Wednesday, April 11, 2007

Growth Theory and Supply Side Economics

Standard growth theory has implications for policy recognizably similar to the prescriptions of "supply siders".

Simple endogenous growth models tend to imply that investment income should be supplemented not taxed. This is, in fact, an implication of assumptions which nobody believes which are made for convenience. Relax the assumption that labor supply is exogenous and the implication of the model is "it depends" see "Can Waste Improve Welfare ?" The Journal of Public Economics. vol. 77 pp 45-79. I think this is true of all alleged implications of economic theory, which, in the abstract has no implications and which moves from the abstract to the concrete via assumptions which no one believes.

Robert Lucas, who knows his rhetoric, seems to have guessed that once one gets away from standard models with perfect competition, complete markets and 0 externalities, anything can happen. Thus he has always insisted on models in which the market outcome is best. It is easy to turn a case for a subsidy into a case for a tax. It is hard to turn a case for Lasciare Fare (sorry in my mind French becomes Italian) into something else.

The standard result in a Solow Swan exogenous growth model are similar to those in a Walrasian model: if there is no need for government spending, optimal taxes are zero and if income is perfectly equally distributed, the optimal tax is a poll tax. There is a further result due to Judd and (speparately) Chamley that, even if there is good reason to tax, the tax rate on capital income should go to zero in the long run. The logic is fairly simple, a tax on capital income is equivalent to a rising tax on consumption. If the economy goes to a balanced growth path (as is standard in the models and as fits the data pretty well) the optimal tax on consumption goes to a constant. This means that the optimal tax on capital income goes to zero as t goes to infinity.

It is standard in economics to assume that t has gone to infinity. This is the sort of reasoning that Keynes refuted when he noted that "in the long run we will all be dead." In fact, the implication of the Chamley Judd result for policy debate is the opposite of what it seems to be.

Imagine that we have rich capitalists and we want to redistribute income. C & J argue that taxes on the capitalists should go to zero. In fact, if there are now limits on allowed tax policies they should be zero at all positive t. The reason is simple. It is best to grab their wealth in one fell swoop (a capital levy equivalent to an infinite tax on capital income) as this is a lump sum tax which does not distort. If that is forbidden it is best to grab their wealth as fast as possible. If there is a limit on the maximum tax on capital income (say 100%) in the simple model it is best to tax capital income at 100 % until income has been redistributed as much one wants (given the trade off between total income and equality due to distortions which aredue to the 100 % limit). The models are simple and do not include any consideration of tax evasion or avoidance. Thus the tax to the max rule does not apply in the real world.

Thus, upon consideration, the implications of the most standard model is that Bush/Reagan type policies are totally horribly aweful. A tax cut with no spending cut is a tax shift. Eliminating, say double taxation of dividends, without reducing spending, makes it very likely that the path of taxes of capital income is low now high in the future. This is terrible policy.

Now if one does not give a damn about income distribution, one can just rely on a poll tax. However, if one does care about income distribution, standard theory says that it is best to redistribute as quickly as possible. That is a policy of gradual redistribution via a tax on capital income is Pareto inferior to an more rapid redistribution of equal magnitude.

Economists have no special insights into the question in ethics about the relative importance of equality and growth. However, the widely accepted arguments about the implications of standard growth models for the capital income tax can be summarized as follows : Bush and Reagan got it backwards.

1 comment:

Mark Thoma said...

Thanks for this.