?) but none of these libertarian open immigration advocates, to my knowledge, has addressed the moral basis for assigning voting rights, nor advocated creating a large permanent class of resident non-voters (although the Objectivist, Harry Binswanger, seems to imply the latter). So, Objectivists excepted, we are talking here about massively increasing the number of people who have the right to vote in US elections. (And are the Objectivists really prepared to argue in favor of a US with 300 million citizens and 3 billion permanent resident aliens? Can we at least do something about birthright citizenship first?) How did libertarianism go from denying the US government's right to exist, to heralding the inalienable rights of the Janjaweed to participate in it and direct its course?
If we define libertarianism as strict adherence to the principle of non-initiation of coercion, we can't use libertarian moral theory to defend the government's existence, let alone specific policies. This includes the right to vote. But practically the whole point of libertarianism is to provide guidance for voters and government officials. We either need to write libertarian moral theory off as pointless or as self-contradictory nonsense, or else we need to define it more broadly, so we don't have to throw it away just as it becomes necessary.
My approach has been to treat it as a heuristic for utilitarianism. The Objectivist approach seems to be to plead necessity and retreat to an endorsement of the "night watchman state," but no further. A third approach might be to regard the government's job as that of maximizing overall liberty, even though this means getting one's hands dirty and initiating coercion more than is strictly necessary for defensive purposes. Perhaps "neo-con" foreign policy might be close to this, although I don't know if it is well enough defined for anyone to be able to say for certain what it means.
I think part of the problem with libertarianism is the failure to make a clean enough break with classical liberalism. Locke and Jefferson couldn't decide whether the government owed its legitimacy to its defense of natural rights or to majority rule. But my having a natural right to life, liberty, and property cannot be compatible with your having a natural right to dispossess me of it. For a libertarian, the correct answer should be defense of natural rights. I have to agree with Ayn Rand on this issue: the right to vote cannot be a fundamental human right in the sense that life, liberty and property are. Democracy may be a means of constraining a government to tend to its legitimate business, but it cannot be the source of the government's moral legitimacy. Majority rule is at best a derived, or secondary principle, a tool that a society uses in order to restrain its officers from overreach. This is why Leonard Peikoff makes such a point of distinguishing between a democracy and a republic.
Mencius Moldbug presents a "reactionary" view of voting. He links to Senator Benjamin H. Hill, who writes,
Suffrage, then, is not a right—it is not a privilege—it is a trust, and a most solemn and sacred trust. It is the trust of preserving society, of securing rights, of protecting persons.
I also want to point out that these trustees are not charged with the responsibility to do missionary work. They are only responsible as trustees to their fellow citizens, and if they are immigrants, to their new fellow citizens.
At the very least, if we want quasi-libertarian moral ideas to shed light on how voters and governments should behave, we need to allow a "night watchman state." We need the state to be able to provide security. What does this imply for border control and immigration?
I'm back to my earlier claim: libertarianism makes sense "only within the context of some basic level of trust in both the sovereignty of the government and in the government's commitment to protect some reasonably-defined set of its citizens' rights."
How, then, do we deal with uncertainty and trust? Am I morally obligated to trust you? From a practical standpoint, can I trust you? The Binswanger article acknowledges the right and need to keep out criminals and terrorists, but how do we identify them? Who should bear the cost of my uncertainty regarding your likely future behavior? In domestic law, we have the principle of innocent until proven guilty; in criminal law, guilt needs to be proven beyond a reasonable doubt, and in civil law, the standard is a preponderance of evidence. We also want our courts to be reluctant, but not infinitely reluctant, to initiate coercion by issuing restraining orders. I note that I am more inclined to want our courts to issue restraining orders if the person in question is judgment proof. I also note that the 3 billion people Carter Van Carter says are living on less than $2 a day would almost certainly be judgment proof.
My answer is that I don't feel that I have an inherent moral obligation to trust anyone, but there are practical reasons why I have to trust some people to varying degrees. Also, we want our collectively hired agents to be systematic about how they treat their employers, so we are forced to oblige these agents to treat all of their employers with a certain amount of deference, despite our frequent and well-justified misgivings about many of our peers. I have practical reasons, but not necessarily moral reasons, why I want my agents to bend over backwards to show that they are not abusing their employers. These practical arguments do not necessarily extend to outsiders.
Roughly, I see three levels of trust:
In an ideal world, my fellow citizens would be a subset of the people I really trust. In practice, they are not. This has gotten bad enough in the US that both Ayn Rand and Neal Stephenson have described the situation as "a cold civil war." In particular, US immigration policy seems to be driven by what is effectively stock dilution, the desire by political elites to import new voters who will be sympathetic to their attempts to violate the property rights of their political opponents, or at least, more receptive to the politicians' appeals to "identity politics" (aka racism) and attempts to buy votes with taxpayers' money. At the local level (e.g. Boston), Edward L. Glaeser and Andrei Shleifer call this "The Curley Effect;" David Henderson undermines his own position on immigration by observing the same phenomenon in California (as Mencius Moldbug does in Detroit):
James Michael Curley, a four-time mayor of Boston, used wasteful redistribution to his poor Irish constituents and incendiary rhetoric to encourage richer citizens to emigrate from Boston, thereby shaping the electorate in his favor. Boston as a consequence stagnated, but Curley kept winning elections.As Peter Brimelow adapts Bertolt Brecht, "the government is dissolving the people and electing a new one." At best, the politicians are pandering to ethnic bigots among the current voting population. In any case, it seems to me that, both morally and practically, the libertarians' and Objectivists' stereotypical suspicion of democracy is entirely justified.
Harry Binswanger at least deserves credit for distinguishing between (non-voting) residency and (voting) citizenship (although he is silent on when we should start trusting someone with a ballot). Bryan Caplan seems utterly cavalier about political risk. Here's a particularly egregious example. Curiously, the word "assimilation" doesn't show up in any of Caplan's blog posts, except in readers' comments (as of early 2012). The far more moderate Megan McArdle, to her credit, does discuss it. Caplan does talk about assimilation and culture in this EconTalk podcast, but he trivializes it. He talks about "culture" in terms of baseball and opera, and TV shows like Friends and Seinfeld. This is the same trivialization that I complain about Unitarian Universalist ministers doing here. In contrast, a friend of mine reports being asked by a visitor from the Philippines what is an appropriate bribe to offer to a policeman at a traffic stop. That's what I think is an important cultural question. Similarly, what if I'm called to court as a juror or perhaps a witness to a crime committed by a member of my family or my ethnic or religious group? Which comes first, friends and family, or abstract principles of justice? That's a cultural question. So is my attitude towards honesty and self-deception. Am I more likely to count you as a friend if you let me get away with face-saving nonsense or if you don't?
As James Madison wrote in The Federalist no. 55,
As there is a degree of depravity in mankind which requires a certain degree of circumspection and distrust, so there are other qualities in human nature which justify a certain portion of esteem and confidence. Republican government presupposes the existence of these qualities in a higher degree than any other form.
Madison wrote "human nature," but culture is crucial here. How quickly and effectively does the process of assimilation work, and what are the parameters that govern it? Part of the argument between the pro-immigration (e.g. Caplan) and con- (e.g. John Derbyshire) camps is about whether to lump various groups of immigrants together for statistical purposes, or to distinguish between, say, Indian graduate students and La Raza activists. Do you look only at the first generation, many of whom are illegal and trying to avoid drawing the government's attention, or do you look at the children and grandchildren? Do you assume that recent immigrant populations and the social environments they find here today are statistically similar to previous populations and the environments they found? Note that La Raza is receiving a government subsidy. Even if these immigrant populations are statistically similar, do you assume that in the modern world of cheap global transportation, they would continue to be statistically similar if our immigration policies were made radically more generous? Have previous waves of immigration really gone all that well when it comes to preserving classical liberalism? One of the recurring themes on Russ Roberts' EconTalk is that it's very easy to cherry-pick social statistics, even if you don't intend to.
For example, John Derbyshire has a nice rant (transcript) about statistical legerdemain in the context of minorities and public education. He warns elsewhere that the government cannot be trusted to report crime statistics honestly where ethnicity is concerned.
I'm picking on Caplan in particular because he, of all people, as the author of The Myth of the Rational Voter (my review here), should be concerned about perverse behavior by voters. If you come from a dysfunctional third-world country, and I let you play on James Madison's playground equipment, are you and your children well-intentioned enough and wise enough not to break it? Caplan writes (p. 124), "One interesting prediction of rational irrationality is that fluctuating incentives make people bounce between contradictory viewpoints." For example, people in poor countries will try to emigrate to rich countries, but will vote against adopting the policies that make countries rich. The private choice is decisive. The public choice is not. But on his blog, Caplan denies that this is a problem. He claims to have answered his critics, but from my reading of him, all he has done is argued that open borders would be good for foreigners, and even that is only probable in the short term. Do the proponents of open borders have a scientific model of assimilation? What differential equation are they using to predict the rates of assimilation of various groups of immigrants? How sensitive is this model to the respective population sizes? Do we have a data base on the effects of open border policies in developed countries that is relevant in the modern world? My impression is that Caplan is strongly attracted to ideas that shock people, and that his practical defense of open borders is a potential textbook example of confirmation bias.
When libertarians do try to make government policy suggestions, we encounter two related problems: slippery slopes and ambiguity about whether or not the people we are advising actually have enough control to be able to implement our suggestions. There is one set of policies that I might choose if I were a benevolent despot. There is another set of policies we might like to see enshrined in an inviolable constitution for a country in which we expected to be part of an unpopular minority. In either extreme case, we either certainly do or certainly do not have control over day-to-day government policy. The reality is that the best we can hope for is a little bit of influence rather than control. Our political and educational efforts have some small probabilities of tipping individual legislative decisions, with no real hope of enacting coherent packages, and we are, or ought to be, worried about things that lower the odds that our actions will be decisive. Thus we need to worry about the sorts of things Eugene Volokh wrote about in his aforementioned paper on slippery slopes. We also don't have a useful theory of rhetoric, of how to persuade voters and whether they are persuadable. Instead, we tend to treat votes mechanically, either as random variables or as determined by public choice theory. Neither of these approaches is realistic. They tend to conceal rather than illuminate slippery slopes.
By "slippery slope," I mean that if the government adopts one policy today, perhaps with my support, this policy will cause economic or social changes that increase the odds that the government will adopt a related policy in the future, perhaps without my support. Libertarians are certainly aware of slippery slope problems. It is only a slight exaggeration to say that libertarians are terrified of slippery slopes. They're the basis for David Friedman's argument in favor of anarcho-capitalism. Libertarians sometimes criticize socialists for not appearing to understand this. As Caplan's fellow EconLog blogger Arnold Kling puts it,
I think that (non-classical) liberals and libertarians see the problem of "special interests" differently. Liberals view special interests as exogenous to the policy process. You have to overcome special interests to create good policy. Libertarians see special interests as endogenous. Policy is what creates them.
But this awareness hasn't translated itself into useful suggestions for government policy. David Friedman invokes fear of slippery slopes to justify anarcho-capitalism, but ignores the problem when discussing immigration. This is particularly odd in light of the fact that in chapter 44 of The Machinery of Freedom, 2nd edition, he argues that medieval Iceland "was subverted by an alien ideology--monarchy." This was the result of too many culturally unassimilated immigrants from Norway.
Step back for a moment and consider the question of whether diversity is good or bad as a general principle. Economists typically take political stability (including secure property rights and the rule of law) for granted, and argue that greater diversity leads to greater gains from trade. (On the other hand, fragmented markets tend to undermine economies of scale.) Political scientists do not take political stability for granted, and worry that heterogeneous populations (e.g. Hutus and Tutsis) are more likely to fight civil wars. Diversity is the reason why Singapore does not have trial by jury. Alex Tabarrok is clearly aware that civil juries are making the practice of obstetrics extremely expensive in Philadelphia. The George Zimmerman trial seems to be almost verbatim out of a Tom Wolfe novel. It's hard to talk about the O. J. Simpson trial without discussing the race and gender of the jury. The Duke lacrosse team trial was also largely racial theater. John Derbyshire emphasizes the political science argument against diversity per se.
I go further than that. Following James Madison, I want to live in a classical liberal society, and I believe that in a democracy, only certain cultures are compatible with this. If I had a choice of living in a society containing a mixture of English, Japanese, and Indians, or a homogeneous society populated by Pashtuns, I would choose the former.
Caplan actually goes so far as to argue that ethnic goodwill is a bad thing because it leads to socialism. He attributes the supposed stinginess of the US social safety net to ill will between US whites and blacks. He thus defends open immigration as a good thing because it increases diversity and hence ethnic strife. I think this is nonsense on stilts, and that he is making a virtue of a vice. I think socialism is a combination of a pretext for looting and an exercise in sanctimony theater. I think socialism has more to do with the breakdown of traditional religion than with lack of diversity.
Recall the discussion of what Mancur Olson called "selective incentives" from chapter 16, or sociological "glue." Rationality is not enough. If we want to analyze the risks associated with slippery slopes, we need to have a better handle on the special kind of goodwill or honor that can motivate a voter to act as a faithful trustee rather than a looter. In other words, we need to fill in the "big human-nature-shaped hole" that John Derbyshire finds in Virginia Postrel's political science schema. We need to take culture seriously, even if it means admitting that the "reactionaries" were right all along.
Also, I'm serious in wondering about the effects of past waves of post-revolutionary immigration. How do we measure scientifically a person's suitability to serve as an unsupervised trustee of other people's lives, liberty, and property? Is that part of the story of the growth of the US government? Bruce Tufts worries about reverse-assimilation:
If the goal is exporting the behaviors of the most successful capitalist nation, open borders is a poor way to do it - the most likely outcome seems to me to be an importation of woe that overwhelms the good: the assumption is that assimilation into US culture is a given (with the freedom loving you're-going-to-reap-the-benefits-and-downsides-of-your-actions-yourself school of culture as the target); this seems a quite over-optimistic basis for a national immigration plan. The risk is cultural dilution...resulting in loss of support for the very target intended for 'export'.
Again, where are the differential equations that predict the rates of forward- and reverse-assimilation? The numbers matter. What do the numbers have to be before the "importation of woe" (e.g. Victor Davis Hanson's experiences) overwhelms the benefits of Tyler Cowen's "cheap chalupas?" Are we already there?
Update, 7-14-2015: History repeats itself. (Ireland, 1167)
3-23-2016: And why am I writing this in English rather than Cornish? (King Arthur and the Saxons)
10-4-2018: Reverse-assimilation reminds me of an F. Scott Fitzgerald quotation:
First you take a drink, then the drink takes a drink, then the drink takes you.