Integration of Fiscal and Monetary Sectors in Econometric Models: A Survey of Theoretical Issues and Empirical Findings

Recent literature on the simulation of existing econometric models of national economies directs attention to the effects of alternative fiscal and monetary policy actions.1 The current debate on economic stabilization policy actions focuses on the impact of the method of financing government spending and, in turn, has led to increased analysis of the government budget constraint. This constraint specifies that the total flow of government expenditure must equal the total flow of financing from all sources. The total flow of financing includes taxes, net government borrowing from the public, and the net amount of new money issued. Budget deficits or surpluses alter the size of public debt, and the method of financing such deficits or disposing of such surpluses affects the composition of private wealth. Hence, any discussion of the effects of fiscal policy actions should distinguish the different monetary repercussions that result from such alternative modes of financing budget deficits or disposing of budget surpluses.

Abstract

Recent literature on the simulation of existing econometric models of national economies directs attention to the effects of alternative fiscal and monetary policy actions.1 The current debate on economic stabilization policy actions focuses on the impact of the method of financing government spending and, in turn, has led to increased analysis of the government budget constraint. This constraint specifies that the total flow of government expenditure must equal the total flow of financing from all sources. The total flow of financing includes taxes, net government borrowing from the public, and the net amount of new money issued. Budget deficits or surpluses alter the size of public debt, and the method of financing such deficits or disposing of such surpluses affects the composition of private wealth. Hence, any discussion of the effects of fiscal policy actions should distinguish the different monetary repercussions that result from such alternative modes of financing budget deficits or disposing of budget surpluses.

Recent literature on the simulation of existing econometric models of national economies directs attention to the effects of alternative fiscal and monetary policy actions.1 The current debate on economic stabilization policy actions focuses on the impact of the method of financing government spending and, in turn, has led to increased analysis of the government budget constraint. This constraint specifies that the total flow of government expenditure must equal the total flow of financing from all sources. The total flow of financing includes taxes, net government borrowing from the public, and the net amount of new money issued. Budget deficits or surpluses alter the size of public debt, and the method of financing such deficits or disposing of such surpluses affects the composition of private wealth. Hence, any discussion of the effects of fiscal policy actions should distinguish the different monetary repercussions that result from such alternative modes of financing budget deficits or disposing of budget surpluses.

This paper has three purposes. First, to review briefly, in the context of a government budget constraint, the theoretical literature on the impact of fiscal and monetary policy. This discussion focuses attention on the long-run effects of money-financed or bond-financed deficit spending. Second, to review the integration of fiscal and monetary sectors of some of the existing econometric models. In particular, attention is given to the IS-LM structures of these models to see how they incorporate an explicit or implicit government budget constraint and whether or not the crowding-out effects are implied by these models under bond-financed deficits. Third, to discuss the results of fiscal and monetary policy simulations as reported by some of these models. This discussion interprets the relative effectiveness of fiscal and monetary policy by analyzing the behavior of dynamic multipliers.

I. The Monetarist Controversy and the Government Budget Constraint

As is well known, the monetarists have challenged the traditional Keynesian view that increases in government expenditure, even if financed by taxes or by borrowing from the public, inevitably lead to increases in income. They contend that the Keynesian view does not fully take account of the displacement of private spending covered by the taxation or borrowing required to finance government expenditure. When government spending is financed through taxation, the displacement or crowding out of private spending through the interest rate effect is not reflected in the balanced-budget multiplier, which in the conventional Keynesian analysis equals unity. Government expenditure financed through bonds also crowds out private spending, and might result in little net effect on total spending.

The crowding-out effects that can occur if the interest rate rises following an increase in government spending can be incorporated in the Hicksian IS-LM framework. In the tax-financed case, the IS curve shifts to the right following an increase in government spending. The initial rise in income generated by a balanced-budget change gives rise to an increased demand for money balances for transactions purposes. With money supply unchanged, the increased demand for money balances causes the interest rate to rise. This depresses private spending, which in turn pushes back the IS curve. The rise in the interest rate also causes a decline in net wealth, which shifts the LM curve to the left. The combined effect of these shifts in the IS and LM curves partially offsets the expansionary effect of a balanced-budget increase in government spending, so that the balanced-budget multiplier is less than unity. Similar shifts in the IS and LM curves take place when expansion in government expenditure is financed by bonds. In this situation, too, the initial rise in income stimulates the demand for money balances, which causes a rise in the interest rate. Although more bonds are now held by the public, the rise in the interest rate reduces the market value of the existing stock of government bonds, and this may cause a decline in net wealth. The increase in the interest rate also leads to a displacement of private spending, and, as a result, the initial expansionary effect of a bond-financed increase in government spending is partially (or even wholly!) offset by perverse wealth effects.2

Much of the force of the monetarists’ arguments on the perverse wealth effects (i.e., the crowding-out effect) associated with bond-financed or tax-financed government spending stems from the fact that, until recently, the Keynesian models of income determination neglected to incorporate the government budget constraint. Such a constraint properly accounts for the injection of new money and bonds into the economy as a result of fiscal policy actions. Moreover, even in the absence of any discretionary fiscal policy actions, a government budget constraint accounts for the repercussions of a pure monetary action on budget deficits or surpluses, In other words, such a constraint closes the relationship between the government and the rest of the economy. In the IS-LM framework, the closure of the system by a government budget constraint affects the IS and LM curves simultaneously.

The government budget constraint reflects the fact that total government expenditure (goods and services, interest payments, and other transfer payments) must equal total financing available from taxes, borrowings from the private sector (i.e., member banks and nonbank public), and the creation of high-powered money. This means that such a constraint restricts the government’s freedom to choose arbitrary values for all the policy variables. That is, given n policy instruments, the government can assign arbitrary values to, at most, n − 1 of them. The remaining one must be determined residually to satisfy the constraint.

A budget deficit or surplus involves the stock/flow changes in net claims held by the central bank and the private sector on the government. Hence, the government budget constraint can be written as

G+BT=dBFC+(dBFB+dBFP)(1)

where G is government expenditure on goods and services and other transfer payments, B is interest payments on outstanding privately held government bonds, T is tax revenues, BFB and BFP are, respectively, the member banks’ and the nonbank public’s net public holdings of government bonds, and BFC is the central bank’s claim on the government, Disregarding foreign transactions, BFC can be expressed in terms of the central bank’s balance sheet as

BFC+Lb=C+R(2)
or,BFC=C+RNB(2)

where C is currency in circulation, R is member banks’ reserves with the central bank, Lb is rediscounts and advances to the member banks, and RNB = R − Lb is unborrowed reserves.

Since changes in the central bank’s net claims on the government equal changes in high-powered money (i.e., dBFC = dC + dRNB), the government budget constraint, equation (1), can now be expressed as 3

G+BT=dC+dRNB+dBFB+dBFP(1)

Now, an increase in G raises income but does not necessarily lead to an equivalent rise in T, which depends on income. Hence, any changes in the budget deficit, which is represented by the left-hand side of equation (1′), must be matched by corresponding changes in the right-hand side of equation (1′). The monetary repercussions of fiscal actions, as captured by the right-hand variables of the government budget constraint, cause changes in high-powered money and privately held government bonds. Changes in the latter aggregates affect the interest rate, which in turn allows the crowding-out effects to take place. It follows then that to the extent that the crowding-out effects can take place following a fiscal action, the government budget constraint provides the necessary link between the fiscal sector and the rest of the economy.4

One of the earliest attempts to analyze the crowding-out phenomenon was made by Musgrave (1959). He employed a Keynesian model of income determination and utilized a simple government budget constraint.5 In the Musgravian system, unless the fraction of claims (β) that people wish to hold in the form of government bonds increases, any sale of bonds to the public raises the rate of interest. Consequently, in such a situation, a government deficit financed by bonds raises the interest rate, which depresses private investment spending, leading to a crowding-out effect on total spending (see equations (3) and (4) in footnote 5). Alternatively, if the budget deficit is financed by money creation, the level of private investment rises. This is so because the increase in the supply of money tends to lower the interest rate (or, at worst, leaves the interest rate unchanged if the ratio V/Ma remains unchanged, which can happen only when the entire increase takes place in the transactions demand for money, i.e., dM = dMt hence, dMa = 0). Also, in the Musgravian framework, the interest-induced wealth effects are not nearly as important as the direct effect of the interest rate on private investment spending. (Notice that, since the wealth variable is not specified in the Musgravian framework, its effects are captured by the introduction of a money supply variable in the consumption function.) As long as the budget deficit leads to an increase in the money supply, government spending has a net expansionary effect on income. When an increase in government spending is financed by a rise in tax receipts, the net effect of such an increase in spending on income is ambiguous. However, in this situation there is a likelihood that the interest rate will rise, since the initial rise in income relative to an unchanged money supply tends to increase the demand for money for transactions purposes (an increase in the transactions demand for money (Mt) reduces asset money (Ma), which causes the interest rate to rise). If this happens, the rise in the interest rate affects investment adversely.6

Among others who have emphasized the importance of including a government budget constraint in standard macroeconomic analysis of fiscal and monetary policy are David Ott and Attiat Ott (1965), Carl Christ (1967, 1968, and 1969), William Silber (1970), Roger Spencer and William Yohe (1970), Bent Hansen (1973), and Alan Blinder and Robert Solow (1973). They have all argued that traditional macroanalysis has not treated bond-financed and money-financed government deficit appropriately, and that this has led to incorrect conclusions regarding the multiplier effects of government spending.7 Except for Bent Hansen (1973), who emphasized the importance of including a government budget constraint to highlight the interdependency of fiscal and monetary policies as stabilization policies, the other authors have emphasized multiplier effects of government spending in the context of such a constraint.

Ott and Ott (1965), employed a standard Keynesian model of income determination in which they introduced a definitional equation for the net financial asset of the public and equated a change in this variable with the government budget deficit in the previous period.8 After solving an IS-LM system incorporating these two equations (equations (a) and (b) in footnote 8), they obtained two important results: (i) the long-run equilibrium government expenditure multiplier (dY/dG) is equal to

the reciprocal of the marginal propensity to tax (1T/Y), which is considerably larger in magnitude than the multiplier obtained by excluding the government budget constraint; (ii) monetary policy (debt management and/or changes in money stock through open market operations in existing government bonds) cannot permanently affect equilibrium income.

The first result can be obtained simply, by an equilibrium analysis. Since budget deficits or surpluses affect financial assets (money and government bonds) held by the public, and since changes in financial assets affect aggregate spending, an equilibrium condition requires that the budget be balanced. Accordingly, given a tax function T = T(Y), this condition is G = T(Y*), where is an equilibrium level of income. A change in government expenditure raises the equilibrium level of income, which in turn induces an increase in tax receipts. But since the marginal tax rate (T=TY) is less than unity, it follows that the increase in government expenditure will be partly deficit financed by an increase in either (high-powered) money or bonds or a combination of both. However, despite the one-shot increase in government expenditure, the budget deficit will persist in subsequent periods until income rises sufficiently to generate tax revenues that will close the budgetary gap. The rise in income thus has to equal the product of incremental government expenditure and the marginal tax rate to generate the necessary increment in tax revenues to rebalance the budget.

This result has been reached by others, in particular, Christ (1968) and Blinder and Solow (1973). It is important to recognize that this result is a direct consequence of the equilibrium condition, which requires a balanced budget. Further, it holds irrespective of the presence of wealth effects and regardless of the interim methods of financing budget deficits.

The second result also follows from the equilibrium condition of a balanced-budget requirement. If an open market policy action increases the money supply, income will rise, because such an operation lowers the interest rate (in the IS-LM framework, the LM curve shifts to the right), resulting in higher tax receipts and, with government expenditure unchanged, a budget surplus. This, in turn, decreases the net financial assets of the public (government debt and money) and has a contractionary effect on income. As long as G remains unchanged, these surpluses will continue to reverse the initial rise in income. Income will continue to decline until the budget is balanced again. Since G = T(Y*), it follows that income, Y, must decline to the original level, Y*.9 Although a pure monetary policy cannot affect the level of equilibrium income, the time path of the movement to the original equilibrium can always be affected by such a policy. Clearly, for a given government expenditure policy and a given tax system, if the current income is below the equilibrium income (or full employment output), an expansionary monetary policy will speed up the rise in income to equilibrium.

Despite its striking conclusions, the Ott and Ott (1965) analysis remained incomplete because it failed to distinguish between the effects of money-financed and bond-financed deficits. It is true, in a comparative static sense, that the multiplier effect of an increase in the level of government spending is independent of the wealth effect regardless of how the interim budget deficits are financed. But this is a long-run proposition and depends entirely on the equilibrium balanced-budget condition (i.e., G = T(Y)). What happens in the process of moving to equilibrium if the interim budget deficits were financed by bonds or money? Would the equilibrium income change be different from that given by dY=1TdG? The Ott and Ott framework is not capable of answering this question, since it has no mechanism to analyze the alternative forms of financing government spending.

In a seminal paper, Blinder and Solow (1973) have tried to provide the answer to the foregoing questions. They pointed out that previous users of the government budget constraint committed an oversight by ignoring interest payments on outstanding government bonds. A proper government budget constraint, therefore, must include interest payments, and it could be written as

dMdt+(1r)dBdt=G+BT(Y+B)(3)

where

article image
article image

Equilibrium is obtained only when dMdt=dBdt=0. Hence, at any equilibrium, budget balance is given by

G+B=T(Y+B)(4)

Assuming an initial equilibrium, an increase in the level of government spending will move income to a new equilibrium. The equilibrium multiplier for government spending as obtained from equation (4) is

dYdG=1+(1T)dBdGT(5)

The multiplier expression (5) allows Blinder and Solow (1973) to analyze the implication of a bond-financed or a money-financed deficit. If budget deficits are entirely money financed, then dBdG=0. This implies that

dYdG|M˙(moneyfinanced)=1T

is the equilibrium multiplier. But if deficits are financed entirely by bonds, then

dBdG>0

so that

dYdG|B˙(bondfinanced)=1+(1T)dBdGT

becomes the equilibrium multiplier. Thus, contrary to the monetarists’ position, the equilibrium multiplier for money-financed deficit spending is smaller than that for bond-financed deficit spending.12

The Blinder and Solow (1973) results are even more striking than those of Ott and Ott (1965). It is well known that, if the LM curve is not vertical, the impact multiplier will be larger if the deficit is financed by creating money. Why then is bond-financed deficit spending more expansionary in the long run? Blinder and Solow give the following intuitive justification. Starting from an initial long-run equilibrium income level with a balanced budget, an increase in government spending if financed by bonds will cause subsequent deficit financing to be larger than if financed by money, for two reasons: (i) income will rise less, so that the induced increase in tax receipts will be smaller, and (ii) the privately held outstanding bonds will increase, and this requires greater interest payments. Additional interest payments in subsequent periods result in higher levels of disposable income. Hence, the increase in income in later periods will be greater under bond financing than under money financing. That is, a given initial budgetary gap is harder to close under bond financing, so that it takes a greater rise in income to induce tax receipts sufficient to close the budgetary gap in a new equilibrium.

Two observations on the Blinder and Solow (1973) results for bond-financed deficit spending are worth making. First, according to their intuitive justification, the impact of a money-financed deficit spending is more expansionary than that for a bond-financed deficit spending in the short run, irrespective of whether a new equilibrium exists or not. Second, the existence of a new equilibrium cannot be decided on a priori grounds.13 A new equilibrium can exist, however, if the wealth effects are nonperverse. The nonperversity of wealth effects is thus an empirical question.14 In addition, even if it is assumed that the wealth effects are nonperverse, theory cannot answer how soon a new equilibrium is reached when the original one is disturbed by a bond-financed deficit spending. If the transition process to a new equilibrium is too long, the fact that bond-financed deficit spending can be more expansionary is not very interesting from the point of view of stabilization, where the length of time elapsed to yield the effects of economic policies is important.

The wealth effects are, however, unambiguously nonperverse for money-financed deficit spending. The existence of a new equilibrium in this case is thus guaranteed on a priori grounds. This result definitely strengthens the monetarists’ position, since they tend to regard money-financed deficit spending as hardly distinguishable from monetary policy. The monetarists’ position could be further strengthened if it could be demonstrated unambiguously that a pure monetary policy was not countered by perverse wealth effects. Blinder and Solow (1973) did not investigate this issue. This omission is somewhat puzzling; quid pro quo, they should have investigated whether monetary policy was effective in the long run.

The following analysis of the effectiveness of monetary policy suggests that a strong case can be made that, contrary to the claims of the monetarists, monetary policy may have perverse effects on income in the long run. The Blinder and Solow framework can be employed for analyzing a pure monetary action, which is defined here as an open market operation.

Let G0 + B0 = T(Y0 + B0) be an initial balanced-budget equilibrium. An expansionary monetary policy action results in a change in the balanced-budget equilibrium by

dB0dM=T(dY0dM+dB0dM)(6)

where dB0dM is the change in interest payments and dY0dM=(1T)TdB0dM is the change in equilibrium income. Clearly, the sign of dY0dM is the same as that of dB0dM, since 0 < T′ < 1. The sign of dB0dM must be negative by definition, since an open market purchase operation is nothing but a swap of government bonds for money. The exact expression is given by

dB0dM=B0rdrdMr<0(7)15

Expression (7) implies that dY0dM<0. That is, a pure monetary expansion has a depressing effect on income at a new equilibrium.

Two comments are in order before giving an intuitive explanation of this paradoxical result. First, a standard IS-LM analysis of an expansionary monetary policy implies that such a policy action shifts the LM curve to the right and lowers the rate of interest (i.e., drdM<0) initially. The fall in the rate of interest raises the value of net wealth, which would have favorable wealth effects on consumption. Also, the increased value of net wealth would shift the LM curve even further to the right at the new equilibrium. This implies that income rises at a new equilibrium. But this has not happened. Second, since income has declined at a new balanced-budget equilibrium, the rate of interest must have risen (i.e., (drdM>0).16 The rise in the rate of interest implies that the wealth effects are perverse for an expansionary monetary policy.17

The reason for this paradoxical result is simply this: Starting from a given equilibrium income level with a balanced budget, an open market purchase policy (increase in money supply) will cause income to rise, as in normal IS-LM analysis, at the beginning of the transitional process to a new equilibrium. The rise in income results in a budget surplus, which is used to retire government debt, so that interest payments decrease. The decrease in interest payments is even greater because of the initial reduction in the private sector’s holdings of government bonds, owing to the open market purchase action. The reduction in the supply of government bonds, given their demand, raises the interest rate, which reduces net wealth. The reduction in net wealth tends to offset the initial rise in income in subsequent periods. This process continues, and the economy reaches a new equilibrium only when income has fallen enough to induce a decline in tax receipts that will match the decline in interest payments so that the budget is balanced again.

The crucial element in the foregoing discussion is the existence of a new equilibrium, irrespective of whether the initial equilibrium is disturbed by a fiscal action or a pure monetary action. Surely, the monetarists’ challenge of the effectiveness of fiscal policy has raised complex issues that cannot be adequately answered on theoretical grounds alone. Although the Blinder and Solow framework attempted to provide answers to these complex problems, their model is too simple. The balanced budget as a precondition for equilibrium envisages a stationary state, since population is given once and for all. Would the effects of fiscal and monetary policy be different if the economy is growing over time? In a growth context, how can one envisage a balanced budget? These are larger and more complex issues, the answers to which are not yet definite, and it is beyond our scope to discuss them here.

The synoptic discussion of the literature carried out thus far suffices to highlight the importance of including a government budget constraint in models of income determination. By including the monetary aspects of fiscal policy with the help of this constraint, new light has been shed on the impact of alternative modes of financing budget deficits or disposing of budget surpluses. Although the theoretical literature focuses on the long-run effects of such methods of budget financing, the fact remains that the analysis of fiscal and monetary policy without a government budget constraint can be faulty, even in the short run.18 With this perspective, it is worth investigating how some of the existing econometric models integrate their fiscal and monetary sectors with the government budget constraint.

II. The IS-LM Structures of Selected Econometric Models

Despite differences in scale, purpose, and internal emphasis, econometric models can be seen as elaborate IS-LM schemes. The IS structures depend on the real product supply and demand equations, while the LM structures relate to the financial sector supply of and demand for assets. The principal structural elements and their behavioral characteristics, as embodied in the IS-LM frameworks, determine the performance of such models.

In this section, references are made to the IS-LM structures of ten econometric models, the first seven of which describe the U. S. economy. The remaining models describe the Canadian, Australian, and New Zealand economies. The ten models are as follows: 19

  • (i) the Brookings Model (James S. Duesenberry, Gary Fromm, Lawrence R. Klein, and Edwin Kuh)

  • (ii) the MIT-PENN-SSRC (MPS) Model (Albert Ando, Franco Modigliani, and Robert Rasche)

  • (iii) the Michigan Quarterly Econometric (DHL III) Model (Saul H. Hymans and Harold T. Shapiro)

  • (iv) the Data Resources (DRI) Model (Otto Eckstein, Edward W. Green, and associates)

  • (v) the Wharton Model Mark III (Michael D. McCarthy)

  • (vi) the Bureau of Economic Analysis (BEA) Model (Albert A. Hirsch, Maurice Liebenberg, and George Green)

  • (vii) the FRB St. Louis (SL) Model (Leonall C. Andersen and Keith M. Carlson)

  • (viii) the RDX2 (Bank of Canada) Model (John F. Helliwell, Fred Gorbert, and associates)

  • (ix) the RBA1 (Reserve Bank of Australia) Model (W. E. Norton and J. F. Henderson)

  • (x) the New Zealand (Reserve Bank of New Zealand) Model (R. S. Deane and associates)

All the models exhibit both the cyclical and trend characteristics of the economies they represent. In other words, the models reflect stock/ flow relations, lag structures, and nonlinearities, which cause fluctuations when subjected to exogenous shocks. The degree of disaggregation of the production sector, except for the Wharton model, is limited, and this sector is completely excluded from the SL model.20

The fiscal sector of all the models treats most government expenditure items and tax rates exogenously. Except for the SL model, which treats “high employment” government expenditure as the only policy variable from the fiscal sector, the other nine models have disaggregated expenditure and tax variables. In several of these models, state and local government expenditure is treated endogenously, and, to the extent that this expenditure is explained by other endogenous variables, the effects of such spending are determined within the models.21 Tax receipts in all the models (the SL model has no taxes) are determined endogenously. In most models, the tax functions are specified in terms of relevant bases and rates and other variables that affect their yields.22 Legal and institutional delays in tax collections are captured by lag-distributed explanatory variables in most models. In short, the fiscal sector in these models, when collapsed in terms of G and T, not only represents the theoretical construct of the government budget constraint but also accounts for cyclical and trend characteristics of the economies that they purport to represent.

In general, the models referred to here (except the SL and DHL III models) explain the behavior of asset holdings (such as currency, bank deposits, and mortgages) by a stock-adjustment theory of portfolio selection.23 Although there is considerable variation between models, the usual practice in models that employ the stock-adjustment theory is to relate the quantity of change in a sector’s holdings of a financial asset to the wealth of that sector, rates of return on the asset itself and on other competing assets, and the level of holdings of the asset at the end of the previous period. Other variables, such as some measures of income, including its lagged values, also influence asset-holding behavior. A few of the models also account for the cross effects of competing assets by employing distributed lags on the asset’s own rate of return as an explanatory variable in explaining a sector’s asset-holding behavior equation.24 In particular, the MPS and DRI models have such indirect cross effects.

A complete and consistent model of a financial sector should be cast, in terms of the balance sheet identities,25 as a general equilibrium sub-model within a complete econometric model. This is so because each sector engages in market transactions that affect its balance sheet and the balance sheets of other sectors through asset demands and liability supplies. Through all such transactions, each sector’s portfolio of assets must balance with its liabilities, irrespective of whether expectations are fulfilled or not. This implies that asset behavioral equations are functionally dependent in a manner that satisfies Walras’ law. Therefore, holdings of an asset in each sector’s balance sheet are always determined as a residual when the remaining assets are determined elsewhere in the system. Consequently, explicit specifications of equations determining the remaining asset holdings imply the unwritten specification of the residual assets. An immediate consequence of this, if cross effects of competing assets on the own asset are ignored, is that the residual asset must adjust appropriately to satisfy the identity. It may happen that the residual asset is unresponsive to such cross effects. In any case, if the explicit asset-holding equations corresponding to each balance sheet’s identity are improperly specified, the burden of adjustment falls inappropriately on the residual assets. This in turn biases the effects of economic policy measures.

To highlight the importance of balance sheet identities in a monetary sector, the following definitions are suggested for purposes of illustration: 26

Treasuryidentity:Gold¯+BF=BFC+BFB+BFP(8)
Centralbankidentity:BFC=C+RNB(9)
Memberbanksidentity:BFB+BP+RNB=D+DT(10)
Memberbankssubidentity:RNB=Rq+Rf(10)
Privatenetwealthidentity:C+D+DT+BFP+KBP=WP(11)27

where

article image
article image

The treasury identity gives the breakdown of the holdings of the stock of government liabilities arising out of gold stock and government debt between the banking system (BFC and BFB) and the nonbank public (BFP). The central bank identity matches its net holdings of government debt to its liabilities: currency plus unborrowed reserves (i.e., high-powered money). The member banks’ identity equates banks’ demand and time deposit liabilities with asset holdings (government debt, loans to private sector, and unborrowed reserves).28 The member banks’ subidentity shows that unborrowed reserves equal the sum of required and free reserves. The private net worth identity relates net worth to money supply, government debt holdings, and nominal value of capital stock.29

These identities form the basis of a monetary system that is linked to the real (and fiscal) sector in the income/expenditure account through the specifications of behavioral equations. For instance, in these econometric models (except the SL model), income influences such financial variables as currency and demand deposits. The influence of income on asset choice is a direct causal link from the income/expenditure accounts to the financial sectors. But linkages run both ways. Some monetary variables (such as interest rates) determined in the financial sector affect various spending flows. In fact, in most of these models consumption and investment are influenced by such variables as wealth (or liquid wealth, as in the Brookings model), bank loans, asset prices, and interest rates. In particular, the housing sectors in most of these models are heavily influenced by net wealth and interest rates. These linkages, even without the incorporation of a government budget constraint, serve to interlock the IS-LM structures in these models.

A complete and consistent linking of the real and financial sectors through the balance sheet identities automatically incorporates the government budget constraint. A budget deficit (surplus) results in a change in outstanding public debt, and the change in the size of debt alters private ownership of claims against the government and the composition of private wealth. Changes in the size of the public debt resulting from budget deficits or surpluses and their modes of financing can be accounted for by balance sheet identities. In fact, the differential of the treasury identity defines its left-hand and right-hand variables, respectively, as the change in the size of outstanding public debt and changes in the stocks of holdings of such debt by the central bank, the member banks, and the nonbank public. This can be shown as

dBF=dBFC+dBFB+dBFP(12)

Since budget deficits or surpluses equal the change in the size of outstanding public debt, a fiscal balance equation can be written as

G+BT(Y+B)=dBF(13)

These two equations can now be combined to specify the government budget constraint proper:

G+BT(Y+B)=dBF=dBFC+dBFB+dBFP=dMdt+1rdBdt(14)30

The exact nature of the specification of these two equations defining the government budget constraint is of importance in defining the effects of alternative forms of financing budget deficits or disposing of budget surpluses. First, the choice of variables in terms of endogenicity or exogeneity in the fiscal balance equation is crucial. In general, econometric models treat G as exogenous and T as functions of appropriately defined tax bases, so that the fiscal balance variable, dBF (i.e., budget deficit or surplus), as shown in equation (13), is determined endogenously as a residual. In the context of the government budget constraint, this implies that at least one of the right-hand variables of equation (14) has to be endogenously determined in the monetary sector. It could be the high-powered money stock (BFC or M) or the privately held government bonds (either BFB or BFP or both).31 The endogenicity of all the right-hand financing items of the government budget constraint in no way implies that the money/bond mix of financing budget deficits cannot be controlled by the authorities. Notice that the high-powered money stock consists of currency and unborrowed reserves. The monetary authorities can always treat unborrowed reserves as a policy instrument and indirectly control the stock of high-powered money.

Second, a change in the size of public debt emanating from a budget will have a different monetary effect depending on what form it takes. For instance, if a change in the size of debt (dBF) is largely in the form of long-term government securities, then the monetary effects of a budget deficit will be less expansionary on money supply and aggregate economic activity than if it takes the form of short-term securities.32 It is not enough therefore to integrate fiscal and monetary sectors with an aggregative net government borrowing variable (dBF) as defined by equation (13). This equation must be sufficiently disaggregated and integrated properly with the monetary sector to account for differential impacts of various components of net government borrowing. Failure to do so implies that the security markets in the monetary sector have to be treated as a single market, because changes in the holdings of short-term and long-term debt by the banking system and the nonbank public cannot be related to short-term and long-term government borrowing unless the latter is disaggregated. Some of the models (e.g., the MPS and the DRI) indeed treat long-term and short-term securities separately.

Finally, if equation (12) is omitted—provided that certain conditions are satisfied—equation (13) is sufficient to imply a government budget constraint. These conditions can be obtained by considering the following case: 33

Suppose that equation (13) is specified and that all but one of the balance sheet items on the right-hand side of the omitted equation (12) are determined elsewhere in the financial sector of an econometric model. In this situation, equation (12) is implicit. This is so because the left-hand variable (dBF) of equation (12) is defined by equation (13) and the single undetermined variable on the right-hand side may now be determined residually.34 The right-hand side variables of the implicit equation (12) are high-powered money stock (BFC) and privately held government bonds (BFB + BFP). Hence, as long as dBF is defined by an equation of type (13) in an econometric model, an explicit determination of either high-powered money or government bonds in its financial sector is sufficient condition for the existence of an implicit government budget constraint. In other words, to allow for the existence of an implicit government budget constraint, one of these variables (either high-powered money stock or privately held government bonds but not both) must be omitted from the financial sector. This is accounted for by Walras’ law, which says that if all other markets are cleared in the financial sector, the omitted asset market will be also. Suppose that the privately held government bond market is omitted in an econometric model. To see how this market is implicitly included in such a model, we trace through the economy’s response to fiscal policy. An increase in government spending raises income both directly and indirectly through the Keynesian multiplier. This rise in income increases tax revenues, but if the high-powered money stock is held fixed, the government must finance the remaining spending by selling bonds to the private sector. The private sector, too, can meet its higher demand for money balances for transactions only by selling government bonds. For both reasons, the interest rate must rise and “crowd out” some expenditure, even in the short run. Thus, the qualitative implications in this sense (i.e., for omitted privately held government bonds) are no different from qualitative implications that would be given by explicit treatment of government bonds in equation (12). It is possible, of course, that mis-specification of behavioral equations in the financial sector will bias the result, but, as a logical matter, if equation (12) is omitted and if either the higher-powered money stock or privately held government bond is omitted in the financial sector, the government budget constraint is specified implicitly.35

The incorporation of the crowding-out phenomenon in the context of an expressed or an implied government budget constraint in these models may now be considered.36 The SL model has no government budget constraint; also, it reflects a monetarist view, since exogenous monetary expansion is the main determinant of total spending (nominal gross national product (GNP) in the model). Again, the interest rate is related inversely to monetary expansion. When a fiscal action is taken (i.e., “high employment” government expenditure is increased) without an accompanying expansion in money supply, the resulting increase in the interest rate has a contractionary influence on GNP, and thus limits the expansionary effects of fiscal action. In other words, crowding-out effects are implicit in the SL model when the budget deficit is financed by bonds.

The DHL III model has a small monetary sector in which only the interest rates for treasury bills, corporate bonds, and commercial papers are determined endogenously. It has an integrated IS-LM framework with a fiscal balance equation determining dBF. The IS-LM structures are interdependent, because real variables affect interest rates and the latter affect consumption, investment, and other macroaggregates, usually with a lag of one period. Although the DHL III model does not have equation (12), the budget deficit (dBF) enters the equation determining the treasury bill interest rate. A rise in government expenditure results in an increase in the budget deficit, which causes the treasury bill rate to rise, and this in turn moves the other two interest rates up. The model allows one period (quarter) to elapse before the effects of increases in interest rates are felt on consumption investment, and hence on income. The DHL III model thus accounts for the crowding-out effects. The model, however, cannot distinguish between bond-financed and money-financed deficits, since these variables do not appear explicitly in the financial sector, and since interest rates are determined endogenously.

In the BEA model, the government budget constraint is specified implicitly. The fiscal balance equation determines the budget deficit endogenously, while the financial sector determines, inter alia, the stock of high-powered money. (Currency demand is determined behaviorally, but unborrowed reserves are a policy variable.) In view of the sufficiency condition for the existence of an implicit government budget constraint, privately held government bonds (BFB + BFP) must be omitted in the financial sector, which, indeed, does occur. The IS-LM structures in this model exhibit the usual interactions; consumption is positively affected by liquid wealth (NP).37 Liquid wealth in turn is behaviorally determined by interaction with the real sector. Interest rates and liquid wealth also affect investment. An interesting feature of the integration of the fiscal and monetary sectors through an implicit government constraint in this model is that the budget deficit affects the treasury bill rate. In the model, an increase in the budget deficit causes the treasury bill rate to rise, which in turn induces a decline in demand for free reserves (Rf). Given the unborrowed reserve subidentity (i.e., RNB = Rq + Rf), it follows that with unborrowed reserves fixed, required reserves (Rq) must rise to maintain the balance. But Rq can rise only if demand deposits (and/or time deposits) rise following an increase in the budget deficit. This implies, however, that the money supply increases. Hence, a pure bond-financed deficit in this model would require a corrective monetary action to offset the induced increase in the money supply; this could be effected by an appropriate reduction in unborrowed reserves. Now, the rise in the treasury bill rate following an increase in the bond-financed budget deficit exerts an upward pressure on corporate bond and commercial paper rates. The real value of liquid wealth is adversely affected by the rise in interest rates. This, in turn, causes consumption and investment to decline. Thus, in the BEA model, an increase in the budget deficit crowds out private spending when it is financed by bonds. A similar analysis would show that if deficits are money financed, it causes an expansion in private spending.

The Wharton model has an implicit government budget constraint. The model has a fiscal balance equation, and its financial sector determines the high-powered money stock but not the privately held government bonds. According to the sufficiency conditions, the latter may be deemed to be determined residually. The budget deficit variable, however, does not enter into any equation of the financial sector. Yet, the model is capable of generating a crowding-out phenomenon when the deficit is bond financed. This is so because interest payments on government bonds (B) are functionally related to the government bond yield. The bond yield, in turn, is inversely related to demand and time deposits. Thus, if the budget deficit is financed by bonds, it exerts an upward pressure on the government bond rate because the interest payments rise. As a result, bank deposits tend to decline, which in turn adversely affects consumption. Also, since the government bond rate affects the long-term interest rate in the same direction and the latter rate affects investment, the IS-LM structures of the Wharton model capture the crowding-out phenomenon when deficits are bond financed.

The Brookings model is an interesting case of an implicit budget constraint. It has a fiscal balance equation but no equation (13). Also, it does not determine the high-powered money stock (since the currency demand equation is suppressed), although it treats government bonds exogenously and regards the unborrowed reserves as the key monetary policy variable. According to the sufficiency condition for an implicit government budget constraint, the stock of high-powered money can be treated as being determined residually. As a result, the implicit government budget constraint sharply distinguishes between money-financed and bond-financed budget deficits. With government bonds treated exogenously, the model assumes that any budget deficits are financed by money. Alternatively, if the budget deficit is accompanied by an appropriate increase in government bonds, the residually determined high-powered money stock remains unchanged, so that deficits are now bond financed. There is, however, one difficulty in the money finance case, because the budget deficit variable does not appear anywhere in the financial sector. Consequently, changes in asset holdings—the model has only two assets: demand deposits and time deposits—are brought about by the interactions with the real sector (variables from the income/ expenditure account appear in the asset demand functions). Apart from determining financial assets, the financial sector also determines three interest rates endogenously, namely, the short-term rate on treasury bills (rS), the long-term rate on government bonds (rL), and the time deposit rate (rT). The gap between rL and rS is determined by an equation that employs the term structure of interest rates of previous periods and government bonds as explanatory variables. Both these variables are exogenous to the model. This equation is written as follows:

rLrS=f(()Στ=1nWτ(ΔrLr),()(dBFB+dBFP))(15)

Clearly, since both the explanatory variables are exogenous, the gap between rL and rS cannot change when budget deficits are financed by money. However, both rL and rS can change because of the shift in the IS curve brought about by a budget deficit. But changes in rL and rs must then be in the same direction and of the same magnitude so that the gap (rL − rs) remains unchanged. Hence, the shift in the LM curve in response to a shift in the IS curve must be such as either to leave both rL and rS unchanged or to have both these interest rates change appropriately to maintain the previous gap after they change. The implication of the effect of a monetary expansion on income in the context of equation (15) in the Brookings model financial sector has been studied elsewhere.38 The conclusion reached in that study was that monetary expansion has relatively less impact on income when government bonds are treated exogenously in equation (15) than it would have if the bonds were treated endogenously. In other words, the money-financed budget deficit in this model may not be as expansionary as one would expect if equation (15) were specified differently.39 On the other hand, the specification of equation (15) ensures that the crowding-out effects are present in the model. This is so because any exogenous change in government bonds (i.e., the bond-financed deficits) alters the gap between rL and rS, and since these interact appropriately with the real sector, the crowding-out phenomenon exists.

The MPS model incorporates the government budget constraint in an interesting manner. The model has the usual fiscal balance equation (13) defining the budget deficit variable, dBF. Instead of equation (12), the model employs an equation in which the change in private net worth is defined as personal saving plus capital gains, that is,

dWP=S+capitalgains(16)

where S is defined in the model in terms of income/expenditure identities. Since S can be related to budget deficit, it follows that equations (13) and (16) implicitly and correctly specify the government budget constraint.40

The financial sector determines, among other variables, all the major balance sheet items behaviorally, excepting the privately held government bonds (BFB + BFP). But this is precisely what is required to meet the sufficiency condition for the existence of an implicit government budget constraint as implied by the fiscal balance equation (13) and equation (16). Although the high-powered money stock is determined by the model, it is the currency demand that is behaviorally determined, while unborrowed reserves are treated as a monetary policy variable. Since government bonds are residually determined and the money supply is controlled by unborrowed reserves, the model is capable of distinguishing between money-financed and bond-financed budget deficits. Needless to say, the IS-LM structure of the model adequately incorporates the crowding-out phenomenon through the interaction of the real and financial variables.

For the remaining four models—the DRI, the Canadian RDX2, the Australian RBA1, and the New Zealand—the government budget constraint is specified explicitly by equations of types (12) and (13) for the first three models,41 while the New Zealand model has an implicit government budget constraint, since it has only the fiscal balance equation and determines money supply behaviorally without determining the privately held government bonds.

Although the crowding-out effects of bond-financed budget deficits are present in all these models, the Canadian RDX2 model has some properties worth mentioning. Unlike most models, the RDX2 does not treat unborrowed reserves as the key monetary policy variable. Instead, the financial sector of the RDX2 has a reaction function determining the short-term interest rate (rS). The main intermediate targets of monetary policy appear in the reaction function in lagged form. In addition, the current period U.S. Treasury bill rate enters the reaction function with a positive coefficient.

The reaction function can be described as follows:

rS=b+b1rS(1)+b2rus+b3BP(1)+b4Lg(1)+b5PC(1)

where

article image

b1 > 0

This shows that, at any time, the level of rs is determined by what happened in the preceding period and by the exogenously given U. S. Treasury bill rate. Other things being equal, any change in rus induces a proportional change in the short-term rate, which in turn alters the composition of asset holdings in the economy through the asset-holding behavioral equations. The influence of rs in the IS-LM structure of the RDX2 ensures that the crowding-out effects take place.

In the RDX2 model, the high-powered money stock is influenced by rS, and privately held government bonds are determined residually in the government budget constraint (i.e., equations of type (12) and (13)). Suppose that the authorities decide to finance the budget deficits by bonds. An increase in the budget deficit will induce an increase in high-powered money that the authorities cannot control unless rS can be changed appropriately. But rS remains unchanged in the current period as long as the U. S. Treasury bill rate does not change. Hence, the model has a built-in mechanism to allow the deficit spending to be partly money financed and partly bond financed. But this is not all. Any change in the current U. S. Treasury rate transmits a change in rS. Thus, even if government could decide on a desired money/bond mix of deficit finance, economic events in the United States could alter the mix. In other words, fiscal and monetary policy making in the RDX2 model is not independent of what happens in the United States.

The foregoing discussion gives enough evidence that the IS-LM structures of these models do allow for the crowding-out effects to take place when budget deficits are bond financed. This is seen in all cases, regardless of whether their government budget constraints are expressed or implied. Hence, if the contractionary effects of fiscal actions without an accommodating monetary policy are important, then empirical findings of these and similar models should bear this out, or at least be indicative, particularly since these models do not employ the balanced-budget equilibrium condition. Besides, as already mentioned, these models have elaborate lag structures and have many institutional features that capture the cyclical as well as trend characteristics of the economies that they purport to represent. Thus, from the point of view of stabilization, these models appear to be adequate in exhibiting rough orders of magnitude of the relative effects of fiscal and monetary actions—albeit with the reservations that arise because of possible misspecification of behavioral relations in these models.

III. Empirical Findings on the Crowding-Out Phenomenon

Most of the econometric models referred to in the preceding section have been subjected to extensive multiplier analysis. The simulated dynamic multipliers are available for fiscal actions with nonaccommodating monetary policy and for pure monetary actions. Despite the nonavailability of multipliers for fiscal actions with an accommodating monetary policy (i.e., money-financed and induced tax-financed deficit spending), the available econometric results do permit a reasonable analysis of the empirical significance of the crowding-out effects of fiscal and monetary actions.42

Before discussing the available econometric results on the crowding-out effects, a few reservations are necessary. First, there is a lack of standardization of exogenous variables and initial conditions between models. For example, some models employ the interest rate to control the money supply, while others use unborrowed reserves for this purpose. Also, the initial conditions for which the multipliers were computed range from the onset of recessions to the middle of booms. Second, all the models explain most of the prices within the models. But the specification of some of the prices (e.g., implicit GNP deflator, consumption deflator) ranges from a simple identity between current and constant values to a highly sophisticated price/output conversion technique, as employed in the Brookings model.43 Third, the models differ significantly in capturing nonlinearities in the relations between real income, employment, and prices, as indicated by the time profile of the distributed lags. Finally, the financial sector of models with an implied government budget constraint (except for the Brookings model) treat privately held government bonds (BFB + BFP) as a single variable.

The implications of these different specifications on the size of the dynamic multipliers are difficult to ascertain. A priori, where the unemployment rate is high or the capacity utilization rate is low, a fiscal stimulus will lead to a greater increment in nominal and real GNP. For similar reasons, dynamic multipliers are expected to have greater magnitudes in recessions than in booms. It is not certain how endogenicity of prices affects the multipliers.44 Although it is generally agreed that a fiscal stimulus causes prices to rise, which dampens the effect of such a stimulus on real GNP, this need not be true to the extent that price equations are inversely related to labor productivity as indeed they are in the Brookings, the MPS, the DRI, and the Canadian RDX2 models.45 Further, when the unemployment rate is high or the capacity utilization rate is low, a fiscal stimulus may have a negligible effect on prices, and, hence, fiscal multipliers may be of greater magnitude in such situations. Similarly, the possibility exists that when privately held government bonds (BFB + BFP) are residually determined in the financial sector, bond-financed government spending may have a greater multiplier effect on GNP than if these bonds were determined by the market mechanism. This is so because residually determined government bonds might not have as much influence on interest rates as the market-determined government bonds.

The preceding reservations should be kept in mind while examining the results of the simulations presented in Tables 1-4. As seen in Table 1, changes in nominal and real GNP resulting from sustained shifts in nondefense government expenditure differ considerably between the models. The complete crowding-out effects (i.e., when a sustained fiscal stimulus eventually produces negative effects) do not occur in nominal terms for a sustained expansion in government nondefense spending with nonaccommodating monetary policy in any of the models, including the monetarists’ SL model (upper half, Table 1). This contrasts with the complete crowding-out effects in real terms (lower half, Table 1). The results indicate that it will take 4 quarters for the SL model and between 16 and 40 quarters for most of the other models before the complete crowding-out effects take place in real terms. In one model (DRI), the complete crowding-out effects take place after 16 quarters of sustained fiscal stimulus; they last for at least 8 quarters after that and reverse themselves thereafter. The crowding-out effects are also affected by the cyclical and trend characteristics of the models. The impact multipliers (i.e., first-quarter multiplier) of the models for real GNP range from 0.5 (the SL model) to 1.8 (the Brookings). In general, the multipliers peak out about the fourth or fifth quarter and maintain this position for another 4 quarters thereafter before the crowding-out effects fluctuatingly dampen the sustained fiscal stimulus and eventually cause real GNP to decline.

Table 1.

Eight Econometric Models: Dynamic Government Nondefense Expenditure Multiplier for Gross National Product—Nonaccommodating Monetary Policy Case

article image
Sources: Gary Fromm and Lawrence R. Klein, “A Comparison of Eleven Econometric Models of the United States,” American Economic Review, Papers and Proceedings of the Eighty-fifth Annual Meeting of the American Economic Association, Vol. 63 (May 1973), p. 391; John F. Helliwell and others, The Structure of RDX2, Bank of Canada, Staff Research Studies, No. 7 (Ottawa, 1971), p. 259.

BEA Model: Period 1962-71. Increase of $1 billion in federal nondefense expenditure; proportion was due to compensation of government employees based on actual data for 1962-71. Decrease of $1 billion (1958 dollars) in personal taxes.

Brookings Model: Period 1956:1-1965:4. Increase of $5 billion (1958 dollars) in government expenditure; decrease of $5 billion in personal taxes. Tax multiplier computed as ratio to deflated and undeflated values of $5 billion, respectively.

DHL III Model: Period 1962:1-1971:4. Increase of $1 billion in nondefense expenditure; decrease of $1 billion in personal taxes.

DRI Model: Period 1962:1-1971:4. Increase of $5 billion in federal nondefense expenditure; decrease of $5 billion in personal taxes.

MPS Model: Increase of $1 billion in exports without accommodating monetary policy; decrease of $1 billion in personal taxes.

FRB St. Louis (SL) Model: Period 1962:1-1966:4. Increase of $5 billion in nondefense expenditure.

Wharton Mark III Model: Period 1965:1-1974:4. Increase of $1 billion in nondefense expenditure with average associated change in government wage bill and employment; decrease of $1 billion in personal taxes.

RDX2 Model: Period 1964:4-1970:4. Increase of $100 million in nonwage government expenditure. Col. (9) assumes no alternative means of financing (i.e., use of money/bond mix). Real gross national product (GNP) nonwage government expenditure multipliers are in constant 1961 dollars.

The RDX2 dynamic multipliers are derived by dividing the effects of a shift of $100 million in gover-

ment expenditure on GNP by 100; that is, dY=YGdG, where dG = 100.

Table 2.

Six Econometric Models: Dynamic Personal Tax Multiplier for Gross National Product

article image

See Table 1, footnote 1.

Table 3.

Two Econometric Models: Government Nondefense Expenditure Multipliers for Real Gross National Product Under Alternative Forms of Financing

(In constant 1958 dollars)

article image
Sources: Vijaya G. Duggal, Lawrence R. Klein, and Michael D. McCarthy, “The Wharton Model Mark III: A Modern IS-LM Construct,” International Economic Review, Vol. 15 (October 1974), Table 12, p. 589; Saul H. Hymans and Harold T. Shapiro, “The Structure and Properties of the Michigan Quarterly Econometric Model of the U.S. Economy,” International Economic Review, Vol. 15 (October 1974), Tables 7-9, pp. 644-47.
Table 4.

Seven Econometric Models: Dynamic Multiplier of Open Market Purchase Policy for Gross National Product

(dY/dM or dY/dRNB in constant 1958 dollars)

article image
Source: Gary Fromm, “Survey of United States Models,” Chapter 9 in The Brookings Model: Perspective and Recent Developments, ed. by Gary Fromm and Lawrence R. Klein (Amsterdam, 1975), pp. 376-418.

BEA Model: Period 1965:4-1970:3. Increase of $1 billion in unborrowed reserves.

Brookings Model: Period unknown. Increase of $1 billion in unborrowed reserves; multipliers obtained from Robert J. Gordon, “The Brookings Model in Action: A Review Article,” Journal of Political Economy, Vol. 78 (May/June 1970), p. 511.

DHL III Model: Period 1957:1-1966:4. Increase of $5 billion in money stock.

DRI Model: Period unknown. Increase of $1.16 billion in unborrowed reserves.

MPS Model: Period unknown. Increase of $11 billion in money stock via increase of $1.9 billion in unborrowed reserves.

FRB St. Louis (SL) Model: Period 1970:4-1980:4. Increase of $5 billion in money stock.

Wharton Mark III Model: Period 1965:1-1970:1. Increase of $1 billion in unborrowed reserves.

The foregoing picture changes somewhat when fiscal stimulus takes the form of personal tax cuts (Table 2). In nominal terms, there is no crowding-out effect in any of the models (the SL has no taxes). In real terms, the complete crowding-out effects in the personal tax-cut case take place in only two models (the BEA and the DRI), compared with five models (excluding the SL and the RDX2 models) where fiscal stimulus came through expansion in government nondefense expenditure. Interestingly enough, the magnitudes of the crowding-out effects differ significantly between these two alternative forms of fiscal stimulus (compare lower halves of Tables 1 and 2). The crowding-out effects are significantly less in magnitude in the personal tax-cut case, compared with those in the expansion of the government nondefense expenditure case. Of course, according to the balanced-budget theorem, the tax-cut multipliers are expected to be a unit less in magnitude than the government expenditure multipliers. This theorem appears to be operative only when crowding-out effects are not complete; although, even in this situation, a few of the models show the difference in magnitude of multipliers to be less than unity between the two forms of fiscal stimulus (compare the corresponding multipliers of the DHL III, DRI, and MPS models).46

In some models where the complete crowding-out effects take place under personal tax-cut fiscal stimulus (the BEA and the DRI models), these effects are significantly less than what can be explained by the balanced-budget theorem.47

Dynamic multipliers for government nondefense expenditure in the accommodating monetary policy case (i.e., money-financed budget deficits) comparable to Table 1 have not been available in published form; however, Fromm and Klein (1973, p. 393) report that such multipliers have been simulated by keeping constant interest rates or expanding unborrowed reserves appropriately to accommodate deficit financing by money creation. According to these authors, some of the models show higher government nondefense expenditure multipliers, but the long-run patterns remain the same or are intensified, compared with the multipliers for government expenditure in the nonaccommodating monetary policy case. According to these two authors, the greatest long-run effects are found in the DRI where money-financed government spending gives 1.5 times greater multipliers after 8 quarters; thereafter, the crowding-out effects begin to assert themselves. The effects for other models are somewhat more uniform over the entire time path of 40 quarters.

A full-fledged comparison of the effects of budget deficit under alternative forms of financing, however, can be made for the Wharton and the DHL III models (Table 3). In both models, the money-financed goverment expenditure multipliers for real GNP are uniformly higher than bond-financed multipliers; the latter are uniformly higher than the tax-financed multipliers. The crowding-out effects fluctuate mildly and are quite small (ranging from 0.1 in the first quarter to 0.6 in the tenth quarter in the Wharton model) in both models. At the end of five to ten years, the crowding-out effects are intensified in the DHL model but even then the complete crowding-out effects do not take place. Although the pure tax-financed fiscal multipliers have the least magnitude, the ranking of multiplier magnitudes under bond-financed and money-financed deficit financing in these two models does not support the theoretical results of Blinder and Solow.

To give a flavor of the effects of pure monetary policy, dynamic multipliers of open market purchase policy on real GNP are given in Table 4. Unlike the effects of fiscal stimulus (Table 1), a sustained monetary expansion exhibits wide variations in its effects. In two of the models (the DHL III and the MPS), a sustained $1 billion expansion in money supply results in less than $1 billion increase in real GNP, while the same policy in the other models results in increases in real GNP that peak, ranging from $1.3 billion (the DRI) to $11.6 billion (the Brookings), at the end of one and a half years.48 However, continued sustained increase in money supply has ever-increasing perverse effects. In four out of the seven models, the complete crowding-out effects take place. Interestingly enough, this happens at the end of five years in all these four models, and the complete crowding-out effects are greatest in the SL model. Moreover, the perverse crowding-out effects do not reverse themselves in any of these models as they do at least in one model (the DRI) under bond-financed fiscal stimulus. In this connection, it is of interest to mention that the bond-financed sustained fiscal stimulus in the RDX2 model does exhibit fluctuating crowding-out effects but the complete crowding-out effects do not occur even at the end of seven years (Table 1, column 8). This contrasts sharply with the crowding-out effects of fiscal stimulus under an accommodating monetary policy (i.e., money-financed deficits).49 In the RDX2 model, the complete crowding-out effects in money-financed deficit spending occur at the end of six years and appear to be intensified thereafter.

What, in general, can be said about these results in the context of an expressed or an implied government budget constraint insofar as the short-run and the long-run effectiveness of fiscal and monetary policies are concerned? First, in the short run, both fiscal and monetary policy actions have significant effects on income. Deficit spending, whether money financed or bond financed, has an expansionary effect on income, although the money-financed deficit spending is more expansionary. Further, the expansionary effect of such policies extends as far as four years in most models. Second, the expansionary effects of both fiscal and monetary policy actions are countered by the crowding-out effects, which are intensified in the long run. Third, the theoretical result of Blinder and Solow that, in the long run, bond-financed deficit spending is more expansionary than money-financed deficit spending holds good in only two models (the DRI and the RDX2—see the lower half, columns 3, 8, and 9, Table 1).50 In most models, the expansionary effects of bond-financed deficit spending are overwhelmed by the crowding-out effects in the long run. Fourth, the balanced-budget theorem is not supported empirically. Depending on the cyclical and trend characteristics of the models, the balanced-budget multiplier can fluctuate widely around unity. Fifth, a monetary expansion, in general, has a weaker effect than a comparable expansion in government spending. In two instances, an increase of $1 billion in money supply has much weaker effects, compared with an equivalent increase in bond-financed government spending. Sixth, despite variations in the relative effects of fiscal and monetary policy actions, most models have similar short-run fiscal multipliers, although these multipliers diverge over long periods. But this is not so for monetary policy, because the cyclical effects of monetary policy actions are more pronounced than those of fiscal policy actions in the long run. Finally, these econometric results show that the economy, in responding to a fiscal or monetary policy disturbance, does not converge to an equilibrium, because the complete crowding-out effects continue to be intensified even after the lapse of ten years.

IV. Concluding Remarks

Three major impressions emerge from this survey. The first is that the incorporation of a government budget constraint in models of income determination properly accounts for the new injection of money and bonds into the economy—a fact often ignored in the past by fiscalists and monetarists alike. As seen in Section I, the recent contributions incorporating such a constraint have shed new light on the monetary repercussions of the government’s budget balance. But critical issues still remain. While these contributions have provided theoretical reasons why bond-financed deficit spending could be more expansionary than money-financed spending in the long run, they have not demonstrated this unambiguously. The basic issue of crowding-out or perverse wealth effects of fiscal policy still remains an open question. Further, by treating prices as exogenous and thus ignoring the possibility of some Phillips curve trade-off, these contributions sidetrack one of the major issues of economic stabilization policies. In addition, there is hardly any literature on the issue of crowding-out effects in the context of a government budget constraint in open economic models of income determination.51

The second is that the integration of fiscal and monetary sectors in an econometric model can be carried out without explicitly specifying such a constraint. As discussed in Section II, under certain conditions a reasonably well-specified financial sector with an explicitly specified fiscal balance equation suffices to guarantee the existence of an implicit government budget constraint. There we saw that while the majority of the econometric models do not explicitly specify such a constraint, it can be done implicitly. An explicit government budget constraint, however, is preferable, on the ground that the omitted financial asset market (in most models it is the market for government bonds) always bears the burden of adjustment in accordance with Walras’ law. If the financial asset markets are estimated by “constrained estimation procedure” so that Walras’ law is satisfied for the residually determined asset,52 then an implicit government budget constraint, in principle, is as good as an explicit one. But this has not been true in most models.53

The final impression, as is evident from Section III, is that, in the short run, fiscal policy actions accompanied by an accommodating monetary policy appear to be more effective compared with pure monetary actions. In the long run, the effects of both fiscal (with or without an accompanying accommodating monetary policy) and pure monetary policy actions are curtailed by the intensification of perverse crowding-out effects.

BIBLIOGRAPHY

  • Andersen, Leonall C. (1971), “A Monetarist View of Demand Management: The United States Experience,” Federal Reserve Bank of St. Louis, Review, Vol. 53 (September 1971), pp. 311.

    • Search Google Scholar
    • Export Citation
  • Andersen, Leonall C. (1973), “The State of the Monetarist Debate,” Federal Reserve Bank of St. Louis, Review, Vol. 55 (September 1973), pp. 28.

    • Search Google Scholar
    • Export Citation
  • Andersen, Leonall C., and Keith M. Carlson, “A Monetarist Model for Economic Stabilization,” Federal Reserve Bank of St. Louis, Review, Vol. 52 (April 1970), pp. 721.

    • Search Google Scholar
    • Export Citation
  • Andersen, Leonall C., and Jerry L. Jordan, “Monetary and Fiscal Actions: A Test of Their Relative Importance in Economic Stabilization,” Federal Reserve Bank of St. Louis, Review, Vol. 50 (November 1968), pp. 1124.

    • Search Google Scholar
    • Export Citation
  • Ando, Albert (1973), “equations in the MIT-PENN-SSRC Econometric Model of the United States” (mimeographed, University of Pennsylvania, January 1973).

    • Search Google Scholar
    • Export Citation
  • Ando, Albert (1974), “Some Aspects of Stabilization Policies, the Monetarist Controversy, and the MPS Model,” International Economic Review, Vol. 15 (October 1974), pp. 54171.

    • Search Google Scholar
    • Export Citation
  • Ando, Albert, and E. Cary Brown, “Personal Income Taxes and Consumption Following the 1964 Tax Reduction,” in Studies in Economic Stabilization, ed. by Albert Ando, E. Cary Brown, and Ann F. Friedlander (The Brookings Institution, Washington, 1968), pp. 11737, especially p. 117.

    • Search Google Scholar
    • Export Citation
  • Ando, Albert, and Franco Modigliani, “The ‘Life Cycle’ Hypothesis of Saving: Aggregate Implications and Tests,” American Economic Review, Vol. 53 (March 1963), pp. 5584.

    • Search Google Scholar
    • Export Citation
  • Bischoff, Charles W., “The Effect of Alternative Lag Distributions,” in Tax Incentives and Capital Spending, ed. by Gary Fromm (The Brookings Institution, Washington, 1971), pp. 61130.

    • Search Google Scholar
    • Export Citation
  • Blinder, Alan S., “Can Income Tax Increases Be Inflationary? An Expository Note,” National Tax Journal, Vol. 26 (June 1973), pp. 295301. See also Bent Hansen and John H. Hotsen, “Fiscal and Monetary Policy Reconsidered: Comment,” American Economic Review, Vol. 61 (June 1971), pp. 44451.

    • Search Google Scholar
    • Export Citation
  • Blinder, Alan S., and Robert M. Solow (1973), “Does Fiscal Policy Matter?” Journal of Public Economics, Vol. 2 (November 1973), pp. 31937.

    • Search Google Scholar
    • Export Citation
  • Blinder, Alan S., (1974), “Analytical Foundations of Fiscal Policy,” in The Economics of Public Finance, ed. by Alan S. Blinder and others (The Brookings Institution, Washington, 1974), pp. 3115, especially pp. 4557.

    • Search Google Scholar
    • Export Citation
  • Brainard, William, “Uncertainty and the Effectiveness of Policy,” American Economic Review, Papers and Proceedings of the Seventy-ninth Annual Meeting of the American Economic Association (hereinafter referred to as American Economic Review, Papers and Proceedings), Vol. 57 (May 1967), pp. 41125.

    • Search Google Scholar
    • Export Citation
  • Branson, William H., “The Use of Variable Tax Rates for Stabilization Purposes,” in Broad-Based Taxes: New Options and Sources, ed. by Richard A. Musgrave (Johns Hopkins University Press, 1973), pp. 26785.

    • Search Google Scholar
    • Export Citation
  • Brown, E. Cary, “Analysis of Consumption Taxes in Terms of the Theory of Income Determination,” American Economic Review, Vol. 40 (March 1950), pp. 7489.

    • Search Google Scholar
    • Export Citation
  • Chow, Gregory C, “On the Long-Run and Short-Run Demand for Money,” Journal of Political Economy, Vol. 74 (April 1966), pp. 11131.

    • Search Google Scholar
    • Export Citation
  • Christ, Carl F. (1967), “A Short-Run Aggregate-Demand Model of the Interdependence and Effects of Monetary and Fiscal Policies with Keynesian and Classical Interest Elasticities,” American Economic Review, Papers and Proceedings, Vol. 57 (May 1967), pp. 43443.

    • Search Google Scholar
    • Export Citation
  • Christ, Carl F. (1968), “A Simple Macroeconomic Model with a Government Budget Restraint,” Journal of Political Economy, Vol. 76 (January/February 1968), pp. 5367.

    • Search Google Scholar
    • Export Citation
  • Christ, Carl F. (1969), “A Model of Monetary and Fiscal Policy Effects on the Money Stock, Price Level, and Real Output,” Journal of Money, Credit and Banking, Vol. 1 (November 1969), pp. 683705.

    • Search Google Scholar
    • Export Citation
  • Christ, Carl F. (1971), “Econometric Models of the Financial Sector,” Journal of Money, Credit and Banking, Vol. 3 (May 1971, Part 2), pp. 41949.

    • Search Google Scholar
    • Export Citation
  • Committee on the Working of the Monetary System [Radcliffe Committee], Report (Cmnd. 827, London, 1959).

  • Corrado, C., “The Long-Run Properties of the MPS Model and Its Relation with Stabilization Policies” (doctoral dissertation, University of Pennsylvania, 1974).

    • Search Google Scholar
    • Export Citation
  • Corrigan, E. Gerald, “The Measurement and Importance of Fiscal Policy Changes,” Federal Reserve Bank of New York, Monthly Review, Vol. 52 (June 1970), pp. 13345.

    • Search Google Scholar
    • Export Citation
  • Data Resources Inc., “The Structure of the DRI Model” (unpublished, mimeographed, undated).

  • Deane, R. S., “Macroeconomic Relationships Within New Zealand: A Preliminary Examination,” in Econometric Studies of Macro and Monetary Relations (Papers presented at the second Australasian Conference of Econometricians held at Monash University, August 913, 1971), ed. by Alan A. Powell and Ross A. Williams (Amsterdam, 1973), pp. 85114. (Hereinafter this book is referred to as Econometric Studies of Macro and Monetary Relations.)

    • Search Google Scholar
    • Export Citation
  • De Leeuw, Frank (1965), “A Model of Financial Behavior,” in The Brookings Quarterly Econometric Model of the United States, ed. by James S. Duesenberry and others (Chicago, 1965), pp. 465530.

    • Search Google Scholar
    • Export Citation
  • De Leeuw, Frank (1969), “A Condensed Model of Financial Behavior,” in The Brookings Model: Some Further Results, ed. by James S. Duesenberry and others (Chicago, 1969), pp. 270315.

    • Search Google Scholar
    • Export Citation
  • De Leeuw, Frank, and Edward M. Gramlich, “The Channels of Monetary Policy,” Federal Reserve Bulletin, Vol. 55 (June 1969), pp. 47291.

    • Search Google Scholar
    • Export Citation
  • Dhrymes, Phoebus J., and others, “Criteria for Evaluation of Econometric Models,” Annals of Economic and Social Measurement, Vol. 1 (July 1972), pp. 291324.

    • Search Google Scholar
    • Export Citation
  • Duesenberry, James S., and others (1965), The Brookings Quarterly Econometric Model of the United States (Chicago, 1965).

  • Duesenberry, James S., and others (1969), The Brookings Model: Some Further Results (Chicago, 1969).

  • Duggal, Vijaya G., Lawrence R. Klein, and Michael D. McCarthy, “The Wharton Model Mark III: A Modern IS-LM Construct,” International Economic Review, Vol. 15 (October 1974), pp. 57294.

    • Search Google Scholar
    • Export Citation
  • Eckstein, Otto, Edward W. Green, and Allen Sinai, “The Data Resources Model: Uses, Structure, and Analysis of the U. S. Economy,” International Economic Review, Vol. 15 (October 1974), pp. 595615.

    • Search Google Scholar
    • Export Citation
  • Eckstein, Otto, and Robert F. Halvorsen, “A Behavioral Model of the Public Finances of the State and Local Sector,” in Public Finance and Stabilization Policy: Essays in Honor of Richard A. Musgrave, ed. by Warren L. Smith and John M. Culbertson (Amsterdam, 1974), pp. 30932.

    • Search Google Scholar
    • Export Citation
  • Eisner, Robert (1967), “A Permanent Income Theory for Investment: Some Empirical Explorations,” American Economic Review, Vol. 57 (June 1967), pp. 36490.

    • Search Google Scholar
    • Export Citation
  • Eisner, Robert (1969), “Fiscal and Monetary Policy Reconsidered,” American Economic Review, Vol. 59 (December 1969), pp. 897905; also, “What Went Wrong?” Journal of Political Economy, Vol. 79 (May/June 1971), pp. 62941.

    • Search Google Scholar
    • Export Citation
  • Evans, Michael K. and Lawrence R. Klein, The Wharton Econometric Forecasting Model, Wharton School of Finance and Commerce (Philadelphia, 1967).

    • Search Google Scholar
    • Export Citation
  • Fand, David I. (1970), “A Monetarist Model of the Monetary Process,” Journal of Finance, Vol. 25 (May 1970), pp. 27589.

  • Fand, David I. (1971), “The Monetary Theory of Nine Recent Quarterly Econometric Models of the United States: A Comment,” Journal of Money, Credit and Banking, Vol. 3 (May 1971, Part 2), pp. 45060.

    • Search Google Scholar
    • Export Citation
  • Fisher, Franklin M., Lawrence R. Klein, and Y. Shinkai, “Price and Output Aggregation in the Brookings Econometric Model,” in The Brookings Quarterly Econometric Model of the United States, ed. by James S. Duesenberry and others (Chicago, 1965), pp. 65379.

    • Search Google Scholar
    • Export Citation
  • Foley, Duncan K., and Miguel Sidrauski, Monetary and Fiscal Policy in a Growing Economy (New York, 1971).

  • Friedman, Milton (1956), “The Quantity Theory of Money—A Restatement,” in Studies in the Quantity Theory of Money, ed. by Milton Friedman (University of Chicago Press, 1956), pp. 321.

    • Search Google Scholar
    • Export Citation
  • Friedman, Milton (1957), A Theory of the Consumption Function, National Bureau of Economic Research (Princeton University Press, 1957).

  • Friedman, Milton (1959), “The Demand for Money: Some Theoretical and Empirical Results,” Journal of Political Economy, Vol. 67 (August 1959), pp. 32751.

    • Search Google Scholar
    • Export Citation
  • Friedman, Milton (1966), “Interest Rates and the Demand for Money,” Journal of Law and Economics, Vol. 9 (1966), pp. 7185.

  • Friedman, Milton (1968), “The Role of Monetary Policy,” American Economic Review, Vol. 58 (March 1968), pp. 117.

  • Friedman, Milton (1970), “A Theoretical Framework of Monetary Analysis,” Journal of Political Economy, Vol. 78 (March/April 1970), pp. 193238.

    • Search Google Scholar
    • Export Citation
  • Friedman, Milton (1972), “Comments on the Critics,” Journal of Political Economy, Vol. 80 (September/October 1972), pp. 90650.

  • Friedman, Milton and Walter W. Heller, Monetary vs. Fiscal Policy (New York, 1969).

  • Friedman, Milton, and D. Meiselman, “The Relative Stability of Monetary Velocity and the Investment Multiplier in the United States, 1897-1958,” in Stabilization Policies, Commission on Money and Credit (Englewood Cliffs, New Jersey, 1963), pp. 165268.

    • Search Google Scholar
    • Export Citation
  • Fromm, Gary, “Survey of the United States Models,” Chapter 9 in The Brookings Model: Perspective and Recent Developments, ed. by Gary Fromm and Lawrence R. Klein (Amsterdam, 1975), pp. 376418.

    • Search Google Scholar
    • Export Citation
  • Fromm, Gary, and Lawrence R. Klein (1969), “Solutions of the Complete System,” in The Brookings Model: Some Further Results, ed. by James S. Duesenberry and others (Chicago, 1969), pp. 362421.

    • Search Google Scholar
    • Export Citation
  • Fromm, Gary, (1973), “A Comparison of Eleven Econometric Models of the United States,” American Economic Review, Papers and Proceedings, Vol. 63 (May 1973), Tables 5 and 6, pp. 39192.

    • Search Google Scholar
    • Export Citation
  • Fromm, Gary, and P. Taubman, Policy Simulations with an Econometric Model (The Brookings Institution, Washington, 1968).

  • Goldfeld, Stephen M., and Alan S. Blinder, “Some Implications of Endogenous Stabilization Policy,” Brookings Papers on Economic Activity: 3 (1972), pp. 585644.

    • Search Google Scholar
    • Export Citation
  • Gordon, Robert J. (1970), “The Brookings Model in Action: A Review Article,” Journal of Political Economy, Vol. 78 (May/June 1970), pp. 489525, especially p. 511.

    • Search Google Scholar
    • Export Citation
  • Gordon, Robert J. (1971), “Inflation in Recession and Recovery,” Brookings Papers on Economic Activity: 1 (1971), pp. 10558.

  • Gramlich, Edward M. (1969), “State and Local Governments and Their Budget Constraint,” International Economic Review, Vol. 10 (June 1969), pp. 16382.

    • Search Google Scholar
    • Export Citation
  • Gordon, Robert J. (1971), “The Usefulness of Monetary and Fiscal Policy as Discretionary Stabilization Tools,” Journal of Money, Credit and Banking, Vol. 3 (May 1971, Part 2), pp. 50632.

    • Search Google Scholar
    • Export Citation
  • Hall, Robert E., and Dale W. Jorgenson (1967), “Tax Policy and Investment Behavior,” American Economic Review, Vol. 57 (June 1967), pp. 391414.

    • Search Google Scholar
    • Export Citation
  • Hall, Robert E., and Dale W. Jorgenson (1971), “Application of the Theory of Optimum Capital Accumulation,” in Tax Incentives and Capital Spending, ed. by Gary Fromm (The Brookings Institution, Washington, 1971), pp. 960, especially pp. 1718.

    • Search Google Scholar
    • Export Citation
  • Hansen, Bent, “On the Effects of Fiscal and Monetary Policy: A Taxonomic Discussion,” American Economic Review, Vol. 63 (September 1973), pp. 54671.

    • Search Google Scholar
    • Export Citation
  • Helliwell, John F., and others (1971), The Structure of RDX2, Bank of Canada, Staff Research Studies, No. 7, Parts 1 and 2 (Ottawa, 1971).

    • Search Google Scholar
    • Export Citation
  • Helliwell, John F., and others (1973), “Some Features and Uses of the Canadian Quarterly Model RDX2,” in Econometric Studies of Macro and Monetary Relations (Amsterdam, 1973), pp. 2647.

    • Search Google Scholar
    • Export Citation
  • Hirsch, Albert A. (1972), “Price Simulations with the OBE Econometric Model,” in The Econometrics of Price Determination, ed. by Otto Eckstein, Board of Governors, Federal Reserve System (Washington, 1972), pp. 23776.

    • Search Google Scholar
    • Export Citation
  • Hirsch, Albert A. (1973), “The BEA Quarterly Model as a Forecasting Instrument,” Survey of Current Business, Vol. 53 (August 1973), pp. 2438.

    • Search Google Scholar
    • Export Citation
  • Hirsch, Albert A., Bruce T. Grimm, and Gorti V. L. Narasimham, “Some Multiplier and Error Characteristics of the BEA Quarterly Model,” International Economic Review, Vol. 15 (October 1974), pp. 61631.

    • Search Google Scholar
    • Export Citation
  • Hirsch, Albert A., Maurice Liebenberg, and George R. Green, The BEA Quarterly Econometric Model, Department of Commerce, Bureau of Economic Analysis, Staff Paper No. 22 (Washington, July 1973).

    • Search Google Scholar
    • Export Citation
  • Hirsch, Albert A., and Gorti V. L. Narasimham, “Simulation Paths and Multipliers in an Econometric Model With and Without Supply Constraints” [actually, the model of the Department of Commerce’s Bureau of Economic Analysis], Paper delivered at North American Meetings of the Econometric Society (Toronto, December 1972).

    • Search Google Scholar
    • Export Citation
  • Homa, Kenneth E., and Dwight M. Jaffee, “The Supply of Money and Common Stock Prices,” Journal of Finance, Vol. 26 (December 1971), pp. 104566.

    • Search Google Scholar
    • Export Citation
  • Hymans, Saul H. (1971), “Prices and Price Behavior in Three U.S. Econometric Models,” in The Econometrics of Price Determination Conference, Board of Governors, Federal Reserve System, and Social Science Research Council (Washington, 1971).

    • Search Google Scholar
    • Export Citation
  • Hymans, Saul H. (1974), “The Inflation-Unemployment Trade-Off: Theory and Experience,” in Readings in Money, National Income, and Stabilization Policy, ed. by W. L. Smith and R. L. Tiegen (Homewood, Illinois, Third Edition, 1974), pp. 16074.

    • Search Google Scholar
    • Export Citation
  • Hymans, Saul H. and Harold Shapiro (1970), The DHL-III Quarterly Econometric Model of the U.S. Economy, Research Seminar in Quantitative Economics (University of Michigan, 1970).

    • Search Google Scholar
    • Export Citation
  • Hymans, Saul H. (1973), “The Michigan Quarterly Econometric Model of the U.S. Economy” [actually, the DHL-Ill model], in The Economic Outlook for 1973, Research Seminar in Quantitative Economics (University of Michigan, March 1973).

    • Search Google Scholar
    • Export Citation
  • Hymans, Saul H. (1974), “The Structure and Properties of the Michigan Quarterly Econometric Model of the U.S. Economy,” International Economic Review, Vol. 15 (October 1974), pp. 63253.

    • Search Google Scholar
    • Export Citation
  • Jorgenson, Dale W. (1963), “Capital Theory and Investment Behavior,” American Economic Review, Papers and Proceedings, Vol. 53 (May 1963), pp. 24759.

    • Search Google Scholar
    • Export Citation
  • Jorgenson, Dale W. (1971), “Econometric Studies of Investment Behavior: A Survey,” Journal of Economic Literature, Vol. 9 (December 1971), pp. 111147.

    • Search Google Scholar
    • Export Citation
  • Katona, George, and Eva Mueller, Consumer Response to Income Increases (The Brookings Institution, Washington, 1968).

  • Keare, Douglas H., and William L. Silber, “Monetary Effect of Long-Term Debt Finance,” American Economic Review, Vol. 55 (June 1965), pp. 51012.

    • Search Google Scholar
    • Export Citation
  • Keynes, John Maynard, The General Theory of Employment, Interest and Money (New York, 1936).

  • Kresge, David T., “Price and Output Conversion: A Modified Approach,” in The Brookings Model: Some Further Results, ed. by James S. Duesenberry and others (Chicago, 1969), pp. 85108.

    • Search Google Scholar
    • Export Citation
  • Kuh, Edwin, “Theory and Institutions in the Study of Investment Behavior,” American Economic Review, Papers and Proceedings, Vol. 53 (May 1963), pp. 26088; see also, in the same volume, Robert Eisner, “Investment: Fact and Fancy,” pp. 23746.

    • Search Google Scholar
    • Export Citation
  • Liviatan, Nissan, “Consistent Estimation of Distributed Lags,” International Economic Review, Vol. 4 (January 1963), pp. 4452.

  • McCarthy, Michael D., The Wharton Quarterly Econometric Forecasting Model MARK III, Studies in Quantitative Economics, No. 6, Wharton School of Finance and Commerce (Philadelphia, 1972).

    • Search Google Scholar
    • Export Citation
  • Modigliani, Franco, “Monetary Policy and Consumption: Linkages via Interest Rate and Wealth Effects in the FMP Model,” in Consumer Spending and Monetary Policy: The Linkages, Proceedings of a Monetary Conference, June 1971, Monetary Conference Series No. 5 (Federal Reserve Bank of Boston, 1971), pp. 2533.

    • Search Google Scholar
    • Export Citation
  • Modigliani, Franco, and Albert Ando, “The Relative Stability of Monetary Velocity and the Investment Multiplier,” pp. 693728; Michael De Prano and Thomas Mayer, “Tests of the Relative Importance of Autonomous Expenditures and Money,” pp. 72952; the reply by Friedman and Meiselman, pp. 75385; and the rejoinders by Ando and Modigliani, pp. 78690, and De Prano and Mayer, pp. 79192, all in American Economic Review, Vol. 55 (September 1965). See also Donald D. Hester, “Keynes and the Quantity Theory: A Comment on the Friedman-Meiselman CMC Paper,” Review of Economics and Statistics, Vol. 46 (November 1964), pp. 36477.

    • Search Google Scholar
    • Export Citation
  • Modigliani, Franco, and Richard Brumbert, “Utility Analysis and the Consumption Function: An Interpretation of Cross-Section Data,” in Post Keynesian Economics, ed. by Kenneth Kurihara (Rutgers University Press, 1954), pp. 388436.

    • Search Google Scholar
    • Export Citation
  • Modigliani, Franco, Robert Rasche, and J. Philip Cooper, “Central Bank Policy, the Money Supply, and the Short-Term Rate of Interest,” Journal of Money, Credit and Banking, Vol. 2 (May 1970), pp. 166218, especially pp. 16869.

    • Search Google Scholar
    • Export Citation
  • Musgrave, Richard A., The Theory of Public Finance (New York, 1959), pp. 52655.

  • Nagar, A. L., “Stochastic Simulation of the Brookings Econometric Model,” in The Brookings Model: Some Further Results, ed. by James S. Duesenberry and others (Chicago, 1969), pp. 42356.

    • Search Google Scholar
    • Export Citation
  • Norton, W. E., and J. F. Henderson (1972), A Model of the Australian Economy: A Further Report, Reserve Bank of Australia, Occasional Paper No. 3G (Sydney, 1972).

    • Search Google Scholar
    • Export Citation
  • Norton, W. E., and J. F. Henderson (1973), “The Structure of a Model of the Australian Economy,” in Econometric Studies of Macro and Monetary Relations (Amsterdam, 1973), pp. 4983.