Acommon macroeconomic feature characterizing many industrial and developing countries in recent years is the growth and persistence of fiscal deficits. Although this is certainly not a new phenomenon, it is apparent that attempts to reverse these developments are drawing increasing attention in the design and implementation of adjustment policies. The analysis of the impact of fiscal deficits on aggregate demand and, through it, on the rate of inflation, the balance of payments, the level of employment, and the real interest rate has become a centerpiece of macroeconomic policy studies.

Abstract

Acommon macroeconomic feature characterizing many industrial and developing countries in recent years is the growth and persistence of fiscal deficits. Although this is certainly not a new phenomenon, it is apparent that attempts to reverse these developments are drawing increasing attention in the design and implementation of adjustment policies. The analysis of the impact of fiscal deficits on aggregate demand and, through it, on the rate of inflation, the balance of payments, the level of employment, and the real interest rate has become a centerpiece of macroeconomic policy studies.

Acommon macroeconomic feature characterizing many industrial and developing countries in recent years is the growth and persistence of fiscal deficits. Although this is certainly not a new phenomenon, it is apparent that attempts to reverse these developments are drawing increasing attention in the design and implementation of adjustment policies. The analysis of the impact of fiscal deficits on aggregate demand and, through it, on the rate of inflation, the balance of payments, the level of employment, and the real interest rate has become a centerpiece of macroeconomic policy studies.

Although the study of the consequences of alternative fiscal policies has been directly tied to the various methods of financing fiscal deficits, it is possible to focus, as many of the recent controversies on the subject have done, on the differential impact of the alternative financing mechanisms of public expenditures. As the government proceeds to absorb resources from the private sector to finance its spending, a question arises about the impact of such absorption on private sector consumption and, therefore, on aggregate demand. In particular, it is possible to inquire about the differential effects on private sector wealth and consumption of financing government sector spending with taxes as opposed to debt.

The central proposition in this context is that, under a specific set of circumstances, it actually makes no difference to the level of aggregate demand throughout the economy if the government finances its outlays by debt or by taxation (see Ricardo (1951), and Buchanan (1958)). This is the so-called Ricardian-equivalence theorem which states that, for a given path of expenditures, it is economically equivalent to maintain a balanced budget or to run a debt-financed deficit, since the substitution of debt for taxes does not affect private sector wealth and consumption. The underpinnings of Ricardian equivalence are based on the premise that the issue of public debt in the current period is always accompanied by a planned increase in future tax collections, which would be needed to serve this higher level of public indebtedness. Thus, because debt financing is perceived only as a change in the timing of taxation, the Ricardian proposition asserts that such a change has no impact on pri-vate sector wealth and consumption as long as the present value of the stream of taxation remains unchanged (see Barro (1974, 1978a)).

The policy implications of this proposition and the trade-offs that it may offer to the policymakers are, indeed, important; therefore, a care-ful assessment of its analytical, as well as its empirical, validity is rele-vant. It is evident that for this equivalence to hold, a number of assumptions and conditions are required. Although many models have been developed that could produce Ricardian results, there is, with only a few exceptions, no unified analytical framework that allows a comprehensive consideration of the relative importance of the specific assumptions and permits, without undue complications, an analysis of the consequences of extending the basic model to cover more realistic circumstances. Furthermore, although a large body of empirical work on the equivalence proposition has accumulated over recent years, there is confiicting and inconclusive evidence that raises doubts about the meth-odology used in some of these studies.

This paper provides a simple unified model that illustrates the implica-tions of Ricardian equivalence and, using the model, reviews the literature on the subject, considers the effects of relaxing the basic assump-tions, and provides a framework within which to study the implications of various extensions. Key among such extensions is the explicit dis-cussion of open, monetary, and growing economies, in addition, we present an updated survey of empirical work on Ricardian equivalence that focuses on those tests that looked at the response of private sector consumption to government budget variables.1

Changes in the ratio of taxes to debt may, in practice, result in non-negligible effects on private consumption and the macroeconomy. Although these effects may reflect a violation of one or several of the assumptions required for equivalence (for example, the assumption of perfect capital markets), another possibility is that changes in this ratio signal to agents changes in future fiscal policies, which in turn have an impact on current consumption. In fact, the equivalence proposition emerges only under a very specific set of fiscal signals conveyed by observed policy actions; that is, that current tax cuts that are accompanied by increases in the stock of public debt imply higher taxation in the future. In the paper, we also characterize the impact of government policies on the current level of consumption under a variety of fiscal signals and review some of the evidence on the type of signals that could have been extracted in practice.

The paper is organized as follows. The basic assumptions and derivation of Ricardian equivalence are presented in Section I, in the context of a simple intertemporal model. Section II discusses the conditions under which equivalence arises in three extended frameworks: a monetary economy, an open economy, and a growing economy. The main channels giving rise to deviations from equivalence are analyzed in Section III; specifically, we focus on the role of borrowing constraints, distortionary taxes, uncertainty, and finite lives. Empirical evidence on the response of consumption to government budget variables is evalu-ated in Section IV. Section V discusses the implications of the fiscal regime and fiscal signals for the behavior of private consumption, and Section IV concludes the paper.

I. Ricardian Equivalence: Statement, Assumptions, and Derivation

The Ricardian-equivalence theorem of government finance states that substitution of debt for taxes does not affect private sector wealth and consumption (see Barro (1974, 1978a)). The conditions and assumptions required for Ricardian equivalence to emerge can be described by specifying a two-period model of the economy that consolidates the intertemporal budget constraints of the public and private sectors.2 As is explained in the next section, most of the results from the two-period model carry over to a multiperiod setup with growth.

Consider a two-period model where period 0 is the “present” and period 1 is the “future.” Period —1 is used to take into account historically given conditions. We use throughout the following notation: G is government nominal spending on goods and services; T is government nominal lump-sum tax collection; B’ is government debt; / is nominal interest rate; Cis nominal private sector consumption; B is private sector debt; yis non-assets income; P is the price level. The lowercase letters g, t,b’, c, b, and y are used to denote the real values of the correspond-ing variables whose nominal values were denoted by uppercase letters. The government budgets for periods 0 and 1 are given in nominal terms by

G0T0+i1B1=B0B1(1)
G1T1+i0B0=B0,(2)

where the left-hand side is the government budget deficit (inclusive of interest payments).3. Dividing the first equation by the price level P0, and the second by P1, and consolidating them into a single equation yields

g0+g1(1+r0)1+(1+r1)b1=τ0+τ1(1+r0)1,(3)

where

l + r0(1+i0)(P0/P1) and 1+r-1≡(1+i-1)(p-1/P0), with r denoting the real interest rate. Equation (3) is the intertemporal government budget constraint. It states that the present value of govern-ment spending plus initial government liabilities must equal the present value of government tax collections. The equation is a solvency requiremerit on the government, in that for private agents to lend to government they would want to ensure that the latter will raise enough revenue to cover both its spending and the repayment of its debt; that is, that government plans to satisfy equation (3).

With respect to the private sector, its budget constraints for periods 0 and 1 in nominal terms are

C0=Y0+B0(1+i1)B1T0(4)
C1=Y1(1+i0)B0T1.(5)

Expressing these equations in real terms and consolidating them yields

c0+c1(1+r0)1=y0+y1(1+r0)1τ0τ1(1+r0)(1+r1)b1.(6)

Equation (6) is the intertemporal budget constraint faced by the private sector. The present value of consumption spending must equal the present value of net income minus the initial debt commitment. Optimal consumption decisions can be described by the solutions to the problem: choose (c0, c1) so as to maximize U(c0, c1) subject to equation (6), where U denotes the consumer’s utility function.4

Ricardian equivalence can be shown to emerge in this setup by substituting the expression for taxes in equation (3), the intertemporal government budget constraint, into the private sector intertemporal constraint of equation (6), to yield

c0+c1(1+r0)1=y0g0+(y1g1)(1+r0)1.(7)

In a closed economy, a debtor position of the public sector must be matched by a creditor position of the private sector, b = —b hence, these debt terms drop out from the analysis. Equation (7) is the inter-temporal budget constraint of the private sector that holds under the assumption that this sector fully internalizes the budget constraints of the public sector. It can be seen that, for a given pattern of government spending (g0, g1), any two debt-tax patterns (b0,τ0)and(b^0,τ^0) that satisfy the government budget constraint will imply the same equilibrium quantities and prices. In this case, these two debt-tax patterns are equiv-alent economically; the timing of taxes and the size of government debt do not influence private sector behavior.

According to equation (7), the government variable that matters for private sector consumption decisions is the present value of government spending, g0 + g1(1+r0)-1, and not the specifics of its financing. Put in Milton Friedman’s (1984) words, “the whole of what government spends is extracted from the community’s resources, not solely that part financed by what are called taxes.” Given this, changes in the ratio of taxes to government debt that are accompanied by changes in current or future government spending will not in general lead to Ricardian equiv-alence. For example, a current tax cut that is accompanied by a decrease in future government spending, such that the government intertemporal budget constraint is satisfied, has a positive effect on the private sectors perceived wealth and consumption. For Ricardian equivalence to hold, a current tax cut must be assumed to signal an increase in future taxes and no change in government spending.

The Ricardian-equivalence proposition requires a number of key assumptions about the economic environment and the behavior of economic agents. These assumptions have been reflected in the previous derivation and include (1) perfect capital markets with no borrowing constraints on consumers, (2) nondistortionary taxes, (3) full certainty about the path of future taxes and government budget policies, and (4) equal planning horizon for private and public sectors. In what follows, we consider the implications of relaxing some of these assumptions for Ricardian equivalence. Before that, however, it is pertinent to maintain the above assumptions and extend the basic framework to consider three additional environments. First, we examine a monetary economy in which the government can also finance the budget deficit through money creation. Second, we consider an open economy whose capital market is integrated with the rest of the world’s. Third, we discuss a multiperiod growing economy.

II. Extensions

This section considers Ricardian equivalence in a monetary economy. It is shown that changes in the ratio of public debt to taxes have no influence on private sector behavior to the extent that these changes are not accompanied by changes in the money supply path. This condition will be met only in the extreme case that the government meets its debt obligations by taxation—that is, bonds are fully backed by direct taxation, and, therefore, there is no monetization of public debt.5

A Monetary Economy

The two-period model developed previously can now be extended to analyze a monetary economy. To simplify matters we define money as the monetary base, whose nominal and real values are denoted by M and m, respectively. It is further assumed that money balances yield utility to individuals (for example, by providing liquidity services); an assumption that in general implies positive demands for money in periods 0 and 1. The government budget equations (1) and (2) now become

G0T0+i1B1=(B0B1)+(M0M1)(8)
G1T1+i0B0=B0+(M1M0).(9)

Government budget deficits can now be financed by issuing debt or money. Expressing these equations in real terms and consolidating them yields

g0+g1(1+r0)1+(1+r1)b1+(1+π0)1m1=τ0+τ1(1+r0)1+(i01+i0)m0+m1(1+r0)1,(10)

where 1+π0= P0/P-1. Equation (10) is the intertemporal government budget constraint, stating that the present value of government spending plus initial government debt and money liabilities must equal the present value of tax collections plus revenue from money creation.

Similarly, the intertemporal budget constraint of the private sector can be expressed as

c0+c1(1+r0)1+(i01+i0)m0+m1(1+r0)1=y0+y1(1+r0)1τ0+τ1(1+r0)1+m1(1+π0)1(1+r1)b1.(11)

When the private sector fully internalizes government budget policies–that is, substituting the present value of taxes from equation (10) into the right-hand side of equation (11)—its wealth is still equal to the right-hand side of equation (7). Under these conditions, Ricardian equiv-alence can be formally stated as follows: any two government policy patterns (g0, g1, M0,M10, τ1) and (g0, g1, M0, M1, τ^0,τ^1) that satisfy the intertemporal government budget constraint induce the same behav-ior by the private sector because the policy change in question does not alter individuals’ budget sets (see Lucas (1984)).

As stressed by Wallace (1981) and Lucas (1984), one way to interpret Ricardian equivalence in a monetary economy is as an irrelevance prop-osition about open-market operations. That is, while the path of money (M0,M1) influences private sector behavior, the specific channel through which money is injected into the economy—changes in taxes or open-market operations with government bonds—is of no independent importance for real economic variables such as consumption.

All these results hold, as indicated previously, for the polar case in which increases in private sector holdings of government securities signal increased future explicit tax collections. Another polar case arises when increased government securities will be paid off not by collecting higher explicit taxes but by issuing base money and thus imposing on the public an inflation tax. In this case, changes in the ratio of public debt to taxes can affect private sector behavior because changes in a distortionary tax are being used. The effects depend on how money is modeled in the system and on the specific distortions caused by the inflation tax. For example, Aiyagari and Gertler (1985) consider an overlapping-generations economy with heterogeneous agents whose utility depends, among other variables, on real money balances. They produced examples in which changes in the ratio of public debt to explicit taxes, which are accompanied by changes in the inflation tax, redistribute the burden of government finance between the young and the oid and thus may have an impact on aggregate consumption. In contrast, these effects do not arise in models with homogeneous agents and separable utility among consumption and real money balances (see, for example, Liviatan (1982)), where the only effect of increases in inflation on private sector behavior is to reduce the sector’s utility, owing to its reduced money holdings, with no effect on its aggregate level of consumption.

An Open Economy

Consider now an open economy facing a given real interest rate in world capital markets.6 Agents in the economy can freely borrow or lend at this interest rate, denoted by r*. To the extent that the international interest rate faced by the public and private sectors is the same, then the same set of assumptions that gave rise to equivalence in a closed economy will also give rise to it in the open-economy under consideration. Specifically, a tax cut that is accompanied by an increase in the government’s foreign debt will have no effect on private sector consumption and wealth. The increase in the government’s external debt is fully internalized by the private sector, which takes into account the taxes to be imposed in the future to finance the flow of payments to foreign lenders. Thus, internal and external public sector debt are treated in the same way by the private sector.

To illustrate these results, we turn to an open-economy version of the real model developed in Section I. for simplicity, that all borrowing by government and consumers in the domestic economy is made from foreign lenders. Under these assumptions, internalizing the intertemporal government budget constraint into that of the private sector yields

c0+c1(1+r0*)1=y0g0+(y1g1)(1+r0*)1(1+r0*)(b1+b1).(12)

According to equation (12), the net present value of consumption expenditures must equal the net present value of real resources available to the private sector minus the initial value of the economy’s external debt commitment. The higher is the value of this commitment, the lower will be the level of wealth and. hence, of consumption. For a given value of this predetermined variable, however, neither taxes nor the government’s subsequent foreign borrowing has an effect on wealth, which is affected by the government spending variables g0 and g1 and not by the form of finance.

An interesting application of the importance of the existence (or lack thereof) of Ricardian equivalence in open economies is provided by Helpman and Razin (1987). They study the effects of exchange rate management, aimed at reducing inflation, on real economic variables. Inspired by the experience of Israel in the early 1980s and of Argentina and Chile in the late 1970s, Helpman and Razin consider a policyinduced slowdown in the rate of devaluation that is not accompanied by government budget adjustment in the form of fiscal contraction. They show that this policy leads to an increase in government’s foreign borrowing and, hence, to an eventual loss of international reserves. To the extent that Ricardian equivalence holds, this form of exchange rate management has no wealth effects on the private sector, which fully internalizes the future implications of government policies. After show-ing that in the countries mentioned above slowdowns of devaluation were accompanied by increases in private consumption, by real exchange rate appreciation, and by a worsening of the trade balance, Helpman and Razin modeled one specific form of deviation from Ricardian equivalence that yields results that in general conform with the evidence. Their model is based on the idea that because of finite lives, individuals face higher effective interest rates than government (see also Blanchard (1985) and Section III, “Different Planning Hori-zons for Private and Public Sectors,” below). A welfare implication of their analysis is that the devaluation slowdown benefits the current gen-eration and imposes a burden on future generations.

Growth

When extending the model to a multiperiod growing economy, it can be shown that the same type of assumptions that imply Ricardian equiv-alence in the simple model considered in Section I yield the same implication for this extended framework; Barro’s (1974) framework, with operative bequests, can be considered as one such framework. Although in the two-period model of Section I the government paid ail its outstanding debt by the second period, an important question that arises in a multiperiod framework is whether the government can contin-uously finance a permanent budget deficit by selling bonds to the public. This question has been analyzed by McCallum (1984) in the context of an optimizing money-and-growth model, extended to include govern-ment bonds. McCallum showed that the answer depends on the definition of the deficit. If the definition includes interest payments, then it turns out that a permanent deficit can be financed with bonds. This is not the case, however, if the definition excludes interest payments. Moreover, an implication of McCallum’s analysis of the former case is that the stock of willingly held government bonds can increase permanently at a higher rate than output growth, provided that the difference is smaller than the rate of time preference. He suggests, however, that govern-ment’s default incentives would grow together with the size of its debt, so that his results do not necessarily imply that unbounded debt growth is likely to be observed in reality. For some empirical evidence on this issue, see the last part of Section IV.

III. Deviations from Ricardian Equivalence

It is likely that, in practice, changes in the stock of government debt and in the timing of taxes will have an impact on private sector behavior as well as on the economy’s equilibrium allocations. What are the main economic explanations for possible deviations from Ricardian equiva-lence? One possibility is that these changes are accompanied by shifts in government spending or in the extent of monetization of government debt, or in both. As mentioned earlier, we set aside this possibility for the meantime and will return to it in the next section. Another possibility is that some of the other basic assumptions required for Ricardian equiv-alence are not actually met. Four main deviations from these basic assumptions have been emphasized in previous work: the existence of borrowing constraints, of distortionary taxes, of uncertainty about future taxes, and of different planning horizons for private and public sectors. In this section we discuss each of these cases.7

Borrowing Constraints

To illustrate how borrowing constraints and capital market imperfec-tions affect the Ricardian-equivalence result,8 we consider, for simplic-ity, an open economy in which it is assumed that the private sector faces higher borrowing rates than those faced by the government. The higher private borrowing rate could reflect risk of default, costs of verifying solvency, or administrative and transaction costs of operating the loan that are higher, from the foreign lender’s perspective, for the private sector than for the public sector.9 One interpretation of these conditions is that government has an advantage over the private sector in carrying out credit market operations, a situation that seems especially relevant for developing countries.10

Specifically, assume that the private sector faces an effective interest rate of (1 + r*)(l + λ), where λ is a borrowing premium that reflects the above considerations, and r* is the international interest rate (which applies to government borrowing from abroad). Under these assump-tions, government’s and consumers’ intertemporal budget constraints are respectively given by

g0+g1(1+r0*)1+(1+r1*)b1=τ0+τ1(1+r0*)1(13)
c0+c1[1+r0*(1+λ0)]1=y0τ0+(y1τ1)[1+r0*(1+λ0)]1[1+r1(1+λ1)]b1.(14)

Incorporating the government’s constraint into that of the private sector yields

c0+c1[1+r0*(1+λ0)]1=y0+y1[1+r0*(1+λ0)]1g0g1(1+r0*)1(1+r1*)b1[1+r1*(1+λ1)]b1+λ0r0*τ1A,(15)

Where

A(1+r0*)1[1+r0*(1+λ0)]1.

It can be seen that, only when there are no borrowing constraints on the private sector (that is, λ-10=0), equation (15) reduces to equa-tion (12), which gives the intertemporal constraint that applies to an open economy that satisfies Ricardian equivalence. A cut in present taxes that signals an increase in future taxes increases private sector wealth as long as λ0 > 0. The reason for this is that when the government cuts present taxes, it finances its deficit by foreign borrowing, which carries an interest rate of r0*.11 Effectively, then, it is as if the private sector has borrowed from abroad at a lower interest rate than the one it faces otherwise. Thus, in contrast with the equivalence proposition, this change in the timing of taxes will affect private sector behavior. More-over, changes in the ratio of government to private foreign borrowing, in this case, effect the economy’s equilibrium.

An important assumption implicitly made in the analysis is that, when collecting taxes in the future to repay debt, the government has lower transaction (and other) costs compared with those of foreign lenders (see Barro (1978a)). In fact, the relatively high borrowing costs for con-sumers may, for example, reflect substantial monitoring required to assure repayment. Then, it is only if the government is more efficient than foreign agents at performing this monitoring that the above argu-ments would hold. Alternatively, if government monitoring costs were the same as the private sector’s and if it charged consumers a premium to cover these costs, then Ricardian equivalence would arise again. It is only to the extent that transaction costs for collecting repayment of private loans are higher than for collecting taxes that the results given in the previous paragraph hold.12

Although the analysis so far indicates that borrowing constraints may-be important in explaining deviations from Ricardian equivalence, the constraints do not necessarily lead to such deviations. In particular, Hayashi (1985) provides examples from the literature on imperfect cap-ital markets in which Ricardian equivalence holds despite the existence of borrowing constraints. His examples suggest that, unless the exact nature of imperfections in loan markets is identified and the types of arrangement that are available for agents to pool risk are specified, the implications of borrowing constraints for the effects of government bud-get policies cannot be determined.

Distortionary Taxes

Another key assumption underlying the derivation of Ricardian equiv-alence is that taxes are lump-sum and nondistortionary. In practice, however, most existing taxes are likely to be distortionary. These taxes may apply to personal income, consumption, corporate income, foreign borrowing, and so forth. Changes in the timing of these distortionary taxes can affect private sector and economy-wide allocations through their induced wealth, redistribution, and intertemporal substitution ef-fects and thus lead to deviations from Ricardian equivalence.13

Consider, for example, an open economy where the government im-poses a tax on interest payments against foreign borrowing by the private sector. For simplicity, lump-sum and other taxes are assumed to be nonexistent. The intertemporal government budget constraint is given by

g0+g1(1+r0*)1+(1+r1*)b1=r1*θ1b1+r0*θ0b0(1+r0*)1,(16)

where θ denotes the tax rates that apply to private sector interest pay-ments on foreign borrowing. The right-hand side of this equation gives the present value of tax collections by the government, and the left-hand side gives the present value of government spending and initial debt liabilities. The private sector constraint is

c0+c1[1+r0*(1+θ0)]1=y0+y1[1+r0*(1+θ0)]1[1+r1*(1+θ1)]b1.(17)

Consolidating these constraints yields

c0+c1[1+r0*(1+θ0)]1=y0+y1(1+r0*)1g0g1(1+r0*)1(1+r1*)(b1+b1)+r0*θ0b0(1+r0*)1(18)

When there is no tax on interest payments against foreign borrowing, θ0 = 0, and equation (18) reduces to equation (12), one that embodies Ricardian equivalence; however, in the presence of taxes, Ricardian equivalence need not prevail. In particular, consider a tax cut, imple-mented through a reduction θ-1; that is, accompanied by a pertinent change in θ0, such that the intertemporal government budget constraint is satisfied. The change in θ0 will in general have both substitution and wealth effects. For example, an increase in θ0, which is the tax rate on interest payments to be paid in the future, alters the relative price of present versus future consumption in a way that will lead consumers to substitute away from future and toward current consumption. This in-crease also has wealth effects. Although the analysis has focused on a specific tax, similar considerations apply to other distortionary taxes. Changes in labor income taxes and corporate income taxes, for example, will typically affect labor supply, production, and consumption incen-tives through substitution, wealth, and distribution effects. Similar con-siderations apply to the case of money finance, discussed in Section II under “A Monetary Economy,” which results in the distortionary infla-tionary tax.14

In addition to the aggregate real effects of distortionary taxes, changes in the level or in the type of taxation are likely to have distribution effects that reflect differential incidence across individuals in the economy. These distribution effects further contribute to possible deviations from neutrality that arise in the presence of distortions.

Uncertainty About Future Taxes

In deriving the Ricardian-equivalence proposition, it is assumed that a current cut in taxes signals a future increase in government tax collec-tions. The nature, amount, and timing of these future increases in taxes are assumed to be known with certainty by consumers. Obviously, this is a strong assumption. In practice, although the current tax cuts may indeed be associated with future increases in taxes, the exact timing, the type of tax to be increased (for example, property, wage, inflation, and other taxes), and the incidence of the tax across individuals are all uncertain. This uncertainty may lead to deviations from equivalence.

One source of uncertainty is the incidence of future taxes. Consider, for example, a two-period consumer whose disposable income in the future (period 1) is given by (1— α1)y1, where y1 is the non-assets gross income and cti is the tax rate that applies to such income. Assume, for simplicity, that future gross income is known in advance, but the tax rate, α1, is uncertain. Clearly, a cut in taxes in the present may signal an increase in future taxes and, thus, an increase in the expected value of α1. To the extent that this change enhances the uncertainty about inci-dence of future taxes, agents would perceive an increase in the uncer-tainty attached to future disposable income, which in the present case is equal to y12 var α1.15 For risk-averse consumers, increased uncertainty about future net income will typically lead to higher saving (and lower current consumption) aimed at smoothing out the path of consumption over time. Thus, this provides an example of how a tax cut may lead to a decrease in current consumption, in contrast with the equivalence proposition.

Although in this example future gross income was assumed to be known, uncertainty about this variable is, in practice, another source of uncertainty about future taxes. To highlight this case, consider a situ-ation where there is full certainty about the future tax rate, a,, and its incidence, yet future gross (and hence net) income is uncertain.16 Then, the variance of future disposable income equals (1 - α1)2 var y1. A tax cut in the current period that signals an increase in the tax rate α1 in the next period reduces the uncertainty attached to future disposable in-come. Hence, this effect works in an opposite direction for the current level of private sector consumption than the one considered in the pre-vious paragraph. What happens is that the income tax is acting as an insurance mechanism. To see this, we can express tax payments as α1y1=α1y¯1+α1(y1y¯1),wherey¯1

article image
is average future income in the economy. An individual whose future income happens to be higher than the aver-age pays an insurance payment; however, an individual with income lower than the average receives an insurance payment. Thus, an increase in ct; increases income risk sharing, and this reduces individual uncer-tainty about after-tax income, which in turn may lead to an increase in consumption. Obviously, this effect would arise only to the extent that government taxation provides insurance that is not available in the pri-vate market, or, if it is available, then for this effect to arise it must be assumed that government has a comparative advantage in insurance provision.

The analysis has illustrated how uncertainty about future taxes may give rise to deviations from Ricardian equivalence. Clearly, more com-plexity (and realism) can be added by jointly considering uncertainty about different types of taxes (for example, income versus excise taxes) and about the future level of taxes per se. Although one could pursue this avenue in detail, the overall results in terms of Ricardian equiv-alence are likely to resemble those of the present analysis. Another useful extension would be to incorporate explicitly in the analysis the probability of the government defaulting or running into arrears on its liabilities to the private sector. Although expected defaults will probably lead to the emergence of a risk premium in interest rates, to the extent that compensation is not full, one can treat default as a form of taxation. If there is uncertainty about future defaults, this probably has effects similar to those considered in this subsection.

Different Planning Horizons for Private and Public Sectors

A necessary condition for Ricardian equivalence to obtain is that households and government have the same planning horizons and use the same discount factor in their present-value calculations. Here we analyze a departure from this condition that arises owing to individuals’ uncertainty about their lifetime. The main result from analyzing this departure is that, in the presence of such uncertainty, and assuming no bequest motive, a tax cut will lead to a rise in perceived wealth and consumption of currently alive individuals. That is, the tax cut enables a shifting of future tax liabilities to later generations, whose welfare is assumed not to affect that of the current generation (see Barro (1978a) and Chan (1983)).

Consider the two-period open economy analyzed in Section II. As-sume now that, owing to mortality, consumers are uncertain about their lives in the future and denote by p the probability of death before the start of the next period. Drawing on Blanchard’s (1985) model, we incorporate this uncertainty into the analysis as follows.’17 It is assumed that loans require the purchase of life insurance. Such life insurance assures that outstanding debt commitments are met regardless of whether the debtor is alive or not. If we assume a large number of identical agents, free entry, and a zero profit condition in the insurance business, the effective interest rate faced by consumers now becomes R0=[(1+r0*)/(1p)]1,wherer0*is the real interest rate of the risk-less world. This is the effective cost of borrowing for consumers. It differs from government’s cost of borrowing, r0*. as long as there is a nonzero probability of death. When the intertemporal government budget constraint is fully taken into account by consumers, their budget constraint is

c0+c1(1+R0)1=y0+y1(1+R0)1g0g1(1+r0*)1(1+R1)b1(1+r1*)b1+pτ1[(1p)(1+R0)]1.(19)

It can be seen that when p = 0 equation (19) becomes identical to equation (12), which is the relevant budget constraint under Ricardian equivalence. When p > 0, however, a current tax cut that is accompanied by an increase in future taxes through an increase in τ1, raises consumers’ perceived wealth and consumption. Owing to mortality, the current tax cut signals a less than one-to-one increase in the present value of future taxes to be paid by currently living consumers. Put differently, the tax cut effectively shifts part of the burden of taxation from current to future generations.18

A crucial assumption in deriving these non-Ricardian results is that the added tax liabilities on descendants are not fully counted in the wealth calculations by current taxpayers. Otherwise, and to the extent that voluntary intergenerational transfers are operative within the pri-vate sector, the shift from tax to debt finance would not represent a new opportunity for the current generation to extract funds from future generations (see Barro (1974)). In such a case, current consumers will react to a cut in their presen? taxes by increasing their voluntary transfers to the next generation so as to restore the balance of wealth across generations to its previously optimal level, and. consequently, current consumption would remain unchanged, as in the equivalence case. These considerations highlight the importance of determining the impact of government debt and taxation policies on intergenerational transfers. In recent work, Cukierman and Meltzer (1986) emphasize the redistribution effects of changes in public debt, both over time and across generations. They do that in a framework that represents an extension of Barro’s (1974) to allow for heterogeneity in agents’ abilities, wage earnings, and initial nonhuman wealth. Cukierman and Meltzer show that some agents may be bequest constrained in that, whereas their optimal behavior would be to leave negative bequests to their successors, they cannot reach such position tn practice because, under the prevailing institutions, they cannot directly obligate the future labor income of their descendants. Such individuals are likely to favor any tax cuts that increase their lifetime income at the expense of their descendants’, and any such tax cuts thus affect the aggregate demand for consumption.19

IV. Empirical Evidence

This section presents a survey of empirical evidence on the impact of government budget variables on private consumption and on the Ricardian-equivalence hypothesis. Although the survey is selective, it covers the main methods that have been used in empirically testing this hypothesis.

A traditional approach in testing the Ricardian proposition with time-series data has been to regress private consumption on government budget, as well as other relevant variables. A prototype of such re-gression is given by the following equation:

ct=a0+a1yt+a2yt1+a3gt+a4wt+a5τt+a6bt+a7trt+a8zt+μt,(20)

where c is a measure of private consumption expenditures, y is personal or national income, g is government spending on goods and services, τ is tax revenue, w is household’s net worth, b’ is net government debt, tr is government transfers, z measures other variables that are not related to the government budget, and p. is a stochastic error term. All variables are in general expressed in real per capita units, and t is a time index. The variables yt and yt-1 are included as proxies of permanent income that, together with beginning-of-period net worth (wt)), are assumed to affect consumption. In some formulations, personal income and government spending are decomposed into permanent and transitory components. The coefficient on government spending is interpreted as reflecting two effects: the impact of this variable on private sector consumption through its direct impact on wealth (for example, as in equation (7) above) and its impact through the substitutability of private sector con-sumption and government spending, which in turn depends on how government spending affects private sector utility.20

To test for Ricardian equivalence, most studies along this approach test the restriction a5 = a6 = a7 = 0. If this restriction is met by the data, the equivalence proposition is not rejected; otherwise, it is rejected. The empirical evidence on this issue is inconclusive. On the one hand, studies by Barro (1978b), Kochin (1974), Kormendi (1983), Seater and Mariano (1985), and Tanner (1979) report evidence that supports the null hypoth-esis. Conflicting evidence, however, has been reported by Blinder and Deaton (1985), Feldstein (1982), Modigliani (1987), Modigliani and Sterling (1985), and Reid (1985). To a large extent, these discrepancies reflect differences in sample periods, econometric techniques, and meth-ods of empirically measuring the different variables.21 For example, Modigliani and Sterling (1985) dispute Kormendi’s (1983) results. They claim that by changing the methods of deflating government private sector expenditures, of measuring real government interest payments, and of estimation (including more lags than Kormendi and focusing on a formulation in levels and not in rates of change), Kormendi’s basic results on equivalence are reversed. Along similar lines, Reid (1985) shows that the results are sensitive to averaging of variables over business cycles and to whether the economy is undergoing a period of business contraction or expansion. And Hernández-Catá (1982) shows that some of the coefficients estimated by Feldstein (1982) are sensitive to cor-rection for multicollinearity.

Even if problems of measurement and estimation were nonexistent, one can object to using estimates of equation (20) to test for Ricardian equivalence. To elaborate on this point, it is useful to compare the equation actually estimated (equation (20)) with the general specifica-tion suggested by a multiperiod version of the intertemporal model developed in the previous sections:

ct=f(yt,yt+1,gt,gt+1,wt,rt,rt+1,τt,τt+1,trt,trt+i,kt,kt+i,...),(21)

for i = l,2.....T. The variable k measures here money creation to finance the government deficit. Consumption in each period is related to current and (expected) future values of its fundamental determinants: income, government spending, interest rates, and so forth.

Within this formulation, Ricardian equivalence amounts to zero re-strictions on the block of variables measuring current and future taxes, transfers, and debt: yet the exact specification to be tested depends on the specific postulated mechanism that is supposed to give rise to non-equivalence. Typically, none of the future variables suggested by the theoretical models is explicitly included in the estimated equation (20)—a surprising feature, given that Ricardian equivalence embodies a strong intertemporal element22.” Consequently, the fact that a researcher finds, for example, that a5, a6, and a7 are significantly different from zero may just be an indication that current taxes, transfers, and govern-ment debt are “good” predictors of future government spending, quite in line with equivalence. Unless an equation like (20) is jointly and explicitly considered with the signaling rote of the explanatory variables for their own future values, this equation is not likely to be informative in a decisive way for the empirical validity of Ricardian equivalence. Moreover, equation (20) abstracts from interest rates, government money creation, and government’s foreign debt, variables that are likely to affect consumption; these omissions may create additional bias in the parameter estimates. Another difficulty with the traditional approach is that it typically does not make explicit the optimality problem that gives rise to the estimated consumption function, thus generating ambiguity in the interpretation of a given set of results.

A more recent approach in empirical tests of equivalence attempts to overcome some of the above-mentioned difficulties by directly deriving the estimated relations from explicit intertemporal optimization frame-works (see, for example, Aschauer (1985) and Leiderman and Razin (1988)). To illustrate how such a test can be constructed, consider first the model analyzed by Aschauer (1985).

Assuming that households maximize the present value of utility from consumption in current and future periods, and that the utility function is quadratic, Aschauer focuses on the Euler equation (or first-order condition):

Et1ct*=α0+α1βct1*.(22)

The variable ct* measures effective private consumption, which is as-sumed to be given by ct*=ct+δ¯gt, where c measures actual private sector spending on consumption, and g measures government ex-penditures. Et-1 is the expectation of a given variable conditional on information up to time, t —1. Thus, this specification allows for govern-ment spending effects on private sector utility: each unit of g is assumed to yield the same utility as δ¯ units of private spending. The parameters α0 and α1 are explicitly derived in the analysis. They are nonlinear functions of the real interest rate, the rate of time preference, and the “bliss level” of effective consumption. Equation (22) can be rewritten as

ct=α0+α1ct1+α1δ¯gt1δ¯Et1gt+vt.(23)

To estimate this equation assuming rational expectations, a time-series process must be assumed for gt so as to generate the expected values Et-1gt. Considering such a process jointly with the consumption equa-tion yields the system actually estimated by Aschauer:

ct=d0+d1ct1+d2(L)gt1d3(L)Dt1+μ1t(24)
gt=e0+e1(L)gt1+e2(L)Dt1+μ2t,(25)

where d(L) and e(L) denote polynomials in the lag operator.

According to equation (23). lagged values of government spending and deficits (D) are used to forecast current values of government spending. The null hypothesis is a set of cross-equation, nonlinear re-strictions on the parameters. Aschauer’s findings, based on quarterly U.S. data for 1948, first quarter, to 1981, fourth quarter, yield an esti-mated value of δ¯=0.23; that is, a dollar increase in government spend-ing leads to a 0.23 cents offset through a decrease in private sector consumption spending.23 Moreover, the data do not reject the cross-equation restrictions, indicating that the impact of government deficits and spending on private sector consumption can be attributed to the channel specified here; that is, through substitutability of public spend-ing for private consumption in consumers’ utility. Although these find-ings support the notion of Ricardian equivalence, it is not clear how statistically powerful they are because the alternative, non-Ricardian hypothesis is not tightly specified in the model.

Leiderman and Razin’s (1988) framework allows for a closer nesting of null and alternative hypotheses on Ricardian equivalence. Their mod-eling of deviations from Ricardian neutrality is primarily based on the assumption that individuals have finite horizons, and their specifications draw on and extend the work of Blanchard (1985). A finite horizon is captured by the assumption that there is a probability γ≤ that an individual will survive to the next period. Assuming that the representa-tive consumer maximizes expected lifetime utility subject to an infertemporal budget constraint, and that utility is quadratic, they derive the following equation for aggregate consumption per capita:

Ctβ0(R1)+(1γ)β1Et1Σj=0(γ/R)t+j(Yt+jTt+1)+ψCt1+εt,(26)

where Ct, denotes aggregate consumption per capita, R is the riskless interest factor (1 + riskless interest rate), β0 and β1 are fixed parameters that include preference parameters, Et-1 is the conditional expectations operator, and Y and T are aggregate per capita gross income and taxes.

Being derived from an intertemporal framework, equation (26) shows how current consumption depends on current and expected future gross income and taxes. Equation (26) can be used to test the Ricardian-equivalence proposition in the present model. The key parameter in this context is γ. When γ = 1, consumers’ behavior satisfies Ricardian neu-trality, and equation (26) indicates that only lagged consumptionCt-1 can be used to predict current consumption (over and above the constant term)—as in Hall (1978). However, when γ<1, expected wealth affects consumption over and above the effect of lagged consumption. In such case, a cut in current-period taxes raises expected wealth and, thus, results in an increase in current consumption. The reason for this effect is that under these circumstances the future tax increases that are needed to intertemporally balance the government budget are given a smaller weight by finite-horizon consumers than the weight they give the current cut in taxes.

To implement equation (26) with time-series data, it is required to specify, under rational expectations, the stochastic processes that govern the evolution of gross income and taxes. Using first-order autoregressive representations for the latter, Leiderman and Razin (1988) construct a system of three equations (two stochastic processes and a consumption equation) whose parameters are subject to nonlinear cross-equation restrictions. Using monthly time series for Israel covering the period of high budget deficits, 1980-85, their findings from estimating this system provide support for the hypothesis of Ricardian neutrality, in that the γ = 1 hypothesis is not rejected at standard significance levels. More-over, the same pattern of results emerges upon extending the model to allow for liquidity constraints as an additional source of nonneutrality and to allow for substitutability between public and private consumption in private sector’s utility.

Although these results are supportive of the neutrality hypothesis, it would be desirable to investigate their robustness in more complex frameworks that allow for (1) other channels of nonneutralities, such as the existence of distortionary taxes or income redistribution effects of government policies; (2) more general specifications of preferences and stochastic processes; and (3) monetary and exchange rate effects on consumption. In any case, this discussion suggests that empirically test-ing Ricardian equivalence using an intertemporal stochastic framework is not oniy a feasible task but also the most appropriate approach to test the theory underlying this hypothesis.

The notion that liquidity constraints are important in generating devi-ations from equivalence has been stressed by Hubbard and Judd (1986). They performed simulations attempting to determine the magnitude of the aggregate marginal propensity to consume out of a temporary tax cut. To do so, they extend Blanchard’s (1985) model by specifying the existence of two types of individuals: those with low productivity and wage who have no access to borrowing against their future wages, and those with high productivity and wage who can borrow. In a Ricardian setup and with perfect capital markets, the marginal propensity to con-sume out of a temporary tax cut is equal to zero. When capital markets are perfect but there is a positive probability of death. Hubbard and Judd (1986) obtain Blanchard’s (1985) consumption function, for which they show the simulated marginal propensity to consume is positive but of a negligible order of magnitude. When that model, however, is ex-tended so that 20 percent of the labor force is assumed to be liquidity constrained, there is a more than quadrupling in the value of the mar-ginal propensity to consume. This result is due to the fact that consump-tion equals the wage for low-productivity workers, so that for them a tax cut is met with a marginal propensity to consume equal to unity. In these calculations liquidity constraints take the form of full credit rationing. In practice, there could be other forms of capturing the relevant constraints—as, for example, through differential interest rates. Existing tests for liquidity constraints have been recently surveyed by Hayashi (1985). According to him, the time-series evidence is not conclusive, and key parameters have not been precisely estimated. One possible reason for this is that time-series studies have looked at economy-wide aggre-gate data, and probably useful information on liquidity constraints of different sectors is lost in the process of aggregation. The most useful evidence on liquidity constraints is likely to emerge from micro data. Although the pertinent micro evidence surveyed by Hayashi (1985) suggests a nonnegligible role for liquidity constraints, the fact that the behavioral parameters are contaminated with measurement errors in the variables and with shocks resulting from changed tastes creates econometric problems of identification that have not yet been fully overcome.

A simulation-based assessment of another source of nonequivalence is presented by Barsky, Mankiw, and Zeldes (1986). They consider devi-ations from Ricardian equivalence that arise owing to uncertainty about future taxes (see Section III, under “Uncertainty About Future Taxes,” above). In particular, they focus on conditions under which a tax cut and debt issue increase risk sharing and thus lead to a reduction in individual uncertainty about after-tax income. Thus, there is a positive marginal propensity to consume out of a tax cut because the cut reduces precau-tionary saving. Obviously, a key assumption in the analysis is that by increasing future taxes (matching the current tax cut) government pro-vides insurance to individuals that is not available in the private market. Under plausible assumptions regarding preferences and the extent of income uncertainty, the authors’ simulations deliver nonnegligible mar-ginal propensities to consume out of a tax cut—for example, 0.3 or 0.5. Thus, they claim that even though consumers are Ricardian in that they fully discount future tax liabilities, their consumption does react to the current tax cut owing to its effect on uncertainty. Again, a key assump-tion used in generating this effect is that there are no markets through which agents can insure against future income risk.

Other empirical studies have directly focused on the intergenerational implications of Ricardian equivalence; for recent surveys of the methods and findings, see Kotlikoff (1984) and Boskin and Kotlikoff (1985). Models in which equivalence holds are in general models of intergenera-tional altruism: consumption of particular extended-family members de-pends on the resources of other extended-family members. Controlling for demographic changes, this implies that consumption should be invar-iant to changes in the age distribution of resources. Boskin and Kotlikoff (1985) take the latter to be the null hypothesis (that is, the altruism hypothesis) for their econometric work on postwar U.S. data. Their results indicate rejection of the altruism model, in that the age distribu-tion of personal income (and some of its components) has significant explanatory power for aggregate consumption beyond that of other more standard determinants of consumption. Although these are unambig-uous results, the authors suggest that more work is required in checking for model misspecifications before one can reach final judgment on the validity of the altruism model. Other work, mostly with cross-sectional data, that has concentrated on the effects of social security and of inter-generational transfers has in general produced results that contradict the altruism or equivalence hypothesis (for details, see Kotlikoff (1984)).

Finally, there is the issue of whether a government can run a per-manent bond-financed deficit, and thus have government debt growth indefinitely (see McCallum (1984)). Hamilton and Flavin (1986) look at this issue with an empirical perspective. They show that the hypothesis that government can accumulate a continuously growing stock of debt, as a resuit of budget deficits, is mathematically equivalent to the hypoth-esis that prices can rise continuously in a self-fulfilling speculative bub-ble. They suggested using empirical tests that were developed for the latter hypothesis to provide evidence on the government borrowing hy-pothesis. After conducting several econometric tests based on postwar tune series for the United States, they concluded thai the evidence supports the idea that government is not perceived by private markets to implement a policy of continuous borrowing over time.

V. The Signaling Role of the Fiscal Regime

The fiscal regime prevailing in an economy, as well as the type of fiscal relationships expected to arise from such a regime, is an important factor in determining the response of private agents to fiscal signals and, in particular, of private consumption to changes in the tax-to-debt ratio. Despite its importance, the issues related to the nature of the fiscal regime have not received, in our opinion, the attention they merit in discussions about the analytical validity of Ricardian equivalence. This section discusses how what is being assumed about the prevailing fiscal regime affects the assessments of the effects of government budget pol-icies on the economy’s equilibrium (see Feldstein (1982) and Sargent (1982a)).

The Ricardian equivalence arises in a specific fiscal regime: one in which government debt is fully backed by taxation. In this case, an increase in government bonds in the hands of the public signals increased future explicit tax collections with a present value that exactly matches the value of existing government’s bond obligations. Following Sargent (1982a), we refer to this as the polar Ricardian fiscal regime. This regime represents a case of fiscal accommodation, in that current debt financing or an open-market sale by the central bank leads to an increase in future taxes so as to service the new debt.

A second polar policy scenario can be referred to as the polar non-Ricardian fiscal regime, in which bonds are backed by implicit infla-tionary taxation in the form of money creation.24 Here, monetary policy is fully accommodating fiscal deficits, in that the central bank issues money so as to finance these deficits.25 An increase in the stock of government bonds signals, in this case, a change in future base money growth so that government debt is eventually monetized. Given, therefore, that the inflationary tax is usually distortionary. an increase in government debt is likely under this regime to affect private consumption.

The cases just discussed are extreme examples of fiscal regimes. In reality, observed regimes usually lie between these extremes, according to the extent of fiscal and monetary accommodation used by the author-ities. Furthermore, although the discussion above implicitly assumed that the time path of government spending is given, in many circum-stances changes in taxes and debt may, in fact, signal future changes in government spending that will induce a completely different set of reac-tions on the part of the private sector.

These considerations suggest that a prerequisite for assessing the validity of the Ricardian equivalence and for analyzing the effects on the economy of a change in the debt-tax ratio is to specify the fiscal signals conveyed by the policy change. In a polar Ricardian regime, and pro-vided that the assumptions required for equivalence are met, a tax cut coupled with a bond issue will have no effect on private sector wealth, consumption, and interest rates. However, in a case in which a tax cut actually signals a decrease in future government spending, in general it will lead to an increase in the private sector’s perceived wealth and will thus have a positive effect on the demand for consumption. Similarly, policy changes that signal future changes in monetizatton and in the distortionary inflation tax will, generally, have nonnegligible effects. In addition, if the economy is mostly open, the analysis of the fiscal regime has to take into account the possibility of resource transfers from abroad. For example, in a regime in which bond financing of deficits is likely to be serviced with future foreign transfers and aid, an increase in govern-ment spending financed with bonds is likely to have a larger impact on consumption than in a regime where bond issues signal the need for future tax liabilities. Clearly, therefore, the results of the analysis are sensitive with respect to the characteristics of the fiscal regime in operation and the assumptions of the public about the stability of such regime.26

In practice, fiscal regimes differ across countries and change over time. Furthermore, at each point in time there is. typically, uncertainty about the regime that will prevail from then on. For example, a high government budget deficit financed by debt can be regarded as un-sustainable and, therefore, may be taken to signal future contractions in the deficit; however, whether these contractions will be effected through cuts in spending or increases in explicit tax collections, and when these actions will be taken, is in general unknown. Economic agents have subjective probability distributions for future behavior of government variables, and these distributions are conditioned by past developments and adjusted to the fiscal signals provided by the current actions of the authorities (see Feldstein (1982)). This uncertainty about the policy regime may certainly affect the behavior of the economy. Specifically, Drazen and Helpman (1986) and Masson (1985) have explicitly shown how the dynamics of inflation during a transition from a period of un-sustainable deficits to one of sustainable policies strongly depend on this uncertainty. The latter may give rise to a positive risk premium on bond interest rates, reflecting inflation uncertainty, and to changes in private sector saving and consumption. Moreover, as stressed by Drazen and Helpman (1986), policy regime uncertainty may imply that looking at contemporaneous correlations between budget deficits and inflation is not a meaningful way of determining the inflationary impact of a given government budget.27

From this discussion, it follows that what is required for meaningful policy analysis is an intertemporal signaling approach, in which the implications of a given policy change for the intertemporal relationship between monetary and fiscal policy are explicitly taken into account before assessing the effects of this policy change on the private sector and the economy.

What is the evidence on fiscal regimes? Consider some relevant fiscal figures that summarize budget performance in industrial countries.28 The evidence for industrial countries as a group, which could provide some information on the regime in operation, is as follows. (1) The ratio of the budget deficit to gross domestic product (GDP) increased from 1.5 percent of GDP in 1972 to 5.75 percent of GDP in 1983; this is almost a quadrupling of this ratio. (2) This increase in the deficit is the result, in almost all countries except the United States, of an increase in total outlays that is twice as fast as that in revenues. Total central government outlays increased from about 26 percent of GDP in 1972 to over 35 percent of GDP in 1983,29 In contrast, total central government revenues of industrial countries rose by more than 4 percentage points from 1972 to 1983, from 25 percent of GDP to 29.25 percent of GDP, primarily through increases in nontax revenue and social security contributions.30(3) The increased deficit was mostly financed by domestic borrowing. Domestic financing rose from 4.5 percent of total outlays in 1972 to 5.75 percent in 1983. Financing from the monetary authorities accounted for a small percentage of total outlays, reaching a maximum of 2.5 percent of outlays in 1975 and declining since then. Similarly, foreign financing played a relatively minor role for industrial countries; it reached an average value of 1.5 percent of total outlays during 1972-83 and ex-hibited a downward trend throughout this period.

What can be inferred from these trends? It certainly depends on the perception about the sustainability of recent policy patterns. If it is assumed that a continuation of these debt-financed deficits is not sustain-able, these facts ought to signal either future decreases in the size of deficits or increased monetization. Judging from recent history, in-creases in monetization of the debt are not likely to be heavily used,31 If that is the case, the question arises regarding the specific mechanisms and channels to be used in implementing future cuts in deficits: Will these take place primarily through decreased spending or through in-creased taxation?

Obviously, the answer to this question is typically uncertain. Yet, in predicting the future behavior of private consumption, different answers to this question will yield different predictive implications. For the United States, one possibility is to assume the persistence of past long-term trends and to use evidence on how large deficits were actually reduced in the past to extract signals about future fiscal actions. As is well known, the largest budget deficits that occurred in the United States in the past can be attributed to war periods such as those of the Civil War, World War I, and World War II. Tax rates and revenue increased in each war episode, but government spending rose more, thus creating these large budgetary deficits. After each war, sooner or later the deficit reversed into surplus, with decreases in postwar government spending absorbing most of the action in this direction (see Peacock and Wiseman (1961) and Tanzi (1985b)). Thus, to some extent, wartime deficits have signaled postwar surpluses, with government spending falling more than taxes after the war. Although the current scenario is one of nonwar increases in deficits, the resolution of these deficits may well take the form suggested by these wartime episodes. Alternatively, the deficits may primarily signal future increases in taxation, an option that has opposite implications for the behavior of private consumption than those of the postwar pattern. The example of the United States, however, may not necessarily be generalized. Government policies usually differ across countries, and our discussion suggests that it would be important to use past and current information on a country-specific basis to detect the signals that are most likely to be conveyed by the current deficit policies in each case.

Key aspects of observed fiscal regimes have been addressed by several studies using econometric techniques. Most such studies focused on the extent to which current and past deficits have been accompanied by monetary accommodation. In some of the studies—for example. Blinder (1983) and Joines (1985)—specific central bank reaction functions are postulated and estimated, with budget deficits appearing as one of the explanatory variables for money creation. Other authors, such as King and Plosser (1985), have investigated this issue in a relatively model-free form by using vector autoregressions. Most of the pertinent evidence has been surveyed by Dwyer (1985), and the main conclusion arising from these studies,32 based on data for the United States and other industrial countries, is that there is no clear and statistically significant link be-tween budget deficits and government money creation. This conclusion is consistent with the actual figures on monetization discussed earlier for the United States.

Quite surprisingly, there are only few studies that provide evidence on the predictive role of budget deficits and other government variables for future values of government spending and revenue. Some evidence is presented by von Furstenberg, Green, and Jeong (1986), who concen-trated on the sequencing of taxes and spending. Using quarterly postwar data for the United States and vector autoregression analysis, the au-thors concluded that there is no evidence in support of the hypothesis that changes in taxes precede (or effectively limit) changes in spending in the same direction. The reverse sequence—spend first and tax later— received some support from the data. This was particularly so for two categories of disaggregated spending: cyclically adjusted transfers and defense spending. In their study of postwar annual data for the United States, King and Plosser (1985) found that government purchases and tax rates do not appear to be predictable from the other fiscal and monetary variables considered, except for some role of the previous year’s real deficit. Although these findings are useful, they are only indicative. Further work using data for different countries and time periods is required before one can characterize the signaling role of cur-rently observed variables for future levels of government spending and revenue,

VI. Concluding Remarks

In this paper we have surveyed some aspects related to the effects of government budget policies on private sector consumption. The dis-cussion has centered on the Ricardian-equivalence theorem of govern-ment finance, which states that substitution of debt for taxes has no impact on private sector wealth and consumption. Although this is a valid and useful proposition, there are likely to be deviations from it in practice. These deviations need not be attributed to irrationality or lack of full discounting of future tax liabilities by the public. Agents may be fully rational; yet. owing to the presence of factors such as borrowing constraints or distortionary taxes, which represent departures from Ricardian assumptions, their optimal behavior will result in the non-equivalence of taxes and debt insofar as aggregate demand is concerned. Also, to the extent that substitution of debt for taxes conveys signals of future changes in government spending or in money creation, or in both, private consumption will not remain invariant. Other sources for non-equivalence were discussed in detail in the paper.

The fact that there exist deviations from Ricardian equivalence im-plies that deficit finance policies can have an impact on private consump-tion and aggregate demand that would be nonexistent otherwise. By exploiting these deviations, however, policymakers may affect the funda-mental sources for nonequivalence.33 Thus, nonequivalence opens up policy trade-offs whose positive and normative implications remain to be explored.

Finally, it should be stressed that we have considered the implications of only a specific set of fiscal and monetary policies within a given institutional framework. In practice, changes in government budget ac-tions may be accompanied by substantial policy-induced changes such as financial liberalization or opening up of the economy, which can result in important effects on the private sector’s real and portfolio decisions.

References

  • Aiyagari, S. Rao, and Mark Gertler, “The Backing of Government Bonds and Monetarism,” Journal of Monetary Economics (Amsterdam), Vol. 16 (July 1985), pp. 1944.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Aschauer, David Alan, “Fiscal Policy and Aggregate Demand,” American EconomicReview (Nashville, Tennessee), Vol. 75 (March 1985), pp. 11727.

    • Search Google Scholar
    • Export Citation
  • Aschauer, David Alan, and Jeremy Greenwood, “Macroeconomic Effects of Fiscal Policy,” Carnegie-Rochester Conference Series on Public Policy (Amsterdam and New York), Vol. 23 (Autumn 1985), pp. 91138.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Barro, Robert J., “Are Government Bonds Net Wealth?” Journal of PoliticalEconomy (Chicago), Vol. 82 (November-December 1974), pp. 10951117.

  • Barro, Robert J., (1978a), “Public Debt and Taxes,” in Federal Tax Reform: Myths andRealities, ed. by Michael Boskin (San Francisco: Institute for Contemporary Studies), pp. 189209.

    • Search Google Scholar
    • Export Citation
  • Barro, Robert J., (1978b), The Impact of Social Security on Private Saving, American Enterprise Institute Studies, No. 199 (Washington: American Enterprise Institute for Public Policy Research).

    • Search Google Scholar
    • Export Citation
  • Barsky, Robert B., N. Gregory Mankiw, and Stephen P. Zeldes, “Ricardian Consumers with Keynesian Propensities,” American Economic Review (Nashville, Tennessee), Vol. 76 (September 1986), pp. 67691.

    • Search Google Scholar
    • Export Citation
  • Blanchard, Olivier J., “Debt, Deficits, and Finite Horizons,” Journal of Political Economy (Chicago), Vol. 93 (April 1985), pp. 22347.

  • Blinder, Alan S., “On the Monetization of Deficits,” in The Economic Consequences of Budget Deficits, ed. by Laurence H. Meyer (Boston: Kluwer-Nijhoff, 1983), pp. 3973.

    • Search Google Scholar
    • Export Citation
  • Blinder, Alan S., and Angus Deaton, “The Time Series Consumption Function Revisited,” Brookings Papers on Economic Activity: 2 (1985), The Brookings Institution (Washington), pp. 465511.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Boskin, Michael J., and Laurence J. Kotlikoff, “Public Debt and United States Saving: A New Test of the Neutrality Hypothesis,” Carnegie-Rochester Conference Series on Public Policy (Amsterdam and New York), Vol. 23 (1985), pp. 5586.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Buchanan, James M., Public Principles of Public Debt: A Defense and Restatement (Homewood, Illinois: R.D. Irwin, 1958).

  • Buiter, Willem H., “A Guide to Public Sector Debt and Deficits,” Economic Policy: A European Forum (Cambridge, England), Vol. 2 (November 1985), pp. 1479.

    • Search Google Scholar
    • Export Citation
  • Buiter, Willem H., and James Tobin, “Debt Neutrality: A Brief Review of Doctrine and Evidence,” in Social Security Versus Private Saving, ed. by George M. von Furstenberg (Cambridge, Massachusetts: Ballinger, 1979), pp. 3963.

    • Search Google Scholar
    • Export Citation
  • Carmichael, Jeffrey, “On Barro’s Theorem of Debt Neutrality: The Irrelevance of Net Wealth,” American Economic Review (Nashville, Tennessee), Vol. 72 (March 1982), pp. 20213.

    • Search Google Scholar
    • Export Citation
  • Chan, Louis Kuo Chi, “Uncertainty and the Neutrality of Government Financ ing Policy,” Journal of Monetary Economics (Amsterdam), Vol. 11 (May 1983), pp. 35172.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cukierman, Alex, “Uncertain Lifetimes and the Ricardian Equivalence Proposition,” Working Paper (Tel-Aviv: Tel-Aviv University, September 1986).

    • Search Google Scholar
    • Export Citation
  • Cukierman, Alex, and Allan H. Meltzer, “A Political Theory of Government Debt and Deficits in a Neo-Ricardian Framework,” Working Paper (Tel-Aviv: Tel- Aviv University, August 1986).

    • Search Google Scholar
    • Export Citation
  • Demopoulos, George D., George M. Katsimbris, and Stephen M. Miller, “Central Bank Policy and the Financing of Government Budget Deficits: A Cross-Country Comparison,” European Economic Review (Amsterdam) (forthcoming, 1988).

    • Search Google Scholar
    • Export Citation
  • Dotsey, Michael, “Controversy Over the Federal Budget Deficit: A Theoretical Perspective,” Economic Review, Federal Reserve Bank of Richmond (Richmond), Vol. 71 (September-October 1985), pp. 316.

    • Search Google Scholar
    • Export Citation
  • Drazen, Allan, and Elhanan Helpman, “Inflationary Consequences of Anticipated Macroeconomic Policies,” NBER Working Paper 2006 (Cambridge, Massachusetts: National Bureau of Economic Research, August 1986).

    • Search Google Scholar
    • Export Citation
  • Dwyer, Gerald P., Jr., “Federal Deficits, Interest Rates, and Monetary Policy,” Journal of Money, Credit and Banking (Columbus, Ohio), Vol. 17 (November 1985), pp. 65581.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Evans, Paul, “Do Large Deficits Produce High Interest Rates?” American Economic Review (Nashville, Tennessee), Vol. 75 (March 1985), pp. 6887.

    • Search Google Scholar
    • Export Citation
  • Feldstein, Martin, “Government Deficits and Aggregate Demand,” Journal of Monetary Economics (Amsterdam), Vol. 9 (01 1982), pp. 120.

  • Frenkel, Jacob A., and Assaf Razin, “Government Spending, Debt, and International Economic Interdependence,” Economic Journal (London), Vol. 95 (September 1985), pp. 61936.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frenkel, Jacob A., and Assaf Razin, “Fiscal Policies in the World Economy,” Journal of Political Economy (Chicago), Vol. 94 (June 1986), pp. 56494.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frenkel, Jacob A., and Assaf Razin, Fiscal Policies and the World Economy: An Intertemporal Approach (Cambridge, Massachusetts: MIT Press, 1987).

    • Search Google Scholar
    • Export Citation
  • Friedman, Milton, “The Taxes Called Deficits,” Wall Street Journal (New York), April 26, 1984.

  • Hall, Robert E., “Stochastic Implications of the Life Cycle-Permanent Income Hypothesis: Theory and Evidence,” Journal of Political Economy (Chicago), Vol. 86 (1978), pp. 97187.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hamilton, James D., and Marjorie A. Flavin, “On the Limitations of Government Borrowing: A Framework for Empirical Testing,” American Economic Review (Nashville, Tennessee), Vol. 76 (September 1986), pp. 80819.

    • Search Google Scholar
    • Export Citation
  • Haque, Nadeem Ul, “Fiscal Policy and Private Sector Saving Behavior: Tests of Ricardian Equivalence in Some Developing Economies,” IMF Working Paper WP/87/51 (unpublished; Washington: International Monetary Fund, July 1987).

    • Search Google Scholar
    • Export Citation
  • Hayashi, Fumio, “Tests for Liquidity Constraints: A Critical Survey,” NBER Working Paper 1720 (Cambridge, Massachusetts: National Bureau of Economic Research, October 1985).

    • Search Google Scholar
    • Export Citation
  • Helpman, Elhanan, and Assaf Razin, “Exchange Rate Management: Intertemporal Trade-Offs,” American Economic Review (Nashville, Tennessee), Vol. 77 (March 1987), pp. 10723.

    • Search Google Scholar
    • Export Citation
  • Hernández-Catá, Ernesto, “The Impact of Fiscal Variables on Personal Consumption: An Interpretation of the Empirical Evidence” (unpublished; Washington: International Monetary Fund, July 1982).

    • Search Google Scholar
    • Export Citation
  • Hubbard, R. Glenn, and Kenneth L. Judd, “Liquidity Constraints, Fiscal Policy, and Consumption,” Brookings Papers on Economic Activity : 1 (1986), The Brookings Institution (Washington), pp. 150.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Joines, Douglas H., “Deficits and Money Growth in the United States, 1872-1983,” Journal of Monetary Economics (Amsterdam), Vol. 16 (November 1985), pp. 32951.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • King, Robert G., and Charles I. Plosser, “Money, Deficits, and Inflation,” Carnegie-Rochester Conference Series on Public Policy (Amsterdam and New York), Vol. 22 (1985), pp. 14796.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kochin, Levis, “Are Future Taxes Anticipated by Consumers? Comment” Journal of Money, Credit, and Banking (Columbus, Ohio), Vol. 6 (August 1974), pp. 38594.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kormendi, Roger C, “Government Debt, Government Spending, and Private Sector Behavior,” American Economic Review (Nashville, Tennessee), Vol. 73 (December 1983), pp. 9941010.

    • Search Google Scholar
    • Export Citation
  • Kotlikoff,Laurence J., “Economic Impact of Deficit Financing,” Staff Papers, International Monetary Fund (Washington), Vol. 31 (September 1984), pp. 54982.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Leiderman, Leonardo, and Assaf Razin, “Testing Ricardian Neutrality with an Intertemporal Stochastic Model,” Journal of Money, Credit and Banking (Columbus, Ohio)(forthcoming, 1988); previously published as NBER Working Paper 2258 (Cambridge, Massachusetts: National Bureau of Economic Research, May 1987).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liviatan, Nissan, “Neutrality of Government Bonds Reconsidered,” Journal of Public Economics (Amsterdam), Vol. 19 (November 1982), pp. 26170.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lucas, Robert E., Jr., “Money in a Theory of Finance,” Carnegie-Rochester Series on Public Policy (Amsterdam and New York), Vol. 21 (Autumn 1984), pp. 955.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lucas, Robert E., Jr., “Principles of Fiscal and Monetary Policy,” Journal of Monetary Economics (Amsterdam), Vol. 17 (01 1986), pp. 11734.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Masson, Paul R., “The Sustainability of Fiscal Deficits,” Staff Papers, International Monetary Fund (Washington), Vol. 32 (1985), pp. 577605.

  • McCallum, Bennett T., “Are Bond-Financed Deficits Inflationary? A Ricardian Analysis,” Journal of Political Economy (Chicago), Vol. 92 (February 1984), pp.12335.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Modigliani, Franco, “The Economics of Public Deficits,” in Economic Policy in Theory and Practice, ed. by Assaf Razin and Efraim Sadka (New York: St. Martin’s Press, 1987), pp. 344.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Modigliani, Franco, and Arlie Sterling, “Government Debt, Government Spending, and Private Sector Behavior: A Comment,” Working Paper (Boston: Massachusetts Institute of Technology, 1985).

    • Search Google Scholar
    • Export Citation
  • Peacock, Alan T., and Jack Wiseman, The Growth of Public Expenditures in the United States (Princeton, New Jersey: National Bureau of Economic Research, 1961).

    • Search Google Scholar
    • Export Citation
  • Poterba, James M., and Lawrence H. Summers, “Finite Lifetimes and the Crowding Out Effects of Budget Deficits,” NBER Working Paper 1955 (Cambridge, Massachusetts: National Bureau of Economic Research, June 1986).

    • Search Google Scholar
    • Export Citation
  • Reid, Bradford G., “Aggregate Consumption and Deficit Financing: An Attempt to Separate Permanent from Transitory Effects,” Economic Inquiry (Long Beach, California), Vol. 23 (July 1985), pp. 47586.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ricardo, David, “On the Principles of Political Economy and Taxation,” Vol. 1 of The Works and Correspondence of David Ricardo, ed. by Piero Sraffa (Cambridge and New York: Cambridge University Press, 1951).

    • Search Google Scholar
    • Export Citation
  • Sargent, Thomas J. (1982a), “Beyond Demand and Supply Curves in Macroeconomics,” American Economic Review Papers and Proceedings (Nashville, Tennessee), Vol. 72 (May 1982), pp. 38289.

    • Search Google Scholar
    • Export Citation
  • Sargent, Thomas J. (1982b), “The Ends of Four Big Inflations,” in Inflation: Causes and Effects, ed. by Robert E. Hall (Chicago: University of Chicago Press, 1982).

    • Search Google Scholar
    • Export Citation
  • Sargent, Thomas J., and Neil Wallace, “Some Unpleasant Monetarist Arithmetic,” Quarterly Review, Federal Reserve Bank of Minneapolis (Minneapolis), Vol. 5 (1981), pp. 117.

    • Search Google Scholar
    • Export Citation
  • Seater, John J., and Roberto S. Mariano, “New Tests of the Life Cycle and Tax Discounting Hypotheses,” Journal of Monetary Economics (Amsterdam), Vol. 15 (1985), pp. 195215.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tanner, J. Ernest, “An Empirical Investigation on Tax Discounting: A Comment,” Journal of Money, Credit and Banking (Columbus, Ohio), Vol. 11 (May 1979), pp. 21418.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito (1985a), “Fiscal Deficits and Interest Rates in the United States,” Staff Papers, International Monetary Fund (Washington), Vol. 32 (December 1985), pp. 55176.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito (1985b), “The Deficit Experience in Industrial Countries,” in Essays in Contemporary Economic Problems: The Economy in Deficit, ed. by Phillip Cagan (Washington: American Enterprise Institute for Public Policy Research, 1985), pp. 81119.

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito “Do Large Deficits Produce High Interest Rates? A Comment” (unpublished; Washington: International Monetary Fund, 1986).

    • Search Google Scholar
    • Export Citation
  • Tobin, James, Asset Accumulation and Economic Activity (Chicago: University of Chicago Press, 1980).

  • Van Wijnbergen, Sweder, “Interdependence Revisited: A Developing Countries’ Perspective on Macroeconomic Management and Trade Policy in the Industrial World,” Economic Policy: A European Forum (Cambridge, England), Vol. 1 (November 1985), pp. 81137.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Von Furstenberg, George M., Jeffrey R. Green, and Jin-Ho Jeong, “Tax and Spend, or Spend and Tax?” Review of Economics and Statistics (Cambridge, Massachusetts), Vol. 68 (May 1986), pp. 17988.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wallace, Neil, “A Modigliani-Miller Theorem for Open-Market Operations,” American Economic Review (Nashville, Tennessee), Vol. 71 (1981), pp. 26774.

    • Search Google Scholar
    • Export Citation
  • Wallace, Neil, “A Legal Restrictions Theory of the Demand for ‘Money’ and the Role of Monetary Policy,” Quarterly Review, Federal Reserve Bank of Minneapolis (Minneapolis), Vol. 7 (Winter 1983), pp. 17.

    • Search Google Scholar
    • Export Citation
*

* Mr. Leiderman is Professor of Economics at Tel-Aviv University. This paper was prepared while Professor Leiderman was a visiting scholar at the Fiscal Affairs Department.

Mr. Blejer is Division Chief of the Special Fiscal Studies Division in the Fiscal Affairs Department of the Fund. He holds a Ph.D. from the University of Chicago and has taught at the Hebrew University of Jerusalem, Boston Univer-sity, and New York University.

The authors thank A. Drazen, E. Helpman, B.T. McCallum, A.H. Meltzer, and their colleagues in the Fund for their useful comments and suggestions.

1

Other tests of Ricardian equivalence have related interest rates to govern-ment budget variables. Overall, the evidence from these tests is not conclusive. For the most recent work along these lines, see Evans (1985) and Tanzi (1985a, 1986). For previous surveys of the analytical aspects of Ricardian equivalence, see Barro (1978a), Buiter and Tobin (1979), Dotsey (1985), and Tobin (1980).

2

This is the simplest framework that can be used to analyze intertemporal aspects of Ricardian equivalence. The model used is similar in many respects to the open-economy specification in Frenkel and Razin (1987), Chapter 7.

3

Notice that, since this is a two-period model, we have assumed that B, - 0; that is, that all the debt is retired in period 1.

4

Throughout the analysis we assume that government spending does not affect private sector utility. Because the Ricardian hypothesis is concerned with how a given path of government spending is financed by taxation versus by bonds’ sales, this assumption does not affect the analysis. Where the assumption is critical, however, is in empirical tests of the equivalence proposition, which have to control for the effects of changes in government spending (see Section IV).

5

Sargent (1982a) refers to this scenario as the polar Ricardian regime.

6

On Ricardian equivalence in the world economy, see Frenkel and Razin (1985, 1986).

7

See also Barro (1978a) and Carmichael (1982) for an analysis of these and other deviations.

8

For recent work linking consumption to fiscal policy and liquidity constraints, see Hubbard and Judd (1986). They assume that low-productivity individuals cannot borrow against their future income (see Section IV) and thus have a marginal propensity to consume equal to unity with respect to a current-period cut in taxes. In this subsection, we use a different type of constraint.

9

Notice, however, that a similar type of result would arise if private trans-actions in foreign exchange are subject to taxation such that the effective cost of foreign borrowing is higher for the private sector compared with the govern-ment. A specific example of this sort of taxes is discussed in the next subsection.

10

In fact, most of the foreign borrowing by developing countries is done by the public sector or through its guarantees.

11

An implicit assumption here is that the interest rate r0*. charged by foreigners does not depend on the application of the funds by the borrowing government. In fact, some of the effects discussed here could be offset if private sector projects have a higher rate of return than those of the public sector. This could reduce the risk of lending to the private sector, which could reduce or eliminate the spread.

12

For a somewhat related analysis of the effects of open-market operations that stresses the role of differences in the technology of public versus private sector intermediation, see Wallace (1983).

13

Lucas (1986) argues for considering distortionary taxes as the main way to generate deviations from equivalence. On the macroeconomic effects of dis-tortionary taxes in an intertemporal framework, see Aschauer and Greenwood (1985). The effects of budget deficits under distortionary taxes in open econo-mies are analyzed hy Frenkel and Razin (1987). Chapter 8.

14

A similar argument could be made for the case of default by the government on its liabilities, which could be considered a type of distortionary tax. This point is discussed in more detail in the next subsection, under “Uncertainty About Future Taxes,”

15

The notation vara, indicates the ex ante variance of aj across individuals

16

See Chan (1983) and Barsky, Mankiw, and Zeldes (1986). The discussion here draws on Dotsey (1985).

17

For an open-economy formulation, see Frenkcl and Razin (1987).

18

What happens if. in contrast to the assumptions here, the government’s planning horizon is shorter than that of individuals? Obviously, this would affect the policy actions taken by a specific administration. The private sector, how-ever, in making its consumption decisions, will probably take into account how future administrations will deal with the commitments left by the present one. Thus, an intertemporal government budget constraint remains even though ad-ministrations do change.

19

Cukierman (1986) analyzes effects that arise when, owing to uncertainty about the length of lifetime, individuals attach a positive probability to their being bequest constrained at some future date.

20

Some studies have also explicitly considered the effects of social security on consumption.

21

That this is the case is also stressed in HernSndez-Cata’s (1982) previous review of empirical evidence.

22

Studies that distinguish between permanent and transitory components of the explanatory variables are a partial exception.

23

See Kormendi (1983) for a similar estimate.

24

This is the scenario assumed in Sargent and Wallace’s (1981) framework.

25

On the validity of some basic monetarist principles under different fiscal regimes, see Aiyagari and Gertler (1985).

26

As discussed by Sargent (1982a), the degree of substitution between de-mands for assets (such as base money and government bonds) also depends on the prevailing regime.

27

This discussion also suggests that, since the probability of change in regime is not constant, the coefficients in consumption equation regressions such as equation (20) are not likely to be invariant through time.

28

All the figures are taken from the Fund’s International Financial Statistics, Supplement on Government Finance Statistics, Supplement Series, No. 11 (Wash-ington, International Monetary Fund, 1986).

28

In the 1972-83 period the highest growing component of total outlays was interest payments. There was a twofold increase in the relative share of this item to 9½ percent of total outlays.

30

On the deficit experience in industrial countries, as well as conceptual and definitional issues when measuring budget deficits, see Tanzi (1985b).

31

For evidence on the United Kingdom, see Buiter (1985), Obviously, this scenario differs from that of hyperinflationary episodes. Sargent (1982b) devel-ops the argument that changes in the fiscal signals toward less monetization of public debt were an important factor in determining the success at stopping inflation in the European hyperinflations of the 1920s.

32

See also the cross-country comparsion present by Demopoulos, Kartsimbris, and Miller (1988, forthcoming,

33

Using as an example the setup with borrowing constraints (see Section III, under “Borrowing Constraints”), foreign lenders to the domestic government may change their method of determining interest rates depending upon the use that the domestic government makes of the borrowed funds.

IMF Staff papers: Volume 35 No. 1
Author: International Monetary Fund. Research Dept.