Polcy, Planning, and Researcb W013KING PAPERS World Development Report Office of the Vice Presid&nt Development Economks The World Bank August 1988 WPS 35 Background Paper for the 1988 World Development Report Improving the Allocation and Management of Public Spending Stephen Lister The reforms most needed to rehabilitate a developing country's planning and budgeting system are generally simple organiza- tional measures, not sophisticated analytical techniques. Mem PoLicy. Planning, and Recarh Canplex distibutes PPR Woiking Papcrs to dissdmiat the findings of work in pogrin and to ancourage the exchange of ideas among Bank naff and all others intrested in developrnen issue. These papes crry the names of the authorss, rflea only their views, and dsould be used and cited accordingly. The fundings, inierpmutazions. and conchsions are the sutbors own. They shonld not be atsuibuted to the World Bank, its Board of DiLctors, its managemma. or any ofits manbcr contries. Polac,Pnning, and Research | World Development Report| When a country's planning and budgeting process should be made public in order to system is dilapidated, an important first step is to educate public opinion about economic altema- rehabilitate basic budgetary and accounting tives, build a consensus, and spread responsibil- functions and to generate public expenditure ity for public spending choices. data that can provide a starting point for rational planning. Other basic measurce. include: * Monitoring development projects and pro- grams requires explicit information on actual ex- The budget process should first determine penditures of individual projects and the entire what resources are available and adopt aggregate program, including simple breakdowns by im- revenue and expenditure targets. Then the entire plementing agency, sector, source of finance, set of public expenditure issues should be looked and so on. For investment programs, it must be at simultaneously. If total expenditure is al- possible to relate expenditure infomiation to a lowed to be the outcome of the aggregation of financial plan for each project. The use of sectoral bids, the overal expenditure is almost standardized project profiles for all approved certain to be unsustainable. projects is an invaluable technique. - The responsibility for reconciling expendi- * The financial woes of developed countries ture bids with agreed overal targets must be de- increase the value of extemal aid and the lever- centralized. Sectoral ministries should be given age exercised by aid agencies. To minimize ceilings within which to prepare their estirdtes. friction with aid agencies, govemments should exercise careful aid management by monitoring * It is vital that recurrent and development project preparation and implementation, working expenditure programs be considered simultane- to minimize project problems, and encouraging ously in each sector, since a country's ability to dialogue between the govemment and aid sustain the recurrent costs arising from new agencies. investrnent may be doubtful. In restructuring recurrent budgets, it is important to establish This is a background paper for the 1988 World "norms ' for the level and distribution of expen- Development Report. Copies are available free diture to provide properly for key services. from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Rhoda * In the long run, the budgeting and planning Blade-Charest, room S13-060, extension 33754. The PPR Working Paper Series disseminates the findigs of work under way in the Bank's Policy, Planning, and Research Complex. An objective of the series is to get these fmdings out quickly, even if presentations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent official policy of the Bank. Copyright 0 1988 by the International Bank for Reconstruction and Developinentjhe World Bank COM2XrL Page I. INTRODUCTION 1 II. PLANNING AND BUDGESING PROCESSES 2 Planning and Budgeting - Ideology and Stereotypes 2 Ministry of Finance vs Ministry of Planning 3 Dilapidation of Planning/Budgeting Systems 5 The Importance of Budget Process and Presentation 12 Budget Methodology - the Importance of Sequence 14 Economic Analysis of Projects and the Preparation of PIPs 16 Monitoring and Managing Plans and Investment Programmes 20 :-v. RECURRENT BUDGET ISSUES 23 Budget Classification 23 "Recurrent costs arising" 24 Recurrent Expenditure "norms" and an approach to the restructuring of recurrent budgets. 25 IV. AID AGENCIES, AID COORDINATION AND TECHNICAL ASSISTANCE 31 Effects of the Increasing Relative Importance of Aid 31 Aid Management and Aid Coordination 32 Technical Assistance 35 Annex A: Main Sources Boxes 1. Deterioration of Planning and Budgeting Systems - A Typical Scenario 7 2. Institutional Contrasts between Countries 8 3. Devaluation of the Budget Process - an Example 10 4. Annual Plans 1) 5. "Objective" Ranking Criteria for Project Selection 17 6. Project Profiles 19 ACRUNYMB CG Consultative Group ERR Economic Rate of Return FY Financial Year M&E Monitoring and Evaluation KF Ministry of Finance MoH Ministry of Health MP Ministry of Planning oaM Operation and Maintenance PBS Planning and Budgeting System PEs Personal Emoluments PER Public Expenditure Review PIP Public Investment Programme SAC Structural Adjustment Credit SSA Sub-Saharan Africa WDR World Development RepoLt ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING I. INTRODUCTION 1. This paper was commissioned by the World Bank as a contribution towards the preparation of the 1988 World Development Report (WDR). I was asked to comment on the points made in a preliminary outline of the proposed Public Expenditure chapter of the WDR and to illustrate my comments with specific examples of problems and attempted solutions. 2. The paper draws mainly on my own practical experience and that of Mokoro Limited in the field of national and public expenditure planning. As will be apparent, this experience is heavily concentrated in Anglophone sub-Saharan Africa (SSA). Readers familiar with other regions will be able to judge which comments may have a wider relevance. Annex A lists some of the published documents to which I referred in preparing this paper, but it is not exhaustive, and I also referred to a large number -f unpublished World Bank documents. 3. The paper began as a series of comments on an early outline of :'.e WDR. It retains its early structure in that it consists of a number of separate but related comments on aspects of planning and budgeting in Africa. I have arranged my observations under three main headings. Section II deals generally with the planning and budgeting process, Section III concentrates on the problems associated with the planning of recurrent expenditure, and Section IV addresses aid agency involvement in and technical assistance to planning and budgeting. 4. One theme of this paper is that many of the necessary reforms to planning and budgeting systems (PBSs) are very simple (which is not to say that they are easily accomplished). There is a tendency on the part of governments and external agencies alike to look for more sophisticated solutions than are actually required. It is not sophisticated economic analytical techniques that are most lacking but simple organisational ones. However, the basic elements of an effective PBS are interdependent; in countries where the PBS is in disarray, isolated improvements to elements of the PBS may be ineffectual. 5. Opinions expLessed are, of course, my own and not necessarily shared by the World Bank or any other body. Page 1 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING II. PLANNING AND BUDGETING PROCER9Z. planning and Budgeting - Ideology and Stereotypep. 6. The WDR outline noted correctly that tnere has been a trend away from comprehensive economy-wide plans towards less ambitious forms of forecasting/planning public expenditure. This trend has not been uncontested, however, and there remain very different concepts of what really constitutes planning. These contrasting attitudes contribute to the climate in which planners and aid agencies have to operate and are worth some attention for that reason alone. Moreover, some of the distrust of the trend towards public expenditure planning is based on a confusion of technical Issues with Ideological ones. Disentangling the two would he,p to promote consensus about the general importance of planning and controlling public expenditures. 7. The basic Ideological divide is a familiar one, between market- oriented, non-interventionist preferences at one end of the spectrum and Interventionist, 'planned economy' preferences at the other. The existence of ambivalent attitudes within government Is particularly noticeable in countries that have adopted a socialist rhetoric and/or where a substantial proportion of economists have had Eastern-bloc training. The association of the capitalist World Bank and IMF with proposed reforms is enough to attach a pejorative label to them in many minds. 8. Ideological predispositions are reinforced In many cases by institutional cleavages between the Ministry of Finance and the Ministry of Planning. (As noted below, the institutional division between Finance and Planning more often that not survives attempts to combine the ministries.) There is, unfortunately, considerable validity in the stereotypes which see finance ministries as concerned with short-term expenditure control more than the long-term effectiveness of government expenditure or the overall developmental and macro-economic effects of the government budget, while Rlanning ministries are too concerned with economy-wide aggregates over which the government has little practical control. Hence the irrelevance of many of the planning documents produced by the latter and the absence of strategic direction in budgetary management by the former. Page 2 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING 9. Each minlstry is, of course, more sensitive to the other's shortcomings than to its own, and there is a corresponding tendency for planning ministries to see emphasis on public expenditure management as the abandonment of "real" planning and the triumph of the "finance" viewpoint. This Is unfortunate because the case for better public expenditure management Is not an Ideological one and Is only coincidentally associated with a trend towards market-orlented as opposed to interventienist approaches. Fiscal prudence is not an exclusively capitalist or socialist trait. Any government will be more successful In pursuing Its goals, whatever they may be, If it can make realistic assessments of the resources available to it and draw up and implement a coherent strategy for their allocation. A government which is unable to plan Its own expenditure Is unlikely to be very effective in plann!.ng for the economy as a whole. 10. The most effective pressure for better public expenditure planning has stemmed from the financial predicament in which SSA governments find themselves, but this also helps to explain the lack of enthusiasm for the trend within these govc-rnments. Ministries of Planning could see the trend as an opportunity to impose developmental views and a strategic perspective on the budgeting process. In practice, they are more likely to view the situationi as one in which power shifts to the Ministry of Finance in order to Impose a series of cuts on a whole range of government programmes. Planning Ministries are especially likely to feel marginalised by this process if their previous emphasis has been on incremental capital investment and the promotion of developuent projects, since development expenditures are likely to be the most severely cut. 11. In the short term the sidelining and alienation of the designated planners within government may make little difference to the implementation of emergency restructuring programmes. In the long run, howevez, programmes which are the product of duress rather thah conviction must be less likely to yield sustained improvements in planning and budgeting systems. Minlstry of Finance vs Ministry of Plannlng 12. A common institutional question is whether the Ministry of Finance (MF) should be merged with the Ministry of Planning (MP). The stress on public expenditure management as a fundamental aspect of planning, if not Its main component, makes this more topical. The issue is not new: there are many examples of mergers between Planning and Finance ministries and almost as many examples of a subsequent demerger. A number of countries have been through several rounds of this process. only rarely is there a genuine fusion of MF and MP; more often a single Minister simnly presides over both the pre-existing institutions. Page 3 ALLOCATION AND MANAOEMENT OF PUBLrC SPENDINt 13. The question of whether to have one ministry or two In the field of finance and planning is, In Itself, not very Important, although the issue of reconciling MF and MP roles is.lll As already noted, there is considerable validity in the stereotypes that Planning and Finance ministries perceive in each other. So long as those stereotypes persist, it Is Inevitable that the two lnstitu . will be uneasy bedfellows, and that the Issue of merger will be pd. tived as the subordination of one institution to the other. In a clime of financial stringencv, it Is also almost inevitable that the Ministry of Plenning will see mer;er as a means of strengthening Ministry of Finance's hand in pursuing expenditure reductions.121 14. It is Important therefore to conduct the debate in terms of the substance of the planning and budgeting process rather than the institutional prerogatives involved. Maybe improvements in that process can be reinforced by changes in the institutional arrangements between MF and MP, but these are not the essence of the problem. For example, a common and fundamental weakness has been a disjunction between recurrent and development budgets. Often these are drawn up by separate processes, even different sets of people, with little or no account taken of the complementarities between recurrent and development activities within a given sector. Stich a weakness may be perpetuated and exacerbated by institutional dichotomy between MF and MP, but "merger" will not necessarily resolve the problem, and resolution of the problem need not necessarily involve merger. 15. A number of current and proposed projects for assistance to planning agencies in SSA have been premised on an assumption that current problems in planning and economic management are a result of the weakness or decline of Ministries of Planning. This Is a misleading interpretation, although it is one naturally favoured by Ministries of Planning. Their preferred inference is that their own status and powers need to be restored. In practice, a more convincing explanation is that largely exogenous economic crises have simply exposed Inherent defects In planning systems and institutions. Nostalgia reveals an Inability to grasp this point. [11 It is also necessary to take account of the role of the ministry responsible for manpower/personnel, especially in view of the importance of planning and rationalisation of recurrent expenditures. (2] It Is significant that the successful merger of Finance and Planning Ministries in Botswana (a) involved, in effect, the takeover of Finance by Planning (the converse is more common) and (b) took place in a context of rapid expenditure growth. Page 4 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING 16. Weakenl:g of Ministries of Planning has undoubtedly occurred, but as an effect rasther than a cause of economic crisis. They have suffered, as have othez government agencies, from general debilitation associated with prolonged budgetary crises. In many cases too their status has been devalued by the sharp diminution In the development budget resources in their gift, at the same time as their morale has been undermined by the palpable Irrelevance of many of the planning documents to which they devote so much of their efforts. See Box 4 below for an example. DilaRidation of Planning/Budget Systems 17. Assessment of institutional performance and capacity is inevitably somewhat subjective, but it can be based on indices such as: - whether, and with what lags, government accounts appear; - the credibility of the expenditure records that are kept; - whether a recognised budgeting routine and timetable is adhered to; - whether Estimates appear on time (i.e. at or near the beginning of the FY to which they refer); - whether EstImates are published and/or adequately circulated within government;[31 - the degree of credibility of the Estimates as a means of controlling the volume and compositien of public eypenditure;[41 - the information available about and the control exerted over public service employment. [31 It is not uncommon to find spending ministries or departments that have never seen the Estimates that are supposed to govern expenditure. In some cases standards have deteriorated to the point where this is no longer considered abnormal by those affected. (41 Aid agencies are very sensitive to the quality of financial administration by governments. In most cases they choose to bypass the government institutions in which they have no confidence. This is understandable, but may exacerbate the problem of rehabilitating government capacities. (For example, the separate routing of aid funds often makes It harder for governments to manage and monltor their development programmes.) Page 5 ALLOCATION AND MANAGZN3NT OF PUBLIC SPRNDING 18. Two related points are worth making about the Ollapidation of planning/budgeting systems and institutions in BSA: (a) It is easily overlooked Just how far the deterioration has gone in some cases. (See Box 1 for a typical scenario.) This has direct implications for the nature of the remedies that are practical and the speed with which they can be effected. (b) At the same time, although the fundamental economic problems faced by SSA countries are strikingly similar, there are important differences between countries in their institutional performance and potential. (See Box 2 for an illustration.) Page 6 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING UUUU33UUU.UUUU.3UZUUUUUUUUU3U3UUUUUmUg-"SUUUa-U3UUUU3E3U---*U1t MUUU2uUs3tU Box 1t Deterioration of Planning and Budget Systems - a Typleal Scenario. The following scenario illustrates the deterioration of planning and budgeting systems that Is common: - Production of government accounts has fallen years behind. As a result budget documents, instead of comparing current estimates wilth previous actual expenditures, show only previous estimates (and not even previous years' actual releases). - Aid-funded expenditures, which form a high proportion of development expenditure, do not generally pass through the government budget. Although attempts are made to note such flows, it is believed that estimat,s of aid funding are often too high, since there is insufficient allowance for slippage and the planned phasing of earmarked expenditures over a number of years, so that there Is a danger of double counting aid flows In the aggregation of estimates from one yaar to another. - Annual project estimates are not linked to planned life-time project costs, and so there is little basis for tracking and monitoring project expenditures. - Control over the public service payroll has virtually lapsed. The government does not know the numbers on the payroll or how many of those on the payroll are "ghosts". Despite falling real wages the number of putilc sector employees continues to rise steadily. - Financial control has also deteriorated, and It Is widely acknowledged that actual uses of funds often differ from authorlsed uses. (Thus, even data on the pattern of releases of funds, when available, has to be treated with reserve.) However, budgetary control has not completely broken down. In some respects management has been quite conL-ervative, with releases of funds being more closely geared to revenue receipts than to the Estimates. This form of expenditure control is quite arbitrary: cuts fall across the board and cannot be anticipated by those affected. Although the payroll is a priority, public sector borrowing has fuelled inflation and both personnel and non-personnel expenditure has fallen in real terms. Growing numbers of public employees have added to the downward pressure on real wage levels. Officials of Planning and Finance Ministries are themselves poorly paid and supported, and there has b'< an exodus of accountants to the private sector. Page 7 ALLOCATION AND HANAOENKNT OF PUBLIC SPENDING *s3muuuumumuUuwuuumauuuuuuumuuuuumm-s-uu-uu-uu--g-uuuu----u--u-u-----u-sa- Box 2z Institutional Contrasts between Countries. To Illustrate the range of institutional environments within SSA it is worth contrasting Tanzania with Ghana. (This is not to imply that either country represents an extreme, and It should also be noted that G;,dna has made considerable progress towards rehabilitating Its planning and budgeting system.) The qualitative directions for Improvement of the planning and budgeting system (PBS) In Tanzania and Ghana are quite similar but the institutional contexts are very different. While the PBS in Ghana, at its nadir, reached almost complete dereliction, the comparable Tanzanian institutions have remained much stronger, despite also experiencing the kind of deterioration described in Box 1. In Tanzania the PBS is thoroughly lnstitutionalised - procedures are well established and their legitimacy (as distinct from theli efficacy) is not seriously questioned by participants. There Is a large cadre of Tanzanian planners and administrators, many of them very experienced in the positions that they hold. Their diligence Is particularly noteworthy against the background of deteriorating material support and declining real salaries that applies across the public service. The documents whose production Is the principal focus of activity appear regularly, in a timely fashion and are printed and circulated. The planning and budqeting process in Tanzania Is genuine and effective In the sense that actual resource allocation does depend on Its outcome. This is not to say that actual allocations match those planned: through failure to anticipate fully the dearth of resources, large and systematic discrepancies between planned and actual allocations have become usual, especially In real terms. But the pattern, If not the level, of allocations Is strongly determined by the planning/budgeting process, and this in turn means that participants are bound to take it seriously. This is not a trivial point, given that the formal planning/budgeting process In some other African countries Is largely superficial, cosmetic and Irrelevant. The relative strength of the Tanzanian PBS obviously owes much to political continuity. It provides a very different base from which to address common public expenditure problems than was available in Ghana. Page 8 ALLOCATION AND MANAGMHENT OF PUBLIC OPENDINO 19. Where dilapidation has gone furthest, it makes the task of imprjving public expenditure management more difficult, in two ways: - first, rehabilitation of basic budgetary and accounting functions becomes a prerequisite; - secondly, there is an absence of the current and historic public expenditure data that should provide the starting point for rational public expenditure planning. The latter point relates not only to whether ba.slc Information (such as recent actual expenditure data) Is available, but also to the credibility of information that emetges trnm a budgetary system that many participants have learned from experience not to take too seriously (see Box 3). Page 9 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING nox 3: Devaluation of the Budget Process - an Example Deterioration In budget processes easily becomes a vicious circle. The following example, based on a real case, illustrates how budgetary systems can be eroded and also how the consequent disillusion Itself becomes an obstacle to the system's rehabilitation. Original Situation: In the 1970s standard budget guidelines were Issued by the Ministry of Planning and Finance. In the case of Health, these were f6llowed by guidelines from the Health Planning Unit to provincial directors and divisional heads. The format of the Budget for the Ministry of Health was standardised, meaning that all Budget preparation forms could be pre-printed and each year the Planning Unit issued a new edition of a complete set of forms for preparing the annual Estimates. Budget preparation began early and followed a well-formulated timetable. There were established procedures for consultation between the different levels of the health system and the national ministry provided all concerned with a list of standard costs, revised annually, for use In the detailed computation of budgets. The list contained standard costs for feeding patients and trainees, for travel and miscellaneous allovances, vehicle purchases and running costs, and for a variety of standard office and health facilities equipment. Those compiling the Estimates had details of expenditure so far during the current year. Initial preparation of the Estimates took place over a two-month period, during which Planning Unit staff would be expected to visit all provinces; this was followed by a month In which the Planning Unit appraised all submissions and prepared the aggregated Ministry of Health Budget. Submission of the draft to the Ministry of Finance then took place three months before the actual start of the financial year In question. Present gituation: Things are now very different. In principle nothing has changed, but the discipline of the timetable has been lost, and the data essential for preparation and evaluation of the budget are no longer available. In the last budget cycle, the Ministry of Finance circular did not reach Ministries until after the deadline for Estimates to be submitted; it contained no resource guidelines but Introduced changes in the procedure for reckoning certain items that nullified preparatory work that the Ministry of Health had taken on Its own lnitiative. Moreover, because statistics were generally lacking and no unit costs were available, provincial directors and heads of departments set about the budget preparation with no sense of purpose. Knowing what had happened in previous years they failed to adhere to the timetable established by the Planning Unit, which In the end had to prepare a consolidated Ministry Budget In the apace of a single weekend without much regard to the few departmental and provincial submissions that had been received. No evaluation was possible and consequently the MoH was unable to provide the type of Justificatlon of Individual Items and sub-items at provincial Page 10 ALLOCATI9N AND MANAGEMENT OF PUBLIC SPENDING level that the MF sought. The budget that eventually emerged bore little resemblance to what had been submitted, and as the year proceeded the Ministry of Finance released only 50S of budgeted provision for non- personnel ltems. No guidance was given as to how the cut in funds should be allocated between Items and sub-items within the budget and expenditure on some Items, such as the feeding of patients, exceeded both the releases and the budgeted provision. Consegqences: Not surprisingly, beleaguered health officials have become rather cynical. They know that the government cannot fall to allocate resources to the health services. They know that whatever they submit as a budget request will not be reflected In the ultimate approval, which will arrive late and be at variance with the funds subsequently released for expenditure. They also know that whatever they submit, or even whether or not they submit a draft, will not be assessed by the Hinistry of Health which, anyway, exercises little authority or sanction. A cycle of mutually reinforcing scepticlsm prevents the budgetary exercise from being taken seriously by any of the participants, and this in Itself can stymie attempts to rehabilitate the system. … -================ =…=====_========w=============== = =P==age =====… = Page 11 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING The Importance of Budget Process and Presentation 20. Managing public finances In SSA has become typically an exercise in crisis control. In a crisis atmosphere, horizons shorten and there is a tendency also to narrow the group of effective decision takers. In the long run, however, better public expenditure management requires a more public process. Indeed, the process of planning may be at least as valuable as the budget/planning documents that are Its visible product. The opportunities for educating opinion about economic alternatives, for building consensus, and for spreading responsibilities for public expen#lture choices more widely are lost If the process Is always conducted in secrecy and in haste. 21. The task of rehabilitating budgetary systems may be long and difficult, especially where current dilapidation is particularly extensive. Tt may be some time, for example, before government accounts can be restored to the point of providing a reliable benchmark for each year's budget exercise. At the same time some Improvements are available that are relatively simple and straightforward to Implement. The way information is presented to (ostensible) decision makers Is often a case in point. 22. Thus, for example, budgets (and Annual Plans - see Box 4) that are presented to parliaments and published are often remarkably uninformative documents. A simple shortcoming (though it may on occasion be less an oversight than a reflection of the state of monitoring and accounting records) Is the failure to show previous years' actuals and/or future yeaLs' projections alongside the current year's estimates. This is particularly common where projects In the development budget are concerned. Estimates commonly give a so-called "one-year time-slice" of a proJect without indicating its fund!'.g requirements in subsequent years or revealing its previous expenditures (or if these are shown, the period covered may be unspecified or Inconsistent among projects). Very often projects exist for yeaxs without more description in the budget documents than a short title. rn such circumstances, parliamentary discussion of the budget is almost bound to be perfunctory and superficial. The less information Is provided In the budget documents, the less effective they are as a means of controlling and directing actual expenditure. Page 12 ALLOCATION AND DWANOSENT OF PUBLIC 8PENDING Box J: Annual Plans The concept of an Annual Plan that translates a medium-term plan Into a practical annual programme is superficially attractive, but does not seem to work well in practice. Typically, the preparation of the Annntal Plan suffers from the absence of a realistic medium-term financial framework. In principle, the Five Year Plan (together with a longer term Perspective Plan if there Is one) provides the broad "resource envelope" to which the annual exercise hqs to conform. Most often, the economy is so far off its anticipated cour , that these documents are no longer relevant. The Annual Plan Is usually prepared as part of the same exercise as the annual budget and there Is therefore no question of an annual budget narrowly concerned with government expenditure being set within the context of a prior, more general, Annual Plan. The detailed tables of the Annual Plan add little to those provided in the budget documents, and nor does the Annual Plan provide projections that extend beyond the budget's horizon. A lot of administrative effort goes Into the preparation of the Annual Plan but to very little practical purpose. To a large extent the document becomes an end in itself. (It is what the Planning Ministry does, while the Budget Is what the Finance Ministry does.) A more relevant approach requires: (a) that medium term Plans be more closely geared to the planning of public expenditure (recurrent as well as capital) within a realistic resource envelope; (b) that the annual budget Itself be seen as the principal mechanism through which the Plan Is effected and, If necessary, adjusted to new circumstances; (c) that the planning horizon for both Plan and budget should always extend two to three years beyond the current fiscal year. Page 13 ALLOCATION AND NANAGENENT OF PUBLIC SPENDING 23. It Is worth paying attention too to the way in which budgetary aggregates are summarised, since this can affect the terms of debate on budgetary policy. A key concern should be the size of the budget deficit (defined as the gap between expenditure and domestic revenue) and there should be special attention to how the deficit will be financed (by external aid and by foreign and domestic borrowing). This essential information is often obscured by arbitrary classifications of revenue (e.g. by treating domestic borrowing as revenue on a par with tax receipts) and, most seriously, by a failure to present reasonable forward projections of the budget. It Is not just a question of failing to present this information publicly: all too often the planners themselves do not compile It or see the need for it.l51 Budget Hethodology - tile Importance of Seguence 24. In principle, the determination of the budget Involves the solution of a set of simultaneous equations that reconcile revenues with expenditures and so forth. In practice, the process of formulating a budget Is as much political as it is technical or economic, and it is bound to be iterative. Whether the outcome of the process represents a sustainable budget path and whether the allocation of resources is optimal, subject to the general constraints on overall resource availability and other specific constraints that may apply, may well depend on the sequence the process follows. 25. A key aspect of the approach advocated in the present paper Is that it Involves looking at the entire set of public expenditure issues simultaneously. A system that allowo major expenditure decisions to be taken separately and at different times Is bound to generate an upward pressure on expenditures and Is very unlikely to achieve a satisfactory overall pattern of expenditure.C61 If total expenditure is not explicitly targeted, but allowed to be the outcome of a process of aggregation of sectoral bids, the aggregate expenditure target that emerges Is virtually 151 A similar sloppiness Is often shown in studies of public expenditure when the allocation of resources among sectors is analysed without making a distinction between discretionary and non-discretionary elements of expenditure. The biggest non-dlscretionary Item Is usually debt-service, and analyses of expenditure that do not separate out public debt can be particularly misleading. A sector's share of discretionary expenditure may have held up even when its share of total expenditure has fallen sharply. [61 But viewing the entire set of public expenditure options simultaneously means that there has to be direct competition among projects/programmes that may be at different stages of definitlon and appraisal. This creates methodological problems that are discussed later in this paper. Page 14 ALLOCATION AND MANAOEMENT OF PUBLIC SPENDINO certain to be unsustainable. The reconciliation of expenditure with resources actually available then takes place piecemeal and often arbitrarily, through more or less selective cuts in expenditure, or withheld provision, or non-compensation for Inflation, or erosion of public-sector salaries. Responsibility for post hoc reconciliation of actual expenditure with resources available usually falls to the Ministry of Finance, which then attracts the odium that goes with It, together with justified criticism for the inefficiency of such methods of expenditure control. 26. It Is crucial, therefore, that the budgetary exercise should bgisn with the assessment of overall resource availability and the adoption of aggregate revenue and expenditure targets. The initial assessment and proposal of targets should Involve collaboration between the main economic agencies of government (noably the Ministries of Planning and Finance and the central bank). The p.,posed targets should then be explained and accepted at tne political level before the detailed elaboration of the budget proceed . 27. The next requirement Is that responsibility for reconciling expenditure bids with the agreed overall targets must be decentrallsed. It is neither effective nor efficient for this role to be left entirely to the central MF/MP. The overall target needs to be disaggregated and sectoral ministries given cellings within which to prepare their estimates. This forces sectoral agencies to reveal priorities and Is often resisted for that reason: ministry headquarters would rather simply aggregate bids and force fF/MP to do any cutting, even though the sector ministry is usually in a better position to judge which cuts are least undesirable. 28. The process of disaggregation of the overall targets among sectors tends to make theoretical economists uncomfortable. In theory, one should chose between all alternative uses of expenditure at the margin, and the methodological basis for segmenting the allocative mechanism is shaky. Nevertheless, the sectoral disaggregation of targets is not In practice as difficult as It might seem In prospect, particularly since the exercise never starts from a blank slate, but begins with the inertia of an existing pattern of expenditure.t71 The best way to test the appropriateness of sectoral ceilings is to draft a budget that complies with them. It is then possible to judge whether there are Imbalances that should be corrected by transferring resources from one sector to another. [71 See also the sections of this paper dealing with project selection and the search for recurrent expenditure "norms". Page 15 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING Economic Analysis of Prolects and the Preparation of PIPs 29. Even leaving aside the relationship between an investment programme and recurrent spending, which is considered In a later section of this paper, the preparation of a Public Investment Programme (PIP) Is considerably less straightforward than the preparation and appralsal of the individual projects of which it is comprised. Among both aid agencies and governments there Is a tendency to assume that the task of PIP preparation consists simply of the simultaneous analysis of a lot of projects, using familiar proJect analysis techniques. 30. A little reflection shows that Lhis cannot be the case, for the following reasons: - The PIP his to take Into account the stock of ongoing projects. Typically, these are numerous and varied, and documentation of them, especially o'- ones that are not aid-funded, Is patcly. Many are more properly recurrent than Investment activities; others are, in principle, Investment projects, but are kept alive at a level of funding that does not move them appreciably nearer to completion. To take existing projects as given and consider only the merits of potential new projects Is not an option because: (a) most often, available resources are Inadequate for the completion of all ongoing projects; (b) the necessary information as to what is involved in the completion of ongoing projects Is not readily available, but has to be sought as part of the exercise; and Xc) the continuation of many existing projects Is not defensible even in spite of the sunk costs argument. - There is the widely recognised, but recalcitrant problem of finding appropriate techniques for comparing dissimilar projects that compete for funds (education vs infrastructure, for example). (See Box 5.) - There Is the less appreciated, but equally awkward problem of asynchronism between projects. The project cycle is a familiar concept. Projects pass through various stages between identification and completion. All but very small and simple projects are likely to pass through successive stages of progressively more rigorous planning and appraisal before Implementation. A full-scale cost-benefit analysis, or equivalent analytical technique, is only applicable when the project has been worked out in detail. Preparing a PIP requires considering all the projects that represent potential claims on resources during the planning period, but they will not all be at the same stage of the Individual project cycle. Page 16 ALLOCATION AND MANA(BMENT OF PUBLIC SPENDINO Box 5: "Obiective" Ranking Criteria for Prolect Selection There Is a recurring tendency for economists/planners to try to escape from the element of subJ%ctivity In deciding between competing claims on expenditure. Devising ranking systems whereby projects are awarded points according to their perceived contribution to various objectives Is one ploy. But it rather gives the game away when the devisers of the system have to reject the results as "wrong". One example was reported in roughly the following terms: In this activity the project has made a substantial and unambiguous contrilbution. In collaboration with senior (Ministry of Planning] staff a set of 18 criteria were established. The development of the computer network allowed all (PIP) projects to be ranked in order of priority in order to prepare a core programme. At the same time, It was noted that mechanistic application of criteria produced some maverick results. The criteria were heavily weighted in favour of directly productive activities and the first round of ranking virtually eliminated projects in the social services sectors. It was therefore necessary to depart from a strict ranking based on economic criteria by ore-selectlng the sectoral composition of the investment programme. (Emphasis added.) In view of the time constraint, this was no doubt Inevitable, but the project team should review the bases of the criteria in order to allow for linkages between the outputs of the social services projects and economic productivity. Why look for a key to the back door when the front door is open? Page 17 ALLOCATION AND HANAGEHENT OF PUBLIC SPENDING 31. For those who start from a paradigm in which rigorous analysis of *n individual project leads to a yes/no decision on its implementation, according to whether It meets some objective criterion, such as a specified ERR, the puzzle of how to compile a defensible PIP Is Indeed perplexing. It need not be. The practical approach Is: (a) to segment the problem by dividing the available resource parameters among sectors, so as to confine direct trade-offs between projects Initially to ones that are broadly similar; (b) to see the process as a matter of screening, In which the Initial acceptance of projects Is conditional on their meeting satisfactory appraisal criteria at the appropriate stage In their life-cycle; (c) to recognise that the proposed phasing of projects Is as Important an issue as outright acceptance or rejection. 32. Within this framework, the importance of maintaining a rolling multi-year perspective becomes clear. Given lead times and implementation periods for projects, freedom of manoeuvre is much more limited in the short term than over several years. 33. The preparation of standardised project profiles (see Box 6) Is a simple but Invaluable technique in preparing and managing a PIP. Page 18 ALLOCATION AND MANAGEMENT or 'UBLIC SPENDING Box 6; Prolect Profiles The documentation that projects ultimately require varies a great deal, depending on the nature, size and complexity of the project, the requirements of the funding agency and so forth. But decisions need to be taken at many stages before full documentation and implementation of a project. These include preliminary judgments whether the project merits the effort of fuller write-up and appraisal, what Its costs and phasing are likely to be, and which potential source of funds may be appropriate. The discipline of preparing simple standardised project profiles at an early stage Is invaluable. (See Botswana's National Development Plan and Planning Officers' Manual.) Both the words and numbers of a profile are Important. It Is Idle to manipulate figures that cannot be related to a basic project description, while It is also vital at an early stage to estimate the likely costs and potential benefits of a proposed project, even though these estimates will require subsequent refinement. The essential components of a project profile are: (a) a concise description of project content, objectives and anticipated effect; (b) an unambiguous Identifying title and project number; (c) an estimate of project costs from commencement to completion; (d) a proposed annual phasing of costs; (e) an estimate of the recurrent costs arising from the project. The profiles need to be on the record and compiled according to the same currency and price conventions. In updating Investment programmes annually, the summary figures for all profiles need to be revised, to reflect both general changes (in price levels, for example) and specific changes In the cost of phasing of Individual projects. ProJect descriptions should need less frequernt amendment but provide an important check on whether the orlginal prolT.nt concept Is being adhered to. Crucially, project profiles and development budget categories must be made consistent with each other, so that the profiles provide the bridge between the medium term plan/investment programme and the annual budget. Page 19 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDINO Monltoring and Managing Plans and Investment Programmes 34. Monitoring of development projects Is rightly a matter of concern. Thls is appropriate both to keep track of projects' implementation and to discover whether they have the effects intended. But the factors that are most Important In the monitoring of a plan/investment programme as a whole are not necessarily the same as for the individual projects of which it is comprised. Different types of information may be required and by different sets of people, and for the programme as a whole there Is a premium on standardised project information that can be aggregated. Very often elementary Information that might be regarded almost as trivial by someone responsible for the M&E of a particular project would actually be extremely useful if consistently available for all projects. When it comes to the aggregate monitoring of a PIP, there is a clear order of priority for the Information that Is needed, and sophisticated but partial information is much less useful than basic but comprehensive data. 35. The first requirement Is Information that should be, but often is not, available from government accounting records concerning actual expenditures by Individual projects and the programme as a whole. Simple breakdowns by implementing agency, sector, source of finance etc. are crucial, as is their prompt availability. Of course financial information of this sort may conceal as much as It reveals. Physical progress may not be commensurate with expenditure, a project whose financial implementation Is going according to plan may nevertheless be totally misconceived and so forth. Financial monitoring Is not enough, but It Is a vital starting point and first priority. 36. As far as Investment programmes are concerned, a second elementary requirement is that it must be possible to relate expenditure information to a financial plan for each project. The development and maintenance of standardised project profiles for all approved projects is an Invaluable technique (see Box 6). Initially, at least, monitoring data should follow the classifications used In project administration, even If these do not correspond precisely to economic or other relevant categories. Page 20 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING 37. Basic monitoring Is thus strongly related to the rehabilitation of government accounting systems, but this may be a long process and substitutes may have to be sought in the Interim. In setting up a system to monitor the investment programme Independently, the emphasis should be on making basic Information rapidly available. It Is Important to pay attention to the relevant lncentive5. Firstly, Information should be collected by those who are intended to use It, since the "customer" for Information has the greatest Interest In It and should be best able to Identify priorities In Its collection and analysis. Secondly, there need to be appropriate Incentives for the providers of the Information. (A simple incentive Is to make releases of funds conditional on the prior provision of information about previous expenditures.) conversely, where the supplier of information does not have a very direct Interest in the exercise, his task can at least be made less onerous by accepting Information In the form that comes most easily to him. 38. Often the pressure for better monitoring comes from aid agencies. This can be useful. For example, the external Incentive provided by the desire to present progress reports to a CG may help to counterbalance a government's natural preoccupation with short term crisis management. On the other hand, it Is Important that the government should be persuaded of the value of the monitoring exercises proposed and able eventually to sustain them without external assistance.181 39. Most Investment programmes are poorly managed. This Is Indicated, Inter alia, by the perpetuation of projects that fall to progress and by the difficulties commonly experienced In rapidly utilising available aid. Improvements depend both on better current monitoring of the programme and on better forward programming. [8] " ... monitoring and evaluation has been largely donor-led, whether at project or national level. The Information generated and the time-frames for its production have almost always been defined by donors. There has consequently been little useful feedback Into the national budgeting and planning processes. This predominance of donor Interest In monitoring and evaluation Is echoed In the frequent -- almost universal -- complaint that government monitoring and evaluation has been Ineffectual on account of a lack of political commitment by governments. Without that commitment and a domestic demand for monitor.ng and evaluation, It seems likely that past failures and inadequacies will only be re-created." " ... government monltorlng and evaluation systems established in some SSA countries have been overly complex, bearing the hallmark of external design and lnattention to local personnel and other constraints. one common result has been that monitoring and evaluation systems set up on donor instigation and principally by means of technical assistance have rarely survived the withdrawal of such assistance." (Both quotations are from Commander, 1987.) Page 21 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDINO 40. A large part of the difficulty of managing an investment programme well arises from uncertainty about (and fluctuations In) the level of resources available, and from the fact that arbitrary expenditure control is usually easier to impose than selective restriction. The concept of a *core" programme Is an attempt to address both these difficulties. The aim is to ensure that the most important projects are not Interrupted by delays in releasing funds on account of general financial stringency. Since funds available can never be forecast with certainty, It will be possible to protect the core only If (a) the total PIP Is not too optimistic in its resource assumptions and (b) the core does not form too large a proportion of the PIP. 41. Furthermore, core projects can be protected only If there are adequate budgetary mechanisms for Identifying them and for releasing funds in a discriminating way. Funds for core projects should be released automatically at the beginning of each accounting period, whereas funds for non-core projects should only be released selectively and after a further review of the financial position. 42. However, non-core projects that require "lumpy" expenditure (e.g. a single construction contract) should also not be interrupted once begun, and may therefore have to be treated in the same fashion as core projects once they have been allowed to commence. The corollary Is that non-core projects should not be allowed to start unless sufficient funding to complete them Is assured. In turn, since implementation of most projects extends over a number of years, this emphasises the Importance of basing budget and PIP on a rolling, multi-year perspective of resource availability. Put another way, designation of core programmes may be a useful expedient, but it should become much less necessary as expenditure planning Is strengthened. 43. An important aspect of PIP management Is to ensure that funds are used for their designated purposes. Control systems can help, but cannot solve the problem In the absence of political and bureaucratic commitment.[91 [91 One development expenditure control unit found that Its own funding was the first to be diverted - by the Ministry of Planning itself! Page 22 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING III. RECURRENT BUDGET ISSUES Budget Classification 44. Criticism of the traditional distinction between recurrent and development (or capltal) budgets Is commonly overdone. The divlsion often works poorly In practice, but there Is more to be gained by Improving than by abolishing It. The distinction between recurrent and capital expenditure corresponds, albeit roughly, to the distinction In economics between cornumption and Investment. It encourages attention to the problem of financing the recurrent costs arising from development expenditures. The biggest drawback of the recurrent/development demarcation In practice Is that It can allow the two budgets to be drawn up separately, by different processes and even by different groups of pecple. Particularly now that the ability of governments to sustain the recurrent costs arising from new Investment Is so often doubtful, It Is vital that recurrent and development expenditure programmes are considered jointly within each sector. But equally, within each sector as well as overall, it Is important to analyse proposed programmes into their recurrent and capital components. 45. In many cases, existing demarcations between recurrent and development budgets cannot be taken at face value: (a) So-called projects funded under the development budget are often locally-funded recurrent activities. (b) The same applies, but In a less straightforward way, to many aid- funded projects and project components; aid agencies have recognised that It Is pointless to finance new Investments In sectors where recurrent activities are grossly underfunded, but they are constrained to dress up their aid as development support. There Is a spectrum ranging from rehabilitation projects which are, In a sense, genuine investments, but ones which would not have been necessary but for the neglect of O&M, to the funding of equipment and materials that constitutes more or less direct recurrent budget support.[101 (c) It is also rightly polnted out that some recurrent activities (health and education provision for example) can represent very important forms of Investment. [101 Also, of course, much aid may fall to be reflected In the government budget at all. Page 23 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDINO 46. The recurrent/development categorisation Is not the only one that Is potentially very helpful for the analysis and preparation of public expenditure programmes. Often, the structure of the budget, particularly the recurrent budget, does not easily yield elementary breakdowns of expenditure which relate to crucial policy issues that arise In expenditure programming, such as the breakdown between expenditures on primary and secondary health care, between different educational institutions, or between different districts or regions. (A classification that easily yields one type of breakdown - e.g. between salaries and other charges - may be correspondincly opaque In other respects.) "RSecurrent costs arlsing" 47. The Importance of anticipating the recurrent costs (and specific O&M requirements) arising from new development projects Is now a commonplace. However, largely because of past neglect of this point, there is hardly anywhere where monitoring the recurrent costs associated with new projects Is more than a small contribution towards the task of assessing the true level of future claims on the recurrent budget. The issue has macro dimensions that cannot be adequately addressed by a micro (project) approach. 48. In the first place, public finances are often In such dlsarray that It Is futile to look to precedents to establish the likely recurrent funding requirements of a new project. Reliable historic figures on actual recurrent costs associated with previous projects are rarely available, and, in a context of chronic underfunding of the recurrent budget, actual experience Is likely to understate true costs. Secondly, preparing an accurate estimate of a proJect's recurrent funding requirements Is Itself no guarantee that they will actually be met (though it may be a valuable aid to project selection). 49. The macro questions that cannot be reached through proJect analysis alone Include: (a) The development Implications of the develogment budaet: development pro,ects, particularly locally-financed ones, are often grossly underfunded and their cost to complete (which may well not be justifiable) is commonly a matter of conjecture. (b) The develonment implicatlons of the recurrent budget: a substantial backlog of maintenance has often built up so that rehabilitation projects may make a substantial prior claim on development funds available. Also, built-in growth of services (e.g. primary education) may have implications for development expenditures (e.g. school construction). Page 2. ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING (c) The recurrent implications of the recurrent budget: this has two components - the expenditure claims Implicit In current underfunding of recurrent services, and predictable growth In those commitments that may not carry corresponding project costs to which the recurrent arising could be attributed. This aspect may be particularly recalcitrant, as discussed in the paragraphs that follow. Recurrent expenditure wnorms" and an approach to the restructuring of recurrent budgets(lll 50. Questions about the appropriate level and pattern of recurrent expenditures are Increasingly being raised In parallel with World Bank PER/PIP exercises. In one exercise related to the framing of a Structural Adjustment Credit, the government was asked to "establish expenditure norms for the main [non-personnell Items of expenditure for the Ministries of Agriculture, Health and Education" and to establish procedures for Implementing the norms. Some of the lessons from the exercise are of general relevance. 51. Thete was considerable confusion as to what this exercise was to comprise, although the background situation was clear enough. During years of financial stringency (a) the real value of resources available to the Health, Education and Agriculture sectors of Government, among others, had fallen and (b) the proportion of those resources devoted to Personal Emoluments (PEs) had increased.[121 One of the aims of the recovery programme had therefore to be to reverse both these trends - i.e. to Increase the levels of resources allocated to these key sectors and at the same time to correct imbalances in resource allocation within each sector. [111 The system advocated in this section reflects that successfully developed In Botswana over many years. 1121 It is not always the case that declining real resources are accompanied by an increase in the share of resources devoted to PEs. The sight of personnel with virtually no material support suggests this conclusion, but It must be remembered that there have often also been drastic declines In real salary levels. The primary problem is the poverty of governments; the maldistrlbution of residual resources is a secondary effect. Page 25 ALLOCATION AND NANAUEHENT OF PUBLIC SPENDINO 52. Such Imbalances are often quite tangible: teachers without text books, extension workers without transport, hospitals without drugs, and so on. But, although the problem is highly visible, years of stringency, ad hoc budgeting and arbitrary expenditure control, combined with poor expenditure records, make it difficult to know what the level and distribution of expenditure now ought to be in order to provide properly for key services. Hence, a search for "norms" to guide the allocation of resources as recovery commences. There are potentially "norms" at several different levels, all of which may be relevant to aspects of the problem. 53. Thus, for example, nozms may be sought at the level of effllency wilthin operatlons. If you have a car, you also need the fuel etc. to be able to run It, and a norm here might simply be a rule of thumb based on experience as to how much ought to be provided for the running cost of each vehicle. Such an approach can be very valuable, both In framing and In evaluating budget estimates, but it also has Its limits. While It may be straightforward to establish a general rule about the true running costs of a vehicle, this still leaves open the question of who actually needs vehicles and how many of what sort. Similarly, rules of thumb are often used to relate provision for the maintenance of a building to Its original construction cost, but such rules do not address the question of whether existing buildings are adequate or appropriate for their purposes. 54. A second level at which norms may be sought concerns the balance between programmes or activities within a sector. For example guidelines could be adopted about the balance between primary and secondary services In both the health and education sectors, or about, say, the balance between agricultural extension and agricultural research. 55. There is a third, more aggregative, level at which one could ask, for example, what would be the level of resources reguired to run entire sector proarammes 2roDerly (how much would it cost to fully fund the existing health service?) Or, going further, attempts could be made to relate flnancing levels for services to external criteria, such as relating services required to the size of the target population (universal primary education, Health for All, etc). 56. Finance Ministries, and especially their staff responsible for administering the recurrent budget, dre likely to Interpret norms rather narrowly. Standardised unit costs are often a familiar concept, though they may have fallen out of use in conditions of rampant Inflation and budgetary chaos. Beyond that, the search for norms may be interpreted mechanistically in terms of rather arbitrary ratios between different subtotals of the budget. Page 26 ALLOCA:'ON AND HANAGEHENT OF PUBLIC SPENDING 57. Two Important points emerged very early on In the exercise: (a) Proper recording and control of expenditure Is a sine qua non: without Improvements In this direction norms of any kind will not be meaningful or enforceable. (b) Thore Is considerable scope for the Introduction (or re-introduction) of standardised unit costs (unit cost norms) as a way of Instilling discipline and ensuring consistency In the compilation of the Estimates. 58. Concerning unit cost norms the following points should be noted: (a) Few are available "off the shelf". Determining appropriate unit cost norms is not always simple (local prices change rapidly as the currency depreciates and historic data, even If available, may not be a reliable guide to true costs), but in time a useful body of such norms could be built up. (b) To be of most use, such norms need to be available at the commencement of each year's budget cycle so that they can be used when the estimates are first drafted. Their availability later, as the Estimates are reviewed and finalised, is also useful but not an adequate substitute. (c) Norms of this kind should not be applied too rigidly, or they may create problems of the type they are intended to help to remove. For example, similar vehicles may quite legitimately have different operating costs when used for different purposes or in different regions. There needs to be an element of flexibility to allow for this. Such flexibility could enter both in the Initial preparatlon of the Estimates and In the scope for subsequent virement. 59. Finally, useful as unit cost norms might be, it would be a delusion to think that the entire recurrent budget could ever be satisfactorily built up from a series of unit cost norms.1131 Broader approaches (within which unit cost norms have a role) are likely to be of more value In helping to restore a satisfactory pattern and an adequate level of recurrent expenditure. The paragraphs which follow set out a proposed budget methodology. [131 The search for mechanistic ways of drawing up a budget has parallels with the yearning for "objective" project selection criteria (see Box 5). Page 27 ALLOCATION AND MANAOEKENT OF PUBLIC SPENDINO 60. How budget guldelines could work. It has to be accepted that the central planning/finance ministry does not always know best, especially when it comes to the details of the effective management of specialised services. For example, it does not make sense for MF/MP to try to stipulate an exact allocation of hospital expenditures as between different items: such a decision should be left primarily to the judgement of professionals in the Ministry of Health, and within the Ministry of Health as well decisions about the allocation of resources may need to be decentralised to varying degrees. The problem is to allow enough decentralisation in the interests of efficiency while retaining enough control to ensure that expendlture patterns correspond to national priorities (and that the overall level of expenditure matches appropriate macroeconomic targets). 61. A practical way of doinq this Is to Issue. at the beginning of each budget cycle, expenditure guidelines on what total resources will be available and how these resources should be distributed between broad categories of activity. For example, the Ministry of Education might be told to budget for an increase in real terms of 10%, but to hold higher education spending constant while increasing primary education expenditures by 20% Within these guidelines, and probably subject to restrictions on recruitment and PEs, It would be left to the Ministry to determine the best allocation of resources within each activity, although their proposals would be subject to review by MF/MP in the normal way. Guidelines could also be Issued on the allocation of expenditures as between regions, etc. It would be necessary to restrict Ministries' freedom to vire funds from one activity/region to another, but virement within activities/regions need not be so tightly controlled. 62. It must be emphasised that, in order for such guidelines to be used as a way of planning expenditures, the framework of Government budgeting must permit the analysis of expenditures by programme. For example, in Health one would want to be able to categorise votes into, say, rural/urban, primary/secondary and by region, etc. Ensuring that this can be done Is more important as a first step than the derivation of the guidelines themselves. 63. The application of guidelines of this sort would have to be Iterative: I.e. there would have to be some scope for adjusting the guidelines in the light of Estimates submissions If submitting agencies were able to present a compelling case for doing so. At the same time It is important that agencies do not avoid hard choices between alternative uses of funds simply by always appealing to MF/MP for more and shifting to MF/MP the responsibility for any cuts or omissions that may be needed. If, therefore, a ministry Is convinced that one of its activities needs more funds than Its guideline allows, the ministry should be required to propose an equivalent saving from one of its other activities. Page 28 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING 64. Also, such guidelines would have to aim at a gradual change In the composition cf expenditure - the "Ideal" pattern could not be achieved in one step but would have to be sought over a period of years. Guidelines could only work at all In a context where funds actually released correspond closely to the provision In the Estimates, and It would be essential to have timely expenditure records from which compliance with the guidelines could be monitored. In many cases, the decentrallsation that Is Implied may be hampered by the fragmentation of sectoral responsibilities among a number of ministrles. The approach also underlines the Importance of the development of economic planning skills within sectoral ministries. (Again the development of effective minlsterial planning units may be hampered by the Institutional fragmentation of sectors.) 65. Derivlng Expenditure Guldelines. There remains the question of how to set such guidelines. The problem is probably not as difficult In practice as It may seem in principle. It would be very difficult to determine a single "correct" level of resources to allocate to any activity, but all that is practically required is a sense of the qualitative directions for change. Thus, If It is apparent that, say, primary health care has been underfunded, expenditure guidelines should be set that allow a faster rate of growth for primary health care expenditures than for others; this should be continued (probably over several years) until the Imbalance ceases to be obvious, but, since the process takes time, it Is not necessary to know at the outset exactly what Its outcome will be - guidelines can be continually revised In the light of experience.l141 66. All the same, It would be Interesting to know, at least approxlmdtely, what would be the cost of operating Government's existing services properly. Indeed, thils Is a crucial piece of Information If Government is to plan Its public expenditures rationally over the medium term, and it has a direct bearing on the balance that should be struck between development and recurrent expenditures. There Is little point In allocating incremental resources to new capital investment by Government if the real need Is for additional recurrent expenditure to enable past Investments to be properly utilised. Thus the appropriate size of the Public Investment Programme (PIP) cannot be determined independently of decisions about the volume of recurrent expenditure required. This is true for individual sectors as well as for the public expenditure programme as a whole. [141 Of courses, to allow for rational adjustment by sectors the guidelines must have a medium-term horizon, although next year's ceiling may be definite while subsequent years' guidelines are more Indicative. Page 293 ALLQCATION AND NANAGENENT OF PUBLIC 6PENDINO 67. This point would apply even If there were no backlog of unmet recurrent expenditure requirements, since new development expenditures generate additional recurrent expenditure commitments that need to be allowed for within recurrent expenditure guidelines. At project level it Is vital to anticipate the recurrent Implications of each project. The PIP proJect profile format (see Box 6) should take this Into account. As a minimum, for each project an estimate should be made of the eventual annual recurrent cost to Government arising from its completion. As the planning system strengthens It should be possible to become more precise and to track the year-by-year impact of each project on the recurrent budget. 68. - Several pragmatic ways to estimate total recurrent expenditure requirements, either for particular programmes/sectors or for Government as a whole, might be considered: (a) An extrapolation of requirements starting from a "normal" period within the country. (b) Cross-country comparisons. (c) Comparison of Estimates requests with actual provision. (d) Building up of an estimate from unit cost norms, as already discussed above. 69. These approaches are not mutually exclusive and each might have more relevance to some sectors than to others. With (a) the dlfficulty Is likely to be finding an appropriate reference point, since It Is so long since services were adequately funded and there have been such changes In price levels and In Institutional arrangements. With (c) there may be a problem that compilers of Estimates requests have become cynical about the exercise, knowing from experience that even funds provided in the approved Estimates are by no means assured (see Box 3). Availability of data may be a problem In attemptIng to employ cross-country comparisons. Also, some actlvities (primary education, say) may be lnstitut onally very similar In different countries, enabling compalIsons to be easily drawn, but for others (agricultural extenslon, perhaps) it may be harder to find good International analogies. Page 30 ALLOCATION AND KANAGEMENT OF PUBLIC SPENDING IV. AID AGlHCTI9J151 AID COORDINATTON AND TECKHTCAL ASSISTANCE Effects of the Increasing Relative Iportance of Aid 70. The financial stringency that affects most SSA governments naturally Increases the relative importance of aid resources. This has a number of effects, Including: (a) It Increases the leverage that aid agencies can exert. Not only do governments need the money worse, they are also less able to mitigate leverage through fungibility. Ability to accept aid that has conditions attached (whether these relate to policy or procurement) without altering the recipient country's overall expenditure profile depends on the recipient having Its own resources that It can direct to the activities that aid agencies are uaiwilling to support. Nowadays many countries find It very difficult even to meet minimal local 'unding requirements that "unlock" external funds, and the scope for simultaneously satisfying aid agency preferences without substantially altering the recipient's resource allocation pattern Is much reduced. (b) As the relative Importance of aid Increases, so responsibility within government for handling aid matters becomes a more valuable prize. This factor underlies the contentiousness of the allocation of institutional responsibilities for aid coordination in many countries (see below). (c) Increasing dependence on aid, combined with deterioration In standards of government administration and accounting, means that a large proportion of development expenditure bypasses the regular government budget and accounts procedures. This Is an understandable reaclion by the ald agencies, but It increases the difficulty of establishing what is actually happening in terms of public expenditures. Similarly, the monitoring of aid-financed expenditures and proJects tends to follow the disparate requirements of various aid agencies, and may make more difficult the government's task of monitoring the overall expendlture programmes. (The tendency for the boundaries between recurrent and development support to become blurred has already been mentioned.) [151 Regardlng terminology, a more restrictive use of the term "donor" would be helpful. The term Is often used loosely to embrace all aid agencies irrespective of the terms on which they provide aid. Even grant aid should not be treated as a free resource (it may for example have. substantial recurrent cost Implications for the recipient) and It Is undiplomatic for the World Bank to style Itself a donor when much of Its aid Is on comparatively hard terms. Page 31 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING 71. Understandably, the effectiveness with which aid is irobillsed and used has become a matter of concern both to aid agencies and to recipient governments. "Aid coordination" Is a recurring topic In the dialogue between aid agencies and governments. Unfortunately, as I explain In the next section, the dialogue is often unproductive because of a failure to appreciate the diversity of tasks that the term "aid coordination" may cover. Aid Management and AId Coordination 72. Typically the dialogue about aid coordination takes place In terms of where the institutional responsibility for aid coordination should be located, and the arguments are often sterile because of failure to appreciate the subtleties of the issues involved. The management of aid has many aspects; once these are clearly distinguished, it becomes clear that there are legitimate and complementary roles for the various government bodies that are usually involved in arguments about the proper allocation of aid-related responsibilities. It Is also important to recognise and strengthen the connections between aid management and development planning In general. Aid coordination is not a discrete activity that can be compartmentalised and assigned to a single agency acting In Isolation. 73. It needs also to be emphaslsed that aid managemint s a wider set of activities that cannot be handled exclusively by whichever agency Is given prime responsibility for aid coordination. There need to be appropriate links between the Planning Ministry and other agencies and, of equal importance, between the aid coordination department and other departments of the Planning Ministry. The basic aims of aid management should be: (a) To optimise external resource Inflows, bearing in mind the capacity of Government and of the economy to service any debt involved. (b) To optimise the allocation of these resource inflows, In support of Government's general development strategy. (c) To ensure the effective utillsation of aid once It is allocated. These basic aims are interconnected, since the volume of inflows is not independent of the uses to which they are put. Funds may be available for one purpose but not for another, the rate at which funds are disbursed will depend on the rate of implementation of the projects to which they are assigned, and so forth. Page 32 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING 74. There is an "aid cycle" analogous to, and closely connected with, the "project cycle". Its main elements are: - obtaining aid commitments In principle - converting such commitments Into substantive agreements - determining the specific projects/programmes to which the aid Is to be applied - commencing the projects/programmes so that disbursement can begin - carrying through and monitoring the Implementation of projects so that disbursement Is fully and rapidly achieved - recording and meeting repayment obligations. 75. Effective aid management requires attention to all the elements of the aid cycle. Much of the dissatisfaction that arises about aid management concerns failures to move smoothly from one stage of the cycle to another - unconsummated commitments and slow disbursements being two typical complaints. Yet too often the debate about aid coordination has been conducted with Implicit reference to responsibilities for only one or two elements of the cycle, notably the general seeking of aid and Its allocation amongst alternative projects. 76. Aid agencies often bemoan the lack of an effective central aid coordinating body on the government side. This seems to relate mainly to two aspects: - At the Xrolect level there .'s need for effective follow-up of problems that arise In the preparation and Implementation of Individual projects (and consequently in the flow of aid disbursements). The Ministry of Planning, as custodian of the development programme and coordinator of the Plan's preparation, is often the appropriate body to assume this role. (Note that this function requires coordination within government as much as with aid agencies.) Page 33 ALLOCATION AND MANAGEMENT OF PUBLIC SPENDING - At aggregate level the nature of aid has been changing. It is now less related to specific conditions attached Independently to "their" projects by the various aid agencies and much more to a general appraisal of the country's macroeconomic management and development strategy as a whole. In this the aid agencies tend to follow World Bank and IMF leads, and the Consultative Group process epitomises this; It creates a need for forums in which the aid agencies can discuss matters of common concern with Government In the Intervals between Co meetings, and this In turn requires that the Government be able to take the initiative in organising the necessary dialogue. (Again, such dialogue requires many participants on the Government side, but there must be a clearly identified agency within Government with the responsibility for organising Government's participation In the dialogue.) 77. Governments are often ambivalent about "aid coordination" meetings with aid agencies, since there is less than universal enthusiasm for the diminution of sovereignty that is Implied in the engagement of aid agencies In "policy dialogue". It could be helpful to distinguish more clearly between meetings that are essentially technical and those concerned with policy (although the boundaries inevitably blur). In some cases government disinclination to create or maintain forums for "policy dialogue" has led to a failure to address technical and bureaucratic obstacles to aid utilisation that government and aid agencies alike have an unambiguous interest In tackling. 78. Standards of aid management are generally very low. Many (most?) SSA governments are too passive In this respect, leaving the initiative mainly with the aid agencies. Preparation of a realistic PIP Is an Important step towards better aid management since it restricts the freedom of manoeuvre of separate aid agencies in identifying and preparing projects that suit their preferences and brings into focus the question of which aid sources should be approached for which projects.[161 It needs to be emphasised that, while attempts at coordination amongst themselves by aid agencies may be helpful, they cannot substitute for effective aid management by the government Itself. 1161 This is an aspect of aid management that is often very poorly handled. For example, each recipient has a fixed allocation of Lome Convention funds from the EEC. The strategy should be to draw down those funds as rapidly as possible, not, as often happens, to assign them to very long-term or Inherently slow-moving projects. Page 34 ALLOCATION AND HANAGEhENT OF PUBLIC SPENDING YahnIeCA AsRistance 79. Public expenditure management and reform Is a very sensitive area In which to involve technical assistance. (Such sensitivity Is reflected In some governments' preference for UNDP or bilateral assistance In this area rather than involve the World Bank.) General experience with technical assistance to strengthen planning performance and Institutions has been disappointing. Short-term assistance has been effective In generating emergency documents, such as PIPs, but parlaying this into durable lmprovements In planning performance is more difficult. 80. As regards sensitivities: (a) It is imoortant to define the allegiance of TA personnel and consultants clearly. In most cases, irrespective of the funding source for the assistance, personnel should be unambiguously responsible to the government they are assisting. Personnel are bound to be less effective If they are regarded by the goiernment as agents of an external agency. (b) A similar point applies to the Bank's conduct of Public Expenditure Reviews. It Is probably unworldly to regard these as joint government/Bank exercises, but they can be of great value In demonstrating the kind of analyses that ought ideally to be undertaken by governments themselves. 81. Past failure of technical assistance has been largely due to the unsatisfactory nature of the systems and institutions being assisted. It Is important to focus first on the desired outcomes of the PBS and allow technical assistance requirements to be inferred from this. Too often it is assumed that the presence of. expatriate personnel per se can achieve Improvements. 82. Tensions may be caused by misperceptions of why technical assistance mav be needed. Governments are more likely to accept TA If the task involved is seen as a highly technical one (say the computation of ERRs, for example). Often, the real value of TA may be in helping to re- establish relatively straightforward planning and organisational techniques of the kind discussed in this paper. Page 35 ANNEX A: DOCUMENTS REFERRED TO This Is a list of some of the published documents referred to in the course of preparing this paper. Botswana 1985: National Development Plan 1985-91. Ninistry of Finance and Development Planning, Republic of Botswana, December 1985. Botswana 1986: Plannina Officers' Manual. Ministry of Finance and Development Planning, Republic of Botswana, June 1986. Commander 1987: Government Monitorina and Evaluation Systtms !n Africa- What are the Questions? Simon Commander, UNDP, New York, April 1987. Howell 1985: Recurrent Costs and Aaricultural Develoomnt. John Howell (ed), Overseas Development Institute, London, 1985. Tanzania 1984: A Manual of the Budget_ Process in Tanzania. Ministry of Finance, Tanzania, 1984. Tanzania, various years: Annual Plan. Uganda 1982: Recovery Proaramme 1982-84. Ministry of Planning and Economic Development, Kampala, 1982. Uganda 1983: Revised Recovery Programme 1982-84. Ministry of Planning and Economic Development, Kampala, 1983. Uganda 1987: Rehabilitation and Development Plan 1987/88-1990/91. Ministry of Planning and Economic Development, Kampala, 1987. World Bank 1984: Public Sector Manacement in Botrwana - Lessons in Pragmatism. N Raphaeli, J Roumani, and A C HacKellar, World Bank Staff Working Papers, Number 709, World Bank, Washirgton, 1984. Zambia 1984: Restructuring in the Midst of Crisis. GRZ/CG Report, Lusaka 1984 (2 Volumes). Zambia 1985: An Action Programme for Economic Restructuring. Government of the Republic of Zambia, June 1985. Zambia 1986: Economic Review and Annual Plan. 1986. National Commission for Development Planning, Lusaka. ------------------------------------------------ PPR Working Paper Series Title Author Date Contact WPS14 The Optimal Currency Composition of External Debt StuJn Claessens June 1988 S. Bertelsmeler 33768 WPS15 Stimulating Agricultural Growth and Rural Development In Sub-Saharan Africa Vijay S. Vyas June 1988 H. Vallanasco Dennis Casley 37591 WPS16 Antidumping Laws and Developing Countrios Patrick Messerlin June 1988 S. Torrijos 33709 WPS17 Economic Development and the Debt Crisis Stanley Fischer June 1988 C. Papik 33792 WPS18 China's Vocational and Technical Training Harold Noah June 1988 W. Ketema John Middleton 33651 WPS19 Cote d'lvoire's Vocational and Technical Education Christiaan Grootaert June 1988 R. Vartanian 34678 WPS20 Imports and Growth in Africa Ramon Lopez June 1988 61679 Vinod Thomas WPS21 Effects of European VERs on Japanese Autos Jaime de Melo June 1988 S. Fallon Patrick Messerlin 61680 WPS22 Methodological Problems in Cross- Country Analyses of Economic Growth Jean-Paul Azam June 1988 E. Zamora Patrick Gulilaumont 33706 Sylviane Guillaumont WPS23 Cost-Effective Integration of Immunization and Basic Health Services In Developing Countries: The Problem of Joint Costs A. Mead Over, Jr. July 1988 N. Jose 33688 WPS24 World Bank Investments in Vocational Education and Training John Mlddleton July 1988 W. Ketema Terri Demsky 33651 PPR Working Paper Series Title Author Date Contact WPS25 A Comparison of Alternative Training Modes for Youth in Israel: Pesults from Longitudinal Data Adrian Ziderman July 1988 W. Ketema 33651 WPS26 Changing Patterns in Vocational Education John Middleton July 1988 W. Ketema 33651 WPS27 Family Background and Student Achievement Marlaine E. Lockheed July 1988 R. Rinaldi Bruce Fuller 33278 Ronald Nyirongo WPS28 Temporary Windfalls and Compensation Arrangements Bela Balassa June 1988 N. Campbell 33769 WPS29 The Relative Effectiveness of Single-Sex and Coeducational Schools in Thailand Emmanuel Jimenez August 1988 T. Hawkins Marlaine E. Lockheed 33678 WPS30 The Adding Up Problem Bela Balassa July 1988 N. Campbell 33769 WPS31 Public Finance and Economic Development Bela Balassa August 1988 N. Campbell 33769 WPS32 Municipal Development Funds and Intermediaries Kenneth Davey July 1988 R. Blade-Charest 33754 WPS33 Fiscal Policy in Commodity- Exporting LDCs John Cuddington July 1988 R. Blade-Charest 33754 WPS34 Fiscal Issues in Macroeconomic Stabilization Lance Taylor WPS35 Improving the Allocation and Manage- ment of Public Spending Stephen Lister August 1988 R. Blade-Charest 33754