Hayek writes in Law, Legislation, and Liberty:
The preservation of a free system is so difficult because it requires a constant rejection of measures which appear to be required to secure particular results, on no stronger grounds than that they conflict with a general rule, and frequently without our knowing what will be the costs of not observing the rule in the particular instance. A successful defence of freedom must therefore be dogmatic.
So what is the general rule to be? There must be no infringements whatsoever on free exchange? Government should stick to the night watchman role? No. Hayek himself accepted numerous infringements on pure economic freedom. Hayek believes old classical liberal principles that once guided action no longer command assent, and we must find something to replace them: “The loss of [belief in principles] and the preference for expediency is in part a result of the fact that we no longer have any principles which can be rationally defended.” He suggests that principles such as “limit government to functions x, y, and z” or “laissez faire” are out, then.*
So what should replace them? Hayek, rather notoriously, never tells us. At least, I haven’t read all of Hayek, but in chapter four of Law, Legislation and Liberty, the source of all Hayek quotations and paraphrases so far in this post, the chapter where he calls for a defense of liberty based on dogmatic principle, he doesn’t tell us what that principle is to be. When I saw the “defence of freedom must be dogmatic” quote on the cafehayek blog, I went to check the context but accidentally looked at page 61 of The Constitution of Liberty (Hayek) rather than page 61 of Law, Legislation and Liberty (Hayek) and found him saying something strikingly similar, and again not providing us with the principle he called for. It seems to me he rather made a career out of this.
What’s needed, apparently, are principles that, unlike past prejudices, can be defended rationally, but not on the grounds that a particular violation will have predictable negative results. Such a principle must allow government to take desirable actions, but not undesirable ones, yet must not attempt to look at each action individually and say what its results will be. It must be justified ultimately on the grounds that it developed as part of an organic order beyond our understanding, so to violate it must have literally unimaginable consequences. Hayek recognizes his basic difficulty: it would be best on Hayekian grounds, he makes clear, to have mere prejudices, whose rationale we cannot and do not try to articulate, limit government action. But such prejudices have been overcome, the lack of articulated rationale being part of their undoing.
Part of the answer is for Hayek to come along and give an articulate justification for the concept of tacit knowledge, which was his great contribution. It now becomes rational to accept that much of our knowledge cannot be articulated. But that being said, what dogmatic principle are we to adopt? Should we go back to “government should only do x, y and z,” or should we now say “besides x, y, and z, government should now do a, b, and c, but should not do d”? By what principle do we exclude d? Is the caricature version of Hayek (food stamps lead inevitably to serfdom) right after all? Now that our organically developed prejudices have sadly been overcome, by what principle do we decide which principle to adopt?
If things are really as radically unknowable as Hayek sometimes, as in LL&L, ch. 4, I don’t think there’s any way out. You can’t tell which prejudices are desirable to keep because of their organic justification, and which ones are desirable to discard as part of further organic development. Even the prejudice in favor of rationally defensible principles may be justified as organically developed. And of course the attempt to articulate the principles behind the organic order becomes fraught with perils. In the course of this same LL&L, ch. 4, it takes Hayek from trying to find the organic basis for the previously existing free society that he saw us losing, to positing the need for a utopian ideal (!), which has never really existed, but which we need to guide us close enough to a free society; while, of course, not telling us what this ideal is to be.*
The solution, I think, is to depart from the extreme claims of our limited knowledge. For instance, we cannot know everything we need to know to determine the market-clearing price of an item, or to make a pencil. But we can know why we cannot know this, and we can know the consequences of price controls, or of attempts to build pencils through a centrally planned economy, or to collectivize agriculture. Price controls will bring about shortages, central planning will suffer from an insurmountable information problem and prevent resources from going to their highest-valued use;** thanks partly to Hayek, we know and understand these things. Protectionism and farm subsidies will create dead-weight losses.
More broadly, we can say that grand plans, comprehensive changes, are likely to fail for some unforeseeable reason. “But World War II! The moon landing! Yes! We! Can!” World War II was the failure of a grand plan, by the Germans and Japanese. It’s not that governments are so good at wars, it’s just that they’re going against other governments; someone has to win. There’s a reason we say “no battle plan survives contact with the enemy,” and why soldiers coined the phrases “snafu” and “fubar.” Ken Burns’s WWII documentary is full of really terrible mistakes by the Allies. Eisenhower, that wise and prudent statesman, had the bright idea that all the troops, everywhere, should be given a turkey, regardless of local knowledge. One group of men was overmatched against powerful German cannons, protected only by darkness, I think in a forest somewhere. The damn Huns were doing heavy damage, but they did not know exactly where they were firing. The major obviously did not want to set a fire to cook turkey, and asked his superior for an exemption; the superior said no, sorry, orders are orders. They set their fire, and the Germans shellacked them mercilessly.
The moon landing is a relatively simple thing- just go to the big rock, plant a flag, say grab some stuff, say something idealistic and go home. The physical engineering is I’m sure very complicated (nobody doubts the human mind is capable of feats of physical engineering and science), but the social engineering is nonexistent. The moon landing is just the type of pointless vanity project, like the pyramids, that government is perfectly capable of succeeding at if it can marshal the necessary resources. The aqueducts and the interstate highway system would are more useful instances of the same core competence.
There might not be a bright line between grand and non-grand plans, or comprehensive and non-comprehensive reforms, but I think it’s pretty safe to say that food stamps are on one side of the spectrum, ObamaCare is on the other. Philips Curve-style fine-tuning might be on the “grand design” end of the spectrum, or perhaps is an instance of micromanagement and tinkering that is open to the same basic objections, while a Friedman-style rules based monetary policy would be closer to the “create an order in which free people can operate” core competence of government.
Finally, our reason for rejecting certain government actions a priori, regardless of expedience arguments, need not be based on organic justifications at all. Hayek thought all moral principles were reducible to their own survival value, but we need not follow him here. The moral theory behind socialized medicine is, basically, that society rather than individuals should make decisions about medical spending. Rationing by well-intentioned experts is morally preferable to rationing by impersonal market forces. If you don’t believe that, you won’t believe in socialized medicine; if you do believe that, you probably will. All wonkery is subordinate to this basic ideological divide, and often gives only a veneer of factual basis to an argument based on speculation, narratives and ideology. Spending should be made efficient in terms of aggregate outcomes, so that society should not spend money on health care if the health benefits are not worth the costs. But it is individual human beings who experience monetary costs and health benefits- who experience at all, in fact. So the meaningful question is whether an individual’s health benefits are worth the costs to him, not whether they are worth the costs to society. And it is individuals, not societies, who have intrinsic worth and meaning and all that sort of thing- at least, so I claim. I’ll not justify that claim here, only note that it is the kind of a priori argument that one can make in opposition to certain state actions.
We might equally conclude that programs like MedicAid are justified. I suppose they involve interpersonal utility comparisons in the final analysis, but they don’t rely on calculations of social cost vs. social benefit. In other words, we see that a poor person is sick and treat medicine as an objective need, not an arbitrary preference.*** It’s true the government has to decide what the person “really needs,” but it’s not taking from them a choice they would otherwise have. The same basic logic applies to food stamps.**** There is no real attempt to calculate aggregate social costs and benefits involved, and the only real utility comparison going on is that a middle class person has less need for a very small amount of money than a poor person has for food and medicine. If government sets about to buy everyone’s medicine, or everyone’s food, or everyone’s shoes, or cover everyone’s retirement, we have a different situation.
* No, really. We shift from treating illiberal actions as erosions of a preexisting organic order, to the claim that “to some extent the guiding model of the overall order will always be an utopia, something to which the existing situation will be only a distant approximation and which many people will regard as wholly impractical. Yet it is only by constantly holding up the guiding conception of an internally consistent model which could be realized by the consistent application of the same principles, that anything like an effective framework for a functioning spontaneous order will be achieved. Adam Smith thought that ‘to expect, indeed, that freedom of trade should ever be entirely restored in Great Britain is as absurd as to expect an Oceana or Utopia should ever be established in it.’ Yet seventy years, later, largely as a result of his work, it was achieved.” So Adam Smith exposed the order of mercantilism to rational criticism, and got it overthrown!
** There’s value judgment involved here, obviously: in the USSR, those who benefited from the system were probably pretty satisfied with the allocation of resources. The value judgment is that subjective individual preferences are the standard of value. The economist’s insight is that voluntary exchange creates results that reflect individuals’ priorities and make these individual preferences part of a social system that exists to fulfill them. The “utilitarian” and “moral” case for free markets thus become identical, and hundreds of thousands of college libertarians can find a new hobby.
*** Here the claim about objective morality in the above paragraph (human beings have intrinsic moral meaning, by virtue of being human) overcomes the subjectivist claim in the same paragraph (individual preferences are the standard.) Claims about social costs and benefits are, however, thoroughly flawed in both conceptions.
**** Though Mayor Bloomberg tried to prove caricature Hayek right by banning the use of food stamps to pay for junk food.