|
Article Excerpt
This paper reexamines the impact that paying interest on reserves has on price level indeterminacy, volatility, and economic well-being. Unlike the previous literature, this model includes an after-tax deficit financed by assets (bonds and reserves) whose returns are linked. I show that the number of steady-state equilibria and the corresponding level of indeterminacy are equal to, or greater than, those arising in the no-interest economy. When the level of indeterminacy is the same, the economic volatility is reduced by paying interest. However, greater indeterminacy in the interest economy results in greater volatility. Finally, paying interest on reserves can enhance welfare. (JEL D6, E3, E5)
I. INTRODUCTION
The issue of paying interest on reserves was introduced by Friedman almost 50 years ago in A Program for Monetary Stability. Friedman's original motivation was to make the 100% reserve requirement of the "Chicago Plan" more palatable to a banking system subject to only a fractional reserve system. The goal of the Chicago Plan and the proposal to pay interest on reserves was to establish greater price level stability and to reduce excessive price level fluctuations. (1)
In the subsequent decades, there has been considerable research regarding the implications of paying interest on reserves. (2) Three studies, in particular Sargent and Wallace (1985), Smith (1991), and Freeman and Haslag (1996), have examined in detail whether Friedman's proposal would bring about the desired reductions in price level indeterminacy and volatility, as well as the welfare implications of switching from a system of not paying interest on reserves to one which did. (3)
However, these works suffered from two specific limitations. First, they did not equate the interest paid on reserves to returns on assets of similar risk and duration. Second, they assumed that the government either ran a balanced budget or had a surplus.
By assuming that the budget was not in deficit, these works sidestepped two important issues: (a) the impact that deficit financing has on the means for financing interest payments and (b) the complications that arise from simultaneously attempting to finance a deficit and set the real return on reserves. One of the key results of this previous literature was that how interest payments were financed was crucial to the likelihood of indeterminacy and volatility arising. However, if the sum of government expenditures and interest payments (on bonds and reserves) exceeds tax revenue, then the issue of how interest payments on reserves are financed (via taxes or earnings on assets) is no longer relevant. Instead, the appropriate concern is whether the government can simultaneously finance the deficit (by issuing bonds or printing money), link the return on reserves to other assets (such as bonds), and maintain sufficient returns on bonds and reserves such that both assets are desired by consumers.
The objective of this paper is to reexamine, in the presence of an after-tax deficit, the impact of switching from a system where reserves earn no interest to one where they do. This is accomplished in the context of a two-period overlapping generations model with multiple assets and an after-tax government deficit that must be financed by a combination of debt and seigniorage income. The primary goal is to compare the level of economic indeterminacy, economic volatility, and welfare gains in an economy where interest is paid on reserves to one where reserves earn no interest.
More specifically, this paper addresses the following three questions. First, in the presence of a government deficit and a return on storage that dominates all other rates of return, does paying interest on reserves reduce potential indeterminacy of equilibria? Second, under the same conditions, does the amount of economic volatility increase or decrease? Third, are there any welfare justifications for switching to a sys tem where reserves earn interest? In addition, given the presence of both debt and seigniorage in financing the deficit, the issue of unpleasant monetarist arithmetic is explored.
The key findings of this paper can be summarized as follows. When there exists an after-tax government deficit and reserves are paid a rate of return equal to that of bonds (and less than the return on storage), the number of steady-state equilibria (in terms of real money balances) is equal to, or greater than, the number that arise when no interest is paid on reserves. Thus, the level of economic indeterminacy is equal to, or greater than, that in an economy without interest payments. This runs counter to what Friedman had envisioned and the results of Smith (1991). In addition, the steady-state equilibrium associated with the highest level of real money balances is a source, while the steady-state equilibrium associated with the lowest level of balances is a sink. Any other steady states will alternate between being sinks and sources. For those steady states that are sinks, convergence is monotonic.
Second, when the number of steady-state equilibria is the same in the interest and no-interest economies (i.e., the level of indeterminacy is the same), the economic volatility is reduced with the introduction of interest payments. However, when greater indeterminacy exists in the interest economy, there also exists greater volatility. Third, when multiple (generically two) steady-state equilibria exist in both economies, the equilibrium associated with low real money balances in the interest economy is welfare improving compared to the no-interest economy. In addition, if the economy is converging to the low real money balances' steady state in the no-interest economy, then after switching to an interest-paying regime, the economy will transition to the low real balances' steady state in the interest economy. At the unstable steady state in the no-interest economy, if the government begins to pay interest, then a new equilibrium can be reached only if there is an accompanying expansionary open market operation. Finally, under a narrow set of conditions, unpleasant monetarist arithmetic may arise.
The key to the intuition behind these results is understanding the constraints that paying interest places on the means for financing the deficit. (4) In the no-interest economy, the government can price discriminate and choose whether to use more expensive (bonds) or less expensive (money) means to finance its deficit. In addition, it can decide whether to use a large seigniorage tax base (and small tax rate) or conversely a large seigniorage tax rate (and small tax base). Its decision regarding which instruments and what size tax rates to use is independent to the extent that in the end, the deficit must be financed.
When the return on reserves is linked to the return on bonds, the government's ability to price discriminate in its financing options is limited. Thus, financing the deficit via seigniorage and bonds is more expensive relative to the no-interest economy. In addition, with the return on money linked to bonds, the government's ability to exercise a trade-off between the tax rate and the tax base is curtailed. As a result, paying interest on reserves shrinks the set of real money balances that is consistent with financing the fixed deficit.
When there exist multiple steady-state equilibria in both the interest and the no-interest economies, the range of real money balances (the seigniorage tax base) will be smaller in the interest economy because the link between the returns on bonds and money reduces the government's options. This results in less volatility in the interest economy. However, because the government can still choose the initial level of real money balances (from an infinite set of possibilities), indeterminacy is unaffected.
The existence of multiple steady-state equilibria is contingent on the deficit-financing options (the quantity of bonds and reserves) being consistent with consumers wanting to hold all assets. Because the no-interest economy faces fewer constraints on financing its deficit, extreme values for tax bases and rates (e.g., a very small seigniorage tax base and a very high tax rate, i.e., high inflation) are more likely to be consistent with financing its deficit. However, while financing the deficit may be possible, the associated tax base and tax rate may not be consistent with individuals wanting to hold all assets (i.e., if inflation is too high, individuals will not want to hold money). In this case, the low real money balances' steady state in the no-interest economy is obviously not consistent with equilibrium. While there might be two candidate steady-state levels of real money balances, only the larger one is consistent with financing the deficit without violating the requirement that all assets earn a nonnegative return. Thus, under certain parameter settings, the interest economy will have two steady-state equilibria, while the no-interest economy only one. Obviously, in this case, both the level of indeterminacy and volatility will be greater on the interest-bearing economy.
Finally, it is assumed that both economies are Samuelson case economies, where savings (and hence consumption) are strictly increasing in the level of real money balances. Since the set of real balances consistent with equilibrium is larger in the no-interest economy, the steady-state equilibrium values of real balances are higher at the high real balances' steady state and lower at the low real balances' steady state than those of the corresponding interest economy. Thus, when comparing the low real money balances' steady states and dynamic equilibria converging to them, paying interest will be welfare improving.
The basic economic model used in this paper is a variation of Bhattacharya et al. (1998) and simply augments it with interest payments on reserves. The structure of the economy is as follows. The economy consists of an infinite sequence of two period lived, overlapping generations, where individuals across generations are identical in all dimensions. Consumers are endowed each period with a given amount of a consumption good which they either consume or invest. Individuals may invest their saving in any of the three different assets. There is a storage technology that pays the highest rate of return, government bonds, and money, whose return is dominated by all other assets. It is assumed that individuals cannot invest directly in the storage technology and that all investment in storage must be intermediated and is subject to a reserve requirement. Required reserves pay a rate of return equal to that of government securities. Thus, individuals save by purchasing bonds and depositing their savings with intermediaries.
In addition, there exists a government that must finance a constant per capita after-tax deficit while also paying interest on bonds and reserves. This deficit and interest payments are funded by some combination of money creation and new debt offerings. Finally, it is assumed that the government conducts policy by choosing (once and for all in the first period) a ratio of bonds to currency. Variations in this ratio can be thought of as permanent open market operations.
The remainder of the paper proceeds as follows. Section II describes the model economy, while Section III states conditions necessary for steady-state equilibrium to exist and examines the dynamic properties of the model. The propensity for unpleasant monetarist arithmetic to arise is also discussed in this section.
Comparisons...
|
