APPENDIX I Incidence Analyses and Income Redistribution
- International Monetary Fund
- Published Date:
- September 1986
Most studies of the income distributional effects of public policy fall within the category of “incidence analyses” of taxation and expenditure. They either analyze the effects of particular measures or, if appraising the overall incidence of the tax system or the net budgetary effects of the tax and expenditure system, rest on the analytical assumption that one is appraising the distributional effect of a set of measures relative to the effects of a proportional income tax increase or decrease (to be able to ignore the macro-economic effects of a particular revenue or expenditure measure).57 They are partial analyses with crude assumptions made on the effective elasticity of response of consumers or factors of production to the effects of a tax or price change. No general equilibrium effects are considered. The focus is on fiscal measures and does not readily take account of the other types of policy measures typically included in a Fund program.
Expenditure incidence estimates are particularly fraught with conceptual difficulties. Analysts are uncertain whether to examine the incidence of the expenditure (e.g., salaries paid to teachers), the monetary value of the service provided by the expenditure (e.g., the cost of the education imputed to the student), the benefit derived from the service (e.g., the benefit derived by the student from the education), or the overall incidence effects from the expenditure (taking account of any general equilibrium effects on relative prices and income) as it affects the distribution of income. How does one treat the incidence of government investment expenditure? The flow of benefits from such expenditure derives more from the present stock of government capital than from the flow of present investment expenditure.58
De Wulf (1980) has noted that incidence analyses are useful when the focus is on smaller expenditure or tax components or on changes in the tax or expenditure system that do not imply wholesale elimination of major programs or revenue sources; in such cases, it can be correctly assumed that the general economic framework will not be affected by assuming away a given expenditure (or tax) program. In the case of many of the microeconomic fiscal measures introduced as part of a Fund program (changes in particular expenditure programs or rates of taxation), such measures are sufficiently at the margin that one is perhaps not distorting reality drastically by accepting the methodological assumptions required to validate traditional incidence analyses. While the data problems alluded to above are likely to preclude specific incidence analyses in the countries recently involved in Fund programs, there exists a large body of incidence literature on which one can rely for making reasonably good inferences on the likely distributional consequences of particular policy measures.
The difficulty in inference becomes greater when a Fund program involves a large number of significant fiscal measures and, more important, and as is generally the case, the adoption of significant macroeconomic and external policy measures. The number of detailed empirical studies tracing the distributional consequences of such policies is more limited, and it is more difficult to separate out the consequences associated with particular policies or the effects that were specific to the countries that were the subject of such studies. Analytically, one is perforce required to decompose particular macroeconomic measures into their micro-economic consequences. For example, to the extent that a devaluation leads to increases in the price of importables, one could attempt to analyze its distributional effects in the same way one would assess the effects of changes in excise or sales taxes. In the case of many macroeconomic measures, such a decomposition is not easy, but there is basically little in the way of an analytical alternative.