I. Traditional Analysis of the Effects of Budgetary Compression ...

The dependent economy model is most frequently used to analyse stabilisation and adjustment policies and macroeconomic equilibrium in the medium-term. This model distinguishes between domestically consumed goods (also called non-tradable) and goods which are exported and imported (tradable), and it allows us to study the impact of changes in expenditures (public and private).

In its simplest version, this model's assumptions are as follows. The external account deficit is identically equal to the government deficit (this is equivalent to postulating that private agents cannot lend to, or borrow from, the government or external sources). There is free access to foreign financing. The economy consists of two sectors of production: an export sector producing primary goods and a sector producing goods which are imperfect substitutes for an imported competitor. Only labour is mobile between the two private sectors.

Generally, public expenditure is viewed exclusively as a component of absorption, i.e. as an element of aggregate demand. Thus, a contraction in government spending results in a decrease in demand and, ultimately, in the price of domestically produced goods and a shift of productive resources into the export sector. The ensuing depreciation of the real exchange rate induces substitution of locally produced goods for imports in the private consumption bundle. Given that the production of tradable goods increases while the demand for imports declines, the trade deficit falls. Consequently, the decrease in the government deficit results in a greater fall in the trade deficit, owing to the joint impact of increased exports and decreased imports. Notice that only two mechanisms act on the re-allocation of resources: demand and price effects.

Neither the original SAP models, nor the critiques which followed their introduction, considered the fact that public expenditure may be a factor in the productivity of inputs an important contributor to the resumption of growth.

Empirical studies (Antle (1983), Chhibber (1988), Lee et Anas (1991), Kessides (1993), Beenhakker (1987)) in developing countries have however revealed that economic activities by the government as a provider of services and public goods have numerous impacts, direct and indirect, on private firms (in terms of factor productivity, output levels, and price formation), and that these impacts may vary between sectors.

These studies have thus shown that expenditures on transportation and communication are significant determinants of productivity differentials between the agricultural sectors of different countries. In fact, in countries with poorly developed infrastructure, improvements in these facilities contribute more to increasing the supply of agricultural output than do price variations.

Similarly, most productive activities use electricity, telecommunications, water and transportation, to name a few, as intermediate inputs. Inadequate infrastructure entails a number of economic costs :

• first, direct costs resulting from production delays, from losses of perishable goods, and from the general underutilization of capacity,
• second, an increase in capital expenditure to compensate for limited access to public goods, and
• third, the spread of these increased costs through the rest of the economy. Analyses have revealed that between 25% and 60% of the final price of food stocks are transportation costs.

Incidentally, we notice that, at least since the beginning of the 90's, there has been a shift in the stance of the Bretton Woods institutions, where increasing attention is paid to defining the role of the government in recognition of its key contribution to allowing economies to operate at full potential.

In light of this assumed impact of government expenditure on supply and on the productivity of inputs, it appears useful to re-examine the appropriateness of budgetary policy in periods of adjustment.

II. ... vs. the Assumption of a Productivity Effect of Public Expenditure

If we accept that government expenditure creates public intermediate goods, i.e. goods which are required for private production, it follows that budgetary restraint translates to a reduced supply of a factor of production, and thus to a negative shock on private sector supply. This implies a reduction in households disposable income.

We can distinguish between three categories of public intermediate goods based on our assumptions concerning the intensity of the externality on private production:

• Public goods which impact identically on all sectors, such that the externality is symmetric. These may include general administrative expenditure, contributing to an improved functioning of the legal system or an increased overall level of security.
• Public goods whose impact is strong on the tradable sector and weak on the non-tradable sector. When this latter sector primarily comprises firms in the informal economy, we may assume that government activity has little relevance to it, and that it is the other sector which is the primary beneficiary. Otherwise, it may be that public expenditure directly benefits the export sector, such as when the government invests in the construction or maintenance of ports and roads.
• Finally, public goods which benefit the non-tradable sector and do not affect the tradable sector. When the non-tradable sector is concentrated in cities while the rural sector is export oriented, the expansion of municipal infrastructure to the exclusion of rural development may illustrate this case.

In all cases, a decrease in government spending induces a depreciation of the real exchange rate as the decline in public expenditure is compounded by the fall in agents' incomes resulting from the deterioration of factor productivity.

If budgetary compression affects the supply of public intermediate goods with limited externalities on overall private production, the price and initial demand effects may dominate productivity effects, resulting in a depreciation of the exchange rate and a modification of production structures favouring the sectors exposed to foreign competition this is the result generally expected from structural adjustment policies. If, however, production externalities effects are significant and counteract the price and demand effects, real exchange rate depreciation is no longer assured. Furthermore, even if the depreciation occurs, it is possible that this relative price change will not provide a sufficient impetus for economic restructuring.

If the decrease in government expenditure primarily induces a contraction in the tradable sector, the economy will only benefit from a fall in the real exchange rate if the decline in the supply of the public good does not result in an excessive contraction of the export producing sector, i.e. if the public good generates weak externalities. On the other hand, if the impact of the externality is substantial, supply in the export sector will deteriorate because of the budgetary policy. Consequently, this simple model allows us to illustrate the conclusions of the aforementioned studies by showing that an improvement in the relative price structure (favouring the prices of export goods) resulting from a decrease in government spending may be accompanied by a contraction in the sector producing tradable goods, owing to the inadequacy of government services.

If the decrease in public spending primarily entails a contraction in the non-tradable goods sector, and if the impact is substantial, then the real exchange rate may appreciate despite the fall in demand. Private demand cannot be redirected to domestic supply since this latter has contracted while the production of tradable goods does not increase because of the fall in its relative price. We would consequently witness an increase in the external deficit.

This preliminary analysis reveals that the mechanisms activated by budgetary cut-backs may vary if we assume that government expenditure creates production externalities on different sectors, and it demonstrates the possible existence of incompatibilities between a tight fiscal policy and the required modification of the economic structure. Nonetheless, so far we have assumed that budgetary restraint only affects intermediate consumption, which in turn generates public intermediate goods. The public payroll however constitutes a large share of government budgets and remains, in many countries, practically the only instrument of stabilisation, given that other expenditures have already been reduced considerably. Consequently, it is of some importance to examine government employment policies and their possible effects on economic adjustment.

III. Government Employment Policies, Production and Unemployment

In order to analyse the impact of layoffs or salary cuts in the public sector, we shall adopt the framework of a dual economy with unemployment (Harris-Todaro model). Developing countries are, in fact, characterised by underemployment of the labour force and by the coexistence of two zones of economic activity one modern, essentially urban, comprising the government and import-substitution producers, the other traditional and consisting largely of producers of agricultural exports and an informal sector producing non-tradable goods.

When layoffs affect a large number of individuals providing unproductive services the increase in the unemployment rate results in a decline in real income and a depreciation of the real exchange rate, reducing the expected real modern wage and inducing a reverse migration of labour toward informal and rural activity. As the traditional wage falls, the production of export goods increases. bsequently, the trade balance improves owing to the fall in demand for imported goods and the increase in exports due both to the depreciation of the real exchange rate and the fall in the wage bill in the traditional zone of activity.

If government layoffs result in a dismantling of public services such that the entirety of the private sector is affected, then the decline in real income and in unemployment will be greater. While the production of importable modern goods and of non-tradable traditional goods is sure to decrease (the negative impact of the decreased availability of the public good is exacerbated by the fall in demand, which is greater for this sector than under the previous scenario), the lower wage rate in the traditional sector attenuates the impact of the reduced supply of the intermediate public good for the traditional export sector. As before, the migration of labour is toward the traditional zone ; its level is however greater because of the fall in production in the private modern import-competing sector.

Clearly, if the export producing zone is to be less affected, the supply of public intermediate goods targeted at the modern zone of activity should be reduced more. Conversely, as in the preceding section, if an appreciation of the real exchange rate results from the fall in the production of non-tradable goods, then the domestic competitiveness of the tradable sectors may decline and conditions will be less favourable for a better trade balance.

Civil service layoffs are politically difficult to implement. Payroll reductions, which have often become unavoidable, often take the form of late payment (wage arrears) or, when political feasible, wage cuts. In some cases, such as Cameroon, these pay cuts have been substantial. In these cases we often observe a sharp fall in the productivity of bureaucrats and high level of absenteeism as they take second jobs in order to maintain a minimum income level. The provision of public services becomes disorganised and sporadic.

The contraction of the government payroll results in a drop in income and hence in demand. Like laying off productive civil servants, cutting wages results in decreased provision of public services resulting in a negative shock to production. As this decline in the wage rate carries over to the rest of the modern sector, i.e. the private import-competing sector, however, this latter benefits from the lower cost of manpower and thus increases supply and hires workers.

In order to simplify and better delineate the induced changes in the labour market, assume that the fall in wages does not alter the productivity of civil servants and, consequently, that it is equivalent to laying off low-productivity public employees. We see that the " tradable " modern sector is in a position to grow (because of the decline in the urban wage rate), and a fall in the rate of modern-sector unemployment may follow. Henceforth, labour's remuneration in the zone of traditional activity is simultaneously driven down by the decline in the modern salary and up by the fall in the urban unemployment rate. These two opposing forces determine the evolution of the "tradable" traditional sector. If we assume, however, that hiring of the unemployed by the modern sector is relatively limited and that, as a consequence, there is a net movement of labour into the traditional zone, production of exports of agricultural goods may increase. As to the traditional sector of non-tradable goods, these opposing movements in the modern wage and in unemployment have an indeterminate impact on demand, and we cannot be sure what the net effect on prices will be. Thus, even if the evolution of the real exchange rate is unclear, it is very likely that the decrease in the modern salary will benefit the "tradable" sectors, both modern and traditional, and not only the traditional sector as was the case under government layoffs.

This theoretical framework may elucidate some of the effects of labour intensive projects funded by international financing agencies and targeted primarily at combating unemployment. These programs are primarily designed to realise the social policy goals of alleviating poverty, being counter-cyclical instruments intended to stimulate demand in poorer regions or in periods of slow growth. For example, road maintenance projects are favourites as they are very labour intensive. Since their main goal is job creation, they basically consist of paying individuals with no resources to perform low-productivity work. If we assume that these projects target the unemployed, and benefit production, in the urban zone, then our model reveals that the impact may reach beyond stimulating demand and lowering the urban unemployment rate to induce a contraction of the agricultural export sector and an appreciation of the real exchange rate. The net result is thus a deterioration of the trade balance and a loss of domestic competitiveness in the "tradable" sectors. These effects may be mitigated if the aid is directed at supplying infrastructure benefiting the tradable export sectors.

IV. Conclusion

In emphasising the productive nature of some government expenditure, this analysis has focused on the implications of political decisions to reduce civil service employment or the supply of public goods. Even though the general equilibrium framework we have adopted does not allow us to consider the lags between initial, short-term recessionary effects due to the decline in government absorption and its medium-term negative effects on sectoral supplies, we conclude that budgetary policy may constitute an important instrument of structural adjustment, and not solely of stabilisation. In fact, even when the fall in domestic demand induced by a contraction of the public sector causes the real exchange rate to fall, it is possible that the production of tradable goods will decline, contrary to the objectives of structural adjustment programs. Furthermore, an appreciation of the real exchange rate is clearly possible when the production of non-tradable goods suffers from the diminished supply of public services. There exist effects which, though secondary, are no less real, and which should motivate a re-examination of the issue of the impact of government cut-backs on the adjustment of the real economy.

Bibliography

• Antle J. M. (1983), Infrastructure and aggregate agricultural productivity : international evidence, Economic development and cultural change, vol 31, n°3.
• Beenhakker H. L. (1987), Issues in agricultural marketing and transport due to goverment intervention, Transportation issues series. Discussion paper n°TRP7. World Bank, transportation department, Washington, D.C.
• Chhibber A. (1988), Raising agricultural output : price and nonprice factors, Finance and development, june, pp. 44-47.
• Dornbusch R. (1980),Open economy macro-economics, Basic Books, Inc. Publi., New-York, 293 p.
• Kessides C. (1993), The contributions of infrastructure to economic development; a review of experience and policy implications, World Bank Discussion Papers n°213, 47 p.
• Lee K. S., Anas A. (1991), Manufacturer's responses to infrastructure deficiencies in Nigeria : private alternatives and policy options, In Economic reform in Sub-Saharan Africa. Chhibber A., Fischer S. (ed by), A World Bank Symposium 1991, 334 p.
• Mesplé-Somps S. (1995) Bien public de production et équilibre général ; une analyse de chocs de dépenses publiques en économie ouverte sous ajustement. Thèse, Université de Paris I, 365 p.
• Mesplé-Somps S. (1996) Politique budgétaire en ajustement et comportements micro-&eacueconomiques : un modèle à facteurs spécifiques avec bien public intermédiaire. upcoming.