Resumé

Quantitative targets are widely used in decremental budgeting. On the macro-level, governments set global limits for the growth of public expenditures, government deficits, or the ratio of government spending to GNP. On the micro-level, reduction targets are frequently assigned to spending ministries or agencies on a selective basis or across-the-board. While this kind of targetry appears to be indispensable to achieve cutbacks, it is also very deficient in that it treats public expenditures as a fungible commodity, consisting of interchangeable units of account. As austerity policies are coming; of age, the inherent weaknesses in the targets approach attract increasing attention, and there is growing concern about the quality of austerity measures. To make sound decisions on such measures, governments must develop a better understanding of the characteristics and properties of various public expenditures. This paper outlines four varieties of expenditure analysis that may contribute to 'the leap from quantity to quality' in decremental budgeting.

Introduction

The governments of the industrial nations may differ on many things, but on one matter they now seem to agree. By 1984, they have all come to the conclusion that the presenit economic predicament requires a policy of budgetary austerity. The ideological trajectories by which they have reached this position vary greatly. Some countries, e.g. Switzerland, are traditionally conservative in their public spending and have practised austerity since long before it was in fashion. In other cases, the pendulum has swung back after a remarkable expansion of the welfare state. Some governments have tried to break out of the recession on their own, only to be forced back into the penance of budgetary rigueur. Other national versions of this stop-go spending rhythm have preceded the present convergence on the need for stringent spending, but bygones can now be bygones. What counts is the present, and at present we are all belttighteners.

---

* This paper was written for the OECD Co-operative Action Programme within the framework of its Joint Activity on Public Management Improvement. I am indebted to Bart le Blanc, Robert D. Bonwit, Rawi Kapil and Henry Verwayen for valuable comments.

In implementing this policy, otherwise divergent governments tend to rely on quite similar methods and instruments. A common feature is the use of quantitative norms. On the macro-level, limits are now frequently established for the highest permissible levels of public borrowing, or public spending as a percentage of GNP, or increase in public spending in relation to the increase of GNP. On the micro-level, there are all kinds of targets which aim to restrict the expansion of individual ministries, agencies, and programmes: caps, ceilings, limits, envelope totals, frames, etc. (For a detailed discussion of such tools, see Tarschys 1985.) A common micro-target, derived from the macro-target, is the across-the-board cut imposed evenly on broad segments of public expenditures. While admittedly crude in many ways, this technique of reductions has the merit of requiring little information at the centre and being relatively acceptable because everybody is hit and nobody is stigmatised. The Japanese call this the principle of baransu (balance), and other nations refer to it as 'the fairness principle⁵ or 'equal pains' or 'shared misery'.

The popularity of across-the-board solutions depends very much on the complex political settings in which cutback decisions are made. An omnipotent minister of finance could easily devise a better way of trimming his budget. Yet given the prevailing balance of forces in modern democratic societies, budget officials tend to view this method as a necessary second best. In some cases, the across-the-board cut may be the only way of arriving at satisfactory reductions, and in other cases it may serve as a useful complement to more selective approaches. Yet nobody is unaware of the grave deficiencies also pertaining to quantitative norms in general and across-the-board targets in particular. By treating public expenditures as a fungible commodity — consisting of interchangeable units of account — the quantitative measures tend to blur all distinctions between various kinds of outlays and inject a good deal of irrationality into the budgetary process. Everyone knows, of course, that there are 'good cuts' and 'bad cuts'; that some reductions are merely optical or transient while others have lasting impact; that some reductions lead to real savings while others are offset by shortfall in receipts or secondary demands on the public purse. Yet this awareness is easily lost in the scramble for quantity. The grand finale of the annual budget drama is often the scene of what Robert Merton calls 'goal displacement'; suddenly numbers matter much more than effects. When it comes to meeting the solemnly-proclaimed norm, make-believe savings are as acceptable as real cuts; and by the budgetary version of Gresham's law, the politically cheap proposals therefore tend to drive out the more cumbersome ones.

This tendency endangers both the efficiency and the legitimacy of the pursuit of expenditure reductions. On the one hand, cuts seldom achieve their declared objectives; deficits grow in spite of bold savings, and government efforts are frequently seen to land embarrassingly wide of their mark. On the

other hand, such repeated failures give birth to many nagging doubts about the rationality of the whole process. If austerity never succeeds, why bother? After several rounds of savings, interest groups and spending departments have become increasingly articulate and sophisticated in demonstrating that cutbacks yield much less than expected. And to the chagrin of the ministry of finance, they are often right.

As austerity policies are coming of age, these problems seem to attract increasing attention. By all appearance both macro-targets and micro-targets are indispensable in decremental budgeting. Ministries of finance do not possess the information required to pass judgement on thousands of different outlays. They can offer stern views on totals and chunks of totals, but rarely on the minutiae of government spending. Without quantitative indicators to portion out the allocation of cutbacks, they stand little change of mobilizing other ministries and getting the job done. Thus, targetry appears to be a sine qua non for the decentralization of hard choices. Yet at the same time there is growing awareness of the: weaknesses inherent in the numbers approach, and many questions are asked about the secondary effects of expenditure reductions. If there are 'good cuts' and 'bad cuts', how does one separate the wheat from the chaff?

In the last instance, thi«; is a matter of political choice. One government might prefer to spend less on defence while another wants to hold back on social security. But within any hierarchy of values, or set of political preferences, one can also rank order various possible cutbacks according to their effectiveness, durability, and dynamic impact. Any government interested not only in advertising cuts but also in accomplishing real savings would do well to distinguish between, e.g.., single-year vs. multi-year reductions, investment vs. consumption cutbacks, and abolished vs. deferred expenditure. It should also have reason to examine the relative impact of different cuts on the rate of inflation, the balance of payments, unemployment, and other crucial variables. The least that such analyses could be expected to achieve would be heightened awareness of the goal conflicts involved in decremental budgeting. A further end, nowhere attained yet, would be to build in certain qualitative dimensions into the quantitative targets and thus make them more reflective of the full spectrum of policy objectives.

Expenditure analysis serves different functions for policymakers. One is to further the understanding of expenditure trends and their underlying social and economic forces. A second function is to strengthen and differentiate the informational basis of budgetary decisions by highlighting probable effects of various options. A third function is connected with the need for forceful presentation and propagation of such decisions. The present economic predicament appears to have reinforced all these motives for looking closely at the properties of different government expenditures. Recent research in social

psychology (Kahneman & Tversky 1977) indicates that the negative reactions to small decrements appear to be much stronger than the positive reactions to small increments. Austerity measures, therefore, are likely to be much more controversial than expansionary policies. Hence an increased need for analysis of public expenditure decisions, whether for understanding them, for making them, for justifying them, or for contesting them.

The following paper purports to discuss four varieties of expenditure analysis and to assess their relevance for decremental budgeting. The first section deals with three classical typologies of public expenditure: the functional, the economic, and the hierarchical. The second section is devoted to a type of analysis that many governments indulge in but seldom advertise in their budgetary documents, namely the calculus of intra-budgetary impact. The third section deals with the impact of cutbacks on various other goals of the government, such as the suppression of inflation, unemployment, and external deficits. The fourth section focuses on a criterion which, theoretically, may be less appealing but which in practice appears to be of great import: that of 'cuttability', or availability for reductions.

The Classical Typologies

Every government budget is, in itself, a categorization of public expenditures. Most budgets contain more than a thousand different categories, organized in such units as chapter, line, and item. For cross-national comparisons, the substantive types are frequently recast into broader functional categories. A classic and elementary distinction is that between military and civil expenditures. With the expansion of the modern state, the latter concept is usually broken down into several functions. The following scheme — one of many variants — has been employed by the IMF Government Finance Statistics Yearbook:

1. General Public Services
2. Defence

3. Education
4. Health

5. Social Security and Welfare

6. Housing and Community Amenities
7. Other Community and Social Services
8. Economic Services
9. Other purposes

A new scheme for the classification of functions of government (COFOG)
has recently been established by the United Nations (1980).

Functional categories would seem to be of rather limited use in the formulation

of austerity programmes. \fet they do appear in some party and government efforts to set out priorities and 'posteriorities' for cutbacks. Particularly at early stages of government savings, certain areas are frequently delimited as 'free zones' protected from reductions or granted milder treatment than others. At later stages of austerity programmes, the opposite is more likely to occur: particular functional categories are set aside for specially intensive screening. While 'early' savers will often vow to spare certain priority areas, 'late' savers are more prone to say that not even politically-favoured areas can be entirely exempted from scrutiny and reductions. Thus, as the economic situation worsens, functional categories give less and less policy guidance.

A second common typology is based on economic categories. One cornerstone of such classification schemes is the Pigovian distinction between real and transfer payments. The latter type of expenditure is included in the national accounts but adds nothing to the gross national product, since it is merely redistributive. Another cornerstone is the dichotomy of capital versus current expenditures. The proportion between these kinds of outlay is often assumed to reveal the extent to which the public household is 'future-oriented'. Because of the crudeness of the two concepts, however, such conclusions tend to be rather weak. A third common distinction is that between wages and purchases.

These concepts are combined in different ways. The IMF Government
Finance Statistics Yearbook employs the following economic categories to
disaggregate public expenditure:

1. Current expenditure:

a) Expenditure on goods and services:
Wages and salaries;
Other purchases of goods and services.

b) Interest payments.

2. Subsidies and other current transfers.

a) Acquisition of fixed capital assets.

b) Capital transfers.

Distinctions of this kind have long been employed not only for descriptive but also for prescriptive purposes. Some countries (e.g. Sweden) have maintained separate budgets for capital and current expenditures. Others (e.g. Japan) have allowed deficit financing of investment expenditures only ('the principle of construction bonds'). Different rules have also been applied to inflation compensation for various types of appropriation. Furthermore, expenditure policies are sometimes based on the observed behaviour of different categories of spending. In recent years, the rapid growth of transfer payments has attracted particular attention. In many countries there is also concern

about the deteriorating ratio of capital to current expenditures, which is partly
explained by the greater inflexibility of consumption as compared to investments.

A third classical distinction is that between levels of government. In analyzing a country's public spending, it is customary to note what parts of it fall on the central government, on state or regional government, and on local government. This classification of expenditures according to hierarchical categories sheds some light on the degree of relative centralization in different nations, but the full picture emerges only when other variables (such as regulation, fiscal autonomy, etc) are taken into account.

The sub-national levels of government have often been affected by austerity measures at the central level, and sometimes been compelled to bear the brunt of the burden. Frequent components of recent budgets and economicpolitical packages have been cuts in transfers to sub-national levels, consolidation of specific grants, and introduction of more complex and selective formulae for the computation of transfers to the sub-national level. Yet there is little evidence that the hierarchical categorization of public expenditures has been of much use in this context. Like the other 'classical' typologies, it can help decision-makers reflect on some broad priorities in the allocation of cutbacks, but it is not an instrument for the fine tuning of decremental policies.

The Analysis of Intra-Budgetary Impact

When austerity budgets are presented to the public, governments tend to emphasize the impressive gross savings that have been accomplished — in relation to last year's outlays, last year's plan for the subsequent fiscal year, or perhaps the extrapolations and requests for appropriations that were on the government's table at the beginning of the budgetary process. Whatever the base for comparison, the figures advertised in such documents differ a great deal from the net savings achieved. Cutbacks come home to roost. By the time that a couple of trimmed budgets have been launched, it is painfully clear to everyone that deficits do not diminish at the same pace as expenditures, if, indeed, they diminish at all. The reason for this is partly the rift between gross and net reductions, or in other words the impact of budget cuts on the budget itself. Some typical ways in which discarded costs reappear include:

(i) Termination costs. When staff are laid off, contracts cancelled or buildings vacated, there are
frequently transition costs to be defrayed.

(ii) Revenue shortfall. If budget cutbacks lead to a decline in economic activity, tax receipts will decline, too. When new higher user charges are employed to take some load off the budget, an astonishing number of users manage to send the bill back to the government: if they are not financed by the budget themselves, they can often deduct the charges from their tax returns.

(iii) Deferred expenditures. You may take it out of this year's budget, but then it is likely to come
back in next year's, and perhaps with a vengeance.

(iv) Cost increases on other items. With the intricate linkages between different types of expenditure, savings on one item may often lead to additional costs on another item. By abolishing preventive measures against juvenile delinquency, there may be more pressure on the prisons; by doing away with free shots against certain diseases, there may be greater demand for sick insurance compensation, etc. The main risk, of course, is an increase in unemployment which imposes a very heavy burden on the budget. A related problem occurs when public financing is replaced by private (e.g. through user charges) and some of those charges may have to be borne by budget-financed units in their capacity as users.

Calculating the exact losses due to these secondary consequences of individual cutbacks is a formidable task which no ministry of finance can undertake. But there are thumb-rules for in-house estimates. Some interviewed budget officials have reported that as a rule they expect net savings to average 0.5 of gross savings. Yet the dispersion should be considerable. Governments interested in real outcomes, rather than handsome budgets, should have an interest in exploring these differentials somewhat further. It would also seem worthwhile to classify cutbacks according to their impact on the budgets of future years. Extrapolations of gross effects are frequently given in multi-year plans or budgets, but assessments; of net figures are seldom performed, much less; publicized. In many cases, ex ante analysis is virtually impossible since the demands or losses generated by a cutback are unforeseeable or unforeseen. Yet crude estimates can always be made and different 'ideal types' of expenditure reduction can be distinguished. Here are five examples.

The quality cut — Australian budgeteers have borrowed this term from the butchers' language — is an expenditure reduction that has few or no adversary intra-budgetary effects in the short term and that yields more for every subsequent fiscal year. Downward adjustments of indexed benefits (such as pensions, sickness compensation, etc.) may belong to this category.

The one-shot cut, by contrast, may give good value immediately but is of no
comfort for future budget-makers. Savings of this sort may occur when a oncefor-ever
investment is cancelled, or a regular expenditure is skipped for one year.

The vanishing cut appears profitable at the outset, but the saving disappears gradually in the following years. This may be due either to resurgent needs and demands for the very expenditure that has been cut or to the fact that the curtailment of government activities in one area entails growing needs and demands in other areas. Effects on the revenue side may also erode the gains once made through expenditure reductions.

The boomerang is an even more unfortunate version of expenditures cut away one year coming back with a vengeance at a later stage. In this cas(j, the net result over time is negative — as when maintenance neglect leads to considerable outlays for capital renewal or when major accidents occur because inexpensive risk-reducing measures have been suppressed for economic reasons.

A cosmetic cut is appealing from a financial point of view but has no effect whatsoever on the real economy. By moving posts between the budget and extra-budgetary domains or from one fiscal year to another, any government can achieve impressive changes in the balance of revenue and expenditures. Decisions to perform such operations are mostly taken shortly before the presentation of the budget, with the purpose of filling out the vacuum created by unachieved savings of a more substantive and durable nature.

These 'pure types' may suffice to illustrate the divergent intra-budgetary effects of different nominal savings. Tax increases can be submitted to a similar analysis. Here, too, the net gains frequently fall short of advertised yields because of secondary effects on other revenues. This is the phenomenon of 'fiscal cannibalism': in high-tax economies, a government raising one tax will often discover that this change eats up the bases of other taxes. Thus, the gross-net gap appears both on the expenditure and on the revenue side of the budget, and places governments practising austerity policy in a double bind: to reduce the budgetary gap by X billion, they must cut expenditures or raise taxes by perhaps twice as much.

The Analysis of Goal Conflicts

A third line of analysis made particularly topical by the present swing towards decrementalism is that of conflicting policy objectives. Yes, governments do want to keep expenditures under control and press down their deficits. But they also have a number of other goals. The tensions between all these ambitions usually come out quite clearly in the continuing dialogue between the spending ministries and the ministry of finance. When the former are asked to rank order their priorities, they can either play adversary politics by holding up the beggar's sore thumb — i.e., propose cutbacks on items particularly dear to the decisionmakers — or make a loyal effort to apply the criterion of minimal damage. Schwarz (1982) has suggested a new version of McGregor's well-known dichotomy of 'theory X' vs. 'theory Y' to capture this choice of response. According to Schwarz's 'theory X_o', budget-financed organizations do not seek to raise their efficiency, do not reveal their slack and do not rearrange their priorities. Therefore, such organizations must be firmly controlled from above. 'Theory Y_o', on the other hand, would imply that budget-financed organizations are responsible, responsive, and cooperative. Given adequate incentives, they are prepared to innovate and rationalize even if this means shedding resources. The experience of different governments in this field appears to vary a great deal, but nowhere have austerity measures failed to generate a heated discussion on the multiple and competing objectives of the government.

To analyze all such goal conflicts on a regular basis is probably beyond the resources of any ministry of finance. In most cases one could expect to obtain sufficient information by a 'mixed scanning' technique that combines general overviews with in-depth investigations of the conflicts between particularly salient objectives. From the recent literature on budgeting and from the country contributions to the OECD meetings of senior budget officials, it is apparent that many governments already undertake such studies, although the findings are relatively seldom reported in publicized budgetary documents. Some goal conflicts that loom large in considerations on austerity policy are the effects of expenditure reductions on (i) short-term economic activity, (ii) medium-to-long-term economic growth, (iii) employment, (iv) the external balance, and (v) the distribution of wealth and income.

The Impact on Economic Activity

Ever since the Keynesian revolution, governments have attempted to estimate how their fiscal and expenditure policies affect the level of economic activity. Whether this is done by a desk calculator or by a sophisticated computerised econometric model, the assumptions are essentially the same: different multipliers are ascribed to different classes of receipts and outlays. To assess the total impact of the public budget on the economy, last year's budget is usually assumed to be a neutral baseline. The practical and theoretical complications involved are considerable, but need not concern us in this context. Suffice it to note this is a way by which public expenditures can be ordered.

The magnitude of the multipliers depends on a variety of spatial and temporal circumstances, such as the marginal propensities for saving and imports in various phases of the business cycle. Table 1 is merely an illustration of what the picture might look like.

The comparison of different policy packages from this particular perspective is a normal undertaking for economic planning divisions within ministries of finance. Yet to what extent do such calculations influence actual choices in the budgetary process? Thus far, governments have often used expenditure

variations to stimulate the economy, but have seldom used them to restrain it. For the latter aim, it has usually been more simple to apply fiscal or monetary measures. With the present shift towards concern for long-term trends in the structure of public expenditures, however, goals and restrictions are tending to change places. In past recessions, the problem was to find a suitable selection of expenditure increases to restore the desirable level of economic activity. With the present emphasis on medium-term and long-term budgetary health, the crucial point is to find the least damaging reductions. The categorisation by multipliers provides one of several answers to that question.

The Impact on Economic Growth

A traditional problem in economic policy is to reconcile the objective of shortterm stabilisation with thsit of medium-to-long-term growth. While some measures may serve both purposes, there is also inevitable choice: if the former goal is given supreme priority, there is little likelihood of making a good score on the latter. The prevailing mood among OECD governments at present seems to be that greater attention should be paid to the problem of growth and competitive capacity. As an expression of this sentiment, the Commission of the European Communities has recently voiced its concern over the growth of unproductive expenditure in government budgets and called for a shift towards more production-related spending. The Commission's Economic Policy Committee has expanded on this theme and emphasized the importance of restructuring public expenditure. In a paper prepared by the Chairman of the Committee, an attempt has been made to distinguish different categories of outlays according to their effects on economic growth (Table 2).

'The vertical axis of the matrix gives a breakdown by economic category, viz. capital expenditure, public expenditure on consumption, and income transfers, while the horizontal axis indicates how productive the expenditure is. Expenditure falling within Category I in the matrix column adds to firms' productive capacity and/or productivity (e.g. certain infrastructure projects, investment premiums, R & D). Such expenditure can be said to be highly productive. Category II comprises expenditure that contributes to or forms part of public-sector production (e.g. public expenditure on consumption excluding R & D, investment in public buildings and schools). Such expenditure does not, or at least not directly, produce an increase in firms' productive capacity and/or productivity. It can be described as low-degree productive expenditure. Category HI comprises expenditure that primarily boosts private consumption. Such expenditure has no direct expansionary effect on firms'productive capacity and/or productivity; nor does it contribute to publicsector production. For this reason, it can be termed non-productive expenditure. This category takes in not only current payments to households but also capital transfers and loans for residential construction'.

In the presentation of this scheme, it is underscored that the categorisation is rough and arbitrary in many ways. The same applies to the following attempt to rearrange Dutch public expenditure data for 1970 and 1980 into the matrix. While the purpose of this is mainly to indicate how the scheme might be used, the figures of Table 3 lend some support to the hypothesis of a shift from 'productive' to 'non-productive* expenditures.

The 'productivity approach' breaks new ground in the analysis of public expenditures by its forceful distinction between the conventional economic categories on the one hand, and the impact-related categories on the other. The important message here is that capital expenditures should not always

be seen as investment and current expenditures as consumption. But there are also several unresolved theoretical problems in this typology. A crucial point is the concept of productivity. Is it true that expenditure contributing to the activity of the private sector can be characterised as highly productive, whereas that forming part of or contributing to activity in the public sector should be classed a priori as being less productive? An influential school in contemporary economics, as exemplified by Bacon & Eltis (1976), would reply in the affirmative, but it is not self-evident that the production of the public sector should be assessed as economically inferior in this manner. One solution to this problem might be to opt for a less controversial notion than 'productivity', as for instance competitive capacity.

The Impact on Employment

The struggle against unemployment is the avowed principal goal of many governments. The efficiency of various public expenditures in the pursuit of this objective is partially highlighted by the two previous typologies. By arranging public outlays according to their multipliers, one will be able to foresee the budget's short-term impact on the general level of economic activity. By grading them according to 'productivity', one gets some underpinnings for predicting the probable level of activity in a more remote future. The likely employment outcomes may be derived from both schemes, yet since a given level of economic activity can be maintained with a smaller or greater workforce, neither the 'multiplier' nor the 'productivity' categorisation is squarely addressed to the question of jobs. To extricate this particular aspect of the economy's performance, one would have to look at the labour intensity of different expenditure items. It would also seem meaningful to distinguish between short-term and long-term effects. In a simple four-field, the picture could look as in Table 4.

In the upper right corner, we find measures used to combat open unemployment. Particularly if made permanent, such jobs may sometimes have a negative impact on long-term employment, either by locking in labour in low-productivity occupations or by pushing up factor costs so as to make exports and import substitution less competitive. In the lower left corner, we find expenditure items that may not create many jobs at the investment stage, but that are much more promising when it comes to exploitation. The items in the lower right corner are less likely to create new jobs but might of course be more successful in defending old ones. It is far from easy to push all kind of public expenditures into this matrix, but the effort might be worthwhile — not least in political settings in which the distinction between immediate and long-term effects is difficult to propagate. Differing views on this matter account for a lot of friction between ministries of finance, on the one side, and 'cabinet wets' or ministries of labour and social affairs on the other.

The Impact of External Balance

In formulating their economic strategies, many governments are severely hampered by an unfavourable external balance. When this is the case, it might be quite important to examine how different public expenditures affect the flow of payments. Again, it would seem reasonable to distinguish between short-term and long-term effects. In the short run, most types of public outlay will probably have an 'average* effect on the balance of payments, i.e. lead to a currency outflow corresponding to the marginal import propensity of the normal consumer. Yet government purchases of foreign goods, foreign aid, and other similar expenditures are likely to have a more negative impact on the external balance — at least in a static perspective.

To assess the effects of public expenditures on the future balance of payments,

one must look at how steady flows of payments are affected by various programmes and investments. In the wake of the oil crises, many countries have tried to restore their trade balances by investing in energy conservation and domestic energy production. In the planning of major industrial projects, it is customary to calculate the expected balance of exports and imports (raw materials, energy, spare parts versus finished goods), but such techniques can be applied to infrastructural projects as well, even if the degree of certainty is normally somewhat lower. A tentative classification scheme is found in Table 5.

The Impact on the Distribution of Wealth, Income, and Welfare

Economic textbooks make a sharp distinction between allocation and distribution as two separate functions of public expenditures. In practice, however, these effects are not easy to disentangle. Like taxes, government outlays affect the distribution of wealth, income, and other forms of welfare. When a service is provided by the public sector, one can normally observe both 'consumer benefits' and 'producer benefits'. One group is positively affected by the increased supply of public goods or services while another is affected by the increased demand. Alliances between these two interests often seem to play a role in expansive policy-making.

In composing their stringency budgets and the 'emergency packages' (or minibudgets) that have become frequent in recent years, governments are certainly aware of the general distributive profile of their proposals. Some ministries of finance make efforts to calculate the net impact of austerity measures on different regions and household categories, but such surveys are rarely publicized. Analysts and politicians are keenly aware of the fact that first-strike effects are much more easily assessed than the distribution of the burden in the long run. Predicting the latter presupposes a multitude of assumptions on

incidence and future economic development. Thus, presented data on the
distributive impact of expenditure reductions tend to be scarce and highly
selective.

The Analysis of 'Cuttability'

A fourth line of expenditure analysis that has become topical through the recent swing to stringency is that of 'cuttability', or availability for savings. The key problem in such studies is to establish the scope of freedom in policy planning. What decisions can be made instantly, what decisions can be made after certain preparations, what decisions cannot be made at all? And furthermore, what are the preconditions for discrete policy shifts: what laws must be changed, what action must be taken, what consent must be secured? Questions of this kind are very sensitive and likely to produce strong feelings in affected constituencies, since even asking them may be taken as a threatening sign that certain cutbacks are planned. Any rational government, however, should be interested in mapping a reasonable number of possible reductions before a few ones are chosen. In incremental budgeting, there is always a full å la carte of proposed expenditure increases before the government; there is no reason why decremental budgeting should be confined to a take-it-or-leave-it decision on a set menu.

In locating the 'cuttable' items of the budget, there appear to be one theoretical and one practical approach. In the former case, the strategy is to make certain assumptions about the structure of public expenditures and then derive 'cuttability' by a deductive method. In the latter case, the starting point is the real world of legal and political obstacles to cutbacks.

One example of the theoretical approach is to examine the jointness of consumption in different types of government expenditures. To a large extent, public services are not collective in the classical Samuelson-Musgrave sense of that term. Education, health care and many other things produced by public organisations benefit the citizens as individuals and can therefore, in theory, be paid for by the beneficiaries. An important trend in current cutback strategies is to replace the tax-financing of public services by private financing through new or increased user charges. To investigate the applicability of this method, it is necessary to examine the divisibility of consumption of the various goods and services provided by the government. The following typology of needs served by the government is proposed by Ehrlicher (1977):

1. Collective needs.

2. Collective needs with an individual interest.
3. Individual needs satisfied by collective production.
4. Individual needs.

As examples of services corresponding to these categories, Ehrlicher mentions: (1) foreign policy, foreign aid, and reparations; (2) security and public administration; (3) road building and maintenance; and (4) health care and education.

Another example of the theoretical approach is to look at the degree to which various policies are related to economic growth. As analysed in the wide-ranging literature on Wagner's law, the rise of government expenditures is strongly related to 'the great transformation' (Polanyi) of our societies. Many services previously provided in the households have now been taken over by the public sector. Others with no counterparts in older times have come forth in response to the new needs of high-technology production and extensive division of labour. Throughout the postwar period, the momentum of government expansion has been kept up by the dynamics of economic development and differentiation. Yet in a period of slower growth or near stagnation one can expect that some forces traditionally propelling growth in government expenditures will become much weaker. However, it is by no means certain that changes in this direction will be brought out as vociferously as were once the demands for new and higher expenditures; opportunities for savings tend to be much more discreet. To spot such stagnation-related potentials for reductions, it might be helpful to look deeper into the causal relations between economic development and the expansion of public expenditures.

The foundations of a classification scheme to this end can be found in Haller (1966). According to this author, one should first distinguish between outlays that are 'constant' in the sense that they are largely independent of economic development and those that are 'variable' or related to economic growth. The latter category can then be subdivided into investment and consumption-related expenditures and into complementary and substitutive expenditures — a distinction depending on whether a given outlay supplements

or replaces market transactions. The examples in Table 6 are supplied by
Haller.

A different, more practical approach may be to look at the legal barriers to policy change. These determine the degree of technical flexibility of the budget. Many types of public expenditure are founded on statutory obligations. Faced with a new economic situation, many governments appear to be increasingly concerned about the relative inflexibility of their public expenditures. Some have set out to analyse this problem with a view to eliminating 'escalator effects' and widening the scope for political discretion. Categorising outlays along this dimension might help policy makers locate savings that can be made in a particular year. That is a good preliminary to the more difficult problem of what should be cut.

The Swedish Ministry of Finance disaggregates the expenditure budget into
the following flexibility categories (Table 7).

The Dutch Ministry of Finance has arrived at the following distribution
(Table 8):

A final type of analysis focuses on the criterion of feasibility. In most situations of budgetary decision-making, political obstacles are much more decisive than legal ones. Statutes can be changed, if only majorities dåre confront the victims of a particular reduction. Thus, differentials in expected resistance become an important element in the break-down of austerity targets. What will pass and what won't? Giving adequate answers to such questions requires a thorough acquaintance with the political scene and a good portion of intuition. Yet beyond Fingerspitzgefuhl there may also be some room for systematic analysis. Programs can be rated according to the size and strength of their constituencies, and opinion polis can be used to gauge the degree of hostility to various possible austerity measures. While no government appears to have moved very far in this direction, the official collection of public opinion data is quite extensive in some countries, as for instance the United States and the Federal Republic of Germany.

Summary and Conclusion

For several decades, rising taxes were the price we paid for getting better public services. Today's politicians face the less enviable task of asking for higher taxes while cutting back on public services. What they can promise is, at best, reduced public deficits, which allegedly are necessary for holding down inflation, interest rates, and unemployment. The final blow to austerity politicians, however, is that even this recipe does not seem to work. Taxes are raised, expenditures are cut — but deficits remain or continue to expand.

Of course, this is not the pattern of events in all OECD countries and in all recent fiscal years. But the same sinister sequence has occurred often enough and in too many nations to be dismissed as an exceptional process. While virtually all governments have sooner or later come to the conclusion that some measure of austerity policy is inevitable in the present economic situation, there is also growing disillusionment about the effects of such policies. The implementation and management of austerity measures appears to be full of pitfalls and stumbling-blocks.

An important question, then, is how to upgrade the quality of austerity policies, or in other words how to get more positive benefits from the pains inflicted. Expenditure analysis may play some role in this pursuit. This paper has distinguished thirteen different categorizations of public outlays in four different classes. Expenditures have thus been divided:

(i) by functions

(ii) by economic categories

(v) by impact on economic growth
(vi) by impact on economic activity
(vii) by impact on employment
(viii) by impact on external balance
(ix) by impact on distribution
(x) by collectivity of consumption
(xi) by dependence of economic growth
(xii) by technical flexibility

(xiii) by political flexibility ('feasibility')

This list is by no means exhaustive, and neither is it likely to be relevant to all governments at all times. The crucial dimensions to examine in a given situation depends very much on the distribution of power, the political agenda, and the hierarchy of policy goals. Levine (1980) has suggested that retrenchment policies advance through certain stages of development, passing from a phase of denial and delay (using one-time revenues, deferral of capital expenditures, postponement of cash payments, etc.) through a phase of stretching (hiring and purchasing freezes, service rationing, minor cuts of the across-the-board type, etc.) to third phase of long-range cutback planning. It is in this latter stage that quantitative targets are widely employed for sharing the burden of reductions between different sectors of the government. Such targetry, however, inevitably leads to a great deal of disillusionment and questioning of the core assumptions behind austerity budgeting. Adding a fourth phase to Levine's three, we might assume that this discovery is followed by a mounting interest in the properties and characteristics of different cutbacks. In this fourth stage — the leap from quantity to quality — expenditure analysis becomes an indispensable tool for managing budgetary objectives.

REFERENCES

Bacon, R. and Eltis, W. 197(5. Britain's Economic Problem: Too Few Producers. London:
Macmillan.

Ehrlicher, W. 1977. 'Öffentliche Sachausgaben', Handbuch der Finanzwissenschaft, vol. 1,
pp. 753-795. Tubingen: J. C.B. Mohr (Paul Siebeck).

Haller, H. 1966. 'Einige Gedanken zum Thema: Öffentliche Finanzen im Wirtschaftswachstum',
Schneider, E., ed., Wirtschaftswachstum und Wirtschaftskreislauf. C. Föhl zum 65.
Geburtstag. Tubingen.

Kahnernann, D. & Tversky, A. 1979. 'Prospect Theory: An Analysis of Decisions under Risk',
Economica, vol. 47, pp. 263-291.

Levine, Ch., et al. 1980. Fiscal Stress and Local Government Adaptions: Toward a Multi-Stage
Theory of Retrenchment. New Orleans.

Schwarz, B. 1982. 'Är en mänskligare syn på organisationer lönsam?'. In von Otter, C, ed.,
Att spara eller inte vara — det är frågan. Stockholm: Arbeitslivscentrum.

Tarschys, D. 1985. 'Curbing Public Expenditures: A Survey of Current Trends', Journal of
Public Policy, forthcoming.

United Nations 1980. Statistical Papers. Series M, No. 70.