بدهی های عمومی و مالیات مطلوب بدون تعهد

Benhabib and Rustichini [Optimal taxes without commitment, J. Econ. Theory 77 (1997) 231–259] study the properties of optimal capital taxes in economies without commitment and no government debt. They find that capital taxes may be different from zero at steady state. This note shows that, once governments have the possibility of issuing debt and smoothing taxes over time, optimal steady state capital taxes turn out to be zero.

The properties of optimal taxation of capital income are well known in environments with fullcommitment.Chamley [4] and Judd [5] showed that, when markets are complete, it is optimal to tax capital heavily in the short run when the distortions on capital accumulation are low and not to tax capital income in the long run when those distortions are large. This result is very robust and has been extended to many different scenarios. 1 However, as Chamley [4] suggests,this argument may break down in environments without commitment. This is so because without commitment future governments are tempted to renege on the announced policy plan and take advantage of a capital levy.This problem is examined by Benhabib and Rustichini [3]. They consider a model with competitive agents and a benevolent government that must tax capital and labor income to finance an exogenous stream of government spending in an environment without commitment. This government selects the optimal time-consistent taxes by choosing the policy that maximizes the individual’s welfare subject to the standard feasibility and implementability constraints and an incentive compatibility constraint that embeds the future governments’ incentives. This constraint says that, for all future governments, the welfare value of continuing with the policy that the current government announces must be at least as large as the welfare value of deviating from that policy. Any policy that satisfies this constraint is clearly time consistent.Benhabib and Rustichini [3] characterize the optimal time-consistent capital taxes. To do so, they first consider an economy with government debt. However, for simplification purposes, they restrict attention to economies where governments cannot issue debt. In this context, they obtain that optimal capital taxes may be different from zero at steady state. In particular, they show that capital may be subsidized. The intuition for this result is that a subsidy to capital accumulation leads to a very high stock of capital. This stock of private capital is so high that the fear to lose its productive gains deters future governments from deviating from its policy announcement and taxing its rents.This notes reexamines the problem of capital taxation without commitment in an economy with government bonds.We obtain that once governments have the possibility of issuing debt and smoothing taxes over time, optimal steady state capital taxes turn out to be zero. The explanation for this result lies in the large distorting effects of a tax/subsidy on capital income in the long run.As Atkenson et al. [2] observe, a positive (negative) tax on capital at steady state is equivalent to an ever increasing tax (subsidy) on consumption. That policy is so distortionary that is not optimal if the government has an instrument to smooth the cost of distortionary taxation over time. Our results have also an intimate connection with those of [3], we show that the issues of government debt during the transition help build a sufficiently high level of capital so that future governments have no incentive to deviate from the long-run zero capital taxes. Thus, in our economy, the subsidies to capital are substituted by government debt as the central commitment device among governments.The rest of the note is organized as follows. Section 2 presents the economy at steady state.Section 3 characterizes the optimal time-consistent taxes. Section 4 concludes.

Benhabib and Rustichini [3] showed that the optimal time-consistent capital taxes may be different from zero at steady state. This notes demonstrates that their result comes from the lack of both a commitment technology and an instrument to smooth taxes over time. Once governments have the possibility of issuing debt and smoothing distortionary taxes over time, steady state capital taxes are zero.This note is related to Phelan and Stacchetti [6]. They study the properties of capital taxation without commitment and no government debt. In particular, they compute the whole set of sustainable equilibria and show that if the incentive compatibility constraints are binding, then capital taxes may be different from zero at steady state. This note shows that, when we allow for government debt, those incentive constraints may not bind at steady state.We believe our results are quite robust and could be extended to more complex frameworks.As commented earlier, if the level of government debt is positive, then the optimal punishment after a deviation,namely, the worst sustainable equilibrium, would be generally characterized by a complete default on government debt. If that is the case, then the sole objective of the optimal management of government debt at steady state is to smooth completely the tax burden over time and, as we have shown, that requires a zero tax rate on capital income.