|
|
|
|
|
|
|
|
|
|
  |
 |
  |
Recurrent Cost Problems in Less Developed Countries
May 1982
>> This Is USAID >> USAID Policy Papers >> Recurrent Cost Problems in Less Developed Countries
Executive Summary III. Causes of Recurrent Cost Problems
V. Solutions to Recurrent Cost Problems
VI. Conclusions and Recommendations
59
[Download original document] IV. Recurrent Costs Analysis
There are two somewhat separable issues in recurrent cost analysis--
- Is a given country presently suffering from a recurrent cost problem?
- Is a given country likely to suffer from a recurrent cost problem in the future?
The first question is easy to answer conceptually. Are there sectors or projects where expansion of the use of variable factors of production will generate a stream of returns greater than the stream of returns associated with new capital investments? Is repair of a road more productive than building another one? Is an increase in teachers likely to have a greater impact on education than an increase in schools or is underutilization of capital, as evidenced by shortages of books, health workers, and maintenance, a rational response due to more profitable opportunities for scarce resources in other activities?
Answers to these questions require analytical skill, experience, and judgment. There is no formula which will provide a ready solution. If, however, existing projects are underutilized, while there is great difficulty in demonstrating positive net present value of new projects, then it is likely that a recurrent cost problem exists. This is presumably the case for many USAID recipients.
Inflationary finance, substantial levels of short-term borrowing, or consistent underfunding of the recurrent costs of development budgets may be signs of a substantial budgetary problem. This budgetary problem is not necessarily identical to a recurrent cost problem, since it includes both revenue and expenditure responses to a deficit. Indeed, it is possible, that a failure to fund the deficit through compensatory borrowing would be offset by reductions on the expenditure side. conversely, short-term borrowing may be a sign of the government's intention to avoid expenditure shortfalls in adjusting to a deficit. How does one identify projected deficits? What are the implications of these deficits for recurrent expenditure shortfalls in the future? The place to begin is by dividing expenditure and revenues into broad categories such as is done in Table XI below. Each of these budget categories can then be associated with critical macro variables. By examining past performance, it is possible to relate each of these budget variables to macro-variables, and then, by estimating the future value of the macro-variable one could estimate the future value of the budget variable. Adjustments must be made for changes in policy or exogenous variables such as the exchange rate.
Projecting Budget DeficitsRevenues Gross Domestic Product Direct Taxes Imports of dutiable items Import Duties Exports of dutiable items Export Duties Consumption Excise Taxes Gross Domestic Product Profits of Government Enterprises Time Trends plus Commitments Foreign Assistance Gross Domestic Product Increases in Domestic Borrowing Output of subsidized sectors Expenditures Total Debt Subsidies Government Plan Interest and Debt Amortization Gross Domestic Product Capital Expenditures Gross Domestic Product Analysis of recurrent costs of Project portfolio Transfers General Administrative including defense Recurrent Expenditures on Development Projects None of this is easy. It may, in many cases, be just as profitable to merely project current trends into the future. Many categories of both revenues and expenditures show remarkable stability. Table XII presents data on temporal variations of key budget categories for a random sample of eight low income USAID recipients. The information in the table was derived from taking the ratio of each budget variable to GDP across each of the seven years and finding the mean and standard deviation of this set of values. If we assume that the random component of each variable is normally distributed, then one can easily calculate a ninety-five percent confidence interval around the mean. For example, education expenditures in Burma have a value of 18% in the table. This means that in projecting into the future, on the basis of 1967-73 data, ninety-five percent of the time the ratio of education expenditure to GDP will be within 18% of the mean value during the 1967-73 period. There are several items of interest from the table. First, certain countries have much more volatile fiscal behavior than others. Sudan, for example, has a much larger confidence interval for most variables than the average, while Sri Lanka has a much smaller one. For countries such as Sudan a more careful and sophisticated analysis is in order. Secondly, certain expenditure sectors are much less volatile then others, education expenditures being more constant then those in agriculture. Lastly, the totals tend to be less variable then the components. Thus one is more. likely to be able to project total revenues than customs duties.
Variation in Key Budgetary Variables, as a Share of CDP,
Selected Countries, 1967-73
(95% Confidence Intervals around mean value in percentages)Current Revenues Burma Burundi Honduras Lesotho Liberia Sri Lanka Sudan Tanzania Average Income Tax 54 30 18 36 16 32 92 28 38 Sales Tax 24 28 26 -- 76 36 52 62 42 Customs Tax 32 22 10 122 16 36 80 36 44 Other Taxes 22 26 32 32 28 14 -- 30 26 Non-tax Revenue 72 22 42 15 50 26 24 36 36 Total 24 10 10 74 12 14 32 18 24 Current Expenditures Defense 12 34 62 -- 34 58 46 84 48 Agriculture 38 28 24 41 96 18 18 28 34 Education 18 16 20 18 68 4 12 20 22 Health 18 28 34 26 42 14 18 44 28 Other 18 20 24 14 22 10 44 28 22 Total 12 8 16 12 10 12 34 20 16 Foreign Grants 40 44 -- 106 34 -- -- 182 82 Capital Expenditures 22 50 70 56 58 14 30 62 46 Total Expenditures 10 10 26 18 15 10 26 32 18 Source: World Bank, World Tables, 1976. Data like those in Table XII can be used either sectorally or globally. For example, over 1967-73, government revenues in Burundi averaged 10.9% of GDP.
One can expect then, with some confidence, that in 1976, revenues would be somewhere between 9.8 and 12% of GDP (given a confidence interval of 10% of the mean). Of course, if there was a change in tax laws or tax collection procedures these estimates would be less firm. In Burundi, over 1967-73, current expenditures exceeded current revenues in only one year, and that by a mere five percent. Consequently, it is likely that total current expenditures will not be allowed to exceed 12% of GDP in 1976. 1 Therefore if an analysis of the current development portfolio suggests that expenditures well in excess of 12% of GDP are required, there is good reason to expect a recurrent cost problem in the future. At the sectoral level, there tends to be much greater variation, particularly in the agricultural sector. For Sri Lanka, where agricultural expenditures averaged 0.55 % of GDP from 1967-1972, and the confidence interval around that average was 18%, or one-tenth of one percent GDP, the funding of a major rural development program during that period that raised recurrent agricultural expenditures to one percent of GDP, would have required a substantial shift in budget priorities.
An exercise of this type represents only the beginning. In order to determine the likelihood of recurrent cost problems in the future, one needs to temper simple projections with a great deal of judgment. Is there a new development plan? Is the plan realistic, and what portion of projected investments are likely to be made and over what time period? What is the expected response of the donor community? Is there likely to be a change in government priorities? In revenue collection or tax laws? What are the best guesses of the rate of growth, over the next five years? Of inflation? What's likely to happen to the terms of trade? How will international price. changes effect the government budget? Is there going to be a change in the level and type of subsidization?
Recurrent cost analysis is not a simple procedure. The HIID-CRDE study of recurrent costs in the Sahel required enormous inputs of consultants time and energies (the seven country study on macroeconomic projections is one thousand pages long).2 The procedures outlined above can be used to indicate where a problem is likely to exist. A quantitative investigation of the size of the problem would require a much more detailed study.
1While data on current expenditures were not available, government revenues for Burundi during 1976 amounted to 11.4% of GDP, within the projected confidence interval. By 1978, however, current expenditures were 14% of GDP, well in excess of the 12% maximum.
[return to text]2CILSS/Club de Sahel, "Recurrent Costs of Development Programs in the Countries of the Sahel."
[return to text]
 |
|||
Last Updated on: July 11, 2001 |
