Pangloss was Right: Reform is Always Bad

Remember Dr. Pangloss? “‘Tis demonstrated that things cannot be otherwise; for, since everything is made for an end, everything is necessarily for the best end…. Those who have asserted that all is well talk nonsense; they ought to have said that all is for the best.” (Candide, p. 4).

Many economists have argued persuasively that the tax system that imposes the least deadweight loss is…whatever tax system we happen to have! The reason is that economic agents adjust their activities, and rewrite contracts, to optimize given the tax system that is currently in place. Changing the tax system, in any way, imposes large costs, costs that will almost certainly swamp the positive long term effects of “reform,” no matter how well intentioned.

This argument is equally persuasive for the reform of congressional institutions, only more so. The potential costs of reform are so large that they may not be measurable.

Consider three specific instances:
Principle 1: “An old tax is a good tax.”[1] Old policies are understood, their deadweight losses capitalized. Sure, these are losses, but they don’t matter at the margin once contracts account for them. Interest groups, citizens, and consumers develop expectations about the way that the policy process works, and gauge tax rates and regulatory costs in making investment decisions. Even if reforms “improve” welfare, in a static sense of approaching optimality, constant fiddling makes expectations diffuse. In short, predictable nonoptimality can easily be better for a society than widely varying, unpredictable reform efforts chasing optimal institutional arrangements.
Regarding the “good tax is an old tax” phenomenon, Buchanan points out:
In the most general terms, the appropriate analogue is the physical law of inertia. It is easier to continue a flow once started than it is to start it in the first place. All that is necessary for this point to be accepted as relevant for an individual decision calculus is some acknowledgement of a temporal sequence of choices.[2]

So the burden of the reformer has to be more than just static improvement. Reform has to be a permanent improvement, with no further action required. Otherwise, it is far better to maintain the current policies. Since reformers rarely have this kind of foresight, and think only of short term improvements, most reforms end up being fantastically expensive.

Principle 2: The “transitional gains trap.”[3] The government can’t even give anything away. Subsidies for an asset result in that asset being overvalued, compared to its market price. Costs are bid up, and owners end up making a normal return, just as they did before the “reform.” Only owners at the time of the policy change get any benefits, and even that gain is realized only if the policy is unexpected. All future owners get nothing.

But if the policy is ever changed back, all owners (those who gained, and those who didn’t) lose large amounts of wealth. Consequently, reforms designed to help a few people rarely accomplish that goal, and end up costing everyone. Ultimately, the costs last forever, because (by principle 1, above), it is cheaper to continue the bad policy than switch to the good one. It would be better still never to implement the reform in the first place, of course, but tell that to earnest young Candide!

Principle 3: “Cost illusion.” Costs of reform are analogous to “renters’ illusion,” the situation where renters underestimate the effect of real estate taxes because renters (unlike homeowners) don’t pay taxes directly. There is some debate about whether renters’ illusion is real[4], but “cost illusion” is rampant among the reformist followers of Candide. The costs of reform are generally imposed on specific sectors; since the reformers don’t have to pay anything, the reform is “free.”

For example, we are told that campaign finance reform requires that television stations give political candidates “free” air time. Well, it really is free to the reformers, but air time is expensive. Prohibiting stations from charges hardly makes it free; reform would simply shift the cost of campaign finance reform onto the stockholders of communication companies. Once thought of that way, the “reform” makes no sense. Why should the stockholders of the media corporations bear all the costs of political reform? What of the alternative, using public funds to pay for the “free” air time? That would be too expensive!

The flaw in this reasoning is obvious. If free air time is too expensive for everyone to pay for, it is certainly too expensive to extort from television stations as a condition of their license. The costs are the same either way.

And that is the rub: reforms are generally too expensive to pay for directly, so we disguise the costs of reform by focusing the expense on a small group, or by pretending the costs don’t exist. Reform, shmeform: Pangloss was right. In equilibrium, things are already for the best.

[1] James Buchanan, Public Finance in Democratic Process: Fiscal Institutions and Individual Choice, Chapel Hill: UNC Press, 1967, p. 58.
[2] Buchanan, ibid., p. 60.
[3] Gordon Tullock, “The Transitional Gains Trap,” Bell Journal of Economics, (1975): 671-678.
[4] For a review, see Wallace Oates, “On the Nature and Measurement of Fiscal Illusion: A Survey.” In Geoffrey Brennan (ed.), Taxation and Fiscal Federalism: Essays in Honor of Russell Matthews. Sydney: Australian National University Press.