The Effect of Tax-Deferred Individual Retirement Accounts on National Savings,
the Government Budget, and Social Welfare
Shinichi Nishiyama
Department of Risk Management and Insurance
GeorgiaStateUniversity
February 4, 2008
Abstract
How much would the introduction or expansion of tax-deferred individual retirement accounts (IRAs) increase the savings of heterogeneous households? How large would the cost of the tax-deferred IRAs be in the short run and in the long run? What are the macroeconomic and welfare effects of those tax-deferred IRAs? The present paper constructs a dynamic general-equilibrium overlapping-generations model with heterogeneous households to address the above questions on tax-deferred IRAs (traditional and 401(k)). Households in this model are heterogeneous with respect to age, working ability, and asset holdings (in IRAs and non-IRAs). Each household receives an idiosyncratic wage shock and mortality shock every year and choose its optimal consumption, labor supply (including the retirement decision), and savings.
The present paper first calibrates the model to the U.S. economy without tax-deferred IRAs (the “baseline” economy). The baseline economy is a steady-state equilibrium, which is assumed to be on the balanced growth path. Then, the paper introduces tax-deferred IRAs to the economy, solves the model for a new equilibrium transition path, and evaluates the macroeconomic and welfare effects of tax-deferred IRAs. Because the policy change reduces income tax revenue especially in the short run, the paper assumes that progressive income tax rates are adjusted proportionately so that possible costs of the IRAs are financed. Finally, the paper conducts several counter-factual policy experiments to show how the prediction of the model would differ depending on the assumptions on the IRAs: the progressivity of current-law income tax, the level of early withdrawal penalty, and the contribution limits.
Because tax-deferred IRAs are analyzed in a heterogeneous-agent overlapping-generations economy, the optimal size and allocation of the IRA savings are shown by age, income, and wealth levels. In addition, solving the model for an equilibrium transition path makes the paper evaluate the welfare change of each age cohort and the government cost both in the short run and in the long run. Regarding the latter, usual steady-state analyses cannot evaluate the transition cost of introducing tax-deferred IRAs.
A substantial literature has analyzed the possible effects of tax-deferred individual retirement accounts. However, most of those papers analyze the IRAs empirically or theoretically in a partial-equilibrium setting. There are some papers that have addressed questions similar to the presentpaper by using dynamic general-equilibrium overlapping-generation models, for example, Imrohoroglu, Imrohoroglu, and Joins (1994), Love (2006), Gomes, Michaelides, and Polkovnichenko (2006). Yet, these papers only analyze the long-run effects of introducing IRAs, due in part to the computational difficulty. The present paper solves the model for equilibrium transition paths, which are in general very time-consuming, by using a complementarity problem method andNewton’s algorithm, and it shows both the short-run and long-run effects of the policy change.
References
Gomes, Francisco, Alexander Michaelides, and Valery Polkovnichenko (2006), “Optimal Savings with Taxable and Tax-Deferred Accounts,” Centre for Economic Policy Research, Discussion Paper 4852, November 2006
Imrohoroglu, Ayse, Selahattin Imrohoroglu, and Douglas H. Joins (1994) “The Effect of Tax-Favored Retirement Accounts on Capital Accumulation and Welfare,” Federal Reserve Bank of Minneapolis, Discussion Paper 92, July 1994
Love, David (2006) “Buffer Stock Saving in Retirement Accounts,” Journal of Monetary Economics, Vol.53, pp.1473-1492
1