Journal Issue

The Planning Process of a Budget Agency: Form and Content

International Monetary Fund. External Relations Dept.
Published Date:
June 1971
  • ShareShare
Show Summary Details

David Klein

All countries have a budget, and even the least developed countries must accordingly have a budget agency, although in different countries it goes under different names and fits in different ways into the government organization. Whatever it may be called, and however it may be organized, its role in developing countries should be changing in tune with their development. At the very early stages of development its role is essentially that of an accountant; it has to take care that all spending of government agencies will be incorporated in the state budget. Quite often government departments try to exclude some expenditures from the budget—sometimes defense and development expenditures as well as other less important items. The budget agency, along with the official who plays the role of accountant general (if there is one), has to fight such attempts and has to establish the budget as an effective control tool. For the ministries this means spending no more than allowed for in the budget, particularly in terms of the manpower they are allowed to employ.

Once these basic requirements are satisfied and the procedures involved become routine, the budget agency ought gradually to assume a new role, that of planner rather than accountant. This is not its only role but it is an important one; many people indeed would how consider it a key role in a country’s economic development and management. How should it be played?

The Budget Cycle

There are two main partners in the budget cycle over the year: the budget agency and the individual government department (for example, the ministry of transportation or health). Both the agency and the ministry aim at the same target, that of formulating the next budget, but their roads are different, as illustrated in Chart 1. The agency looks at the budget as a tool of macroeconomic planning. It is the embodiment of the government’s fiscal policy for attaining national economic goals. For the ministry, on the other hand, the budget is an administrative tool reflecting the ministry’s detailed policy in the particular field for which it is responsible.


Almost always these different viewpoints lead to different budgets. The state budget, composed of all the various ministries’ proposals, is almost certain to be larger than the one prepared by the budget agency. The agency’s budget is naturally smaller, since in many economies resources are fully utilized and an attempt to execute the ministries’ proposals would only cause or accelerate inflation. Even if general unemployment exists, skilled labor or foreign exchange may also be scarce. Thus, the agency has to devise a rational way of allocating resources among the competing requests of the ministries.

Phases in the Cycle

This method of budget preparation is based on macroeconomic and microeconomic considerations. The macroeconomic considerations are derived from the national goals, which indicate priorities in allocating resources. If, for example, the goal is to reduce a, balance of payments deficit, one means of attaining it may be to contract government activities in construction and expand its help to industry. If insufficient secondary education is deemed to be an obstacle to development, then spending will be accelerated in this area rather than for welfare purposes. The macro considerations are, therefore, external to the proposal of the ministry. The micro considerations, on the other hand, are based on the merits of the case and can be applied only when the budget is discussed with the ministry.

This being so, it seems more efficient if the budget agency and the ministry do not start planning simultaneously, as they are shown to be doing in Chart 1. It may be more productive if the agency starts first and, on the basis of macroeconomic considerations, provides to the ministry a framework within which the ministry will plan its next annual budget. While the ministry is doing this in detail, the agency at the same time will prepare its general position as to the structure of the ministry’s budget for the next year.

Chart 2 reflects the first three phases in planning the budget:

  • The agency starts by accepting or determining national goals and deriving from them tentative spending ceilings for the various ministries.

  • The next phase is carried out simultaneously by the agency and the ministry, where each prepares its position with regard to the structure of spending within the given framework. The ministry’s part is much more detailed.

  • When the proposed budget is submitted, micro considerations are discussed to determine the optimal structure of the budget.


The Government’s Role

Two related issues, parts of the process that are external to the relationship between the bureau and the ministry, may be referred to now. First: who determines the national economic goals that are at the basis of the annual budgetary planning process? They should be defined by the government as the basis for long-range planning, which, where it exists, is usually the responsibility of a planning authority. In some instances, where formal long-range planning does not exist, long-run planning considerations may be taken into account by the budget bureau on the basis of which the bureau may suggest short-term goals to the government. However they may be presented to it, it is desirable that the government should adopt short-range goals so that the bureau will be able to use them to indicate spending ceilings for the ministries.

Second, at the other end of the cycle it is possible of course that the agency and the ministry, discussing the proposed budget, will not agree either on the ceiling or on the budget structure—perhaps will agree on neither. The issues are then submitted to the government for decision. The government has now either to change the goals, or the priorities to attain the given goals, or to decide against the ministry’s proposal, or to adopt some mixture of these alternatives.

In any event it is the government that starts the planning process and also ends it. The complete planning cycle is given in Chart 3.


Quantitative Terms

The discussion up to now has focused on the administrative aspects of the planning process and ignored several questions of substance. They all stem from one common feature of the planning process: it is done in quantitative terms. One cannot exaggerate the importance of this feature and the difficulties it creates. Sometimes it necessitates the use of advanced quantitative techniques developed only in the last decade or two. But there are many instances in which the quantitative approach simply means an orderly and rational way of formulating, analyzing, and solving a problem. For analytical purposes one may identify two major instances in which this quantitative approach ought to be applied rigorously.

The first major instance is where the budget agency uses a model designed to simulate the impact of alternative fiscal policies on the attainment of the national goals, to which both long-term and short-term planning are directed.

This may not be the place to describe in detail such a model, but having a general idea about its nature is necessary to evaluate its role in the planning process. Analytically, three different parts may be discerned, as given in Chart 4. The first part defines the overall objective of the planner and the constraints on attaining this goal. The goal, for example, may be maximum growth rate over the planning period and constraints on attaining it may be the availability of labor and foreign exchange.


The second part takes into account the input-output coefficients as they are observed or, if possible, as they are expected to be during the planning period. The input-output matrix is classified on an industrial basis and specifies the amount of inputs each industry or sector buys from every other industry to produce one unit of output.

The third part includes the behavioral functions and accounting identities. A prominent role in this part is given to policy variables such as tax rates, public expenditure levels, and the quantity of money. This part determines the interrelationships between the policy variables and the level and changes in consumption, investment, and the balance of payments.

This model of the economy can be compared to a model that the management of a big corporation may have: the difference is in content, not in form. The objective for a corporation will be maximizing profits subject to constraints on capacity, at least in the short run, of buildings, machines, and skilled manpower. The technological coefficients are obvious, and the behavioral functions are the demand function for the corporation’s products, reaction of competitors, and government policy.

At this stage of the agency’s planning, computers may be useful for three purposes:

  • Basic research to estimate the parameters of the model;

  • Constant updating of the model on the basis of quarterly data, if possible;

  • Simulating the impact of alternative fiscal and monetary policies on the national goals. This simulation is necessary in order to present to the government meaningful alternatives based on the trade-off between goals and means.

Input-Output Analysis of the Ministry’s Budget

The second major instance of the use of a quantitative approach occurs when the budget agency addresses itself to the structure of the ministry’s budget and analyzes the ministry’s proposals. This analysis starts with the policy objectives of the ministry which should be formulated, as far as possible, in physical terms. Some common examples are:

  • Providing elementary education for a given number of pupils;

  • Treating a given number of welfare cases;

  • Maintaining a given mileage of roads;

  • Hospitalizing an expected number of patients.

A ministry may have more than one goal, and not every goal can be quantified; the most conspicuous case is that of military spending. But in many cases it is possible, and here is where we would apply our input-output analysis.

The output is, of course, the quantified goal. Sometimes it is relatively easy to provide; for example, the number of pupils to receive elementary education. If an age distribution of the population is available the number of next year’s pupils is readily available too, although we have to make some assumptions about drop-out rates. Elsewhere the forecast may be more difficult, as in the number of patients to be hospitalized. In any case we need a model that is capable of explaining changes in output and thus of serving as a forecasting tool for the planning period.

Such a model ought to include, as one of its variables, government policy. Is every citizen eligible to be hospitalized? Is every child eligible for secondary education? Does the government maintain all the roads in the country? Including those policies that are relevant to output is important since they might have changed in the past, and it may be that a change is recommended for the future.

On the basis of the output forecast, input requirements are determined. In government spending the most important physical input is, of course, labor. It ought to be subdivided, therefore, into various categories of skill. Where possible, other important types of input should be specified in physical terms. Thus, for example, where machine input is important, the number of machine hours needed to attain the stated output should be specified.

The result is an input-output matrix, similar in concept to the one used in the context of the macroeconomic model. These input-output ratios, worked out year after year, inform us about the production function of government services. They answer the question of how many units of which input are needed to produce a unit of the relevant output. Changes in the input-output ratios reflect another component of fiscal policy. To educate a given number of pupils, do we need X classes or more? Y teachers or more? To treat the number of patients we had over some past period, how many physicians and nurses were required? Were these ratios different in different years? Is there an upward or downward trend? A downward trend may mean increasing returns to scale or higher productivity of labor alone, or a policy change. An upward trend may mean increasing inefficiency, but may also be caused by the fact that capacity does not always increase smoothly, and the average input-output ratio increases in a certain range. An upward trend may also be the result of a policy change.

Difficulties in analyzing trends in the input-output ratios arise also because of changes in the quality of output. The content of “education” may change over time, as may the concept of curing a patient or rehabilitating a delinquent. The qualitative changes usually come very slowly and thus should not pose serious difficulty over the short run. Nevertheless, the possibility should not be ignored, and where feasible the output time series should be adjusted for consistency.

Once the input requirements are determined, one can translate the physical plan into a monetary plan using unit prices of inputs. Here again there is an element of fiscal policy. Should we, or can we, keep input prices constant in real terms? Do changes in the supply of certain skills, needed for providing government services, necessitate changes in the wage structure?

The monetary translation is, of course, the ministry’s budget. The stages of planning this budget are summarized with the help of Chart 5:


  • Formulate objectives in terms of physical output and use a model to forecast this output for the planning period.

  • Analyze real input-output ratios of previous years and deduce the ratios for the coming year.

  • Apply these ratios to the forecast output to obtain the various input requirements in physical terms.

  • Express input requirements in monetary terms to attain the proposed budget.

Two consistency tests should now be carried out. One is in the nature of a feedback, examining the proposed budget to determine whether it fits under the ceiling imposed as a result of the macroeconomic considerations. If not, another iteration through the four phases may be necessary. The second test is comparing the resultant budget with that of other ministries, especially those operating in the same general area. Possible questions in this context may be:

  • Do social grants to low-income families mesh with the unemployment insurance policy; and are both meshed with the government employment policy?

  • Does government investment in agriculture mesh with its agriculture subsidy policy, and does the combination of the two produce a desirable relative share for agriculture in national product, imports, and exports according to the macro-economic plan?

  • Is the planned development of new towns and industries complemented by development of communications?

Once these consistency tests are completed the bureau is ready to negotiate the proposed budget with the ministry, and later present it to the government for approval.

The input-output analysis may be facilitated with the help of a computer; essentially the following operations can be computerized for every ministry:

  • A model for forecasting output, including policy variables;

  • Continuous updating of the input-output ratios over time, with an analysis of the determinants of change;

  • Computing the monetary translation of the input requirements and printing the resultant proposed budget.

Installing the System

It might be thought that so elaborate a system is applicable only to highly developed countries. There are two reasons why this is not so.

First, various parts of the system can be installed, and yield benefits, independently of the others. In particular, even without having a full-scale macroeconomic model, one can introduce the input-output analysis in the discussion of the budget. Some ministries are more advanced than others, and the relevant data can be collected and used, even on a partial basis.

Second, the degree of complexity of every part is not fixed and can be increased gradually. A simple system of national accounts with an economic classification of the public sector accounts is a good start. The macro-economic model does not have to include several hundred equations, based on quarterly data. A model consisting of 5 to 10 equations, based on annual data, may be extremely useful. The input-output analysis of the ministry’s budget may apply even if the most naïve model stands behind the forecast of output.

Budgets are approved by executive and legislative authorities in every country whether they are well planned or not. But budgets are also going to have a certain impact on the economy, whether it was intended or not. It seems, therefore, better to plan the budget to meet a desired set of goals. This will force the planner to formulate explicit goals and design the budget in such a way as to help in attaining them.

World Bank Occasional Papers


by Herman G. Van Der Tak and Jan De Weille

“Reappraisal of a Road Project in Iran” originated in a decision by the World Bank to evaluate the actual economic effects of a project for which the Bank made a loan to the Government of Iran in 1959. After reviewing the general transport setting and historical background of the project, the authors discuss construction costs and delays, the structure and growth of traffic, road user savings, development benefits that were expected to result, and the impact of the road on agriculture. Finally, the authors proffer a cost-benefit analysis of the various road sections in the project, the possible effect of the results of changes in the estimate of benefits, and the opportunities for alternative investment.

127 pages/S3.00 paperback

Available in English from

The Johns Hopkins Press

Baltimore, Maryland 21218, U.S.A.

Other Resources Citing This Publication