The reality of chronic, and at times accelerating, inflation in the past decade has significantly complicated the policy problems of public sector decision makers. While fiscal policy presumably serves as an important element of price stabilization policy, the principal instruments of fiscal policy—government expenditure and revenue—are themselves affected by the inflationary process. Inflation raises the cost of government services and investment and increases the budgetary demands for distributional transfers while simultaneously increasing the amount of revenues collected. What is important to note is that these fiscal effects of inflation do not occur in tandem; different types of revenue and expenditure may respond at different speeds. In fact, it has recently been suggested (Aghevli and Khan, 1978) that one of the dynamic forces sustaining inflation in developing countries is inflation-induced fiscal deficits. Total government expenditure is said to adjust more rapidly than revenue to a change in the price level and in such a way that a procyclical bank-financed budgetary deficit emerges.
The Aghevli-Khan hypothesis is important for, if valid, it suggests the need for a serious re-examination of the current fiscal policy response to inflation. Beyond that hypothesis, the literature is very thin on the short-term dynamics of public revenue or expenditure growth in developing countries, particularly in response to inflation. 1
This paper attempts to provide a theoretical and empirical analysis of the factors underlying the relative speed of adjustment of different types of revenue and expenditure to inflation. In so doing, it provides an empirical test of the accuracy of the Aghevli-Khan hypothesis in describing the fiscal experience of developing countries over the past two decades. Section I suggests a conceptual framework for evaluating how alternative categories of revenue or expenditure are likely to respond to inflation. There are many different forms of expenditure and revenue; each is likely to respond to inflation in a particular way. With respect to expenditure, a distinction is made between the impact of inflation on the cost of providing government services and realizing the objectives of the public sector and the actual effect on the nominal level of expenditure. The government has a considerable amount of discretion in determining the magnitude of the budgetary response to such cost increases and may influence the extent of certain price increases. The reaction of total revenue or expenditure will reflect partly the initial structure of revenue and expenditure, partly the extent to which the rate of inflation has been anticipated, and partly the government’s discretionary response to it. In effect, one cannot accept the Aghevli-Khan hypothesis for a country without a more detailed analysis of the country’s fiscal structure and the characteristics of the inflationary situation.
Section II contains an econometric analysis of the factors influencing the rate of adjustment of revenue and expenditure to inflation. Initially, the fiscal segment of the Aghevli-Khan model for 24 developing countries is tested, and then it is demonstrated that there exists a significant degree of variability in the relative adjustment rates of total expenditure and total revenue. An attempt is made to explain this variability, and the explanation offers some support for one of the Aghevli-Khan hypotheses, viz., that the higher the mean rate of inflation, the greater the tendency for expenditure to adjust more rapidly than revenue to inflation. It also demonstrates that the Aghevli-Khan effect reflects an adjustment by decision makers to anticipated inflation; the adjustment is considerably weaker for unanticipated inflation. Also, estimates are provided of the specific adjustment properties of different categories of revenue and expenditure and of the extent to which such adjustment rates vary according to the character of the inflationary environment and according to the overall fiscal adjustment pattern. Some concluding remarks are in Section III.
Aghevli, Bijan B., and Mohsin S. Khan (1977), “Inflationary Finance and the Dynamics of Inflation: Indonesia, 1951–72,” American Economic Review, Vol. 67 (June 1977), pp. 390-403.
Aghevli, Bijan B., and Mohsin S. Khan (1978), “Government Deficits and the Inflationary Process in Developing Countries,” Staff Papers, Vol. 25 (September 1978), pp. 383-416.
Beck, Morris (1976), “The Expanding Public Sector: Some Contrary Evidence,” National Tax Journal, Vol. 29 (March 1976), pp. 15-21.
Feltenstein, Andrew, Tomás J. T. Baliño, and Ke-Young Chu, “Real Wages and Inflation: An analysis of the Argentine Experience” (unpublished, International Monetary Fund, November 21, 1978).
Heller, Peter S. (1975), “A Model of Fiscal Behavior in Developing Countries: Aid, Investment, and Taxation,” American Economic Review, Vol. 65 (June 1975), pp. 429-45.
Heller, Peter S. (1980), “Diverging Trends in the Shares of Nominal and Real Government Expenditure in GDP: Implications for Policy” (unpublished, International Monetary Fund, June 10, 1980).
Price, R. W. R., “Public Expenditure: Policy and Control,” National Institute Economic Review, No. 90 (November 1979), pp. 68-76.
Tanzi, Vito (1977), “Inflation, Lags in Collection, and the Real Value of Tax Revenue,” Staff Papers, Vol. 24 (March 1977), pp. 154-67.
Tanzi, Vito (1978), “Inflation, Real Tax Revenue, and the Case for Inflationary Finance: Theory with an Application to Argentina,” Staff Papers, Vol. 25 (September 1978), pp. 417-51.
Mr. Heller, economist in the Fiscal Analysis Division of the Fiscal Affairs Department, is a graduate of Trinity College and Harvard University.
Exceptions to this are important papers by Vito Tanzi (1977 and 1978) and Aghevli and Khan (1977) in which the consequences of alternative adjustment lags of revenue to inflation are discussed. See also Heller (1975).
Note that equation (2) could be reformulated as
If one assumes that t1 represents the desired (as opposed to the actual) tax elasticity, then if t1 is more than one, an increase in the price level will cause a greater than proportional increase in the desired R level and thus an increase in the real revenue level; an elasticity of less than unity implies that an increase in the price level yields a decline in the level of desired real tax revenues.
R. W. R. Price (1979). In the mid-1970s, New Zealand similarly decided to absorb the impact of inflation in the form of deliberate cuts in real expenditure.
A detailed discussion of the relative price effect for government services may be found in Heller (1980).
This is hardly a phenomenon restricted to developing countries. In the United States, the ceiling on the pay of senior bureaucrats has severely increased the private sector/public sector wage differential.
“Agencies should be able to offset price increases through increased productivity.”—U. S. Government (1980).
The only reasonably consistent cross-country government finance data base for developing countries is the Fund’s Government Finance Statistics Yearbook. Since this data base contains no more than seven data points per country, cross-sectional time-series analysis is required in order to obtain an adequate number of degrees of freedom. Such analysis necessitates efforts to ensure the credibility of the assumption of homogeneous behavior for the countries within the sample.
The seasonal adjustment procedure uses the X-11 variant of the Census Method II seasonal adjustment program as modified by the Federal Reserve System.
Beck (1979) has shown that there has been a marked difference in the growth rate of the GDP deflator and the CPI relative to the government consumption deflator. If, however, one takes account of both government transfers and investment, the overall deflator for total expenditure does not diverge substantially from the GDP deflator. See Heller (1980).
The algorithm used was developed by Amemiya and was extended and implemented by Berndt, Hall, and Hausman. It is described in Vol. 2 of the Fund’s RAL Statistical Routines (November 1976), pp. 9.151–9.172.
In only 4 of the countries—Brazil, Kenya, Malaysia, and Mexico—does the choice of the deflator affect the relative size of the adjustment coefficients.
The estimation procedure used is a simple error-components model, where additive dummies are used to estimate country-specific intercept terms. The model was also estimated with multiplicative dummy terms in order to take account of possible country-specific variations in the specific adjustment parameters. However, the multiplicative dummies proved significant in only a few cases. The estimations reported in Table 6 relate only to the estimations with additive dummy terms.
It is precisely for this reason that separate estimations were made with multiplicative dummy terms for country-specific variations in the adjustment parameter. As has been noted, there were only a few countries where such coefficients proved significant.