Measurement of the Public Sector Deficit and Its Implications for Policy Evaluation and Design
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

This paper studies budgetary, financial, and monetary policy evaluation and design in a framework of comprehensive wealth and income accounting. Although the focus is on the public sector accounts, inevitably some attention is paid to the private and overseas sectors. Construction of stylized comprehensive balance sheets for the public sector and for its “flow” counterpart (the change in real public sector net worth) forms the basis for a comparison of these balance sheets with the conventionally measured balance sheet and the flow of funds accounts. The conventionally measured public sector balance sheet typically contains only marketable financial assets and liabilities. On the asset side, it omits such items as the value of the stock of social overhead capital, the value of government-owned land and mineral rights, and the present value of planned future tax revenues. On the liability side, it omits the present value of social insurance and other entitlement programs.

Abstract

This paper studies budgetary, financial, and monetary policy evaluation and design in a framework of comprehensive wealth and income accounting. Although the focus is on the public sector accounts, inevitably some attention is paid to the private and overseas sectors. Construction of stylized comprehensive balance sheets for the public sector and for its “flow” counterpart (the change in real public sector net worth) forms the basis for a comparison of these balance sheets with the conventionally measured balance sheet and the flow of funds accounts. The conventionally measured public sector balance sheet typically contains only marketable financial assets and liabilities. On the asset side, it omits such items as the value of the stock of social overhead capital, the value of government-owned land and mineral rights, and the present value of planned future tax revenues. On the liability side, it omits the present value of social insurance and other entitlement programs.

This paper studies budgetary, financial, and monetary policy evaluation and design in a framework of comprehensive wealth and income accounting. Although the focus is on the public sector accounts, inevitably some attention is paid to the private and overseas sectors. Construction of stylized comprehensive balance sheets for the public sector and for its “flow” counterpart (the change in real public sector net worth) forms the basis for a comparison of these balance sheets with the conventionally measured balance sheet and the flow of funds accounts. The conventionally measured public sector balance sheet typically contains only marketable financial assets and liabilities. On the asset side, it omits such items as the value of the stock of social overhead capital, the value of government-owned land and mineral rights, and the present value of planned future tax revenues. On the liability side, it omits the present value of social insurance and other entitlement programs.

The conventionally measured financial surplus of the public sector, even when evaluated at constant prices, presents a potentially misleading picture of the change in the real net worth of the public sector. One reason is that capital gains and losses on outstanding stocks of government assets and liabilities are not included in the flow of funds. For example, the following are omitted: capital gains or losses that are due to changes in relative prices (e.g., changes in the real value of mineral rights), changes in the real value of nominally denominated public sector debt that are due to inflation, and changes in the real value of foreign-currency-denominated assets and liabilities that are caused by changes in exchange rates.

A second reason is that changes in tax and entitlement programs, in the future revenue base, and in discount rates, etc., may significantly alter the planned or expected future streams of taxes and benefits and their present value. Capital gains and losses on such implicit nonmarketable assets and liabilities are part of the Hicks-Simon concept of income, but they are excluded from the flow of funds accounts.

The differences between the conventionally measured accounts and the comprehensive accounts can be very large. In inflationary periods, large public sector deficits (conventionally measured) may be more than offset by the inflation-induced reduction in the real value of the government’s nominal liabilities. Changes in the current account deficit of the balance of payments (conventionally measured) may be offset or enhanced by changes in the value of external assets and liabilities associated with exchange rate changes. Changes in social security legislation may alter the future flows of benefits and contributions. With efficient forward-looking financial markets, such policy changes will not merely alter future rates of return when the financial implications of current legislation become visible and directly measurable—for example, through changes in the amount of public sector borrowing. They will have an effect on current financial asset prices and rates of return; larger anticipated future deficits may raise current interest rates.

After presenting the comprehensive and conventionally measured accounts for the public sector, the private sector, and the overseas sector, the paper proposes some general rules for policy design. These rules derive from a reasonable policy norm or objective and from rather minimal and uncontroversial assumptions about private sector behavior. To translate these general (and, indeed, perhaps rather vague) rules into concrete policies is a task that is well beyond the scope of this paper because a wealth of country-specific knowledge would be required in each case.

The essence of the argument is that, in a first-best world, private agents, governments, and international organizations would decide on spending, saving, lending, production, and portfolio allocation programs, constrained only by comprehensive wealth or permanent income. Single-period or other short-run “budget constraints” would not represent further effective or binding constraints on economic behavior. The perfect internal and external capital markets required to implement the first-best solution, however, do not exist. Private agents are constrained by the illiquidity and nonmarketability of certain assets (e.g., pension rights, human capital, and expected future tax cuts). Dearth of suitable collateral often renders infeasible the borrowing required to spend in line with permanent income. These cash flow constraints, illiquidity, credit rationing, lack of collateral, non-marketability of certain assets and liabilities, and a host of other capital market imperfections force the actions of private agents and national governments to depart from the behavior that would be optimal if comprehensive net worth or permanent income constraints alone had to be taken into account.

Flow of funds accounting on a cash or transactions basis and the analysis of balance sheets consisting only of marketable claims are useful precisely because they will help to identify the conditions under which the behavior of economic agents is likely to be constrained by factors other than comprehensive net worth.

Within a national economy, conventional accounting helps to decide when and how the national authorities, through appropriate fiscal, financial, and monetary measures, can help private agents to avoid or overcome obstacles to spending and saving in line with permanent income (for households) and impediments to production in pursuit of long-run profit or social net benefit (for enterprises). Within the international economy, conventional accounting serves to identify the conditions under which international organizations should extend or restrict credit to national governments to enable them to develop in line with their long-run potential. Exercises in financial evaluation, such as the Fund’s financial programming, should, therefore, start from two sets of accounts. The first set contains the conventional cash-based flow of funds accounts, the income expenditure accounts of the United Nations System of National Accounts (SNA), and the conventional balance sheets of marketable assets and liabilities. The second set contains the comprehensive balance sheets or wealth accounts outlined in the paper and their flow counterparts, describing the changes in real sectoral net worth over time and thus permanent income—that is, the ultimate accrual-based accounts.

Both national governments and international agencies should design fiscal, financial, and monetary policies so as to induce an evolution of the conventionally measured balance sheet and flow of funds accounts that permits private agents and national economies, respectively, to approximate the behavior that would be adopted if either comprehensive wealth or permanent income were the only binding constraint on economic behavior.

The approach developed in this paper implies that conventional financial planning is an essential input into optimal (or even merely sensible) policy design. It also suggests that a set of comprehensive wealth and permanent income accounts (or the best practicable approximation to them) should complement the conventional data base. Without the conventional accounts, of course, analyses based just on the comprehensive wealth and permanent income accounts will fail to take into account many of the actual binding constraints on economic behavior.

“Stabilization policy,” as viewed in this paper, is potentially useful and effective even if goods and factor markets clear continuously. The existence of capital market imperfections that prevent private agents from spending in line with permanent private disposable income and nations from spending in line with national permanent income is necessary before there can be scope for stabilization policy—that is, policy actions or rules designed to permit smoothing of consumption over time by removing or neutralizing constraints on spending other than permanent income. Successful stabilization policy keeps disposable income in line with permanent income and ensures an adequate share of disposable financial wealth in comprehensive wealth. Another necessary condition for potentially desirable stabilization policy is that governments have access to capital markets on terms that are more favorable than those faced by private agents, or, more generally, it is necessary for governments to have financial options that are not available to private agents. Mutatis mutandis, the same condition applies in an international setting for certain international agencies vis-à-vis national governments. The existence of Keynesian effective demand failures that are due to disequilibria in goods and factor markets would, of course, strengthen the case for stabilization policy.

This view of stabilization policy implies that the government’s financing policies (changes in its tax, transfer, borrowing, and money creation mix) should be used for stabilization rather than for variations in its spending program on goods and services. The spending program should aim to achieve the best feasible public/private consumption mix based on national permanent income.

I. A Stylized Set of Public Sector Accounts

Table 1 presents a stylized and simplified “comprehensive” balance sheet for the public sector. Many definitional problems are ignored; for example, throughout this paper the terms “government” and “public sector” are used interchangeably. (See Boskin (1982).) It is assumed that an extremely heterogeneous set of assets and liabilities can somehow be expressed in common value terms, despite the fact that some of the assets are not marketable (Ksoc) or, even if potentially marketable, may lack a current observable market price (KG). Some assets and liabilities are neither marketable nor tangible and merely represent implicit noncontractual (and reversible) political commitments (T and N).

Table 1.

Comprehensive Consolidated Public Sector Balance Sheet at Current Market or Implicit Prices

article image

Referring to T, N, and AM as present discounted values of future streams of payments or receipts involves a rather cavalier use of certainty equivalence; the conditional mathematical expectations of the uncertain future revenues or outlays are discounted by using “risk-adjusted” discount rates. If, for example, future tax revenues are highly uncertain, T would be correspondingly small. The relevant horizon is, in principle, infinite.

For many purposes, it is better not to attempt to reduce marketable and nonmarketable, implicit and explicit, claims to a common balance sheet measure of value. Instead, each of the items in the balance sheet would be modeled as having potentially distinct behavioral effects. The proper way of handling this will depend on the specifics of the model and the application under consideration. For a preliminary examination of the problem of comprehensive wealth and income accounting in the public sector, the balance sheet in Table 1 is, however, useful.

Most of the items in the balance sheet are self-explanatory. Public sector overhead capital is assumed to yield an implicit rental rsocPksocKsoc, which corresponds to the item pGsoc (public sector consumption of social overhead capital services) on the debit side of the public sector current account. The symbol pGKG represents the balance sheet counterpart of the operating surplus of the public enterprise sector in the public sector current account. This may well be a negative item for some of the secular public enterprise loss makers, in which case it should be moved to the liability side of the balance sheet. The present value of current and capital grants is not entered separately; it can be viewed as subsumed under N or T. Net foreign exchange reserves (E*) are entered separately as an asset rather than netting them out against B*ForB*F+BFe+B˜FPe. For simplicity, only nominal capital-certain bonds and real capital-certain bonds are considered. (See Miller (1982).)

The treatment of money in this exposition of the comprehensive wealth accounting framework is somewhat unusual. The reason for adopting this approach is that it represents the simplest way of introducing a nontrivial role for money. Specifically, it keeps the economy from becoming isomorphic to a barter economy when, in Section V, the accounts of the public and private sectors are consolidated in the investigation of debt neutrality. Money, as a social asset that produces liquidity and convenience services, does not disappear when private and public sector assets and liabilities are netted out. The usefulness of the framework of comprehensive wealth accounting does not depend on the acceptability of this approach to modeling money.

Money has value to the private sector because it yields a flow of imputed nonpecuniary liquidity and convenience services. Let ρM be the nonpecuniary rate of return on money. The value to the private sector of its money holdings is given by VM in equation (1)

VM(t)=1p(t)tαH(t)ρ^M(u,t)etui^(s,t)dsdu1(1)

The assumption that the pecuniary and nonpecuniary yields on money and bonds are equalized at the margin yields

ρM=i=r+p˙^p(2)

Equations (1) and (2) imply that

VM=HP(3)

Let ΠM be the present discounted value of the expected future flow of profits to the government from operating the printing presses. Assuming that cash can be produced without cost, the result is

ΠM(t)=1p(t)tαH˙^(u,t)etui˙^(s,t)dsdu2(4)

Integrating equation (4) by parts produces

ΠM(T)=H(t)p(t)+AM(t)3(4)

where

AM(t)=1P(t)tαi^(u,t)H^(u,t)etui^(s,t)dsdu(5)

Thus, AM(t)—the net value of the government’s cash monopoly—can be interpreted as the present discounted value of the interest income that the central bank expects to earn at each future date on a portfolio of government bonds that is equal in value to the stock of high-powered money at that date.

The conventionally measured public sector balance sheet typically omits from Table 1 all nonmarketable and nonfinancial assets and liabilities—that is, Ksoc, KG, RG, T, N, and AM.

The current and capital accounts of the public sector whose balance sheet is given in Table 1 are represented in Table 2. (See Ott and Yoo (1980).) They are stylized SNA accounts and have a number of significant shortcomings when used uncritically as a guide to the changes over time in the balance sheet—especially as regards the evolution of the real comprehensive net worth of the public sector and its components.

For simplicity, it is assumed that government consumption (Gc) and the imputed rental services from social overhead capital have the same price (p).4 uniform depreciation rate (δ) for different types of capital is also imposed. Foreign exchange reserves are assumed to pay the same interest rate as other foreign-currency-denominated financial claims. All of these assumptions serve only illustrative purposes.

The “public sector budget constraint,” which was rediscovered by macroeconomic theorists in the early 1970s, is obtained by consolidating the current and capital accounts of Table 2. Imputed income and consumption are netted out. Deflating by the general price level yields the conventionally measured public sector financial surplus (at constant prices) given in equation (6)

Table 2.

Public Sector Income and Expenditure and Capital Finance Accounts

(At current market or implicit prices)

article image
τnnpGcPKsocPδKsocPGPδKGi(BH+BFp)epi*(B*H+B*FE*)(rB˜H+B˜F)+rGPGPKG+rRpRPRGPKsocPK˙soc+PGPK˙G+PRPR˙G1P(B˙H+B˙F)ep(B˙*H+B˙*FE˙*)(B˜˙H+B˜˙F)H˙P(6)

Even this “real” surplus, however, is likely to be a poor indicator of the change in the real net worth of the public sector, as defined from the balance sheet in Table 1. This change in the real net worth of the government is given in equation (7)

ddt(WGP)PksocPK˙soc+PGPK˙G+PRPR˙G1P(B˙H+B˙F)eP(B˙*H+B˙*FE˙*)(B˜˙H+B˜˙F)H˙P+1P(T˙N˙)+A˙M+(P˙KsocPKsocP˙P)PKsocPKsoc+(P˙GPGP˙P)PGPKG+(P˙RPRP˙P)PRPRG+P˙P(BH+BF+HP)(e˙eP˙P)ep(B*H+B*FE*)P˙P(TN)5(7)

A comparison of the right-hand sides of equations (6) and (7) reveals that the difference between the “real,” or constant, price surplus and the change in real net worth is due to capital gains and losses (Ω) and to changes in the value of the implicit assets and liabilities (Δ) where

Ω=(P˙KsocPKsocP˙P)PKsocPKsoc+(P˙GPGP˙P)PGPKG+(PR˙PRP˙P)PRPRG+P˙P(BH+BF+HP)(e˙ep˙p)ep(B*H+B*FE*)p˙p(TN)(8)

and

Δ=1P(T˙N˙)+A˙M(8)

As regards Ω, the statement that the change in wealth or net worth equals saving plus capital gains is not surprising. The importance of fully accounting for capital gains and losses on existing government assets and liabilities to obtain a correct understanding of the short-run and long-run implications of past, present, and prospective budgetary, monetary, and financial policies has not, however, been appreciated universally.

Considerable interest attaches to behavior by an economic agent, sector, or group of sectors that leaves real comprehensive net worth unchanged. Such agents or sectors consume their permanent income, and their behavior is ex ante permanently sustainable. For policy design, policies aimed at keeping total-national (public plus private) consumption in line with national permanent income—that is, policies focusing on the consolidated public and private sector comprehensive balance sheet accounts—are of special relevance. These are considered in Section V. While there are certainly valid reasons for optimal consumption to depart from permanent income, such divergences must necessarily be temporary, with overshooting and undershooting of the permanent income benchmark canceling each other in terms of present value. The focus on spending behavior that is consistent with constant real comprehensive net worth should, therefore, come naturally in policy evaluation and design. It is noted that equations (7), (8′), and (8′′) represent ex post or realized measures only. For planning, including consumption planning, the ex ante measures are relevant. They are obtained by replacing actual changes in prices with anticipated changes in prices in equations (7) and 8′), and by substituting anticipated changes in the value of implicit assets and liabilities for actual changes in equations (7) and (8′). In what follows, anticipated capital gains and losses replace the ex post measures whenever planned private or public sector behavior is discussed.

II. Amortization of Public Debt Through Inflation and Currency Appreciation

Let us consider, first, changes in the public sector balance sheet that are due to “pure” or general inflation, which is defined as a situation in which all money prices (including the prices of real capital assets) change at the same rate—that is,

p˙KsocpKsoc=P˙GPG=P˙RPR=p˙p

For reasons of space, ignore capital gains or losses on the implicit assets and liabilities T and N that are caused by inflation.

Inflation-induced changes in real public sector net worth (Ω’) are given by

Ω=P˙P(BH+BF+HP)+(P˙P-e˙e)(B*H+B*F-E*)eP(9)

The Closed Economy

In a closed economy, the last term on the right-hand side of equation (9′) can be ignored, BF = 0 and the reduction in the real value of the outstanding stock of nominally denominated government liabilities is given by Ω’.

Ω=p˙p(BH+Hp)(9)

Proper wealth accounting requires that the amortization of public debt through inflation should be put “below the line” in measuring the financing of the government’s net “real” borrowing.6 Above the line, a higher rate of inflation will (if interest rates are free) swell the measured deficit as nominal interest rates rise with the rate of inflation. If the Fisher hypothesis holds and real interest rates are invariant with respect to the rate of inflation, the increased nominal interest payments associated with a higher rate of inflation will be exactly matched by the reduction in the real value of the government’s stock of nominally denominated interest-bearing debt (Ω’’’), defined by

Ω=p˙pBHp(9)

Subtraction of Ω’’’ from the conventionally measured deficit yields the deficit “at real interest rates”—what the conventionally measured deficit would have been if all interest-bearing debt had been index linked. In models that do not exhibit “pre-Ricardian” debt neutrality, changes in the real value of the stock of government interest-bearing debt are the major proximate determinant of “financial crowding out”—the displacement of private capital formation by government borrowing, holding constant the size and composition of the government’s real spending program. The exact nature (degree, scope, and time pattern) of financial crowding out will, of course, be “model specific.” A number of simple examples are analyzed in a sequel to this paper. (See Buiter (1982 c).) The central (and obvious) point is that, ceteris paribus, private agents (whose portfolio demands are for real stocks of assets if agents are free from money illusion) will absorb additional issues of nominal government bonds equal to the erosion in the real value of their existing holdings caused by (anticipated) inflation, without requiring any increase in the real rate of interest. Such government borrowing, therefore, does not raise the degree to which the public sector competes with the private sector for real investible resources.

The ceteris paribus clause of the preceding paragraph includes a given stock of real money balances. Additional monetary financing equal to the inflation tax on existing money balances, (p˙pHp), leaves real money balances unchanged. A conventionally measured deficit equal to Ω’’, financed by borrowing an amount, p˙pBHp, and by money creation equal to p˙pHP is, therefore, consistent with constant real interest rates and a constant degree of aggregate financial crowding-out pressure.7 Note that subtracting Ω” from the conventionally measured deficit yields a somewhat wider concept of the deficit at real interest rates, since the real rate of return (ignoring nonpecuniary liquidity and convenience services) on high-powered money bearing a zero nominal interest rate is minus the rate of inflation.8

The argument for public sector inflation accounting in the closed economy can be summarized succinctly by using a simplified version of equations (1) and (2). Ignoring Gsoc, Ksoc, and RG, let us assume that PG = p and define GI = KG (net investment by public sector enterprises) and τ˜=τnp (real taxes net of transfers and other benefits). If it is assumed, in addition, that r=ip˙p, then the conventionally measured government budget constraint is given by

H˙+B˙Hp+B˜˙HGc+G1+δKGτ˜+(r+p˙p)BHp+rB˜HrGKG(10)

The change in the real value of the stock of interest-bearing debt is given by

ddt(B˙Hp+B˜H)Gc+GI+δKGτ˜+r(BHp+B˜H)rGKGH˙p(11)

The deficit measure that is relevant for aggregate financial crowding-out pressure on private capital formation, given in equation (11), will depend on the amount of monetary financing permitted by the authorities. Useful benchmarks are (a) monetary financing sufficient to keep the real money stock constant: HP=P˙PH˙P; and (b) monetary financing consistent with a zero trend rate of inflation: H˙P=γHP, where γ is the natural rate of growth.9

Equation (11) answers the questions as to whether the fiscal stance (defined by Gc, GI, and τ˜) and the monetary target (defined by H˙p) imply aggregate financial crowding-out pressure (ddt(BHp+B˜H)>0) or crowding-in pressure (ddt(BHp+B˜H)<0). This issue can be addressed in the short run (for a single period), in the medium term (by applying equation (11) sequentially for as many periods as one is interested in), or in the steady state. Note that inflation-induced capital gains or losses on nonindexed bonds cancel the inflation premium in the nominal interest payments; in equation (11), all debt service is evaluated at real rates of interest.10

For aggregate crowding-out pressure on total national (private plus public sector) capital formation, a useful simple measure (noting that GI=K˙G) is

ddt(BHp+B˜HKG)=Gcτ¯+r(BHp+B˜HKG)+(r(rGδ))KGH˙p(12)

The conventional deficit measure is further modified in equation (12) by subtracting net investment by public sector enterprises. Interest payments on net nonmonetary liabilities (BH + BH – KG) are evaluated at the real interest rate, r. If the net rate of return on public enterprise capital (rG – δ) exceeds the opportunity cost of borrowing (r), the “corrected” deficit is further reduced. If the opposite prevails, the corrected deficit is larger by an amount (r – (rG – δ)) KG.

The decline in the real value of total tangible net worth of the public sector is given by

ddt(H+BHp+B˜HKG)=Gcτ¯+r(BHp+B˜HKG)+(r(rGδ))KGp˙pHp(13)

This could be called the inflation-corrected government current account deficit. Debt service payments and receipts on all assets and liabilities (including money) are evaluated at real rates of return.11

Some idea of the magnitude of the overstatement of the government’s true borrowing by the conventionally measured deficit under inflationary circumstances is provided by Table 3 for the United Kingdom and Table 4 for the United States.

Table 3.

United Kingdom: Correcting the Public Sector Deficit for Inflation, 1967–81

article image
Source: Marcus Miller, “Inflation-Adjusting the Public Sector Financial Deficit: Measurement and Implications for Policy,” in The 1982 Budget, ed. by T. Kay (London, 1982).

GDP = gross domestic product.

Inflation correction (1) = annual rate of inflation times market value of public sector debt (midyear).

Inflation correction (2) = annual rate of inflation times nominal value of public sector debt.

Inflation correction (3) is based on the assumption of a long-run real interest rate of 2 percent.

Table 4.

United States: Federal Deficits and Debt, Fiscal Years, 1967–811

(Cols. 1–4 in billions of U.S. dollars)

article image
Source: Economic Report of the President (Washington, February 1982).

Col. (3) = col. (2) deflated by consumer price index. Col. (4) = col. (2) times proportional rate of change of consumer price index.

In 1981 the public sector borrowing requirement in the United Kingdom was £10.6 billion, and the public sector financial deficit rose to £7.5 billion. The inflation correction in that year amounted to about £11 billion, using a variety of estimates. The inflation-corrected deficit was actually a surplus. If it is noted that during 1981 the U.K. economy was also experiencing the worst recession since the 1930s, there can be no doubt that the inflation-corrected and cyclically adjusted (trend or permanent) deficit was actually a sizable surplus. It is a matter of some practical importance whether that constitutes wise countercyclical fiscal policy. The United States during the period 1979–81 also had an inflation-corrected balanced federal budget. Any reasonable cyclical correction for 1981 produces a large inflation-corrected cyclically adjusted surplus. High U.S. real interest rates in 1981 can be explained by the fiscal stance only if large anticipated future inflation-corrected cyclically adjusted deficits are postulated.

The Open Economy

In an open economy, governments can borrow and lend domestically or abroad. Their financial assets and liabilities can be denominated in foreign or domestic currency or can be index linked. Consider equation (9′). The real value of public sector debt denominated in domestic currency is reduced by domestic inflation whether this debt is owned by the private sector or the rest of the world. While, ceteris paribus, inflation also reduces the real value of foreign-currency-denominated financial claims, exchange rate depreciation increases it. If purchasing power parity holds (p˙pe˙e=p˙*p* and through choice of units, ep* = p), equation (9′) becomes

Ω=p˙p(BH+BF+Hp)+p˙*p*(B*H+B*FE*p*)(9)

With purchasing power parity, reductions in the real value of foreign-currency-denominated public sector debt can be calculated by multiplying the foreign rate of inflation by the real value of net foreign-currency-denominated liabilities.

Consider the following stylized representation of the position of a number of small, open developing countries that lack a significant domestic capital market. Government debt is largely placed abroad and tends to be denominated in foreign currency (typically U.S. dollars). In such countries, BH = BF = B˜H = B˜F = B*H = 0. The conventionally measured public sector deficit is12

H˙p+ep(B˙*FE˙*)=Gc+GI+δKGτ¯+ep*(B*FE*)rGKG(14)

If, in addition, only the government borrows overseas, ddt(B*FE*) equals the current account deficit (in terms of foreign currency) of the balance of payments, as shown in equation (15)

ep(B˙*FE˙*)=X+epi*(B*FE*)(15)

Here, X denotes real net exports of goods and services (excluding debt service) plus net transfers and grants from abroad.

Compare the current account balances of two countries, identical in real terms but facing different rates of world inflation. If r* is the world real rate of interest, i*=r*+p˙*p*

or

ep(B˙*FE˙*)=X+ep(r*+p˙*p*)(B*FE*)(15)

If the world real rate of interest is independent of the inflation rate and if purchasing power parity prevails, the current account deficit of the country facing the higher rate of world inflation (p˙*p*)1 will exceed that of the country facing the lower rate of world inflation (p˙*p*)2 by an amount ((p˙*p*)1(p˙*p*)2)e(B*FE*p) that is equal to the difference in external debt service payments. This difference in current account balances should, however, have no real consequences, since the higher debt service item above the line is matched below the line by the larger reduction in the real value of its external liabilities; higher world inflation means faster amortization of external indebtedness. Thus, ddt(ep(B*FE*p)), or the change in net real external liabilities, is the same in the two economies. The country facing the larger current account deficit owing to higher world inflation should be able to borrow to finance its higher external interest payments. (See Sachs (1981).)

What has occurred in recent years is an increase in world real interest rates (r*). This increase does require adjustment rather than, or in addition to, mere financing, with the relative weights on adjustment versus financing depending on the extent to which the increase in world real interest rates is perceived as permanent rather than transitory. Also, to the extent that countries have borrowed on a long-term rather than a short-term basis (or at variable interest rates), unanticipated changes in interest rates will result in once-and-for-all real capital gains or losses on external debt. Finally, significant departures from purchasing power parity have been the rule, especially since the breakdown of the Bretton Woods system of par values. Thus, even with a given world real interest rate (r*), a country’s real external indebtedness will increase whenever p˙*p*(p˙pe˙e)—the excess of the world rate of inflation over the domestic rate of inflation minus the percentage depreciation of the exchange rate—increases.

Many other kinds of open economy can be analyzed, starting from the general framework of equations (6), (7), and (9′), but the general principles should be clear from the simple example that was just analyzed.

III. Budgetary Policy and Monetary Growth: Eventual Monetization of Deficits

If bond financing of deficits causes concern about the crowding out of private capital formation and, in the open economy, about possible adverse consequences for external indebtedness, monetization of deficits is a source of concern because of its inflationary implications. It has been seen that it was necessary to correct the conventionally measured budget deficit for the effects of inflation and exchange rate appreciation on the real value of outstanding stocks of public sector financial assets and liabilities in order to assess changes in the extent to which the public sector competes with the private and overseas sectors for investible resources.

Similar adjustments are required to understand the monetary implications of the deficit, as is shown in this section.

The Closed Economy

From the simplified government budget constraint in equation (10), the following expression is derived for the proportional rate of growth of the nominal money stock.13

H˙H =V[Gc+GI+δKGτ¯Y+(r+p˙p)BHpY+rB˜HYrGKGYB˙HpYB˜˙HY](16)

VpYH is the income velocity of circulation of money. To evaluate the implications of the fiscal stance for monetary growth, it is necessary to specify paths both for public spending and taxation and for nonmoney financing. A particularly useful benchmark for financing policy is one that keeps constant the real values of all government assets and liabilities (other than money) per unit of output. This policy would be one of constant crowding-out pressure per unit of output. These constant liability/output (or asset/output) ratios need not be the historically inherited ones. The exercise can be applied to evaluating the longer-run implications for monetary growth after the debt/output ratios have acquired some desired long-run (or even steady-state) values. Given this rule,

GIKG=B˜˙HB˜H=γ

and

B˙HBH=γ+p˙p

Equation (16) then becomes

H˙HV[Gcτ˜Y+(rγ)[BHpY+B˜HYKGY]+(r(rGδ))KGY](17)

Defining the longer-run fiscal stance by given constant values of BHpY,B˜HY,and KGY and by given, but not necessarily constant, paths of GcY and τ˜Y, it can be seen from equation (17) that longer-run monetary growth is governed by a deficit concept that differs from the conventionally measured deficit in a number of ways. First, the reduction in the real value of the stock of nominal government bonds owing to inflation is subtracted from the conventional measure. Second, in a growing economy the real stocks of government assets and liabilities can increase at the natural rate γ while leaving the asset/output or debt/output ratios constant. The net debt service term in equation (17), therefore, involves the real growth-adjusted interest rate (r – γ). Under inflationary conditions, this rate can be significantly less than i=r+p˙p—the nominal interest rate. To infer the long-term implications for monetary growth (and thus for inflation) of the fiscal stance, a correction for inflation is applied only to the interest-bearing component of the government’s nominal liabilities. The conventionally measured deficit should not also be reduced by the erosion of the real value of the nominal stock of high-powered money balances (p˙pHp) because constancy of the real value of all (monetary and nonmonetary) government debt per unit of output is consistent with any deficit and any rate of inflation.

Large conventionally measured deficits (even if cyclically adjusted) that correspond to small inflation-corrected deficits (or even surpluses)14 reflect current high inflation. They do not indicate the inevitability of high crowding-out pressure or high rates of monetary growth in the future. Even without correction for real growth, an inflation-corrected or “trend” surplus means that (a) with zero money financing, there would be (aggregate) crowding in, and (b) with a bond-financing policy of zero (aggregate) crowding in, there would be negative monetary base growth.

Equation (17) alone does not lead to conclusions about the effects of, say, changes in fiscal stance on monetary growth. It is necessary to use positive economic models to incorporate the effect of any parameter changes on endogenous variables, such as velocity (V), real rates of interest (r and rG), and even the natural rate of growth (γ). Such an analysis is simplest in classical monetarist models, such as that of Sargent and Wallace (1981), in which velocity, the real interest rate, and the natural rate of growth are constants, but equation (17) can be incorporated in models of any type. (See also Buiter (1982 a and 1982 b).)

The Open Economy

From the budget constraint of the simplified open economy, the expression for the percentage growth rate of the nominal money stock given in equation (18) can be obtained, as follows:

H˙H=V[Gc+δKGτ˜Y+(r+p˙p)(BH+BFpY)+r(B˜H+B˜FY)+i*epY(B*H+B*FE*)rGKGY+K˙GY1p(B˙H+B˙FY)(B˜˙H+B˜˙FY)ep(B˙*H+B˙*FE˙*Y)](18)

To evaluate the longer-run monetary implications of the fiscal stance, it is again assumed that all stock/flow ratios on the right-hand side of equation (18) are kept constant. Equation (18) then reduces to

H˙H=V[Gcτ˜Y+(rγ)(BH+BFpY+B˜H+B˜FYKGY)+(i*(p˙pe˙e)r)(B*H+B*FE*pY)e+(r(rGδ))KGY](19)

With purchasing power parity, this simplifies to

H˙H =V[Gcτ˜Y+(rγ)(BH+BFpY+B˜H+B˜FYKGY)+(r*γ)(B*H+B*FE*p*Y)+(r(rGδ))KGY](19)

The evaluation of the long-term monetization implied by the fiscal stance requires the consideration of a deficit measure that has nominal debt service payments “corrected” for the effects of domestic inflation, exchange rate appreciation, and real growth.

In any particular period, the economy may well be far removed from the long-run trend captured in equations (17) and (19) or (19′). Actual monetary growth in the short run will be given by equations (16) or (18). If current inflation is a function only of current monetary growth, as would be true, for example, if velocity were constant, the price level were perfectly flexible, and output grew at its exogenously given trend rate γ, then p˙p=H˙Hγ. Authorities concerned with inflation in the short run may not be gratified to know that the long-run rate of inflation implied by their fiscal stance is low, if current monetary growth and inflation are high. If, as seems more likely, current inflation is a function of current and past monetary growth, and a fortiori if current inflation depends also on anticipated future monetary growth (as it does in models with forward-looking rational expectations), then the long-run monetary growth expressions in equations (17), (19), and (19′) become relevant even for short-term and medium-term policy.

IV. Role of Implicit Assets and Liabilities

On the asset side of the public sector balance sheet, there are T, the present value of planned or anticipated future tax revenues, and AM, the imputed value of the government’s cash monopoly. On the liability side is N, the present value of future transfers and benefits under various entitlement programs. This section considers how the value of these implicit assets and liabilities changes over time, with the focus on N. The treatment of T, AM, and private sector human wealth (in Section V) is analytically identical. The symbol N is defined in equation (20) as15

N(t)tαetui^(s,t)dsn^(u,t)du(20)

The change in the present discounted value of expected future benefits is given by

ddtN(t)=i(t)N(t)n(t)+tαetui^(s,t)ds[tn^(u,t)n^(u,t)tuti^(s,t)ds]du(21)

The first two terms on the right-hand side of equation (21) show how the present value of future benefits changes if all expectations concerning the future flow of benefits and future interest rates remain the same. The last term shows the effect of changes (at time t) in expectations concerning future benefits (tn^(u,t)) and future interest rates (ti^(s,t)). As expected, upward revisions in future benefit entitlements raise the value of N, while higher future expected interest rates lower its value.

The only item on the right-hand side of equation (21) that appears in the cash-based public sector deficit or flow of funds accounts is n(t), current benefit payments; i(t)N(t) does not appear because future entitlements are not a marketable interest-bearing liability of the authorities. Changes in planned or expected future benefit entitlements appear in the accounts only if and when they actually become payable in the future, yet such “revaluations” of N are of considerable policy interest. Even if financial markets are not forward-looking—even if government borrowing affects market rates of return only when it actually occurs—increases in N unmatched by increases in T (or by cuts in other spending programs) imply increased future borrowing or money issues and thus mean trouble for the future. Financial markets, furthermore, appear to be linked intertemporally (as formalized, for example, by models of efficient asset market equilibrium that incorporate forward-looking rational expectations). A larger anticipated future borrowing requirement therefore affects asset prices and rates of return today. An unanticipated increase in future expected (inflation-corrected) deficits crowds out private spending today. The intangible items in the public sector balance sheet must be taken into account.

V. Public Sector Accounts and Private Behavior

The Private and Overseas Sectors’ Accounts

Comprehensive balance sheets analogous to the public sector balance sheet of Table 1 are drawn up for the private sector and the overseas sector. (See Tables 5 and 6.) For reasons of space, the private sector balance sheet consolidates the household sector, the corporate sector, and the private financial sector. For practical applications, further sectoral disaggregation is often required. The balance sheets need little further explanation. Consumer durables and private residential housing can be viewed as included in Kp, and their imputed service flows are subsumed under private income and consumption in the budget constraint.

Table 5.

Private Sector Balance Sheet

(At current prices)

article image
Table 6.

Overseas Sector Balance Sheet

(At current prices)

article image

For simplicity, it is assumed that all claims on, or debts to, the rest of the world take the form of interest-bearing financial claims. Direct foreign ownership of domestic real capital or of domestic resources is not considered but could be added without difficulty. Human wealth (L), the present discounted value of expected future labor income, is a nonmarketable asset in the household balance sheet. The total national stock of land and mineral rights is assumed to be given by R¯.16

The conventionally measured financial surplus of the private sector (at constant prices) and the change in real private net worth are given in equations (22) and (23), respectively:

p+rpPKpPKp+rRPRpRp+(r+p˙p)(BHp+FHp)+ei*p(B*H+F*H)+rB˜H+npτpCδKp(B˙H+F˙Hp)+B˜˙H+ep(B˙H+F˙H)+H˙p+PKpPK˙pPRPR˙G(22)
ddt(Wpp)(B˙H+F˙Hp)+B˜˙H+ep(B˙*H+F˙*H)+H˙p+PKppK˙pPRPR˙G+1P(L˙+N˙+T˙)+(p˙Kp˙pKppp)Kp+(P˙RpRp˙p)(R¯RG)p˙p(BH+FH+Hp)+(e˙ep˙p)ep(B*H+F*H)p˙p(L+N+T)(23)

The conventionally measured financial surplus of the overseas sector (at constant prices) and the change in the real net worth of the overseas sector are given in equations (24) and (25), respectively:

X+epi*(B*FF*HE*)+(r+P˙P)(BFFFp)+rB˜Fep(B˙*FF˙*FE˙*)+(B˙FF˙Fp)+B˜˙F(24)
ddt(WFp)ep(B˙*FF˙*FE˙*)+(B˙FF˙Fp)+B˜˙Fp˙p(BFFFp)+(e˙pp˙p)ep(B*FF*FE*)(25)

These equations require little explanation. For the private sector, the difference between the financial surplus (at constant prices) and the change in real net worth reflects capital gains and losses on existing marketable assets and liabilities (including capital gains and losses that are due to inflation and exchange rate changes) and changes in the value of the intangible and nonmarketable items L, N, and T. On the left-hand side of equation (22), because only cash transactions are included, the implicit liquidity and convenience yield on money balances (ρMHp=iHp) as an item private consumption and of private income is omitted.

Positive Irrelevance and Normative Relevance of Debt Neutrality

The simplest theory of the interaction of the private and public sectors is based on the so-called pre-Ricardian debt-neutrality hypothesis. (See Barro (1974), Carmichael (1979), Buiter and Tobin (1979), Buiter (1980), and Tobin and Buiter (1980).) This hypothesis holds that, given the level and composition of the public sector’s real spending on goods and services, private sector behavior is invariant with respect to changes in the taxation/borrowing mix that finances this spending. Most of the formal models dealing with this issue concern closed barter economies, and the formal invariance propositions tend to be stated in terms of borrowing versus taxing without explicit consideration of monetary financing. For all three financing modes, the informal literature on the subject does, however, assert the irrelevance for real outcomes of the way in which governments finance their spending. The argument underlying this Modigliani-Miller theorem for the public sector vis-à-vis the private sector runs as follows. Spending must be financed (in a closed economy) by taxation, by borrowing, or by printing money. Borrowing is merely deferred taxation. A switch between taxation and borrowing should, therefore, not affect the permanent income and consumption behavior of rational well-informed private agents. Monetary financing implies the imposition of an inflation tax, which under restrictive conditions has the same effect on permanent income as explicit taxes.17

With debt neutrality, private sector spending behavior (for a given program of public spending on goods and services) is constrained only by the consolidated national balance sheet, as shown in Table 7. The distribution of the ownership of the nation’s resources between the public and private sectors is irrelevant. The national flow of funds account, including nonmarketable imputed income and consumption streams, is given in equation (26)

Table 7.

Consolidated Public and Private Sector Balance Sheet

article image
1p{+rsocpKsocKsoc+rGpGKG+rpPKpKp+rRpRR¯+i*e(*+F*HB*F)+i(FHBF)rpB˜F+iH}{Gc+Gsoc+C+δp(pKsocKsoc+pGKG+pKpKp)+ρMHp}1p{pKsocK˙soc+pGK˙G+pKpK˙p+e(E˙*+F˙*HB˙*F)+F˙HB˙HpB˜˙F}SP(26)

The first bracketed term on the left-hand side of equation (26) contains current income, including the imputed return from the government’s cash monopoly iHp. This item is matched in the second bracketed term, containing current consumption, by ρMHp, the imputed value of the nonpecuniary services of money consumed by the private sector. It is possible to omit both items if desired. The change in real national comprehensive net worth is given by

ddt(Wp)ddt(Wp+WGp)SP+(p˙KsocpKsocp˙p)pKsocKsocp+(p˙GpGp˙p)pGpKG+(p˙kppkpp˙p)pkppKp+(p˙RpRp˙p)pRpR¯+(e˙ep˙p)ep(E*+F*HB*F)p˙p{FHBF}+(L˙/p)+A˙M(27)

The change in real net worth equals saving (Sp) capital gains on marketable assets plus changes in the imputed or implicit value of nonmarketable items of wealth. A program of total national consumption in line with permanent national income means a choice of the value of the second bracketed terms in equation (26) so that the expected value of ddt(Wp)=0. Such a consumption program is ex ante indefinitely sustainable and serves as a useful benchmark for consumption planning in this debt-neutral economy.

Debt neutrality is bad positive economics. It requires private agents to be infinite lived or to have operative intergenerational bequest and child-to-parent gift motives in every generation. Perfect capital markets are another necessary condition: future labor income is a source of current spending power on a par with current disposable income and current holdings of government debt.18

The economic behavior that would be generated under debt neutrality is, however, a useful guide to what the aims of policy should be in a world in which a variety of capital market imperfections prevent the “unaided” private sector from acting according to permanent income principles.

It is, for example, well known that, in the absence of operative private intergenerational transfer motives, changes in the borrowing/taxation mix can redistribute the burden of financing a given government spending program between generations, even without the existence of capital market imperfections. If government is motivated by a concern for the utility (lifetime consumption patterns) of future generations as well as of the current generation, it can use the budgetary and financial mechanism to induce the current generation to act as if it were constrained by permanent private sector income rather than merely by the present value of its own lifetime resources.

The endowments listed on the asset side of Table 7, the nation’s technology (broadly defined) and the international trading and lending or borrowing conditions that it faces, represent the unavoidable constraints on the nation’s intertemporal transformation of resources.19 The purpose of financing policy—that is, the choice of the tax, transfer, borrowing, and money creation mix for a given real public spending program on goods and services—should be to keep additional constraints, such as cash flow shortfalls, inadequate liquidity, insufficient collateral, nonmarketability of assets, and credit rationing, from becoming binding or, failing that, to minimize their incidence and consequences.20

Through their budgetary and financing policies, governments (within a national economy) and international organizations (within the international economic system) can act as superior financial intermediaries in changing the composition of private sector portfolios and nation-state portfolios, respectively. Well-designed policy interventions of this kind can minimize the extent to which disposable income, current cash flow, and the portfolio of liquid marketable financial assets become binding constraints on consumption, investment, production, and portfolio allocation, enforcing undesirable departures from behavior according to permanent income principles. Governments, through their unique ability to impose taxes, through their monopoly of legal tender, and through the superior quality of their debts, have a “comparative advantage” over the private sector in borrowing to smooth out income streams.21 The same holds, although perhaps to a lesser extent, for certain international organizations vis-à-vis nation-states.

In what follows, a few examples are given to illustrate this role of the government as the natural borrower and the unique ability of the government to restructure the conventionally measured sectoral balance sheets, flow of funds accounts, and income expenditure accounts so as to permit the economy as a whole to approximate more closely behavior that is constrained only by comprehensive wealth or permanent income.

Fiscal aspects of discovery of a natural resource

Consider the effects of an oil discovery on public sector and private sector balance sheets. This discovery can be represented by an unexpected increase in PR, the value of property rights in land and mineral assets by, say, dpR > 0. To the extent that these property rights are privately owned and marketable, disposable private net worth increases by (RRG)dpR. Following permanent income principles, private agents would consume the perpetuity equivalent of this capital gain in each period. If spending were constrained by a dearth of marketable financial wealth to begin with, a temporarily larger increase in private consumption spending would result. The value of public sector assets increases by RGdpR. The government could choose to increase its own consumption spending in line with the permanent income equivalent of this capital gain. If it chooses not to do so, it faces the problem of enabling the private sector to raise its spending by the perpetuity equivalent of RGdpR.

One way to approach this situation would be to distribute to the private sector (in the form of tax cuts or increased transfer payments) the stream of actual additional oil revenues rR(t) RG(t) dpR (t) as and when they accrue. The present value of such anticipated future tax cuts or increases in transfer payments is, however, a nonmarketable highly illiquid asset that is singularly poor collateral for private borrowing. If there is a gestation period before the new oil comes on stream and a fortiori if development costs have to be incurred before the oil starts to flow, the additional cash flow to the government, and thus to the private sector, may well be negative for several years.

Private agents whose current spending is constrained by current disposable income or other forms of illiquidity will therefore be unable to raise their spending in line with their permanent income. A superior fiscal option is for the government, as soon as the new oil wealth is discovered, to cut taxes or raise transfers by an amount equal to the perpetuity equivalent or annuity value of the discovery. (See Flemming (1982).) This option will require additional government borrowing until the moment when actual revenues exceed their permanent value, at which time the authorities will be able to retire the temporary debt issues, whose function is merely to relax the spending limits on cash-flow-constrained households. With this transformation of future tax cuts into present tax cuts, the nation can consume in line with its new and higher permanent income; the government has transformed future tax cuts into disposable income.

An alternative proposal to handle the same problem has been made by Samuel Brittan of the Financial Times. His proposal amounts to a capital gift to the private sector by the public sector—the equity in the newly discovered oil riches is transferred to the private sector. If this newly privatized wealth takes the form of marketable financial claims, private spending in line with permanent income is again likely to be encouraged, relative to a policy of cutting taxes in line with current oil revenues—the government has transformed future tax cuts into disposable financial wealth.

In this paper, the same symbol T is used both for the present value to households of expected future tax payments and for the present value to the government of expected future tax receipts. Similarly, N represents both the household asset and the government liability corresponding to the stream of future benefits (n).

The presence of an impact on private spending of offsetting changes in, say, T, N, and BH that would prima facie appear to leave household net worth unchanged is then attributed, in a rather ad hoc manner, to differences in the liquidity, marketability, and usefulness as collateral of T, N, and BH. An alternative (but still ad hoc) way of avoiding the debt-neutrality conundrum is to assume that households discount future taxes and benefits at a higher rate than the market rate of return on bonds (and at a higher rate than the government discounts its tax revenues and benefit payments). To avoid a longer list of symbols and notation, this approach was not adopted here. A truly satisfactory treatment of these issues requires the tools of the new microeconomics of credit rationing, collateral, and other capital market imperfections, whose beginnings can be found, for example, in the work of Jaffee and Russell (1976), Benjamin (1978), Stiglitz and Weiss (1981), and Webb (1981; 1982).

“Cyclical” corrections to public sector deficits

Consider an economy in which the level of economic activity—as measured, for example, by output and employment—cycles around a trend. It is not assumed at this stage that these cycles represent Keynesian departures from full employment and normal capacity utilization. They could be regular swings in the natural rate of unemployment.

If the economy represented by equation (10) is simplified even further by ignoring public sector capital and index-linked bonds, the government budget constraint becomes

H˙+B˙HpGcτ˜+(r+p˙p)BHP(28)

Y, the trend level of output, grows at a proportional rate, γ. Actual output (Y) cycles steadily around this trend. If the demand for debt is a demand for real debt per capita and if population (in efficiency units) and Y grow at the same rate, then government financing tends to exercise upward pressure on the real interest rate when ddt(BHpY)>0 at the given real interest rate and the given real per capita stock of money balances. From equation (28), it can be seen that

ddt(BHpY¯)=Gcτ¯Y¯+(rγ)BHpY¯H˙pY¯(29)

It is a stylized empirical fact that, while exhaustive public spending (Gc) tends to grow in line with trend output, taxes net of transfers (τ˜) tend to vary positively with the current level of economic activity. These two relationships can be summarized by

Gc=gcY¯1>gc>0(30)
τ¯=θY1>θ>0(30)

Substituting equations (30’) and (30”) in equation (29) yields

ddt(BHpY¯)=gcθYY¯+(rδ)BHpY¯H˙pY¯(31)

Similarly, the proportional rate of growth of the money stock, assuming that the authorities keep constant the stock of real bonds per capita or per unit of trend output, is given by

H˙H=V{gcY¯Yθ+(rδ)BHpY}(32)

Thus, the current change in BHpY¯ overstates (understates) its trend or long-run average rate of change, and the current rate of growth of the nominal money stock overstates (understates) its trend or long-run average rate of growth whenever output is below (above) its trend value.

Even if it is only the current values of ddt(BHpY¯) and H˙H that matter for current crowding out and current inflation, respectively, the trend or long-run behavior of ddt(BHpY¯) and H˙H obtained by evaluating equations (31) and (32) with output at its trend value Y, will still be of interest to all but the most short-sighted governments.

Furthermore, if current crowding out is a function of anticipated future changes in BHpY¯ and current inflation depends on anticipated future monetary growth (as well as possibly on past monetary growth), current ddt(BpY¯) and H˙H will be a poor proxy for future developments if there are transitory swings in the deficit. From this perspective, cyclical corrections are a simple, if ad hoc, way of approximating the long-run implications of the fiscal stance for crowding out and monetary growth—that is, a quick method for calculating the permanent deficit.

Evaluation of Y at Y in equations (31) and (32) yields a reasonable approximation to the long-run averages only if the positive and negative deviations of Y from Y cancel each other out in the long run, as would be true, for example, if output followed a regular sinusoidal motion about trend, such as Y(t)Y¯(t)=1+Acos(ωt+ε). If positive and negative deviations of Y from Y do not balance on average, the simple cyclical correction gives a biased estimate of the long-run crowding-out pressure and monetary growth implications of the deficit. Such estimates will have to be replaced by an explicit averaging of equations (31) and (32) over long periods.

There are good reasons for letting taxes net of transfers vary with the current level of economic activity rather than making them functions of long-run or permanent income. It is assumed, as seems reasonable, that during the downswing a significant number of private agents are constrained in their spending by current disposable income.22 By reducing taxes and increasing borrowing during the downswing, public spending during that period will be financed to a larger extent by private agents who are not constrained by current disposable income (the purchasers of the bonds). Total consumption, therefore, declines by less than it would decline if taxes (which are assumed to fall equally on disposable-income-constrained and permanent-income-constrained private agents) had been kept constant. In the upswing, the additional debt incurred during the downswing can be repaid out of higher than normal taxes.23 The net result is that consumption is smoothed out over the cycle—a desirable result on grounds of intertemporal allocative efficiency even if product and factor markets cleared continuously. If wage or price stickiness exists, Keynesian problems of the failure of effective demand can also occur. Exogenous shocks to demand can set in motion contractionary or expansionary multiplier processes if private agents are constrained in their spending by current disposable income. The usefulness of automatic stabilizers and of countercyclical budget deficits derives from private spending that is constrained by current disposable income and from other capital market imperfections. It is reinforced by output and labor market disequilibrium.

Constraints of current disposable income on private consumption need not be absolute. Regular anticipated cycles in real income do not, of course, imply corresponding cycles in consumption, even for individuals who can borrow only on very unfavorable terms in order to consume in excess of their current disposable income. They have the option of accumulating a stock of liquid savings that can be run down and built up again cyclically. Even with uncertain stochastic swings in the level of economic activity, a buffer stock of liquid financial assets may permit a measure of income smoothing. Such private saving strategies are, however, likely to be inferior substitutes for access to borrowing on the terms available to the government.

A further option available to the government is to choose partial money financing of increases in cyclical deficits rather than borrowing. This option is more attractive the smaller the number and the less the wealth of private agents that are not constrained by current disposable income and liquidity. The more inelastic the demand for government bonds, the larger is the increase in interest rates that is required to unload additional bond issues on the private sector. (Access to international capital markets may make the total demand for domestic government bonds considerably more interest elastic than private domestic demand alone.) Such countercyclical money issues and withdrawals need not imply any increase in the trend rate of growth of the money stock.

It is to be noted that this view of stabilization policy suggests that taxes and transfers, rather than “exhaustive” public spending on goods and services, should be used to dampen fluctuations in economic activity. Public consumption spending, like all consumption spending, should be smoothed over time in line with permanent income. Public sector capital formation should have its time profile determined largely by the optimal public sector consumption program. Public works and other public spending on goods and services can be effective in regulating the overall level of demand and of economic activity, but they are likely to distort the optimal private sector/public sector consumption mix, unlike well-designed changes in the taxation, borrowing, and money financing mix.

Sales of public sector assets and cosmetic changes in PSBR

Sales of existing public sector financial assets do not appear in the SNA public sector financial surplus, but they do appear in the public sector borrowing requirement (PSBR) and similar transactions records. A “stock-shift” sale of government-owned rights to natural resources (-dRG) or of claims to public enterprise capital (-dKG) to the private sector would not in itself alter public sector or private sector net worth. If it is assumed that the government wishes neither to reduce the level of the money stock nor to acquire private sector capital, the counterpart of a reduction in RG or in KG would be a reduction in BH,B*H, or B˜H with PRdRG+pkGdKG=dBH+edB*H+pdB˜H.

Nationalizing or denationalizing may, of course, be desirable for reasons of efficiency. Total national net worth is altered by such transfers of ownership if the efficiency with which the resources are managed differs between sectors. The financial consequences, however, are virtually nil, because bonds in private portfolios are replaced by other financial claims. If the government gradually sells its assets to finance a flow of spending (PRddtRG+PGddtKG<0),, the difference between this policy and one of conventional financing by borrowing is also largely cosmetic.24 When it borrows, the government incurs an obligation to service the additional debt. When it sells assets, it loses the future income from the assets that it sells. It makes little sense, therefore, to attribute economic significance to the distinction between sales of public debt (below the line) and sales of government financial assets (above the line) as is done with the PSBR in the United Kingdom.

VI. Conclusion

The general conclusions have been stated in the introductory paragraphs. This section contains some more specific and practical remarks.

Comprehensive wealth and permanent income accounting requires explicit judgments concerning expectations about the future because of the need to evaluate nonmarketable, often intangible, and merely implicit assets and liabilities, such as future tax and benefit streams. This requirement is considered to be a salutary aspect of comprehensive wealth accounting. It brings out the distinction between mechanistic bookkeeping and recording of transactions, on the one hand, and accounting for economic policy evaluation and design, on the other hand.

Inflation accounting in the public sector is long overdue. Money illusion in the public sector should cease to be an obstacle to sensible budgetary policy. Taken alone, the public sector financial deficit and the PSBR (at current or constant prices or as a proportion of gross national product) are not very informative statistics. They must be corrected for the change in the real value of the outstanding stocks of interest-bearing public debt to evaluate either the implications of the deficit for financial crowding out or the “eventual monetization” implied by the government’s fiscal stance. Analogous corrections should be made to the conventionally measured external current account deficit or surplus; it is necessary to allow for changes in the real value of external assets and liabilities owing to changes in the price level and the nominal exchange rate.

Omission of government-owned capital and public sector property rights in land and natural resources from the public sector balance sheet can give a misleading picture of the net worth of the public sector and of its present and future fiscal and financial options. This holds true especially for countries where the government owns significant mineral rights (such as Norway, the United Kingdom, the United States, and many of the oil producing nations) and countries in which the nationalized sector accounts for a large share of economic activity (such as the United Kingdom and many developing countries). The sign of the effect on public sector net worth of including publicly owned capital is not self-evident; virtually open-ended commitments to subsidize loss-making public enterprises depress net worth.

The implicit assets and liabilities of the public sector represented by the streams of future tax revenues and of future benefits and transfer payments may well dwarf the marketable financial assets and liabilities in the government balance sheet.

Transitory (e.g., cyclical) deficits and surpluses are mechanisms that enable private agents who are constrained by current disposable income to smooth out consumption and to keep it more closely in line with permanent income. By permitting consumption to be maintained in the face of a transitory decline in income, they also mitigate unemployment and excess capacity if price and wage rigidities prevent an instantaneous market-clearing response to demand shocks. It is sound fiscal management for governments to borrow in the downswing “on behalf of private agents with less favored access to capital markets and to retire these countercyclical debt increases during the upswing, regardless of the rate of inflation. Alternatively, cyclical increases in the deficit could be partly or wholly financed by money creation, to be reversed during the upswing. The optimal financing mix of cyclical (or transitory and reversible) deficits need not be the same as that of permanent deficits. A consideration of this important issue would require the analysis of specific detailed models, which are well beyond the scope of this paper. The focus here is on general propositions that rely on as few detailed model-specific properties as possible.

APPENDIX List of Symbols

article image
article image
article image

REFERENCES

  • Barro, Robert J., “Are Government Bonds Net Wealth?” Journal of Political Economy, Vol. 82 (November-December 1974), pp. 1095117.

    • Search Google Scholar
    • Export Citation
  • Benjamin, Daniel K., “The Use of Collateral to Enforce Debt Contracts,” Economic Inquiry, Vol. 16 (July 1978), pp. 33359.

  • Boskin, Michael J., “Federal Government Deficits: Some Myths and Realities,” American Economic Review, Papers and Proceedings, Vol. 72 (May 1982), pp. 296303.

    • Search Google Scholar
    • Export Citation
  • Buiter, Willem H. (1980), “Crowding Out of Private Capital Formation by Government Borrowing in the Presence of Intergenerational Gifts and Bequests,” Greek Economic Review, Vol. 2 (August 1980), pp. 11142.

    • Search Google Scholar
    • Export Citation
  • Buiter, Willem H. (1982 a), “Comment on T.J. Sargent and N. Wallace: Some Unpleasant Monetarist Arithmetic,” National Bureau of Economic Research, Working Paper No. 867 (March 1982).

    • Search Google Scholar
    • Export Citation
  • Buiter, Willem H. (1982 b), “Money, Deficits, Crowding Out, and Inflation: The Economic Foundations of the MTFS,” paper presented at the Conference on Monetary Policy, Financial Markets, and the Real Economy, International Center for Monetary and Banking Studies, Geneva (June 1982).

    • Search Google Scholar
    • Export Citation
  • Buiter, Willem H. (1982 c), “Deficits, Crowding Out, and Inflation” (unpublished, September 1982).

  • Buiter, Willem H., and Marcus Miller, “The Thatcher Experiment: The First Two Years,” Brookings Papers on Economic Activity: 2 (1981), pp. 31567.

    • Search Google Scholar
    • Export Citation
  • Buiter, Willem H., and James Tobin, “Debt Neutrality: A Brief Review of Doctrine and Evidence,” Ch. 2 in Social Security versus Private Saving, ed. by George M. von Furstenberg, Series on Capital Investment and Saving, Vol. 1 (Cambridge, Massachusetts, 1979), pp. 3963.

    • Search Google Scholar
    • Export Citation
  • Carmichael, Jeffrey, “The Role of Government Financial Policy in Economic Growth” (unpublished doctoral dissertation, Princeton University, 1979).

    • Search Google Scholar
    • Export Citation
  • Flemming, John S., “U.K. Macro-Policy Response to Oil Price Shocks of 1974–75 and 1979–80,” European Economic Review, Vol. 18 (May/June 1982), pp. 22334.

    • Search Google Scholar
    • Export Citation
  • Jaffee, Dwight M., and Thomas Russell, “Imperfect Information, Uncertainty, and Credit Rationing,” Quarterly Journal of Economics, Vol. 90 (November 1976), pp. 65166.

    • Search Google Scholar
    • Export Citation
  • Miller, Marcus, “Inflation-Adjusting the Public Sector Financial Deficit: Measurement and Implications for Policy,” in The 1982 Budget, ed. by T. Kay (London, 1982).

    • Search Google Scholar
    • Export Citation
  • Ott, Attiat F., and Jang H. Yoo, “The Measurement of Government Saving,” Ch. 5 in The Government and Capital Formation, ed. by George M. von Furstenberg, Series on Capital Investment and Saving, Vol. 2 (Cambridge, Massachusetts, 1980), pp. 177241.

    • Search Google Scholar
    • Export Citation
  • Sachs, Jeffrey D., “The Current Account and Macroeconomic Adjustment in the 1970s,” Brookings Papers on Economic Activity: 1 (1981), pp. 20168.

    • Search Google Scholar
    • Export Citation
  • Sargent, T.J., and Niel Wallace, “Some Unpleasant Monetarist Arithmetic” (unpublished, October 1981).

  • Siegel, Jeremy J., “Inflation-Induced Distortions in Government and Private Saving Statistics,” Review of Economics and Statistics, Vol. 61 (February 1979), pp. 8390.

    • Search Google Scholar
    • Export Citation
  • Stiglitz, Joseph E., and Andrew Weiss, “Credit Rationing in Markets with Imperfect Information,” American Economic Review, Vol. 71 (June 1981), pp. 393410.

    • Search Google Scholar
    • Export Citation
  • Taylor, C.T., and A.R. Threadgold, “‘Real’ National Saving and Its Sectoral Composition,” Bank of England Discussion Paper No. 6 (October 1979).

    • Search Google Scholar
    • Export Citation
  • Tobin, James, and Willem H. Buiter, “Fiscal and Monetary Policies, Capital Formation, and Economic Activity,” Ch. 3 in The Government and Capital Formation, ed. by George M. von Furstenberg, Series on Capital Investment and Saving, Vol. 2 (Cambridge, Massachusetts, 1980), pp. 73151.

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

    • Search Google Scholar
    • Export Citation
  • Webb, David C. (1981), “The Net Wealth Effect of Government Bonds When Credit Markets Are Imperfect,” Economic Journal, Vol. 91 (June 1981), pp. 40514.

    • Search Google Scholar
    • Export Citation
  • Webb, David C. (1982), “Default Risk in a Model of Corporate and Government Finance,” Journal of Public Economics, Vol. 17 (April 1982), pp. 289306.

    • Search Google Scholar
    • Export Citation
*

Mr. Buiter, a visiting scholar in the Fiscal Affairs Department when this paper was prepared, is Cassel Professor of Economics at the London School of Economics and Political Science and Research Associate of the National Bureau of Economic Research. The author is indebted to John Makin and Marcus Miller, in addition to Fund staff members, for their helpful comments. All views expressed are strictly the author’s and do not necessarily represent the views of the Fund.

1

Or, equivalently, by

VM(t)=tαH(t)p^(u,t)ρ^M(u,t)etur^(s,t)dsdu

For any variable x, x^ (s, t) is the value of x expected at time t to prevail at time s.

2

Or, equivalently,

ΠM(t)=tαH˙^(u,t)p^(u,t)etur^(s,t)dsdu

3

It is assumed that for any variable x, x^ (t1, t2) =x (t1) for t1t2: the past and present are assumed to be known.

4

Consumption of the imputed services from social overhead capital can be viewed as a transfer (in kind) from the public sector to the private sector rather than as an item of public sector consumption. Alternatively, the services from the stock of public sector overhead capital could be an input into private production.

5

No behavioral significance should be attached to the specification of T and N in nominal terms.

6

Clear statements of this proposition can be found in Siegel (1979) and in Taylor and Threadgold (1979). See also Buiter and Miller (1981) and Buiter (1982 b).

7

It is assumed that borrowing and money creation, per se, do not affect determinants of the demand for public debt other than expected real rates of return.

8

This is the ex post measure. The ex ante real yields are defined in terms of the expected rate of inflation.

9

Money demand is assumed to be unit elastic in income and wealth.

10

The accounting framework does not indicate whether or not the real interest rate varies with the inflation rate.

11

For certain purposes, crowding-out pressure per unit of capacity output or crowding-out pressure per unit of efficient labor is of interest. (See, e.g., Sargent and Wallace (1981).) This would involve replacing equation (11) by the following:

ddt(B˜+BHp1Y)=Gc+GI+δKGτ˜Y+(rγ)(BHpY+B¯HY)rGKGYH˙pY

12

Simplifying assumptions about the public sector accounts made earlier in this section of the paper are maintained.

13

The money stock throughout this paper is the high-powered money stock. Addition of a private banking sector will, in general, be required for practical applications but does not significantly alter the conceptual framework outlined here.

14

That is, deficits corrected for the reduction that is due to inflation in the real value of the stock of nominal government bonds.

15

The appropriate discount rate may include a risk premium.

16

If Table 5 represents the balance sheet of those private agents currently alive, the horizons involved in N, T, and L would be finite if operative intergenerational bequest motives are absent. The symbols N and T in the private balance sheet would, therefore, be smaller than the corresponding items in the public sector balance sheet, even if public sector and private sector discount rates were identical. If there are operative intergenerational bequest motives, or if the private sector is viewed abstractly as containing both current and future generations, an infinite horizon for T, N, and L in Table 5 is appropriate. Even with common horizons, different discount rates for the public and private sectors could lead to changes in private net worth resulting from changes in the public sector balance sheet that leave public sector net worth unchanged. These issues are discussed later in this section.

17

The Modigliani-Miller theorem for money financing has been established formally for models in which money serves only as a store of value. Such “money” has only the name in common with what economists have always meant by money—that is, a means of payment or a medium of exchange. (See Wallace (1981).)

18

Debt neutrality—that is, invariance of the solution trajectories of real economic variables under changes in the borrowing/taxation mix of the government—also requires lump-sum taxes. With nonlump-sum (or distortionary) taxes, transfers, and subsidies, public sector claims on the private sector and private sector claims on the public sector still are netted out in the balance sheet. Real behavior is altered when the borrowing/taxation mix changes because the familiar allocative effects of nonlump-sum taxes, etc., will alter equilibrium prices and rates of return.

19

While it might, for example, be possible for an individual to consume today by borrowing against the present value of future labor income, a closed economic system cannot effect intertemporal shifts of future labor endowments. In an ideal market economy, these and other technological constraints are reflected in the sequence of demands and supplies over time and thus in equilibrium prices (including the asset prices that enter the balance sheets) at each point in time. In an ideal planned economy, the programming of material balances and the use of shadow prices would ensure the same outcomes.

Consider, for example, a simple two-period economy. The trade-off along the intertemporal consumption possibility frontier of c1 (consumption in period 1) for c2 (consumption in period 2) is given by AA in Figure 1. It is defined by the initial endowment of capital (K1), the labor endowments in periods 1 and 2 (L1 and L2), the well-behaved production functions in the two periods [f1 (K1, L1) and f2 (K2, L2)] and the constraint [0 ≤ c1f1 + K1, oc2f2 + K2, c1 = f1(K1, L1) + K1K2]. (See Figure 1.)

The permanent income in this closed system is given by 0D = 0C, determined by the intersection of the consumption possibility frontier with the 45° line OF. The opportunity for international lending and borrowing at a rate r would raise the permanent income of this system unless the slope of the international capital market constraint BB (given by –(1 + r)) equals the slope of the closed economy locus at E. For the figure, it is shown how a low external interest rate raises permanent income to 0D’. Figure 2 shows when a closed economy should not consume its permanent income in each period. Very favorable intertemporal transformation possibilities (Figure 2(A) and 2(B)) suggest consumption in excess of permanent income in period 2; the opposite applies in Figure 2(C) and 2(D). It is still the comprehensive balance sheet that matters for consumption, but constant net worth is unlikely to be optimal. Even with international lending and borrowing, the presence of nontraded goods, whose production can be augmented only slowly and at considerable cost, can make a program of consumption equal to permanent income infeasible or suboptimal.

20

The most desirable policy—eliminating capital market imperfections as far as possible—should be pursued to the full. Budgetary policies should aim to neutralize the imperfections that cannot be eliminated.

21

Because governments have the unique power to impose taxes (unrequited transfers to itself) and because of their ability to declare certain of their liabilities legal tender, the risk of default on government bonds is less than that on private debt. Total current and future natural income is in a sense the collateral for government borrowing. National income tends to be much less variable and uncertain than the income of individual private agents.

Governments effectively pool individual risks and thus eliminate diversifiable risk. An obvious question is why this risk-sharing cannot be done equally well through private insurance markets. One answer is that, even if this were possible, it would be more costly than making minor alterations to a tax structure that is required in any case. A second answer relies on familiar problems of moral hazard in insurance markets. It may be possible to devise efficient private insurance schemes for “bad luck” default. Private insurance markets will operate inefficiently (or may not exist at all) if there is frequent “voluntary” or “dishonest” default and if lenders and insurers cannot differentiate between honest and dishonest borrowers. If it is easier and less costly for the government to levy taxes on reluctant taxpayers than it is for private lenders and insurers to compel performance by dishonest borrowers, then governments have a role as financial intermediaries and government debt is not “neutral.” (See Webb (1981; 1982).)

22

It may be desirable to replace the phrase “spending constrained by current disposable income” by the following: the effect of current disposable income on spending exceeds that of permanent income multiplied by the share of current disposable income in permanent income (allowing for the effect of changes in current income on expectations about future income streams).

23

These higher taxes during the upswing fall on a population that, on average, is likely to be less constrained by current disposable income than it was during the downswing.

24

The earlier caveat about differences in the efficiency with which the assets are managed also applies here.

IMF Staff papers: Volume 30 No. 2
Author: International Monetary Fund. Research Dept.