Summary Measures of Fiscal Influence

“If the executive branch does not provide a suitable set of summary figures, others will produce figures which may or may not be sensible.” 2

Abstract

“If the executive branch does not provide a suitable set of summary figures, others will produce figures which may or may not be sensible.” 2

“If the executive branch does not provide a suitable set of summary figures, others will produce figures which may or may not be sensible.” 2

—Richard Goode

“. . . we believe that developing a single number to measure fiscal influence is a legitimate scholarly exercise.” 3

—Alan S. Blinder and Robert M. Solow

In evaluating the macroeconomic impact of a budget, policymakers and economists frequently use summary measures, such as the size of government expenditure or various notions of budget balance. These measures are employed to answer questions ranging from (1) the short-run/cyclical effects of budgets (typically about the nature of the thrust supplied—i.e., whether it is in the right direction and of a sufficient magnitude to achieve the policy objective) to (2) the medium-run/secular effects where the concern is usually whether the budget will support the full-employment equilibrium of the economy with balanced growth, stable prices, and stable international reserves.

Obviously, any summary measure that is selected to represent the influence of a budget will involve a number of abstractions about the structure of the economy and the ways in which the budget interacts with the economy. It is, therefore, possible to construct a wide variety of summary measures, depending on the underlying assumptions made regarding the dominant objectives of fiscal policy (whether the control of unemployment, inflation, or some other variable); the nature of the economy (whether industrial or developing); the theory chosen to explain the macroeconomy (whether Keynesian, monetarist, or some other); and diverse other reasons such as the availability and use of data or the mode of analysis.4

There are several alternative measures currently available that could give rise to disparate indications of fiscal policy when taken at face value. This is unfortunate, since the measures are all designed to assess budgets in a primarily industrial environment where the dominant objective is one of full employment. Moreover, the measures are based on a number of limiting assumptions and cannot be readily applied to situations where different assumptions hold.

In this paper, some of the issues involved in the selection and interpretation of measures are examined with a view to establishing guidelines for the use of alternative summary measures in the appraisal of fiscal policy. The paper begins (in Section I) with some preliminary definitions and a consideration of the budget balance rule. Three major national measures that have been designed to indicate the effect of the budget on the economy are reviewed in Section II, and an attempt is made to establish (1) the reasons for divergences and (2) the scope for rendering the measures equivalent. The three measures generally ignore financing, an issue that is taken up in Section III, where liquidity balance alternatives are also proposed for use in monetarist environments. Section IV considers the applicability of measures preponderantly concerned with the full-employment objective to inflationary situations. The argument is presented that, in some environments (stagflation and/or a balance of payments constraint), both unacceptable inflation trade-offs and biased assessments of the inflationary impact of budgets may result. An alternative measure that avoids these problems is outlined.

Section V examines the scope for applying summary measures to developing economies, while Section VI draws on the measures presented earlier and illustrates an approach to the evaluation of fiscal policy in an industrial and a developing economy, respectively. Concluding comments follow in Section VII. Detailed calculations for the two case studies are shown in the Appendix.

I. Preliminary Issues and the Budget Balance Rule

The complex of expenditure and financing transactions that make up a budget is conveniently grouped into three major categories: (1) government expenditure, G, taken here to cover purchases of goods and services, transfer payments, and net lending; (2) revenue, T, including both tax and nontax revenue; and (3) borrowing proceeds, B, which, in turn, are usefully subdivided into borrowing from the nonbank public, Bn; borrowing from the banking system, Bb; and borrowing from nonresidents, Bf.5 Such a grouping permits the construction of two major notions of budget balance that have been frequently employed, namely:

theoverallbudgetbalance:GTBBn+Bb+Bf(1)

and

theliquiditybalance:GTBnBb+Bf6(2)

Broadly speaking, Keynesians have tended to focus on the overall budget balance as a summary budget measure, while monetarists have emphasized the liquidity balance. In addition to their uses as budget summaries, the sizes, signs, and variations in the budget balances have frequently been employed for purposes of drawing inferences about budgetary effects and fiscal policy.7

Given the objectives of stabilizing fiscal policy—traditionally those of curbing unemployment in the industrial economies, although others may be considered—and a theory of the interaction of the budget with the economy that delineates the possible role of the instrument, the fiscal analyst seeks to establish the need for fiscal policy and the role of the budget (as an expression of fiscal policy) in serving that need. In order to establish the appropriateness of a budget, a number of issues need to be resolved. Suppose some stimulus is required in order to offset a downturn in a target variable—say, employment. What should the stance of the budget be (neutral, stabilizing, or expansionary), taking into account uses of nonfiscal instruments and any lagged effects of previous years’ budgets? Assuming that the proposed budget stance is appropriate, is it adequate in terms of meeting the objective? Would continuation of the proposed budget policy preserve full-employment equilibrium over the longer run? How does the budget compare with some other budget—for example, the previous year’s budget—in terms of fiscal impact? What would be the influence of the budget on some other target variable (e.g., inflation), and how acceptable are the tradeoffs? Since the policy objective may be met by varying one of the budget components, but other budget components could vary in directions different from those required for meeting the particular objective, measures are needed that summarize divergent movements in budget items and help to provide answers to some of the questions raised above.

Whatever the notion of budget balance, its direct use as a gauge of fiscal influence relies on suppositions that may not be valid. Thus, the inference, drawn from the budget balance rule, that an increase in the budget deficit is indicative of an expansionary policy, presumes that the change in the deficit is an exogenous (discretionary) consequence.8 Usually, however, certain components of the budget will be affected by the level of economic activity and, accordingly, the budget deficit will refect both discretionary and automatic effects.9 A mere widening in the budget deficit cannot, therefore, be interpreted as necessarily indicative of a more expansionary fiscal policy10: the larger deficit may be the result of a recession that reduces revenues or increases expenditures (e.g., for unemployment benefits).11

II. A Review of Three National Measures

Summary measures of fiscal influence that attempt to separate the effects of the budget on the economy from those of the economy on the budget have been constructed in a number of countries. Three of the more prominent of these measures are the U. S. full-employment balance (FEB),12 the Federal Republic of Germany’s cyclically neutral balance (CNB),13 and the Dutch structural budget margin (SBM) and budget impulse (BI) techniques.14 All three have full employment as their target and generally ignore budget financing. A brief description of these measures and an evaluation follow.

1. the u. s. full-employment balance

The U. S. measure, broadly stated, seeks to abstract from the influence of the economy on the budget by standardizing endogenous components at the full-employment level of income.15 In a simplified presentation, the measure is the difference between (1) revenue (the endogenous element) that would be generated by the assumed tax structure at full employment and (2) government expenditure G.16

FEB=T(YP)G17(3)

where T(YP) is revenue at full-employment income for the assumed tax structure.

The principal use of the FEB concept has been to demonstrate that, in a period of recession, the operation of fiscal stabilizers could result in a deficit balance. Consequently, efforts to balance the budget could render the budget more restrictive, which might aggravate the recession. However, reliance on fiscal stabilizers alone might not be sufficient to minimize the cyclical fluctuation, and a more active fiscal policy might be required. In this connection, the FEB concept may be employed for a point-in-time comparison of alternative budgets. Ranking the FEBs for different budgets permits the drawing of inferences regarding fiscal influence; the smaller the FEB, the more expansionary is the budget, and so forth.

For dynamic comparisons, some adjustment needs to be made for the fiscal drag generated by increases in revenue. If expenditure grew at the same rate as full-employment income, but revenue grew at a different rate, the FEB would vary as a proportion of full-employment income and the budget might become either more or less restrictive. An adjustment in taxes or expenditure might be necessary to ensure the constancy of the share of the FEB in full-employment output. If these adjustments were assumed, then the point-in-time ranking of FEBs would also extend over time.

The use of the FEB concept as a norm, for purposes of appraising the posture of the budget when the economy is moving along its full- employment path, has been controversial. At issue is the appropriate level of the FEB over the longer run (when cyclical factors have been eliminated), with some proponents arguing for a zero balance while others argue for a deficit or surplus, depending on whether the private sector will be in surplus or deficit.18 If the normative level of the FEB is not restricted to zero, the norm would correspond to those employed by the Federal Republic of Germany’s Council of Experts and the Dutch authorities.19 While it is not necessary to assign a normative value to the FEB when ranking budgets in terms of their expansionary effects, a norm is useful, both because of the guidance provided regarding the direction in which the budget may need to be changed in order to render it compatible with the medium-term growth trend and, as is shown below, because it will also facilitate countercyclical assessments of budgets.

A fundamental criticism of the FEB concept’s use in assessing the expansionary effects of budgets outside full employment is the implicit assumption that the ranking of budgets at the full-employment level is invariant with respect to the degree of capacity utilization. If, however, the tax structure underlying one budget is more progressive than that underlying another budget, the ranking at actual output levels could involve reversals from the full-employment ranking, the likelihood of which increases, the further the economy is from its full-employment level. Thus, what might appear to be a more expansionary budget at full employment could, in reality, be a more contractionary budget at actual income levels.

2. the federal republic of germany’s cyclical effect of the budget

The formulation of the German measure requires two steps.20 First, there is the concept of the cyclically neutral balance (CNB) defined as

CNB=g0YPt0Y(4)

where t0 is the base-year ratio of revenue to actual GNP and g0 is the base-year ratio of expenditure to potential GNP.

Second, there is the determination of the cyclical effect of the budget

CNB=(GT)CNB(5)

(CEP), where a positive (negative) CEB indicates an expansionary (contractionary) effect.

Starting from a base-year normative level of the budget balance, the CNB measure develops a neutral profile of the budget balance over time that allows for the effect of the economy on the budget. The selection of the base year is guided by the notion that there is some ratio of the overall budget balance to potential output that will support the balanced-growth equilibrium of the economy, in which major targets—such as employment, reserves, and the inflation rate—are at satisfactory levels. In their use of the measure, the Federal Republic of Germany’s Council of Experts determines the base year as that full-employment year in the recent past that was characterized both by reasonable stability and a budget that was judged to be neither excessively expansionary nor excessively contractionary.21 Ratios of government expenditure to potential output, and of revenue to actual output, are computed for the base year. Over time, as the degree of capacity utilization varies and actual GNP deviates from its potential level, the base-year normative balance will fluctuate in the manner prescribed by the CNB rules (see equation (4)) whereby neutral government expenditure grows from its base-year level at the same rate as potential output, while neutral revenue is defined to vary equiproportionately with actual GNP.

The actual budget balance in each year is tested against the relevant neutral balance for that year. Suppose it is in deficit, exceeding the neutral deficit. The budget is, consequently, judged to be expansionary. In order to establish whether this expansionary budget is countercyclical, the situation of the economy needs to be considered. If the economy were in a recession, the cyclically neutral deficit balance would be greater than its base-year level. In these circumstances, a countercyclical budget is indicated by a positive CEB (see equation (5)). However, a positive CEB in a boom year suggests a procyclical budget posture.22

It should be noted that use of the equiproportionate revenue rule in a context where the actual income elasticity of revenue differs from unity will not facilitate the correct separation of discretionary from automatic fiscal policy. However, drawing on a distinction noted earlier,23 the revenue norm permits a differentiation between stabilizing and offsetting fiscal responses. Thus, suppose the elasticity of revenue with respect to GNP is greater than unity, and the budget posture is that of the normative budget. Because of the larger revenue response, a recession would cause the actual balance to widen more than the normative balance. The CEB would be positive, indicating that the budget is exerting an offsetting (and, in the circumstances of a recession, a countercyclical) effect as a result of automatic factors. If the budget responded in exactly the manner prescribed by the CNB rule, the budget would have been merely stabilizing.24

The reliability of the CNB test as an indicator of fiscal impact depends on the structural continuity or continued validity of the base-year normative balance.25

3. the dutch budget impulse measure

Instead of embedding the test for the cyclical effects of a budget directly in a framework that uses the medium-term (secular) adequacy of the budget as its point of reference, the Dutch approach involves a separation of the two aspects. A distinction is drawn between structural budgetary policy that aims at promoting longer-run equilibrium, and cyclical policy, the objective of which is short-run stabilization (viewed as a temporary deviation from structural policy).26 The approach to assessing medium-term budget policy—and the associated structural budget margin technique—uses as its norm a budget deficit that is adapted to the medium-term development of savings and investment in the economy. This notion corresponds to the Federal Republic of Germany’s secular norm (the base-year balance).

In order to evaluate what is also referred to in the German method as the cyclical effect of a budget, the Dutch approach first determines the impulse of the budget using the following formula:

BudgetImpulse(BI)=GT(G1T1yT1)G1(6)

where the subscripts denote the previous year’s value of current variables, and y is the actual rate of growth of GNP. The next step establishes the cyclical effect of the budget as the difference between the impulse and the medium-term (or potential) rate of growth of GNP, y. Symbolically, the Dutch CEB is

DCEb=BIy¯(7)

with a positive DCEB indicating an expansionary effect and a negative DCEB indicating a contractionary effect.

The impulse calculation in equation (6) is a summary attempt at estimating the growth in the budget’s contribution to current aggregate demand. This is easily demonstrated by rearranging equation (6) in the following manner:

BI=(GG1)(T(1+y)T1)G1(8)

If revenue in the current period equals the amount that would have resulted had the previous year’s revenue grown at the same rate as current GNP, BI reduces to the rate of growth of government expenditure. If revenue grows faster than current GNP, the excess over the proportionate increase is deducted from government expenditure on the grounds that it offsets expenditure’s contribution to aggregate demand. The procedure thus amounts to defining an effective rate of growth in government expenditure. This effective growth rate, or BI, is then compared with the growth in potential output (see equation (7)). Equality of the two growth rates implies a neutral budget in the sense that, if the rate of growth in the component of aggregate demand constituted by the budget is the same as the growth rate of potential output, then the budget is not the source of any imbalances between increments in aggregate demand and supply. The essential feature of the approach is that it takes the previous year’s budget as given, and focuses on how deviations from that budget (as defined by the new budget) relate to the emerging supply situation.

The BI procedure, it must be emphasized, does not pass any judgment as to whether the underlying budgetary situation is excessively expansionary. Thus, it is possible for a budget to continue to be inappropriate, even though its changes are assessed to be neutral.27

4. relationships between measures

The comparison of the BI technique with the German CEB is facilitated by expressing equation (7) as follows:

(DCEB)G1=(GT)(G1T1yT1)G1y¯(9)

This simplifies to

(DCEB)G1=(GT)[(1+y¯)G1(1+y)T1](10)

which is also equivalent to

(DCEB)G1=(GT)(g1YPt1Y)28(11)

where g-1 and t-1 are the previous year’s ratios of government expenditure to potential output, and revenue to actual GNP, respectively.

Both equations (10) and (11) exhibit a formal similarity to the Federal Republic of Germany’s CEB (stated in equation (5)). The cyclical effect of the budget under the BI technique amounts to a test of the actual balance against a balance deemed to be neutral. While the growth rates employed in determining the neutral development of the budget are the same, the Dutch method applies these to the previous year’s budget levels (as the assumed neutral base), whereas the German method applies them to the previous year’s neutral budget levels (determined as a neutral continuation of the base-year normative budget). The Dutch and German estimates of the CEB will, therefore, be identical only for the year after the base year. Thereafter, they will differ to the extent that the previous year’s proportionality factors deviate from the base-year ratios.

As a consequence of the Dutch procedure of rolling forward the reference point each year, the appropriate comparison is between the year-to-year changes, or between the first differences of the German CEB stated in equation (5) and the absolute levels of the DCEB as stated in equation (9). Taking first differences for CEBs, calculated for each year with reference to a fixed point, approximates the estimate of the DCEB that is obtained when the base year is always the preceding year. The differences in results are of the second order. This may be seen by taking first differences of the German CEB (equation (5)) as follows:

ΔCEB=CEBCEB1=(GT)(g0YPt0Y)(G1T1)+(g0YP1t0Y1)(12)

which reduces to

ΔCEB=(GT)(G1ny¯T1ny)29(13)

Subtracting from equation (13), the DCEB estimate, in equation (11), yields

ΔCEBDCEBG1=(G1G1n)y¯(T1T1n)y30(14)

In other words, the comparison of changes in the Federal Republic of Germany’s CEB with the actual DCEB (multiplied by G-1) will show differences if the actual levels of expenditure and revenue in the preceding year (which count as neutral in the Dutch procedure) diverge from the neutral levels for the previous year, computed under the German definition. However, since these differences are multiplied by the growth rates of either potential or actual GNP, the magnitude of the differences will be greatly reduced, and the two alternative measures (in the forms specified) will tend to yield similar results.31 Identical results will obtain if the actual budget balance moves in accordance with the cyclically neutral profile.32

A relationship between the FEB and the German CEB may be established by first noting that the latter involves a test of the budget balance against its cyclically neutral level at actual capacity levels. Valuing the budget balance at the full-employment level of the economy generates the FEB, while valuing the CNB at full employment results in the base-year normative balance CNB*.33 If a normative interpretation is placed on the FEB and the appropriate norm, FEB*, is identified with CNB*, then the Federal Republic of Germany’s CEB may be interpreted as the actual income analogue of the difference between FEB and FEB* at full employment. On standardizing budget balances by dividing actual balances by the actual level of GNP, and dividing full-employment balances by the full-employment level of GNP, the following relationships can be postulated:

FEBYPFEB*YP=CEBY(15)

Since (by assumption) the normative FEB* = CNB*, equation (15) may be rearranged as

FEBYPCEBY=CNB*YP(16)

It is a straightforward matter to demonstrate that these relationships hold exactly whenever the income elasticity of response of endogenous budget components is unity.

From equation (16), first differences in the FEB term will equal first differences in the CEB term (since the first differences in the normative balance term are equal to zero, in view of the constancy of the ratio of the normative balance to full-employment GNP). The relationship stated earlier (equation (14)) between first differences of the German CEB and the absolute level of the Dutch CEB, therefore, also holds for first differences of the FEB. The latter, however, need not be given a normative interpretation. Indeed, it is possible to establish a direct relationship between first differences of FEBs and the DCEB without the intermediation of a normative FEB*.

These relationships between measures are illustrated in Table 1 for the United Kingdom. Using the 1972 budget balance as the norm, both the FEB, when given a normative interpretation, and the German CEB yield virtually identical results. First differences of the FEB, together with those of the German CEB, yield assessments of changes in fiscal stance that are similar to those provided by the Dutch CEB.

Table 1.

United Kingdom: Relationships Between Summary Measures1

(Per cent of gross domestic product at current market prices)

article image
Source: Table 2 in the Appendix.

Positive (negative) figures indicate an expansionary (contractionary) effect.

1972 is the base year selected for the purpose of defining the normative balance.

FEB denotes full-employment balance. FEB* denotes normative full-employment balance.

CEB denotes cyclical effect of the budget.

G denotes government expenditure.

III. Financing and Some Monetarist Measures of Fiscal Influence

An assessment of budgetary impact based on an overall balance approach may be modified by a liquidity balance approach, insofar as domestic nonbank borrowing is regarded as akin to revenue. Thus, the assessed contribution of an overall deficit balance to net aggregate demand is reduced to the extent that the deficit is financed through domestic nonbank borrowing since, according to monetarist theory, the latter displaces private expenditure.34

Three issues need to be resolved when deciding on a liquidity balance approach in the evaluation of budgets from the full-employment point of view. First, under what circumstances would domestic nonbank borrowing have effects similar to those of an increase in revenue? The answer will determine which of the two balance concepts would be employed. Second, if the liquidity balance approach is adopted, what adjustments might be necessary to separate the effects of the economy on the budget from those of the budget on the economy (just as with the overall balance)? Third, what is the relationship between the liquidity balance and the monetary operations of the authorities?

Which of the two alternative balance concepts should be employed will ultimately depend on the effects of domestic nonbank borrowing on private sector behavior, especially with regard to expenditure. Consider a sustained increase in government expenditure financed by higher domestic nonbank borrowing. The overall deficit balance would be assessed as more expansionary. The liquidity deficit balance, however, would remain constant; monetarists would conclude that the impact of the budget would not change, since the increase in budgetary demand for loanable funds would crowd out private expenditures, so that the net effect on aggregate demand would be unchanged. Obviously, if the environment were one of monetary ease, the monetarist effect would not occur, since additional loanable funds would be made available to finance private expenditures.

In order to discriminate properly among the alternative concepts, the more appropriate assumption is that of constant monetary conditions—for example, a fixed base money supply and restrictions on the drawing of external credit. In these circumstances, some crowding out of private expenditure may occur, although its magnitude will depend on the state of the economy and the situation of the private sector. Under boom conditions, with shortages in the supply of loanable funds, a higher degree of crowding out is likely to occur than in a recession, when the private sector is unwilling to spend.35

Another issue concerns the adjustments that may be necessary when evaluating fiscal policy through the use of a liquidity balance concept. The simple use of liquidity balances, without any adjustment for feedback between the target variable (GNP or employment) and the budget as an instrument, can only be justified if the liquidity balance is a discretionary variable. As is evident from the budget constraint set out in equation (2) above, this requires that one of three variables—government expenditure, revenue, or borrowing from the domestic nonbank public—be residually determined. Since government expenditure and revenue are specified in plans subject to legislative approval, the residually determined variable would have to be domestic nonbank borrowing. However, it is unrealistic, especially in less developed countries (LDCs), to assume that governments do, in fact, treat domestic nonbank borrowing as a residual variable, or to assume that any shortfall in revenue is automatically made up through higher domestic nonbank borrowing. Any exigencies affecting cash flows tend to be mediated via the accounts maintained with the banking system. Variations in liquidity balance are, therefore, likely to be the result of both discretionary and automatic forces. The parallels with the overall balance are obvious, and rules similar to those considered earlier in Section II may be developed.

The first requirement is to specify a secular norm for the budget as some ratio of the liquidity balance to GNP that is (or would be) supportive of balanced growth equilibrium.36 The budget need not be a source of liquidity (in which case the normative liquidity balance is zero), but it frequently is, especially in LDCs. Through the examination of some past year of stability, or analytically, a normative ratio of the liquidity balance, denoted here as FELB*, may be selected. The comparison of any actual liquidity balance, when evaluated at full employment, with FELB* indicates whether the underlying budget posture (which may be required for some other reason) will be secularly expansionary or contractionary if retained at full employment.37

The next step is to construct a profile of the normative balance when the economy is not at its full-employment level. An analogue of the German cyclically neutral balance is thereby generated, against which the actual liquidity balance is tested to draw inferences about the countercyclical role of the budget. The construction of the cyclically neutral liquidity balance (CNLB) depends on the treatment accorded to domestic nonbank borrowing.38 As before, the normative response in revenue is defined to be equiproportionate to GNP. Regarding domestic nonbank borrowing, two situations may arise, depending on how the government treats this form of financing. The amount of nonbank borrowing may be specified anew with each budget and raised irrespective of fluctuations in GNP. In this instance, the variable should be treated as a discretionary component of the budget.39 Alternatively, the amount of such borrowing depends on private savings behavior, the willingness of the private sector to acquire government securities, and/or on institutional arrangements such as borrowing from a provident fund.40 Such borrowing is influenced by variations in the level of GNP, and is more appropriately handled as an endogenous variable.

A specification of the cyclically neutral liquidity balance (CNLB) in which domestic nonbank borrowing is viewed as a discretionary variable would be as follows:

CNLB=g0YPb0YPt0Y(17)

The only component of the normative balance that would vary with GNP would be revenue.

In the case of endogenous domestic nonbank borrowing, a definition of the CNLB would be

CNLB=g0YPb0Yt0Y(18)

Both the domestic nonbank borrowing and revenue components of the normative balance now fluctuate with GNP. By assumption, domestic nonbank borrowing is assumed to fluctuate equiproportionately with GNP.41

The cyclical effect of the liquidity budget (CELB) is then determined as the difference between the actual budget and either one of the CNLBs defined above:

CELB=(GTBn)CNLB(19)

with a positive CELB indicating an expansionary effect of the budget. If the economy is in a recession, so that the cyclically neutral liquidity balance is widening, an actual liquidity balance that exceeds this neutral level would indicate a countercyclical policy.

Finally, for purposes of assessing whether the budget is becoming more or less expansionary, a measure similar to the Dutch BI may be constructed. Once again, the choice of an endogenous or discretionary treatment of domestic nonbank borrowing will affect the definitions. For the endogenous (equiproportionate) case, the liquidity balance version of the budget impulse (BIL) may be defined as

BIL=(GTB)(G1T1B1)+yT1+yB1G1(20)

or, equivalently, as

BIL=(GG1)[T(1+y)T1][B(1+y)B1]G1(21)

where, as before, y represents the actual rate of growth of GNP.

As in the earlier interpretation of the Dutch BI, the definition here may be viewed as a statement of the effective rate of growth of government expenditure. If revenue and domestic nonbank borrowing are both growing at the same rate as GNP, the effective rate of growth of government expenditure will equal its actual rate, GG1G1. A faster rate of growth in either domestic nonbank borrowing or in revenue will reduce the expansionary effect of the actual rate of growth in government expenditure, and the effective rate defined by BIL will be lower.

If domestic nonbank borrowing is treated as a discretionary (exogenous) magnitude, the expansionary thrust of government expenditure will be reduced by the full amount of such borrowing, irrespective of variations in GNP. Consequently, the appropriate interpretation to be placed on the BIL is that of the effective rate of growth in government expenditure that has been adjusted for such discretionary borrowing. The following definition results:

BIL=(GTB)(G1T1B1)+yT1G1B1(22)

or, equivalently,

BIL=(GB)(G1B1)[T(1+y)T1]G1B1(23)

If revenue grows at the same rate as GNP, the effective rate of growth in adjusted government expenditure reduces to its actual rate of growth.

For an evaluation of whether the budget, employing the liquidity balance concept, is becoming more or less expansionary, the alternative definitions of the effective rates of growth in government expenditure (equation (20) or equation (22)) may be tested against the rate of growth of potential output, y. Thus, the Dutch version of the cyclical effect of the liquidity budget would take the form

DCELB=BILy¯(24)

with a positive value indicating that the budget was becoming more expansionary.42

As long as a consistent treatment of domestic nonbank borrowing is adopted throughout, equivalences similar to those found earlier for the overall balance measures can also be demonstrated for this group of liquidity balance measures. Table 2 in the Appendix presents an application of liquidity balance measures to the United Kingdom and illustrates the equivalences when the measures are appropriately interpreted. A comparison of the assessments, employing the liquidity balance measures and those resulting from the overall balance measures in Table 1, shows a surprising degree of qualitative similarity, even though the absolute magnitudes differ. The matter is investigated further in Section VI.

The final issue to be examined is that of the relationship between the liquidity balance and the operations of the monetary sector. It is particularly important, in a monetary analysis, to establish whether the liquidity implications of the budget are accommodated or sterilized by the monetary sector. Thus, government borrowing from the domestic banking system could be accompanied by a reduction in credit to the private sector, while the proceeds of government borrowing abroad could likewise be sterilized through further reductions in domestic credit or through open market sales of bonds. While there are obvious limits to bank sterilization policies, the point needs to be made that evaluations of fiscal policy and of the ultimate effects of the budget may differ, depending on whether the monetary authorities accommodate the liquidity implications of the budget.

IV. Taking Account of Inflation

The measures considered thus far have as their central concern the objective of full-employment output. Moreover, the measures are formulated in real terms and assume away problems that may arise when the price level varies. Two topics are examined here: first, the conversion of measures from real into nominal values; second, the scope for employing the measures and the modifications that may be required for the assessment of fiscal policy when its objective is that of price stability. A simple measure is presented for assessing the inflationary impacts of budgets, on the assumption that aggregate supply is constrained.

1. nominal valuation of measures

The conversion from real into nominal values is usually rendered by applying a common deflator to the budget items, potential output, and GNP that are stated in real terms in the measures. Questions may be raised about the propriety of such a procedure when the composition of government expenditure on goods and services differs from that of GNP. This is essentially an index number problem and, if data on the alternative deflators are available, it may be appropriately handled. A more difficult issue concerns the discretionary/automatic distinction when budget magnitudes are affected by inflation. To what extent should the effects of inflation on the budget be viewed as automatic and, therefore, isolated in the evaluation of the effect of the budget on the economy?

The problem arises particularly with expenditure. Concerning revenue, the continued use of the equiproportionate rule in determining the neutral response of revenue would simply assign any excess revenue generated by inflation43 to the discretionary category. However, inflation raises issues regarding the appropriate separation of discretionary and automatic effects, both for nominal expenditure and for real expenditure, which has been treated as wholly discretionary up to this point.44 Among countries, a wide variety of practices are found regarding the components of real or nominal expenditures over which discretionary power is exercised. Even wages and salaries—which typically account for about two thirds of public expenditure, and which provide, perhaps, the best example of a component over which discretionary power may be exercised, both in real terms (numbers employed) and in nominal terms (the wage and salary structure)—are subject to qualification. The country may have adopted a scheme of wage indexing, resulting in some loss of control over the nominal level of wages and salaries. Alternatively, nominal expenditure may be subject to cash limits, with the result that real expenditure ceases to be a purely discretionary variable. Or, again, components of real expenditure may be subject to fixed-price contracts, so that both real and nominal outlays are discretionary.45

If a tendency is to be detected in these practices, it would seem to be that, in periods of low inflation, discretionary control over both real and nominal magnitudes is more feasible; such control becomes less feasible for nominal values when inflation is intense and prolonged. Government planning of expenditure is frequently done in real terms, and, in the presence of inflation, is grossed up (i.e., raised) by the expected rate of inflation. In these circumstances, the procedure of grossing up the expenditure component of the neutral yardstick (stated in real terms) by the inflation factor is to be expected. Given common deflators, the nominal valuation of yardsticks relating to the full-employment objective reduces to multiplication of the computed base-year ratios of expenditure (to potential output) and of revenue (to actual output) by the nominally valued potential and real outputs, respectively, in each year.

It should be emphasized that this procedure of nominal valuation, which takes the past rate of inflation as a bygone, and the current or prospective rate of inflation as an exogenously determined datum, is valid only where the assessment is being undertaken in terms of the full-employment objective.46

2. the control of inflation as an alternative objective

The rider—”with reasonable price stability”—is frequently attached to the full-employment objective. Attempts have been made, especially in the statement of the Federal Republic of Germany’s and the Dutch measures, to take account of this additional objective.47 Essentially, these attempts involve amendments to the grossing-up factor that were considered earlier in the discussion of how to determine the nominal value of neutral expenditure. Grossing up by the full extent of inflation is viewed as implying a permissive budget that tends to validate the current rate of inflation.48 The amendments discussed above instead use some notion of normal, cyclically neutral, or acceptable rate of price increase.49 Accordingly, the nominal component of expenditure would rise less rapidly than the rate of inflation, thereby reducing the rate of growth in real expenditure. As the cyclical fluctuation in autonomous variables becomes reversed over time, the procedure assumes that the pursuit of a neutral policy will result in the restoration of full employment, but with an acceptable rate of inflation.

The justification for such procedures rests implicitly on the assumption that there is a stable Phillips curve relationship between the rate of price change and unemployment. Where, however, the Phillips curve relationship breaks down, as in a period of stagflation, it becomes increasingly inappropriate to overload the budget (viewed as a single instrument) with two target variables. The neutral policy may result in neither the attainment of full employment nor of the desired inflation rate. Reliance on the price norm will dampen the inflationary pressure originating from the budget, but will also cause it to be less expansionary in terms of the employment objective. Secularly higher rates of both inflation and unemployment may result.

The emergence of stagflation casts doubt on the appropriateness of measurements, previously described, that focus on the full-employment objective (even where it is modified by price norms).50 These measure-means—for example, through increases in government expenditure, constancy of shares in potential levels of output. This type of neutral policy permits the budget to exert an automatic stabilizing influence whenever actual output deviates from its potential level because of demand factors. The pursuit of this neutral expenditure policy, however, could be inflationary if the implicit excess aggregate demand that is generated when actual output falls short of its potential level (while expenditure keeps in step with the potential level) is not being eliminated through corresponding supply increases.51 An appropriate neutral definition is, then, one where government expenditure grows in step with actual supply. Together with the application of the equiproportionate revenue rule,52 the budget balance will now respond in a manner that does not increase the prevailing rate of inflation.53

These definitions imply a neutral balance that is some fixed proportion of GNP. If the actual budget balance, taken as a proportion of GNP, exceeds this normative proportion, the budget will have contributed to excess demand under the assumed circumstances, resulting either in a higher rate of inflation or a deterioration in the balance of payments, or both. This excess demand effect of the budget (EBL) is given, in liquidity balance terms, as follows:54

EBL=(GTBn)(g0b0t0)Y55(25)

with the normative balance determined either by reference to some past year of stability, or analytically by solving the appropriate model. If EBL is positive, the inflationary pressure generated by the budget will continue as long as supply remains sluggish and the other sectors do not acquiesce in a higher demand by government for national output.

The measure in equation (25) is stated in real terms, but it may be readily converted into nominal terms by the procedure of grossing up real magnitudes through the application of common deflators that was considered earlier. In particular, it should be noted that no price norms are invoked. The rationale is straightforward. In the inflation alternative to the unemployment case, inflation will result whenever there is an excess demand for output by the various sectors. The norm for neutrality requires that government maintain a fixed share of available output, which may be nominally valued by grossing up. Inflation affects the budget, and these effects need to be separated from the effects of the budget on inflation. This separation is undertaken here by comparing the liquidity balance with the nominally valued (using actual deflators rather than price norms) neutral budget norm. Thus, if the real normative balance is in deficit, a 10 per cent inflation will cause the deficit to widen in nominal terms by 10 per cent. An increase in the actual liquidity balance (deficit) of more than 10 per cent would indicate that the budget (fiscal policy) is contributing to inflation.56

While the presumption is strongly in favor of the liquidity balance measure when one is evaluating the inflationary impact of a budget (especially in a situation of sustained inflation), an overall balance measure may be more appropriate in certain stagflationary contexts. If the stagflation is accompanied by increasingly excessive private sector liquidity (relative to the lower aggregate level of supply), financing a higher level of government expenditure by borrowing from that sector would, according to the liquidity balance measure (EBL), indicate an unchanged inflationary impact. However, private sector expenditures may not be fully crowded out, and the inflationary impact of the budget is then more accurately reflected in the higher overall deficit balance.

V. Measures for LDCs

The distinction, which was drawn previously, between the objectives of policy and the appropriateness of measures is particularly relevant to LDCs. Because of differences between their economic environments and those of the industrial economies, additional distinctions must be noted. Thus, the concept of stabilization policy as the minimization of fluctuations around a trend may be problematic for many LDCs, for whom no well-defined trends are discernible and whose chief problems are those of overcoming acute structural maladjustments. Nevertheless, if a distinction is drawn between temporary factors (such as crop failures) and longer-term problems, a case may be made for viewing stabilization policy as one of promoting orderly adjustment to structural imbalances without sacrificing the momentum of growth. Possible objectives of stabilization policy would be those of minimizing unemployment, ensuring stable levels of absorption (consumption plus investment), and controlling inflation.

However, the full-employment objective is problematic for many LDCs. To begin with, the concept of potential output or full employment is difficult to define for an LDC, where resources are frequently underutilized owing to the presence of severe structural bottlenecks. Moreover, the dependence of LDCs on primary production subjects them to the vagaries of extraneous factors such as weather. Is potential or full-employment output to be defined as actual output in a year when production was affected by inclement weather and fell below normal, or as the level that would have been attained had weather conditions been normal? Obviously, stabilization policy cannot concern itself with the reconciliation of actual with potential output if the latter cannot be defined. Neither a crop failure nor a deterioration in the terms of trade can be reversed in the same way that excess capacity may be reduced.57 As a result of various specificities of domestic production, increases in aggregate demand are much more likely to spill over into imports or cause domestic inflation in LDCs than in industrial countries. The more appropriate role of stabilization policy would, therefore, be that of moderating some of the effects of variations in real income on absorption and inflation.

The goal of preserving absorption58 in the face of a calamitous (but temporary) decline in income may be accomplished through several means—for example, through increases in government expenditure, either on government’s own account, with the rise in government expenditure offsetting falls induced in private consumption and investment, or by means of higher transfers to the affected income groups (accompanied by attempts to reduce their tax burdens). Of course, the ability to undertake such a stabilization function depends on the size of a country’s international reserves and the availability of international credit (for example, the use of Fund resources). If reserves are adequate, it will be possible to maintain absorption while minimizing inflationary pressures.

Interestingly, the type of measures that would be required to appraise fiscal policy, when its objective is to minimize cyclical or temporary fluctuations in absorption, are formally similar to those constructed for the full-employment objective. The same set of measures may be used, with the proviso that trend income is to be substituted for potential output in the former case. If a normative share of the budget in trend income were postulated, neutral government expenditure would rise at the same rate as trend income. Any deviations of actual income from the trend level would be reflected in revenue collections, with an equiproportionate response regarded as neutral.59 The actual budget deficit would then be tested against this analogue of the Federal Republic of Germany’s cyclically neutral balance. If the cyclical effect of the budget were positive, the budget would be stimulating absorption and, if the underlying situation was one of real income decline, the policy would be countercyclical. When actual income coincided with its trend level and the normative budget was restored, the budget would not be a source of excessive absorption and would support the balance of payments objective.60

First differences in the cyclical effects of budgets or use of the Dutch BI calculation, with the rate of growth in trend income substituted for that of potential output, would likewise indicate how the impact of the budget was changing in terms of the absorption objective. The analysis conducted would thus be formally similar to that for the full-employment measures, and would be subject to the various qualifications that were considered earlier.61

Inadequacy of reserves would render the balance of payments constraint binding. A budget that was stimulative in the face of a real income shortfall could not be accompanied by an increase in the imported component of aggregate supply, and, therefore, would tend to be dissipated by an increase in domestic inflation. In these circumstances, the more appropriate definition of a neutral budget would be one that preserves the normative share in actual output rather than trend income, the objective of policy being to control inflationary pressures that might accompany a real income decline.62 The EBL measure discussed in the preceding section would be of immediate relevance in the present context.63

VI. Some Illustrations of the Use of Measures

In order to provide concrete illustrations of possible uses of measures, including that of assessing fiscal adequacy, case studies are drawn from the period 1964-75 for both the United Kingdom and Malaysia. The evaluation of fiscal policy in the United Kingdom is undertaken primarily with respect to the full-employment objective, while the evaluation for Malaysia (for which the reserve constraint did not appear to be binding) is done in terms of the absorption objective. The bulk of the analysis for the United Kingdom uses measures based on the overall balance concept, but it also takes account of any contrary indications provided by the liquidity balance measures. The analysis for Malaysia is conducted primarily in terms of the liquidity balance concept.

1. fiscal policy in the united kingdom

The overall balance (also the liquidity balance) has exhibited wide swings during the period under review. This is shown in Chart 1, which also plots the full-employment balance that corresponds to the overall balance in each year. The wide swings in the FEB line point to fiscal policy as the major cause of fluctuations in the overall balance. This is further borne out by Chart 2, which shows that the swings in a cyclically neutral balance (CNB) profile are of much smaller amplitude. It is, therefore, of particular interest to establish the appropriateness of fiscal policy in each year. This is undertaken by drawing on the three national measures in turn.

Chart 1.
Chart 1.

United Kingdom : Fluctuations in Full-Employment Balance (FEB), 1964-75

Citation: IMF Staff Papers 1977, 002; 10.5089/9781451947557.024.A006

1. Overal balance2. Normative full-employment balance.
Chart 2.
Chart 2.

United Kingdom: Fluctuations in Cyclically Neutral Balance (CNB), 1964-75

Citation: IMF Staff Papers 1977, 002; 10.5089/9781451947557.024.A006

1 Normative cyclically neutral balance.2 Overall balance.3 Overall balance for a hypothetical countercyclical policy.

From Chart 1, the position of the FEB line relative to the OB line reflects the direction in which output deviated from its full-employment level. With the exception of 1975, when output fell below its full-employment level, thereby dragging the FEB line below the OB line, the full-employment deficit (surplus) was either always larger (smaller) than the actual deficit (surplus) or else equal to it, as in 1971 and 1972.64 On the face of it, this setting of overly full employment called for a contractionary fiscal posture throughout the period up to 1974. Without recourse to a norm, however, it is not possible to establish from Chart 1 whether budgets were contractionary or not. All that can be inferred are the changes in fiscal stance—that is, whether fiscal policy was becoming relatively more expansionary or contractionary. The swings in the FEB line indicate that the budget moved in a contractionary direction between 1967 and 1969, but thereafter pursued a consistently expansionary direction.

The explicit reliance of the Federal Republic of Germany’s CEB on a budget norm provides a means for assessing the appropriateness of budgets. Employing the overall deficit (2.6 per cent of GNP) in the full-employment year of 1972 as the norm, the CNB line in Chart 2 traces the variations in this normative balance as a consequence of fluctuations in the economy. Except for 1975, the CNB was, in all years, below the normative deficit, reflecting the situation of overfull employment. In these circumstances, an active countercyclical policy to restore full-employment conditions would have required budget balances smaller than those indicated by the CNB line (except for 1975).65 Such a policy is illustrated by the OB’ line. The assessed cyclical effects of budgets (CEB), shown as the difference between the OB’ and CNB lines, would then have been negative. The figure, however, shows that, with the exception of the years 1969-72, the CEBs with respect to the actual OB line have all been positive. The analysis based on the German measure thus indicates that budgets were procyclical from 1964-68, and between 1973 and 1974, but countercyclical from 1969-72 and in 1975.

A similar assessment is obtained using the normative interpretation of the FEB. Employing the 1972 budget balance as the norm, which is indicated in Chart 1 by the FEB* line, a countercyclical policy would require, in years of overfull employment, that the FEB lie below FEB*, and conversely. This countercyclical conjunction of budgets occurred only in the years 1969–72 and in 1975.

Obviously, the assessment of the countercyclical stance of budgets will depend on the norm. Thus, if the normative deficit was arbitrarily set at 5 per cent of full-employment GNP, all the budgets between 1964 and 1975 would appear to be countercyclical, whereas setting the norm at a surplus balance of 1.5 per cent of GNP would have rendered all the budgets procyclical.66 Such flexibility is reduced by the procedure of examining the balances in relatively normal years such as 1971 (when the budget deficit amounted to 0.9 per cent of GNP) or 1972 (when it amounted to 2.6 per cent). The arbitrariness that remains may be further reduced by considering the situation, both with regard to international reserves and to the desired relationships between sectoral balances in the basic national income accounting identity.67 Casual empiricism, together with the manifest need for improving the net foreign asset position of the banking system of the United Kingdom, points in the direction of a normative overall balance in surplus. The conclusions regarding the procyclical stance of budgets would, therefore, be reinforced.

Establishing whether changes in fiscal stance are in the right direction requires an assessment both of the extent to which the underlying budget’s posture is inappropriate and of the cyclical needs of the moment. The appraisal of the latter is, perhaps, best undertaken through recourse to the Dutch method.68 As noted earlier, this method focuses on the effective rate of growth in government expenditure, and does not take any position on the underlying budget posture. The cyclical need may be established by considering impulses from the rest of the economy, which will determine the need for a counteracting impulse from the budget. Thus, if the rate of growth of exports declines, ceteris paribus, increasing the budget impulse would be offsetting. Broadly (and over a period of a year), the different sectoral impulses may be regarded as comparable, in which case an estimate of fiscal adequacy can be obtained.

On the assumption that exports and private fixed investment account for the bulk of the autonomous fluctuations, excess deviations in these variables are plotted in Chart 3.69 Two versions of the Dutch BI are also shown in the figure, based, alternatively, on the overall balance and liquidity balance concepts. In Chart 3, fiscal offset is indicated by movements in the budget lines opposite to those of the autonomous variables. Interestingly, indications similar to those shown earlier also hold for changes in fiscal stance. Between 1964 and 1966, budget impulses (overall balance measure) were procyclical; they became countercyclical in the years to 1971, but once again became procyclical from 1972-74.70

Chart 3.
Chart 3.

United Kingdom: Changes in Fiscal Stance and Fluctuations in Major Autonomous Variables, 1964-75

Citation: IMF Staff Papers 1977, 002; 10.5089/9781451947557.024.A006

With the exception of the few years following the 1967 devaluation, both the underlying budget posture and the directional changes appear to have been inappropriate. However, the assessment of the procyclical character of the changes in fiscal stance should be tempered by the consideration that correcting a structurally incorrect budget may conflict with the cyclical needs of the moment. Indeed, it is quite possible for a budget to be excessively expansionary initially and to continue being expansionary despite the downturn in autonomous variables. A contractionary budget impulse may be required, although the change in fiscal stance will be shown to be procyclical. This example points to the dangers of relying wholly on fiscal impulse-type measures.

It is of interest to contrast the assessments based on the alternative overall balance and liquidity balance concepts. As is indicated by Table 2 (rows A 4-5 and B 5-6, respectively) in the Appendix, the assessment of the cyclical effects of budgets differs only for 1966. The overall balance approach assesses the budget as one of neutrality, whereas the alternative liquidity approach indicates the budget was expansionary. The latter conclusion is probably more valid in view of the sustained overfull employment context. Regarding assessment of changes in fiscal stance, differences occurred in 3 of the 11 years considered: 1968, 1971, and 1975. (See Table 2 in the Appendix and also Chart 3). Unlike the presumption in favor of the liquidity balance measure for the earlier years, the stagflationary situation of 1975, with private savings rising, suggests that the crowding-out effect may have been relatively weak. The budget impulse, instead of being contractionary as shown by the liquidity balance measure, would have been expansionary, as indicated by the alternative (overall) balance measure.

Finally, it is also of interest to evaluate the budget in terms of the inflation objective. This requires the selection of some normative ratio of the liquidity balance to GNP. If the ratio of 1.2 per cent (deficit) of GNP for 1972 is selected as the norm, the inflation measure EBL (see equation (25)) indicates that, except for the postdevaluation years of 1968-71, the budget was exerting an inflationary impact in the period 1965 onward, and that it was particularly pronounced in 1974 and 1975 (see Table 2, row B 4).71

The conclusion appears inescapable from the application of the alternative measures, both with regard to the full-employment objective and particularly from the point of view of the control of inflation, that for much of the period between 1964 and 1975 the increase in the relative size of the government sector was accompanied by budgets that were excessively expansionary.

2. fiscal policy in malaysia

Export earnings, which presently account for well over 40 per cent of GNP, constitute the major autonomous factor underlying the fairly regular cyclical fluctuations of the economy. Since international reserve holdings have been substantial and the level of official international indebtedness low (the debt service ratio has not exceeded 2 per cent of export earnings), the circumstances appear favorable for a countercyclical fiscal policy that would minimize the effects of export fluctuations on domestic absorption. The Malaysian authorities have consciously employed the budget as a countercyclical instrument since 1969.72

Liquidity balance measures are used for the evaluation that follows, partly for illustrative purposes but also because the bulk of domestic nonbank borrowing is from a captive fund—the Employees’ Provident Fund—which, for present analytical purposes, may be treated in the same way as a social security tax. Using a method analogous to the construction of the full-employment balance, trend income liquidity balances (TYLBs) were estimated (see Table 3 in the Appendix) and are shown in Chart 4. The chart demonstrates that the wide fluctuations in the liquidity balance were largely the result of fiscal policy. Fluctuations in the cyclically neutral liquidity balance (CNLB), shown in Chart 5, were of much smaller amplitude.

Chart 4.
Chart 4.

Malaysia: Fluctuations in Trend Income Liquidity Balance (TYLB), 1964-75

Citation: IMF Staff Papers 1977, 002; 10.5089/9781451947557.024.A006

1Liquidity balance.2Normative trend income liquidity balance.
Chart 5.
Chart 5.

Malasia: Fluctuations in Cyclically Neutral Liquidity Balance (CNLB), 1964-75

Citation: IMF Staff Papers 1977, 002; 10.5089/9781451947557.024.A006

1Liquidity balance.2Liquidity balance for a hypothetical countercyclical policy.3Normative cyclically neutral liquidity balance.

The countercyclical posture of budgets may be established either through a normative interpretation of the TYLB or through recourse to the Federal Republic of Germany’s CELB. A constant liquidity balance deficit norm of 1.1 per cent of trend GNP is assumed, which is the same as was observed in 1966, when GNP was on its trend line.73 In those years when GNP falls below its trend level, the CNLB is wider than its normative trend GNP level (see Chart 5). A countercyclical stance is then indicated by a positive CELB (actual liquidity balance wider than the CNLB level).74 A possible countercyclical profile of the budget is illustrated by the dashed LB’ line in Chart 5. An examination of this figure reveals that budgets were procyclical between 1964 and 1970. Although budget balances were sensitive to the cycle, as is evidenced by fluctuations in the LB line (implying that the budget was stabilizing), it was not until 1971 that the budget exerted a countercyclical effect. Much the same result is indicated by Chart 4, which shows that the TYLB did not exceed its normative level until 1971. The downturns (relative to trend) in export earnings and GNP that occurred in 1971 and 1972 were met by budgets that were expansionary and countercyclical. In 1973, there was a sharp upturn and the deficit liquidity balance narrowed considerably, but apparently not by a large enough amount to be judged countercyclical according to the norm. As is suggested by Chart 5, the CELB should have been negative for a countercyclical policy, both in 1973 and in 1974, instead of being positive. However, in the severe recession of 1975, the continued excess of the deficit liquidity balance over its cyclically neutral level resulted in the budget posture being assessed as countercyclical.

Once again, these assessments of budget posture depend on the norm selected. If the normative liquidity balance was in surplus, say, the budget would be termed countercyclical in some of the earlier years, but the assessments mentioned previously for the later years would be reinforced. Alternatively, if a larger liquidity balance deficit were chosen as the norm, the conclusion that budgets were procyclical in the earlier years would be strengthened, but the assessment that the budget was procyclical in 1973 would be reversed. Conceivably, two different norms, one for the earlier period and one for the later period, could be employed in order to generate the result that budgets were mostly countercyclical.

In the present circumstances, where the permissible range of normative balances gives rise to opposite assessments and, moveover, where doubts exist as to the structural continuity of norms, it is especially desirable to consider the fiscal impulse-type measures. The two versions of the Dutch measure,75 based on the alternative balance concepts, are plotted in Chart 6, together with the excess deviations of export earnings (in terms of trend).76 The liquidity balance version shows that budget impulses were procyclical from 1964-68, countercyclical from 1969-74, and procyclical in 1975. Similar results follow from the overall balance version, except that it dates the period of countercyclical impulses from 1970 up to and including 1975.77 The conflicting result for 1975 is attributable to the sharp increase in domestic nonbank borrowing from the previous year. Once again, choosing between the two measures becomes a question of establishing whether the increase in that form of borrowing displaced private expenditure. Partial evidence that this may have been less of a factor in reducing private expenditure is provided by the higher liquidity of the banking system and the limited demand for credit, despite the policy of monetary ease.

Chart 6.
Chart 6.

Malaysia: Changes in Fiscal Stance and Fluctuations in Export Earnings, 1964-75

Citation: IMF Staff Papers 1977, 002; 10.5089/9781451947557.024.A006

Adding the budget impulses to the excess deviations provides indications that the restraint applied to the economy in 1969 was mild; more complete offsets were provided to the cyclical downswing in exports between 1970 and 1972, especially in 1971. Impulses were mildly restrictive in the boom that followed and, at best (i.e., according to the overall balance measure shown in Chart 6), slightly stimulative in the severe contraction of 1975. Interestingly, the evaluation of budget impulses largely confirmed earlier assessments, based on the Federal Republic of Germany’s CELB, that budgets were procyclical in the first half of the period, but became countercyclical in the second half. It should also be noted that the earlier assessment that budgets were procyclical in 1973 and 1974 is not entirely contradicted by the fact that budget impulses were countercyclical in these two years, since their offsetting effects were relatively small. Moreover, with regard to 1975, it is possible for the underlying budget posture to be judged countercyclical, even though the impulses were, at best, mildly stimulative.

VII. Concluding Comments

An essential prerequisite for the evaluation of fiscal policy is the isolation of the effects of the target variable on the budget instrument. Since the simple budget balance rule fails to provide the necessary separation, it has limited usefulness. Therefore, slightly more complex measures of the type examined in the paper are required. However, because of their summary nature, these measures are also subject to several limitations.78 Nevertheless, they do facilitate quick and simple appraisals of budgets and serve at least as a convenient starting point.

Major issues that arise in the evaluation of a budget from the standpoint of full-employment or absorption objectives concern the budget’s profile when cyclical factors are eliminated, the posture of the budget vis-à-vis the cycle, and the changes in fiscal stance or the impulses emanating from the budget. The three national measures reviewed in the paper each focus on a different one of these major issues. Comparison of these measures suggests that each has a comparative advantage with regard to the purpose for which the measure was initially designed. This paper has, however, attempted to demonstrate that, when they are appropriately interpreted, the measures are interchangeable. In principle, any one of the measures may be used for a comprehensive evaluation of a budget. Thus, the U. S. FEB calculations can be employed to indicate the extent to which the restoration of full employment would change the budget profile; to rank alternative budgets at full employment in order to draw inferences regarding their relative expansionary effects (both away from and at full employment); to shed light on the countercyclical posture of budgets when the FEB is given a normative intepretation; and to suggest the impulses that are transmitted by the budget when year-to-year changes in the FEB are considered. Aside from the issue posed when budget components do not respond equiproportionately to changes in GNP (thereby raising the possibility that the evaluations at full-employment levels of GNP may sometimes differ from those made at other levels), the deciding factor in choosing between the three measures is one of statistical convenience. In the latter respect, the German (Fed. Rep.) and Dutch measures appear to be less demanding than the U. S. FEB, which requires the estimation of functions for taxes and other endogenous components of the budget.

Specific limitations of the three national measures with regard to the financing of the budget and also their use for the evaluation of fiscal policy in terms of the inflation control objective were considered. Parallel alternatives drawing on the liquidity balance concept were proposed for use in circumstances where financing considerations are of importance—for example, in an excess demand context—while a simple liquidity balance measure was set up for evaluating the inflationary repercussions of a budget in certain environments. Possible uses of measures in LDCs for the evaluation of budget impacts on objectives other than full employment were also examined. In addition, an approach to the systematic use of alternative measures for the evaluation of fiscal policy, including its adequacy, was illustrated by case studies.

Certain of the uses of measures—for example, to provide countercyclical assessments of budgets—involve normative evaluations. Notwithstanding the element of arbitrariness in selecting a budget balance norm, normative assessments are of considerable value. This is particularly true where, in setting the level of the norm, account is taken of possible linkages between the budget on the one hand, and target variables—such as the net foreign asset position of the banking system, the inflation rate, the rate of growth in output—on the other.79 While the experience of past periods of stability will provide some guidance, the norm may be established as an integral component of a medium-term credit program. Whenever target variables deviate from their programmed values, the normative profile of the budget will be influenced—for example, the widening in the cyclically neutral balance caused by a recession. Taking account of the policy objectives of the moment, a systematic assessment of the possible contributions of alternative budgets to a corrective program may be obtained by examining divergences from the normative profile. The assessment will also indicate how the budget stands in relation to the desired medium-term profile and, hence, the changes in budget that may (eventually) be necessary to restore the norm. The normative evaluation should be supplemented by an examination of budget impulses. Exclusive reliance on the latter, however, could be misleading.

Finally, the scope provided by the various measures for the international comparison of budgets should be noted. Those uses of budgets that depend on specific country norms are obviously less readily adapted for international evaluations. However, the assessment of changes in fiscal stance does not depend crucially on country norms, and would more appropriately allow international comparisons. Indeed, when budget impulses are standardized (estimated as first differences of either the full-employment balance or the Federal Republic of Germany’s cyclical effects of the budget or, alternately, using the Dutch budget impulse method) as percentages of each country’s GNP, a composite index of fiscal influence may be constructed for assessing the impact of budgets in a wider setting.

APPENDIX

Table 2.

United Kingdom: Summary Measures of Fiscal Impact (Central Government)1

(In per cent of GDP at current market prices)2

article image
Sources: International Financial Statistics and Fund staff estimates of potential output.

Positive (negative) figures indicate an expansionary (contractionary) effect or a deficit (surplus) balance. Asterisks (*) refer to normative levels.

Except for full-employment measures, which are given as percentages of the current valuation of potential GDP. All measures are nominally valued by grossing up (raising) real magnitudes (see Section IV).

Potential output at current prices is real potential output multiplied by the GDP deflator. The normative base year is 1972.

Defined as (g72. YP - t72- Y)/Y.

Determined as (G - T*)/Yp, where T* = [1 - (Y - YP)/YT. Empirical studies indicate that the income elasticity of the revenue system is close to unity. See Baas and Dixon, op. cit.

FEB - FEB* = OB* for 1972.

Defined as (Δ OB + yT-1 - yG-1)/Y. See Section II.

Defined as [g72 - b72) Yp - t72Y/Y, with domestic nonbank borrowing treated as a discretionary variable. See Section III.

(FEB - Bn)/YP.

LB - LB* for 1972 as the assumed norm norm See Section IV.

FELB - FELB* = LB* for 1972.

Defined as [ΔLB + yT1 - y(G-1 - B-1)/Y. See Section III.

Table 3.

Malaysia: Summary Measures of Fiscal Impact (Federal Government)1

(In per cent of GNP at current market prices) 2

article image
Source: International Financial Statistics.

Positive (negative) figures indicate an expansionary (contractionary) effect or a deficit (surplus) balance. Asterisks (*) refer to normative levels.

Except for trend income measures, which are stated as a percentage of the current valuation of real trend GNP. In the absence of appropriate deflators, the latter is estimated first by finding the time trend of nominal GNP deflated by the consumer price index, and then grossing up by that index. All measures are nominally valued by grossing up of real magnitudes (see Section IV). The normative base year is 1966.

Defined as (g66

article image
- t66Y)/Y, where
article image
is trend income (TY)

Determined as (G - T*)/

article image
, where T*=[1(YY^)Y]. T. The income elasticity for the Malaysia revenue system has been estimated as about unity. See Chand, op. cit.

TYB - TYB* = OB* for 1966.

Defined as (ΔOB + yT-1 - yG-1)/Y. See Section II.

G denotes government expenditure.

Defined as [g66.

article image
- (b66 + t66)Y] with domestic nonbank borrowing treated as an endogenous variable. See Section III.

(TYBBn*)/Y^whereBn*=[1(YY^)Y]Bn.

LB - LB* for 1966 as the assumed norm. See Section IV.

TYLB - TYLB* = LB* for 1966.

Defined as [Δ LB + y(T-1 + B-1) - yG-1)/Y. See Section III.

*

Mr. Chand, economist in the Fiscal Analysis Division of the Fiscal Affairs Department, graduated from Oxford University and undertook postgraduate studies at the University of Western Ontario. He formerly worked in the Ugandan Government’s Central Planning Bureau.

1

This paper synthesizes and extends some of the studies that have been undertaken in the Fund’s Fiscal Affairs Department on the measurement and analysis of fiscal impact. Several of these studies formed part of a series on the techniques of fiscal analysis in selected industrial countries, with the following drawn upon for the present study: Daryl A. Dixon, “Techniques of Fiscal Analysis in the Netherlands,” Staff Papers, Vol. 19 (November 1972), pp. 615-46; Daryl A. Dixon, “The Full Employment Budget Surplus Concept as a Tool of Fiscal Analysis in the United States,” Staff Papers, Vol. 20 (March 1973), pp. 203-26; Thomas F. Dernburg, “Fiscal Analysis in the Federal Republic of Germany: The Cyclically Neutral Budget,” Staff Papers, Vol. 22 (November 1975), pp. 825-57. The paper has also benefited from the study by John W. Nevile, “Measuring the Impact of Fiscal Policy” (unpublished, International Monetary Fund, February 13, 1976).

2

Richard Goode, p. 402 in “Comments,” in Staff Papers and Other Materials Reviewed by the President’s Commission, U. S. President’s Commission on Budget Concepts (Washington, 1967), pp. 401-404.

3

Alan S. Blinder and Robert M. Solow, p. 12 in “Analytical Foundations of Fiscal Policy,” in The Economics of Public Finance, The Brookings Institution (Washington, 1974), pp. 3-115.

4

Some of the possible diversity of measures is illustrated in the comprehensive analytical reviews by Joergen Lotz, “Techniques of Measuring the Effects of Fiscal Policy,” OECD Economic Outlook: Occasional Studies (July 1971), pp. 3-31, and by Blinder and Solow, op. cit.

5

A more detailed classification and specification of definition can be found in A Manual on Government Finance Statistics (International Monetary Fund, draft, June 1974).

6

Alternative schemes of classification would generate additional notions of budget balance. See Raja J. Chelliah, “Significance of Alternative Concepts of Budget Deficit,” Staff Papers, Vol. 20 (November 1973), pp. 741-84.

7

In the Hansen-Lindahl tradition, fiscal policy is defined as those budgetary transactions that affect the size of public debt but not its final composition, the latter being the responsibility of monetary policy (assuming that the monetary authorities have an independent function). See Blinder and Solow, op. cit., p. 4, for an elaboration. However, the monetarist emphasis on the liquidity balance extends the concept of fiscal policy to include those transactions that concern the portion of the public debt held by the domestic nonbank public.

8

This inference that the budget is expansionary requires that the domestic component of the budget be in deficit, in which case the net contribution of the budget to the demand for domestic output is positive.

9

Early arguments for drawing the distinction in order to appraise more accurately the discretionary uses of fiscal policy were made in Jan Tinbergen, On the Theory of Economic Policy (Amsterdam, 1952); E. Cary Brown, “Fiscal Policy in the Thirties: A Reappraisal,” American Economic Review, Vol. 46 (December 1956), pp. 857’79; and Bent Hansen, The Economic Theory of Fiscal Policy (London, 1958).

10

At this juncture, a distinction between stabilizing and offsetting (expansionary or contractionary) effects should be noted. A variation in the budget balance that is the result of economic fluctuation usually represents a stabilizing response of the budget. In the absence of a budget, and specifically of the tax leak, the multiplier would be larger. Making tax revenue endogenous introduces a tax leak which dampens, via a reduced multiplier, the effects of fluctuations in autonomous variables. However, should the built-in elasticity of the tax system with respect to gross national product (GNP) differ significantly from unity, the size of the multiplier will not be independent of the level of economic activity, since the effects of progression, for example, will cause the tax leak parameters to take values that vary with the level of GNP. Hence, in addition to the dampening effect of the smaller multiplier, an offset to the fluctuation in the autonomous variable could result, insofar as the multiplier changed from what it was before the autonomous disturbance. If the latter effect is not sufficiently offsetting, discretionary actions may be required.

11

It should also be noted that the measures considered in the remainder of the paper do not discriminate between budgets whose overall (or liquidity) balances are identical, but whose components are set at different levels. A more refined analysis would require the application of weighting schemes. Since weights are estimated by solving the reduced form of a model, they should not be directly incorporated in the summary measures presented here that arise from considering equilibrium condition balances. Attaching weights to the expenditure and financing components of a budget will result in an assessment of the total effects of a budget (where the comparison is with a situation of no budget), and not of fiscal policy. The (weighted) evaluation of the latter requires the assignment of weights to the reduced-form components of budgets. On this, see Richard A. Musgrave, “On Measuring Fiscal Performance,” Review of Economics and Statistics, Vol. 46 (May 1964), pp. 213-20 and Bent Hansen, Fiscal Policy in Seven Countries, 1955-1965 (Organization for Economic Cooperation and Development, Paris, 1969).

12

The concept was initiated by the Committee for Economic Development in their Taxes and the Budget: A Program for Prosperity in a Free Economy (New York, 1947) and first applied by the U. S. Council of Economic Advisers in 1962 in the Economic Report of the President (Washington, 1962).

13

Developed by the German Council of Experts. The most elaborate presentation is to be found in the Council’s report for 1970/71, Konjunktur im Umbruch—Risiken und Chancen, Jahresgutachten 1970/1971 (Stuttgart and Mainz, 1970).

14

A full account is provided in The Netherlands Budget Memorandum 1970, Ministry of Finance (The Hague, 1970).

15

The measure has been widely discussed. Recent expositions are provided in Blinder and Solow, op. cit., and in Lotz, op. cit.

16

Full-employment government expenditure may vary if components—social security benefits, for example—are affected by the level of income.

17

In what follows, full-employment income and potential output are assumed to be synonymous. All measures are stated in real terms unless otherwise noted. Problems involved in the definition of full-employment income or output are not considered.

18

The controversy is discussed in Blinder and Solow, op. cit., pp. 38-42. For the current account of the balance of payments to be in equilibrium over the medium term, the budget balance must exactly offset the private sector balance between saving and investment.

20

A discussion of the concept is found in the paper by Dieter Biehl and others, “Measuring the Demand Effects of Fiscal Policy,” in Fiscal Policy and Demand Management, ed. by Herbert Giersch (Tübingen, 1973). See also Lotz, op. cit.

21

In principle, the base-year normative balance need not be restricted to being a historical magnitude, and could be determined otherwise—for example, as part of the solution of a planning model. The matter is discussed further in Section IV.1.

22

Assessing the cyclical posture of a budget thus requires two steps: first, a comparison of the actual budget balance with the CNB for that year; second, an evaluation of how that CNB differs from its base-year level. A budget can, therefore, be classified in any one of six modes, depending on whether it is expansionary or contractionary (the CEB test) and on whether it is procyclical, neutral, or countercyclical (depending on the evolution of the CNB).

23

See footnote 10.

24

Since automaticity is ultimately the result of discretionary policy, the selection of an appropriate revenue response elasticity to separate an automatic from a discretionary revenue response may be viewed as a definitional matter. The Federal Republic of Germany’s rule employs a unit elasticity as the criterion.

25

The discussion of the measure presented above does not take account of certain revisions of the concept made in 1974 by the Federal Republic of Germany’s Council of Experts. The major revision appears to be the incorporation of a procedure whereby neutral changes in the base-year ratios may be made. The tax ratio is permitted to change as long as resulting changes in revenue are offset by changes in expenditure. However, aside from conferring flexibility in the use of the base-year notion, the essentials of the analysis appear to be retained.

26

The notions are explained by Hendrik Burger in “Possible Concepts for Better Planning and Evaluating Fiscal Policy—Experiences in the Netherlands,” pp. 210-22 in Giersch, op. cit., and in “Structural Budget Policy in the Netherlands,” De Economist, Vol. 123 (No. 3, 1975), pp. 329-51.

27

The analogy with a ship’s rudder that always compensates for changes in current, but does not correct for the vessel not being on course, is, perhaps, appropriate.

28

This is derived using the relation 1 + y = Y/Y-1.

29

Using the property YY1=ΔYY1Y1. The superscript n refers to neutral levels, under the German definition, that are generated by the application of the base-year proportionality factors to the preceding year’s levels of potential and actual output.

30

See Biehl and others, op. cit., equation (10), p. 229 for a similar derivation.

31

See Table 1 later in this section. The equivalence is at variance with the following observation of Biehl and others, op. cit., p. 229: “. . . in spite of the insertion of a medium-term target value, the impact analysis still adheres to a considerable extent to the balanced budget rule and to the comparison with the previous year. . . .reservations have to be made against . . . the impact analysis.” It is particularly difficult to see why the impact analysis is viewed as adhering to the budget balance rule, given that it defines for each year, as does the German measure, a neutral growth in the budget, against which the actual budget is tested.

32

Once the normative CNB is specified, it is a simple matter to reconstitute the German measure from the Dutch measure.

33

Asterisks denote normative full-employment levels. Standardization in the indicated manner ensures comparability over time and over different levels of GNP.

34

See, for example, the debate between the two authors in Milton Friedman and Walter W. Heller, Monetary Versus Fiscal Policy (New York, 1969).

35

The above analysis assumes that the conflict between Keynesians and monetarists over the relative efficacy of fiscal and monetary policies in influencing GNP arises from an empirically resolvable difference of opinion, regarding the extent to which private sector expenditure is crowded out by direct government borrowing. If the crowding-out response is known, the conflict is reducible to a semantic one in a wider (general equilibrium) setting. Suppose that crowding out is perfect, and that there is an autonomous increase in government expenditure financed by the (nonbank) private sector. The increase in government expenditure would displace an equal amount of private expenditure, with no effect on GNP (monetary conditions are assumed constant in the manner defined above). Depending on the definitions selected, the overall macroeconomic effect of no change is compatible with the finding that fiscal policy is expansionary (Keynesian definition), in which case the so-called constant monetary conditions imply a contractionary effect, or the finding that fiscal policy is neutral (monetarist definition), in which case constant monetary conditions would be appropriately neutral.

36

The norm implies a particular pattern of the use of output by government, financed by a mix of domestic and foreign credit, with the latter influencing the process of liquidity creation. For a goods market analysis, the focus would be on items above the line, whereas for an equivalent analysis in terms of monetary flows, the emphasis would be on the liquidity repercussions of the budget. It should be noted that the statement of the norm here abstracts from varying price-level considerations that are dealt with in the next section.

37

As with the use of the FEB, the comparison of an actual liquidity balance with its full-employment level (FELB) shows the effects of automatic stabilizers. Alternative budgets’ liquidity balances may also be ranked at their full-employment levels to indicate their relative expansionary or contractionary effects.

38

This is also relevant to the estimation of the full-employment level of the liquidity balance.

39

This treatment was accorded to domestic nonbank borrowing in the U. K. illustration, since no systematic relationship with GNP was observed over the period examined (see Appendix, Table 2).

40

As was essentially the case in Malaysia.

41

The definition will need to be modified if private saving is a stable function of permanent income, rather than of actual GNP. In the former case, any temporary deviation of income from its permanent level would result in saving being more than proportionately affected.

42

The comparison in terms of growth rates may be re-expressed as a percentage of GNP by multiplying equation (24) by G-1 (or G-1 - B-1) and standardizing the resulting expression by dividing by GNP.

43

Revenue will respond faster if, for example, the income tax structure is progressive and brackets are specified in nominal terms. Inflation will push taxpayers into higher brackets even if their real incomes remain constant.

44

Lotz, op. cit., p. 13 has argued that “For practical purposes it may be generally accepted that all changes in government expenditures on goods and services are treated as discretionary . . .”

45

As was generally the case with government procurement in the United States before the onset of higher inflation in 1974.

46

The imposition of cash limits will have the effect of reducing real expenditure when the expected rate of inflation falls short of the actual rate. This practice is, however, more appropriately viewed as an anti-inflationary device and should not, therefore, affect the definition of the neutral yardstick where the policy objective is full employment.

47

With regard to the FEB, the concern appears to have been more with what rate of inflation to assume for the calculation of full-employment estimates, rather than the normative issue of the appropriate level of FEB* for price stability. See Arthur M. Okun and Nancy H. Teeters, “The Full Employment Surplus Revisited,” Brooking Papers on Economic Activity (1970:1), pp. 77-110.

48

Biehl and others, op. cit., p. 226. “The government will not be awarded good marks for its stabilization policy if, in order to guarantee full employment, it accepts, or produces, relatively high rates of inflation.”

49

Burger, “Possible Concepts,” op. cit., p. 218. “A reasonable solution could be to measure the impulse without using a deflator and to confront the result with a growth rate of GNP, of which both the volume and the price element is put at an acceptable medium-term percentage.”

50

In essence, the problem is how to apportion the impact of a budget between output and prices. One extreme is that of the unmodified full-employment measure, which assigns the entire effect of the budget to output. The other extreme is that of the monetarist-type liquidity balance measure, stated in equation (25) below, where the budget effect is assumed to be on both prices and imports.

51

The notion of potential output can be misleading in a stagflation context.

52

Once again, the purpose of the norm is to isolate the interacting effects of the target variable and the instrument. Inflation will cause nominal revenue and expenditure to rise. The norm ensures that only that part of the increase consistent with maintenance of fixed shares of actual output can be viewed as automatic effects.

53

If the rate of inflation increases, this would be attributable to other sectors.

54

In order to avoid excessive duplication, the discussion assumes that domestic nonbank borrowing is a discretionary variable; however, the alternative assumption of endogenous borrowing may easily be employed.

55

In order to see that such a measure logically precedes the measures considered earlier, consider the definition of the Federal Republic of Germany’s CELB, stated in equation (19), CELB = (G - T -Bn) - (g0Yp - b0Y p - t0Y). If the identity g0Yp = g0Y + g0(Yp - Y) is substituted in this expression, the CELB takes the following form: CELB = (G - T - Bn) - (g0 - t0 - b0)Y -(g0 - b0)(Yp - Y). In other words, the CELB equals the EBL stated in equation (25) above, less a proportion of the gap between actual and potential output. Where output is constrained, the incipient supply response that forms part of the automatic stabilizing response of the budget will not occur, and the measures that are based on potential output will provide biased indications of potential inflationary effects.

56

Changes in fiscal impact may be found by taking first differences of equation (25). Interestingly, but only when the measure is standardized by dividing through by GNP, the first differences will equal the (percentage) changes in the ratio of the liquidity balance to GNP. This should not, however, be viewed as validating the budget balance rule, which assumes that the entire budget balance is discretionary (see Section I), whereas here the automatic effects of inflation on the budget are allowed. The Dutch measure, with the BIL definition retained but tested against the growth in actual supply rather than its potential level, will also yield similar results. This paragraph abstracts from any need of government to reconstitute cash balances to maintain a given real level, which might result in a further increase in the size of the nominal liquidity balance deficit.

57

These distinctions are developed in Richard Goode, “Efficacy of Fiscal Policies in Developing States” (unpublished, International Monetary Fund, June 25, 1965).

58

Stabilizing absorption is important, both because of its direct welfare connotations and its implications for the medium-term growth of aggregate supply. Real income fluctuations may be excessively disruptive to investment, for example. The measures stated here do not distinguish between consumption and investment, but assume that the appropriate balance will be maintained in order to ensure adequate growth momentum.

59

The justification for the equiproportionate rule here would be in terms of ensuring that all sectors share equally in the GNP fluctuation; otherwise, private absorption would be affected not only by, say, the fall in GNP but also by the public sector’s demand.

60

The selection of the normative balance would be guided by the need to ensure that the absorption trend corresponded to the income trend; otherwise, an initial situation of adequate reserves might be reversed over time.

61

It should be noted that while the primary objective here is one of absorption, the budget will also have effects on employment, although these are likely to be of a secondary order of importance.

62

These would result if private demand did not decline as rapidly as real incomes and, in the binding balance of payments case, exacerbated demand pressures on domestic sources of supply.

63

The choice between an overall balance or a liquidity balance approach is likely to be academic for many LDCs, insofar as domestic nonbank borrowing is insignificant. However, where this is important, the rule that was stated earlier in Section III may be employed for the absorption objective—namely, the use of an overall balance in recessionary phases of the cycle characterized by excessive private sector liquidity, and a liquidity balance otherwise.

The device of splitting the budget into a domestic component geared to the absorption objective and a foreign component directed to the balance of payments objective may be employed for a partial resolution of the problems posed by reserve inadequacy. See Goode, “Efficacy of Fiscal Policies in Developing States,” op. cit., for a discussion. The measures discussed previously may be applied to each component of the budget balance.

64

Robert Neild and Terry Ward illustrate the same phenomenon in their full-employment calculations that assume an unemployment figure of 650,000. See their article, “The Budget Deficit in Perspective,” p. 16 in The Times (London, July 12, 1976) that summarizes some of the findings presented in their paper, “The Budgetary Situation: An Appraisal,” Cambridge University, Department of Applied Economics (mimeographed, c. 1976). A fuller analysis would take into account variations in normal levels of capacity utilization and the natural rate of unemployment. Results suggesting that the natural rate of unemployment may have reached over 3 per cent of the labor force by 1973, corresponding roughly to the actual unemployment rates in the recession years of 1971 and 1972, are reported in David Laidler, “Inflation in Britain: A Monetarist Perspective,” American Economic Review, Vol. 66 (September 1976), pp. 485-500.

65

Empirical studies indicate that the overall built-in elasticity of the U.K. tax system is close to unity. See, for example, Hessel J. Baas and Daryl A. Dixon, “The Elasticity of the British Tax System” (unpublished, International Monetary Fund, September 23, 1974). Consequently, the use of the CNB rule here will serve to discriminate between automatic and discretionary effects.

66

In terms of Chart 2, changing the norm will cause an approximately parallel displacement in the CNB line of the same amount as the change in the normative balance.

67

The identity may be expressed as follows: (G - T) = (S - /) - (X - M). If the private sector is, on balance, a net saver, the current account will tend into surplus to finance that sector’s accumulation of foreign assets. The actual outcome will, however, depend on the financing needs of the budget. An overall deficit that exceeds the private sector’s surplus balance will cause a current account deficit. Deciding on the size of the normative overall balance also requires that consideration be given to the changing preferences of the private sector and the rest of the world for domestic government securities vis-á-vis foreign assets. An increasing preference for the latter, together with a need to ensure that interest rates do not choke off domestic investment, may require a policy of retirement of government debt and, hence, an overall balance in surplus. The size of the latter will, moreover, depend on the speed with which it is desired to improve the net foreign asset position of the blanking system.

68

Alternatively, first differences in the FEB or the Federal Republic of Germany’s CEB may be employed for a similar assessment.

69

The excess deviation is defined here, analogously to the Dutch BI, as the actual rate of growth in the autonomous variable less the rate of growth in potential output, namely, D=ΔJJ1y¯, or, on standardizing to ensure comparability with the budget impulse DJ-1 = Δ J-1 y. This definition assumes that if autonomous demand variables maintain their share in potential output, no correction from the government sector is required.

70

As an indication of the adequacy of changes in fiscal stance (in the limited sense considered here), adding the impulses from the autonomous variables and those of the budget indicates that net effects were excessively expansionary in 1964, excessively contractionary in 1969, and highly expansionary from 1971-74. A more comprehensive evaluation of fiscal adequacy would take into account the contribution of nonfiscal instruments as well.

71

One way of gauging the appropriateness of this norm is to consider the implications of preserving a given liquidity/GNP ratio. Suppose, as was roughly the case during the 1960s in the United Kingdom, that the ratio is one third. Assume also that potential output is growing at 3 per cent per annum and the average rate of inflation is 6 per cent. Preserving the liquidity/GNP ratio implies an annual increment in liquidity of 3 per cent of GNP. Assuming an unchanged net foreign asset position, the corresponding amount of credit creation would also amount to 3 per cent of GNP. This would have to be allocated between the public and private sectors. The assumption made here—that government acquires credit of 1.2 per cent of GNP, partly from abroad—is, therefore, within the range of acceptability. However, if the assumed target rate of inflation were lower and, moreover, if allowance were to be made for the need to improve the net foreign asset position of the banking system, the normative liquidity deficit balance might need to be converted into a surplus. This would reinforce the above assessment that budgets were, on the whole, inflationary.

72

See Economic Report, 1972/73, Treasury, Government of Malaysia (Kuala Lumpur, 1972).

73

Taking into account normal levels of foreign financing of the budget, the liquidity/GNP ratio, and the rate of growth of the economy, this norm is consistent with permissible levels of domestic credit creation.

74

Since the income elasticity of the revenue system of Malaysia is close to unity, the assessment based on the CNLB rule made here will approximately reproduce the automatic/discretionary distinction. The built-in elasticity estimates are found in Sheetal K. Chand, “Tax Revenue Forecasting: An Approach Applied to Malaysia” (unpublished, International Monetary Fund, March 12, 1975).

75

Alternatively, first differences of the trend income measure or the Federal Republic of Germany’s cyclical effect of the budget could be employed.

76

Excess deviations are defined using the formula of footnote 69. As before, the assessment of countercyclical offsets is made in terms of opposite movements of budget impulses and other impulses.

77

Qualitative assessments of underlying budget posture, based on the alternative balance concepts, were unanimous (see Table 3 in the Appendix).

78

As was stated earlier (footnote 11), the measures do not discriminate between budgets whose overall (or liquidity) balances are identical, but whose components are set at different levels. Nor do the measures take account of any lags in the interaction between the budget and the economy. The use of these measures should, therefore, be restricted to situations where the budget composition or size does not change drastically and where, as is frequently the case, the bulk of the budget impact occurs within the fiscal year.

79

See, for example, the case studies in Section VI.

IMF Staff papers: Volume 24 No. 2
Author: International Monetary Fund. Research Dept.
  • View in gallery

    United Kingdom : Fluctuations in Full-Employment Balance (FEB), 1964-75

  • View in gallery

    United Kingdom: Fluctuations in Cyclically Neutral Balance (CNB), 1964-75

  • View in gallery

    United Kingdom: Changes in Fiscal Stance and Fluctuations in Major Autonomous Variables, 1964-75

  • View in gallery

    Malaysia: Fluctuations in Trend Income Liquidity Balance (TYLB), 1964-75

  • View in gallery

    Malasia: Fluctuations in Cyclically Neutral Liquidity Balance (CNLB), 1964-75

  • View in gallery

    Malaysia: Changes in Fiscal Stance and Fluctuations in Export Earnings, 1964-75