Responsibility for planning and budgeting is often dispersed among several institutions, in a manner depending largely on history and tradition. While a variety of organizational structures can work in practice, coordination is essential. In the absence of such coordination tensions arise, such as that which often exists between the ministries of finance and planning. A common institutional question is whether the two should be merged, but the important issue is not so much whether to have one ministry or two, or a budget office separate from both, but that the two functions should be reconciled. Short-term budgets need to reflect a well-considered, longer term perspective, and medium-term plans need the accountability and relevance provided by direct links with the budget. Medium-term expenditure planning can provide the link between the two. However, few countries have managed to integrate planning and budgeting functions well. The medium-term plan is often disregarded as the budget is prepared. This may be because the detail is insufficient to provide guidance in budgeting, because the budget process is too rushed and too subject to short-term pressures to allow time to consider plan input, or because planners may have less influence than budgeters (the budget is the authoritative legal document, while the plan does not typically have the force of law). Frequent organizational changes can also diminish the influence of planners by undermining credibility. Finally, traditional stereotypes have sometimes hindered coordination. Budgeters are often depicted as being concerned with short-term expenditure control more than long-term development, while planners are depicted as being overly concerned with economic aggregates over which the government has little practical control.
A comprehensive medium-term expenditure plan has several facets: (a) it sets out the macroeconomic framework linking the growth of national income, savings, investment and the balance of payments to public expenditure and revenue; (b) it projects current spending obligations on debt service, public administration, defence, the operation and maintenance of investment, etc; some of these items - debt service on existing debt, civil service pensions, to some extent operation and maintenance - are less flexible than others; (c) it defines a phased public investment programme over several years, distinguishing between high-priority projects and those that will only be undertaken if resources are sufficient; (d) it forecasts the revenue from tax and non-tax sources and the resources needed from domestic and external borrowing and grants. Such a comprehensive plan could only be formulated by few countries, but wider progress is possible on individual components.
Although in practice the link between planning and budgeting has often been tenuous, the annual budget is ideally a one-year slice of a medium-term expenditure plan. Certain key characteristics are important for effectiveness as a policy instrument:< (a) It should be comprehensive, including all central government spending whether financed by general taxes, earmarked sources of revenue, borrowing or grants; a central review should be carried out of major investment projects of subnational levels of government and of state-owned enterprises and this might be included in the budget or a related document. In practice, extrabudgetary accounts abound in both developed and developing countries and the investments and even the borrowing of state and local government or state-owned enterprises are often excluded from central review. This may be to avoid cumbersome and often highly politicized budgetary procedures in funding essential services and may be inevitable in the short-run when budgetary processes are severely dilapidated, but longer term these processes should be improved and all revenues and expenditures consolidated into the budget. Independent funds undermine overall budgetary control, put added pressure on those items of spending that are included in the budget and exacerbate inefficiency and inconsistency in the allocation of public resources.
(b) Government objectives should be linked to traditional budget categories such as salaries, equipment, and supplies. Traditional line-item budgets used to track spending in a narrow accounting sense cannot provide an adequate picture of the extent to which public objectives are being achieved. An alternative is "programme" or "performance" budgeting which reclassify the budget to reflect objectives and programmes and also attempt to monitor government performance by relating input to outcome. Although these are hard to implement because of institutional and informational difficulties in programming and in measuring performance, they lead to improved performance measurement and auditing, broadened managerial attitudes and improved resource allocation. Thus, although "performance budgeting" as such may be impractical, its elements - categorization of spending by programme, emphasis on monitoring performance the viewing the budget as a planning and policy instrument - remain central to better public budgeting.