The Ambiguous Relationship between Government Deficits and Interest Rates, Study notes of Economics

Download The Ambiguous Relationship between Government Deficits and Interest Rates and more Study notes Economics in PDF only on Docsity!

The Effect of Deficits

The Office of the Assistant Secretary

for Economic Policy

U.S. Treasury Department

For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402

TABLE OF CONTENTS

PART III: INTEREST RATES AND THE FEDERAL DEFICIT:

• BACKGROUND AND COVERAGE....................................... OF FINANCIAL ASSETS
• SOME ASSERTIONS ABOUT THE EFFECTS OF DEFICITS ON THE ECONOMY...
  • Short-term financial effects................................
  • Effects of deficits on cyclical recovery....................
• EFFECTS OF DEFICITS ON INTEREST RATES.........................
  • Deficits and interest rates in a simple Keynesian framework..
  • The role of bonds...........................................
  • Some modifications of the Keynesian framework...............
  • The treatment of wealth....................................
  • Comments on empirical evidence.............................
• DEFICITS AND EXCHANGE RATES..................................
• DEFICITS, INFLATION AND THE MIX OF OUTPUTS...................
  • A digression on crowding out...............................
• INCENTIVE EFFECTS............................................
• POTENTIAL GROWTH AND THE STRUCTURAL DEFICIT..................
• A SHORT NOTE ON THE LONG RUN.................................
• CONCLUSIONS..................................................
• INTRODUCTION................................................. A SURVEY OF THE LITERATURE
• SOME RECENT SURVEYS OF THE DEBATE............................
• MEASURING THE REAL DEFICIT AND DEBT..........................
• MEASURING THE REAL INTEREST RATE.............................
• ECONOMETRIC TESTS OF THE EFFECT OF DEFICITS ON INTEREST RATES.
• THEORETICAL RESULTS FROM BASIC KEYNESIAN ANALYSIS............
• THE KEYNESIAN INVESTMENT AND CONSUMPTION FUNCTIONS...........
• THE EFFECT OF WEALTH IN DEFICIT/INTEREST RATE ANALYSIS...... (Cont.)
• DEMAND FOR MONEY IN KEYNESIAN MODELS........................
• QUANTITY THEORY CRITICISM OF KEYNESIAN MODELS...............
• LOANABLE FUNDS ANALYSIS.....................................
• CROWDING-OUT................................................
• IMPLICATIONS OF GROWTH IN THE FEDERAL DEBT..................
• POTENTIAL INSTABILITY.......................................
• MONETIZATION OF THE DEFICIT.................................
  • RATE ANALYSIS.............................................. EFFECT OF WAGE-PRICE BEHAVIOR IN DEFICIT/INTEREST
  • RATE ANALYSIS............................................. RELATIVE PRICE EFFECTS OF TAXES IN DEFICIT/INTEREST
• RATIONAL EXPECTATIONS.......................................
• CONCLUSION..................................................
• REFERENCES..................................................
• INTRODUCTION................................................ SOME EMPIRICAL TESTS
• EXTENDING THE FELDSTEIN-ECKSTEIN ANALYSIS...................
• RESPECIFICATION OF THE DEFICIT/INTEREST RATE MODEL..........
• EMPIRICAL RESULTS...........................................
• CONCLUSIONS.................................................
• REFERENCES..................................................

PART I

GOVERNMENT DEFICIT SPENDING AND ITS EFFECTS

ON PRICES OF FINANCIAL ASSETS

are the merits of these rather restrictive interpretations of the role of budget deficits as automatic or discretionary stabi- lization tools, questions about the effects of government deficit spending on long-term real economic growth recently have become a focus of attention and controversy.

These effects of government deficits are by no means unam- biguous, for even on a most rudimentary level of analysis the answer would depend, for instance, on whether deficits are caused by spending increases or tax cuts, or whether they are financed by monetization of the debt or by sale of government debt to the public. Similarly, conclusions may vary with such considerations as the composition of government spending that the deficits in question are supposed to finance; the kind of taxes contemplated as a substitute for deficit financing; the openness of a country's capital markets to foreign investors; public expectations generated by a prospect of continuing deficits; behavioral attitudes as reflected in, among others, saving habits; and a host of institu- tional arrangements determining the adaptability of labor, product and asset markets to changing economic conditions, all of which influence the effects of deficits on the allocation of resources within the private sector.

Although the relationship between budget deficits and eco- nomic growth is complex, the problem may be made analytically and empirically tractable by phrasing the discussion in terms of prices. Thus, often the analysis is reduced to the question of the link between budget deficits and the rate of inflation and the prices of financial assets, as exemplified by interest or exchange rates. Such analyses imply that the connections between interest rates and investment or saving (or between the exchange rate and exports or imports), and between real capital accumula- tion and economic growth are thought to be fairly well understood. Therefore, if a link between budget deficits and prices of finan- cial assets could be established, a conceptual short-cut supposedly would allow the analyst to deduce the effects of budget deficits on selected macroeconomic aggregates themselves.

The main purpose of this paper is to review the issue con- cerning the effects of government deficit spending on interest rates, and to some extent on exchange rates. Frequently encoun- tered assertions about the causal links between deficits and prices of financial assets will be critically examined and evalu- ated. More specifically, an attempt will be made to demonstrate that theoretical conclusions about these links have no universal validity but depend crucially, instead, on the time horizon of the analysis, the institutional and behavioral assumptions under- lying the analytical model used, the accompanying circumstances and policies postulated and the size of various economic parameters estimated or assumed. In reviewing assertions about the economic effects of budget deficits, some of the concepts frequently (and rather loosely) used in popular discussion will be clarified,

empirical evidence, to the extent that it exists and is germane to the issues discussed, will be presented and the relationship between budget deficits and a number of economic variables will be examined rather extensively within alternative frameworks of economic analysis.

SOME ASSERTIONS ABOUT THE EFFECTS OF DEFICITS ON THE ECONOMY

Assessments of the impact of budget deficits on interest (and exchange) rates vary from "crucial" to "none." As indicated earlier, contradictory assessments can result from a number of causes.

For example, one analytical framework maintains that there is absolutely no difference between higher deficit spending and spending fully financed by additional taxes. According to this line of argument government borrowing is a perfect substitute for taxation: personal income that is not taxed enters the saving stream, rather than being consumed, thus giving rise to an increase in supply of loanable funds equal to the incremental increase in demand for such funds attributable to additional government spending. The argument is, of course, symmetrical. An increase in taxes accompanied by a reduction in government borrowing requirements of the same amount shifts both the supply and demand curves for loanable funds to the left equally. Therefore, there is no impact on the interest rate whether government spending is financed by taxes or borrowing.

At the other extreme is the claim that there is no substitu- tion whatsoever between taxes and government borrowing. This assertion relies on the supposition that personal income that is not taxed is devoted in its entirety to increased consumption. As a consequence, additional government borrowing is not accom- panied by increased private savings. Thus, an incremental demand for loanable funds in conjunction with their unchanged private supply inevitably results in an upward pressure on interest rates.

In the same vein, an assertion is frequently heard that the existence of arbitrage in international financial markets ensures that capital flows respond instantaneously to incipient interest rate differentials among otherwise similar financial instruments denominated in various currencies. Therefore, to the extent that government borrowing does exert upward pressure on interest rates, it must also contribute to an appreciation of a currency generated by interest-rate-induced capital inflows.

A competing line of reasoning, which introduces expectational elements into the analysis, leads to the opposite result. Since deficit spending, as a reflection of lax fiscal discipline, gives rise to fears about future monetization of public debt, expectations

or perhaps even the direction of the effect of increased government borrowing on interest rates.

The same is true, perhaps even more so, with respect to the short-run determinants of exchange rates. For example, even if one should uncritically accept that increased government borrowing does contribute to higher interest rates, it is by no means self- evident that a currency appreciation follows. The theory of international financial arbitrage (as reflected in the so-called "Fisher open" formula) recognizes only that interest rate differ- entials among currencies tend to equal the corresponding annual- ized forward exchange rate premiums or discounts. If the interest rate differential, say, between the dollar and the yen, widens in favor of the dollar, the only thing certain is that the dollar forward premium will increase (forward discount will contract). This very definitely does not mean that the dollar will appreciate relative to the yen. In fact, in order to satisfy the interest parity condition, while the forward dollar appreciates, the spot dollar may have to depreciate relative to the yen. But, in any event, the short-term impact of increased government borrowing on the exchange rate cannot be unambiguously established by theoretical reasoning alone.

Direct examination of data on deficits and interest and exchange rates has not helped much to establish the effects of government borrowing on the prices of financial assets. There is simply no discernible correlation between changes in government borrowing and changes in either interest or exchange rates. This lack of correlation is not particularly surprising. One reason is that, in fact, things do not remain equal for very long. While changes in government borrowing requirements are relatively mild and occur rather slowly, a variety of constantly shifting factors influence interest and exchange rates, thus accounting for their much greater volatility. Furthermore, monetary authorities customarily try to suppress or moderate the volatility of prices of financial assets by intervening in money and foreign exchange markets, thus rendering the task of discerning a short- term empirical relationship between budget deficits and interest or exchange rates even more difficult.

Finally, whatever these short-run effects are, they have minimal influence on the longer-term evolution of real economic variables. While clearly of great significance to participants in financial markets, the causal link between short-term changes in government borrowing requirements and transitory responses of prices of financial assets is of a relatively minor importance for formulation of economic policy.

° Effects of deficits on cyclical recovery

Some analysts assert that high current deficits will prevent or abort the ongoing economic recovery. The argument behind this assertion is that big deficits cause high interest rates;

high interest rates depress expenditures for business investment, housing, autos, and output of other interest sensitive industries; and the economy cannot recover unless those industries recover. The conditions under which big deficits do or do not cause high interest rates will be examined at length in later sections.

But even if big deficits cause high interest rates, this argument is very questionable because inadequate demand for some categories of output need not prevent a recovery if expenditures for other categories of output (such as consumption of nondurables or defense spending) are sufficiently large. Recovery depends on total production and sale of goods and services, rather than par- ticular categories of goods and services. Large deficits do not reduce total economic activity. Depending on economic conditions (including the rate of money growth), the current deficit may put some upward pressure on interest rates or other prices, but this would indicate that there is more than enough, rather than too little, demand for the available supply.

There is no economic theory to support the assertion that a large current deficit will depress the economy. At most, a large deficit that puts upward pressure on the interest rate may contribute to a bias in the composition of total demand against the output of interest sensitive industries. The extent to which this bias will be pronounced is an empirical matter.

Another assertion is that large expected future deficits will prevent the recovery. The argument behind the assertion is as follows. Future deficits make expected future interest rates high. That keeps present long-term interest rates high, because today people will not lend long term at rates that are below the rate they expect to obtain several years from now. This argument implies that interest rates are higher than the level required to finance the current deficit, given current available loanable funds (savings).

One version of the argument is that prospective deficits result in higher expected inflation, which results in expected higher nominal interest rates in the future, thus causing higher nominal rates now. But even if future deficits cause higher expected inflation (which is by no means self-evident), this argument claims that nominal -- not real -- interest rates rise. However, in a rational world high nominal rates should not restrain investment unless expected real rates also rise.

Another version of the argument, in terms of real interest rates, is rather convoluted. It goes as follows. The current (i.e., FY 1984) deficit does not depress the 1983 economy, and the expected 1988 deficit will not depress the 1988 economy. But the expected 1988 deficit is so large, given the expected 1988 private demand for loanable funds, that it results in an expected interest rate in 1988 that is so high it impedes a return to full

through an increase in real GNP, while if the economy is near full employment of resources the increase will primarily be in prices. With a higher nominal GNP the volume of economic trans- actions in nominal terms is greater, with the result that people need more money to carry out the transactions. Hence, the expansion of the deficit increases the demand for money.

Assuming that the central bank does not accommodate this increase in money demand by increasing the growth rate of the money supply, it is necessary for the velocity of money to rise to meet the enlarged transactions demand for money. This comes about through a rise in interest rates. The enlarged transactions demand for money causes interest rates to rise as transactors are willing to pay more for the use of money. At the same time, an increase in interest rates makes money less attractive as an asset relative to other interest-bearing assets, because the interest rate on money is generally less than that on other assets, so money demanded for asset holdings falls. This decline in money demand induced by higher interest rates offsets the increase in money demanded for transactions, and so interest rates stop rising when the demand for money is brought into balance with the money supply.

° The role of bonds

The discussion presented above shows that in the most basic Keynesian framework an increase in the deficit brought about by a more expansionary fiscal policy without an increase in the money supply tends to raise interest rates. The basic Keynesian frame- work can be elaborated by introducing government bonds into the analysis in at least two ways. In both cases the bonds are regarded as wealth and the bond effect reinforces the tendency of the higher deficit to raise interest rates.

First, the bonds are assumed to be a form of wealth which substitutes for the wealth embodied in real capital. Under this assumption, additional government bonds issued to finance an additional deficit thus are perceived to increase wealth. As wealth (substitutes for capital), the new bonds have the effect of increasing aggregate private consumption spending (reducing saving). This increase in consumption is another addition to final demand, and following the same logic as before, the increase in aggregate demand raises the demand for money and causes an increase in interest rates. This bond effect reinforces the increase in government spending or the reduction in taxes to raise demand and thereby raise interest rates.

A second way in which the increase in bonds can raise inter- est rates is that the bonds can affect money demand directly. The presence of additional bonds in the economy increases the ratio of bonds to money in investors' portfolios. In response,

people attempt to increase their money holdings relative to their bonds by selling bonds. This drives up interest rates, and interest rates continue to rise until the bonds have become so attractive that people are willing to hold them.

The foregoing analysis shows that the typical Keynesian result of an increase in the deficit is a rise in interest rates. However, a special case in which the fiscal expansion does not raise rates is the case of the liquidity trap. The liquidity trap is a situation in which people believe that interest rates are so low that they cannot fall further. Indeed, in this situa- tion, interest rates are expected to rise and the prices of assets (such as bonds) are expected to fall so low that an asset purchaser can expect to sustain a capital loss which counter- balances the interest earned on the asset. Fearing capital loss, people hold money and other very liquid assets rather than long-term assets. Thus, an increase in the demand for money for transactions purposes can be met simply by drawing down enlarged holdings of money without any rise in interest rates. Hence in this case an increase in the deficit does not raise interest rates. The practical significance of the liquid- ity trap, which is believed to occur mostly in depressions, is a subject of dispute.

° Some modifications of the Keynesian framework

The value of the Keynesian paradigm for practical policy analysis depends upon the extent to which it accurately and completely models economic reality. To the extent that the Keynesian model abstracts from important relationships, it may offer inaccurate predictions about the effect of deficits upon interest rates.

Indeed, it appears that the Keynesian model excludes impor- tant economic effects that may well dominate the results in certain circumstances. For one, the demand for money may decline when there is a rise in inflation expected in the near future. This is because inflation reduces the real value of money holdings. Given this effect, an increase in aggregate demand brought about by expansive fiscal policy and higher deficits need not raise interest rates, since people may perceive the additional aggregate demand as potentially inflationary and reduce their demand for money to be held as an asset. In essence the rise in expected inflation has the same effect as an increase in the money supply.

Another effect upon the demand for money is the effect of the business cycle. An increase in demand for output stimulated by fiscal policy may induce a cyclical expansion. In an expansion people have more confidence in their immediate future; hence they are more willing to invest in long-term capital and they have

The argument that government bonds are not wealth is based upon the fact that the bonds must be redeemed or refinanced at a later date. If the bonds are redeemed by a general increase in taxes, taxpayers, on average, face a future tax liability, and this liability offsets, at least in part, the wealth embodied in the bonds. Similarly, if the bonds are monetized in the future, the money created to redeem them will create future inflation, and this will reduce the future purchasing power of money and offset the wealth embodied in the bonds. In these cases rational individuals will adapt their saving behavior to achieve their desired accumulation of real assets. It is only if the bonds are indefinitely refinanced by more bonds that the future tax lia- bilities or the inflation and its attendant loss of purchasing power are avoided.

While in the aggregate government bonds are certainly not wealth, many researchers argue that in practice, for a number of reasons, bonds may be perceived as wealth by their holders and therefore the bonds should be regarded as wealth for the purpose of analysis. For one thing, people may not recognize the future tax liability implied by the bonds. Or they may consider it to be so far into the future that they either discount it substan- tially, or they presume they will not be alive and future genera- tions will have to bear the burden of paying off the liability. Moreover distributional effects may be important; people other than those who own the bonds may have to redeem them. In parti- cular, those who hold bonds may have a higher propensity to save and invest than those who will pay future taxes to redeem the bonds. To the extent that government bond-holders do not face a future liability, they will tend to regard the bonds as a form of wealth substituting for real capital, and in the aggregate the bonds will elicit behavioral responses having the same effect as an increase in wealth.

When the idea that government bonds are not considered wealth is incorporated into the Keynesian model the results change sig- nificantly. For example, if bonds are not viewed as wealth, the effect (discussed earlier) of additional bonds in increasing consumption spending, and thereby increasing overall spending and interest rates, disappears, since this effect is based upon the bonds being perceived as wealth. Similarly the effect of addi- tional bonds in raising directly the demand for money and interest rates also disappears, since if bonds are not wealth they do not affect people's portfolios, and there is no need for individuals to adjust their portfolios when the number of bonds in the economy increases.

If bonds are not perceived as wealth by their holders, the basic Keynesian conclusions about tax cuts unaccompanied by spending reductions also change. For if bonds are not considered wealth, a tax cut has little effect upon aggregate demand, and

its effects are felt almost entirely on the supply side. The logic of this result follows from the fact that the bonds issued to finance the increased deficit brought about by the tax cut create an equal offsetting future liability. A cut in current tax liability, accompanied by a future tax liability of equal present value and a current bond purchase equal in amount to the tax cut leaves financial positions unchanged in the aggregate. Since aggregate financial positions have not changed, aggregate demand will be little affected, and so interest rates will also be little affected through this channel. The primary effect of the tax cut is through incentive effects on the supply side. As analyzed above, such supply-aide effects can lead under different assumptions to either a rise or a fall in interest rates.

The assumption that government bonds are not wealth similarly alters the Keynesian conclusion about the effect of an increase in the deficit brought about by an increase in government spending without a balancing increase in the level of taxation. However, in the case of government purchases there are additional effects, since the government demand preempts real output, and that real output is not available for private consumption or investment. Insofar as that output is no longer used for private capital for- mation, capital intensity will be lower, and this will tend to raise the productivity of each unit of capital and raise real interest rates. On the other hand the productivity of capital and real interest rates are also affected by the uses made of output bought by the government; so depending upon these uses the pro- ductivity of private capital can be either enhanced or diminished.

° Comments on Empirical evidence

The theoretical analysis presented so far indicates that the effects of an increase in the deficit upon interest rates are ambiguous; a situation of rising deficits can coincide with a situation of either rising or falling interest rates. In addition to the reasons given so far, deficits cannot be expected unam- biguously to be causally related to interest rates, because a deficit is a residual obtained by subtracting two items, govern- ment expenditures and revenues, which usually have very different effects upon the economy.

The same deficit can arise with many different levels of expenditures and revenues, and the economy will behave differently when expenditures are large than when they are small even if the deficit is the same in either case. Similarly the effect of the deficit depends on whether it arises from a tax cut or an expen- diture increase. An increased deficit brought about by a tax cut targeted toward stimulating investment may lower pre-tax real interest rates while the same deficit increase brought about by new unproductive government expenditures would probably raise the pre-tax real interest rates. Similarly, the same deficit can