Introduction

The purpose of this chapter is to examine the macroeconomic impact of foreign aid to three Southeast Asian countries. This will be done not just for the sake of understanding the connection between foreign aid and economic development in Southeast Asia, though this is one of the goals. More important, however, is the goal of learning something about the donor policies, allocation of aid and some of the institutional factors related to the macroeconomics of aid so that African countries can benefit from the experience of the Southeast Asian economies. The three Southeast Asian economies chosen for this purpose are Malaysia, Indonesia and Thailand (the MIT economies from hereon). The MIT economies have been among the most rapidly growing parts of what a widely cited World Bank Study has called "the East Asian Miracle". Until the financial crisis of 1997 these economies were very much the vanguard of Asia’s economic march to prosperity.

A combination of factors are often cited for explaining the dramatic transformation of the "miracle" economies of Asia. These factors include openness to foreign trade, high savings rates, stable macroeconomic policies, high literacy rates and favorable demographic characteristics. One might also wish to include an institutional structure — certainly far from perfect as the financial crises in these countries have shown — which was flexible enough to mobilize domestic resources and to utilize available foreign resources, including development assistance for promoting economic growth.

While some of the factors cited above may be region - or even country-specific, other factors are amenable to attainment through policy changes. The key question is: should the African countries make certain changes in their macroeconomic policies in light of the Southeast Asian experience? One needs to be cautious here. The history and institutional complexities of Africa are in many respects different from those in the MIT economies. Furthermore, the initial conditions in Africa today are in some respects worse than the initial conditions faced by the MIT economies — with the possible exception of Indonesia— in the late 1960s. In some respects Africa needs to engage in institution building on a much more massive scale than the gradual progress made by the MIT economies over several decades. Keeping such caveats in mind will help us not to overdraw the relevance of the Southeast Asian experience for the African economies. At the same time, I will try to show that there are some lessons to be learned in the area of aid allocation and macroeconomic policies that can be useful for the African economies in the future. However, given the serious problems Africa faces, the institutional aspects of both the giving and spending of aid money will need to be appropriately designed. Only aid that can reach large numbers of genuinely needy people will work in Africa.

I will proceed by first sketching an analytical framework for the study. I will then outline the aid scenario for the MIT economies in the 1970s and 80s. I will also try to show the linkages between foreign aid and other macroeconomic variables such as saving, investment and the government budget in these countries. Next, I will examine some institutional aspects of aid allocation and its macroeconomic impact. It is not possible to give an exhaustive account of these institutional linkages without expanding this chapter into a book. Therefore, it will be necessary to be selective. However, I hope this brief examination will reveal some linkages between aid and various growth-inducing factors which will help define some policy alternatives for Africa in this area. The concluding section will summarize these lessons. One crucial finding is that there are domestic factors that are complementary to foreign financing (called Foreign Aid Complementarity Elements or FACE in an abbreviated form) such as complementary private investment, human capital and governance structures etc. Thus the basic message of this chapter can be summed up as: in order to be genuinely effective, foreign aid must have the right FACE.

2.Some Analytics of the Macroeconomics of Foreign Aid

In this section I discuss the analytical approach followed here in assessing the effectiveness of foreign aid from a macroeconomic perspective. There exists in the literature on foreign aid something of a paradox. On the micro project level many projects financed by aid are deemed to have been successful; however, at the macro level the evidence is much more ambiguous. Here, as Mosley (1986) has observed, one can indeed find a micro-macro paradox. Mosley himself has offered three explanations for this seeming paradox. The first explanation has to do with inaccurate measurement in micro or macro studies (or both). Second, the fungibility of aid within the public sector may also be invoked to find diversion of aid from investment to consumption. I will have more to say about this later. Finally, backwash effects from aid-financed activities may adversely affect the private sector - for instance displacement of foreign borrowing.

According to the approach taken here, the complexity of the macroeconomics of aid arises from the opportunities and constraints that a typical LDC faces. For example, on the opportunities side there is the possibility of increasing investment and consumption in the public sector; there is also the ‘opportunity’ for reducing domestic revenue-raising efforts. These are only some of the possibilities. On the constraint side, the aid-giving mechanism may institute tied aid disbursements, making aid much less fungible. It is difficult to predict a priori the macroeconomic impact of aid, since many of the opportunities and constraints may work at cross-purposes. Ultimately, it is really an empirical issue where studies of individual countries are the main sources of our knowledge. At the same time some theoretical structure is necessary in order to prevent a completely ad hoc procedure.

What follows from the above considerations is a blend of two complementary frameworks. In the first place the three gap modification by Bacha (1984, 1990) of the earlier two gap models pioneered by Chenery and Bruno (1962) and Chenery and Strout (1966) is the starting point for analyzing the need for understanding the impact of foreign aid in a general way. As a second complementary step, the approach developed by Heller (1975), Gang and Khan (1986, 1991), and Khan (1994, 1996), and applied by Khan to the MIT economies (Khan 1995a, b, c) is also adopted in order to examine the macroeconomic impact of aid on the investment, consumption and revenues of the recipient governments. The second class of models, especially those developed by Khan are sensitive to the specific institutional configurations of the recipient of aid. Here I can only offer a brief sketch of both these approaches and explain why these are complementary. The interested reader could consult the references cited here. Also, at the end of this essay the possible welfare impacts of aid to Africa will be evaluated by using the social capability approach developed by Sen (1992), Nussbaum (1995) and others.

In the original Chenery-Strout model, the savings gap is initially the binding one and therefore determines the foreign exchange requirements. Once the investment constraint (and hence the savings gap) is no longer binding, the targeted growth rate can be achieved by foreign aid that can finance extra capital goods import. Within the dual gap formulation, it is also usually the case that aid is more productive when the foreign exchange constraint rather than the savings gap is binding. With the help of these models, aid requirements can also be calculated. Thus for the year 1970, according to Chenery and Strout:

As Lance Taylor (1993) has pointed out, Bacha (1984) was the first to see that the two gaps really correspond to internal and external balances of open economy macroeconomics. Taylor also underlines the humane dimension with respect to foreign aid flows and fulfillment of basic needs in development thinking during the initial phase of the development of the ‘gap models’.

This last point about the utilization of excess capacity that has developed in many LDCs is crucial in the context of the 90s for both African and Asian economies. For most countries, fiscal, foreign exchange and financial constraints have become even more stringent after the recent Asian financial crises. In this context the proper utilization of excess capacity may require increased public borrowing precisely when under the standard "Washington Consensus" structural adjustment programs (SAPs) fiscal contraction is the norm. It may help to point out explicitly the link between public sector borrowing requirements (PSBR) — that is to say, the funds that the government must raise in domestic financial markets in order to pay for its expenditures net of taxes and

other revenues — and capacity utilization plus further investment. A general expression for PSBR is given by the following:

Here the last term is crucial. As Taylor (1993) puts it:

The approach that results from this addition of the third gap focuses on the creation of aggregate demand for increased capacity utilization. Also, an important feature is the complementarity of public and private investment. Thus public spending on infrastructure and utilities are positively related to private investment. Even public financing of manufacturing can, under some circumstances, lead to complementary private investment.

This ‘crowding’ in of private investment through an increase in appropriate types of public investment leads one naturally to a search for models where the relationship between foreign aid and public investment (and more generally, development expenditures) can be explored. This search, in our case, ends up in the utilization of a class of models pioneered in the development literature by Heller (1975). In the following exposition, Khan’s (1994, 1996) model which takes into account explicitly institutional variations and bounded rationality of policymakers in recipient countries is used.

In this model, it is possible to examine the relationship between bilateral and multilateral aid on the one hand, and development and non-development expenditures on the other. An important component of development expenditures is public investment.

Existing work on the impact of aid on the recipient countries under the two different aid regimes is also not conclusive. Heller (1975), and Khan and Hoshino (1992) find no difference between the two sources as far as impacts on the recipient nations are concerned. The pooled time-series cross-section data used in these studies may partly account for this finding. Gang and Khan (1991) using time-series data for India find statistically significant differences between the two sources of aid. However the model used there is limited by insufficient asymmetries in the loss function. In order to capture the asymmetries of policy makers' evaluation function (e.g. if consumption exceeds a preset level, the loss may not be the same as when it falls short by the same level) an explicitly asymmetric loss function is required.

An equally important aspect of policy making in the real world is the endemic uncertainty and institutional bounds to rational behavior. Departures from strict neoclassical utility maximization leads us to a bounded rationality framework. In this framework development and fiscal targets may not be known with certainty and are the outcomes of a complex negotiation process.

Consider the decision-making process of boundedly rational policy-makers who consider ex ante in their budgetary planning certain indicators of the "proper" level of (planned) expenditures and revenues. Although these levels are treated as targets ex ante the assumption of an asymmetric loss function implies that these are not the utility maximizing values. In fact, the policy-makers possess a loss function in which they try to minimize upward and downward deviations which are weighted differently. The indicator levels from which such deviations are measured can be thought of as outcomes of bureaucratic negotiations within the state and between the recipient and the donors.

It is important to use an explicitly asymmetric loss function because policy-makers may weigh the overshooting and the undershooting of these indicator levels differently. For some policy-makers the under-achievement of some indicators may be more significant than overshooting. For others the opposite may be the case.

By this theoretical and modeling strategy, it is possible to estimate the marginal impact of aid on budgetary expenditure and revenue categories. Earlier works such as Heller (1975), Mosley, Hudson and Horrell (1989), Gang and Khan (1991), and Khan and Hoshino (1992) employed linear-quadratic or quadratic representations of the objective function. But recent work uses an objective function with higher degrees of both non-linearity and asymmetry.

A version of the model describes how foreign aid influences the recipient's expenditure and revenue-raising behavior. In meeting preassigned values of indicator levels of expenditures and receipts the decision-makers respond in a predictable manner to any flows of aid from abroad.

The model takes into account the potential affect of aid on development and non-development expenditures. The former type of expenditures include the public sector's contribution to capital formation. Human as well as non-human capital are included. A third component of development expenditures is the government's contribution to social and economic services, e.g. expenditure on health and general welfare. Non-development expenditures are the expenditures on state administration. These two types of government expenditures are financed by internal and external means. Domestic revenues include taxes, public enterprise surpluses and borrowing. External assistance comes in the form of bilateral and multilateral aid.

Much of the literature on the macroeconomic effects of foreign assistance focuses on aid's effect on economic growth. Our modeling approach is to analyze the impact of aid on public sector variables. Since aid funds pass through policy-maker's hand prior to reaching their destination, understanding where these funds are allocated by policy-makers is a prerequisite to understanding the long-term effects of aid. The distinction made here is between current development and current non-development expenditures. As a rule, the former will contribute to the long run health of the economy while the latter will not.

The policy-makers minimize a loss function subject to expenditure constraints. In most general terms, the (quadratic-ratio) loss function, L, is given by

"j" and "k" are related in the following way: if j (respectively k) represents the indicator value (symbolized by *) then i^k (respectively, i^j) equals i. "i" and "j" can be R, D, or N (domestic revenues, development expenditures and nondevelopment expenditure, respectively). The simplest non-linear model which is also asymmetric and economically meaningful, is obtained when b = 2. Note that for exact fulfillment of chosen indicator levels, L = a ₀ + (a _R/2) + (a _D/2) + (a _N/2). The policy-maker is making decisions on various categories of public expenditures. Each decision will reflect on her abilities, possibly her status, or even her job. In an uncertain environment, the best she can do is to reach the stated chosen indicator value.

The loss function stated in equation (2) has the advantage of allowing for asymmetries in loss when the policy-maker over- or undershoots the chosen indicator level. It also allows us to examine different assumptions about the "type" of the policy-maker. For example, writing the loss function explicitly as

illustrates a policy-maker who is "developmentalist" in orientation: undershooting the development expenditure indicator value is worse than overshooting it. At the same time, the above policy-maker is a "fiscal liberal" since overshooting the revenue raising indicator value is worse then undershooting. Such policy-makers are not very anxious about the emergence of the inflationary gap. These bureaucrats are also "non-statist" in that overshooting nondevelopment expenditures is worse than undershooting. Statist bureaucrats who seek to maximize the resources which the state uses to reproduce itself would have loss functions that are asymmetric in exactly the opposite direction with regard to the composition of public expenditure. All in all, there are eight possible characterizations. Part of our problem is to explore which of these characterizations captures the behavior of policy-makers "best" in an empirical setting.

Given the type of policy-maker, the decision making problem can be described as the minimization of a specific form of equation (2). The economic and institutional constraint to which this minimization problem is subjected is the following:

The above, of course, is the accounting identity that expenditures equal receipts. To capture the distribution of foreign aid and domestic revenues into budgetary categories we instead write,

and,

(1 - r _R), (1 - r _B), and (1 - r _M) are the fractions of domestically raised revenues, bilateral aid and multilateral aid, respectively, allocated to government development expenditures. These two constraints reflect alternative uses of government revenues augmented by foreign assistance. The first constraint allows for the possibility that D can be financed partly by domestic revenues and partly by different sources of foreign aid. The second constraint assumes that domestically raised revenues, and foreign aid not used for development purposes, go towards nondevelopment government expenditure. The model thus involves a trade-off between development and other spending by the government. It is a theoretical model of the implications of recipient preferences that can be used to determine the fiscal behavior of the government in the presence of foreign aid.

Solving the constrained loss minimization problem leads to a set of nonlinear simultaneous equations. The direction and extent of the impact of bilateral and multilateral foreign aid on N and D can be estimated.

In sum, the purpose of this model is to determine (a) what effect aid has on the development efforts and fiscal behavior of the recipient; and (b) to what extent the type of donor makes a difference. In determining the effect of aid ([a] above) the type of policymaker in the recipient country turns out to be crucial. I now turn to a discussion of these issues with reference to the MIT economies.

3.Foreign Aid and the MIT Economies

What role did aid play in the economic development of the three Southeast Asian countries? Tables 1, 2 and 3 describe the quantitative dimensions of aid flows to these countries during the 1970s and 1980s. Figures 1 and 2 show graphically the total flow and compositions respectively for Malaysia. Figures 3 and 4 show the same information for Indonesia. The situation for Thailand is depicted in Figures 5 and 6 in a similar fashion.

Of the three countries, Indonesia seems to have received more aid in absolute terms. However, what is similar in all three cases is the increase in non-aid foreign capital flow in the late 80s and early 90s. This is a luxury that African countries did not have and probably will not have in the near future, inspite of their desperate attempts to attract foreign private capital. Thus the first observation, in a comparative sense is that aid will be more important for the African countries than it has been for the MIT economies. The proper use of this aid, as I will argue, can make the difference in the coming years between growth and stagnation.

But how important has aid been to these Asian economies? A study of Indonesia which is an oil-rich — and generally rich in natural resources — is instructive. Table 4 shows the current account situation for the Indonesian economy in the 70s and early 80s. Inspite of earning revenues from oil which

Table 1

Table 2

Table 3

Table 4

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

increased in price during this period, Indonesia was running a current account deficit. Figure 7 shows the budget revenues and expenditures as a proportion of GDP from 1969/70 to 1983/84. Foreign aid made up an (slightly) increasing portion of the shortfall in budget from domestic revenues alone. Later we will discuss whether or not the aid flow dampened the revenue raising efforts of the Indonesian government. For the moment, however, the main point to note is that the volume of aid did enable the government to relax the constraints that are part of the two (or three) gap models.

One important consequence of the steady flow of aid was that it ensured macroeconomic stability without major pressure on foreign exchange reserves. However, it also ensured dependence on aid. As a recent book on the Indonesian economy puts it:

It is also not clear what precise effect foreign aid per se had on development expenditures from the raw data. But it is clear that in some ways aid did play a key role. During the oil boom period, aid funded a smaller but still significant part of the development budget. However, it is during the time when oil prices were low that the role of aid in the development budget became truly remarkable. By late 1980s, the part of development budget financed by aid rose to more than 70 percent with the amazing figure of 81.6 percent in 1988. In the 1990s it became less than 50 percent following strong growth and relatively successful tax reform; still the significance of aid for development financing is quite readily apparent.

At least as significant during this period has also been the flexible manner of delivering the aid. There was for example a shift back to program aid. Some financing of local "rupiah items" was also made possible. Program aid made up over 50 percent of the total before the oil boom. The percentage fell to near one in early eighties. After the mid-1980s crisis, it rose quickly to 34 percent. Benedict Anderson has called this aid dependence the annual "IGGI fix".

The role of aid, it can be stated without going into repetitious details, has been similar for both Malaysia and Thailand. Without the benefit of oil revenues, both these countries have made up shortfalls in their development budgets at various times ranging from 25 percent to 75 percent. However, aid dependence has been less in absolute terms for both of these countries. This was the result of deliberate policy as well as success in attracting foreign investment. After the recent financial crisis Thailand, in particular is a big recipient of IMF loans (as is Indonesia). I now turn to the question of the macroeconomic impact of the aid to these three countries on their development expenditures. On the revenue side, the impact of aid on the domestic revenue raising efforts will also be discussed. As mentioned in the previous section, these exercises can be done by estimating the model presented before by econometric methods.

It is important to remember that policy-makers work with actual budgetary data, and not with theoretically defined entities. In the budgets of these countries, however, a distinction is made between development and non-development expenditures. It has been estimated that a large component of the development expenditures is actually non-investment expenditure. By and large, public administration and defense claim the lion’s share of non-development expenditures. If one includes subsidies for food and other items given to the military, the figure may indeed be even higher.

Development budget includes expenditures on education, health, housing and social welfare. These expenditures are counted here as genuine development expenditures, since they are directly or indirectly related to well-being of the people and human capital formation. Needless to say, public investment is counted as an important component of development expenditures.

As mentioned in the previous section, the "boundedly rational" nature of the policy-makers means that the chosen indicator levels of budgetary targets, are not exact but are only roughly accurate. Since there is very little empirical evidence of the policy-makers’ actual chosen indicator levels for these targets, it becomes an important problem to estimate these. The planning documents are not adequate since they are drawn up at infrequent intervals and represent longer term targets. The categorizations are also different from those required by the approach adopted here. Therefore I try to approximate the chosen indicator levels by regressing the actual ex post values on a series of instrumental variables and then forecasting the indicator values. As Sargent has recently pointed out in the context of rational expectations, the economist or the econometrician actually works in a bounded rationality sense when predicting these values from models such as the ones I have used.

Each indicator level is estimated by specifying an equation relating the actual variable to some instruments. I then regress the actual variable on the chosen instruments (with correction for auto-correlation). Planned D is obtained by estimating an equation where D is a linear function of GDP and total gross domestic investment in the private sector together with proxies for investment in human capital. The fitted values of the dependent variable serve as indicator levels. Planned R is found in a similar manner, by regressing R on GDP and lagged imports and then using the fitted values of the dependent variable as the indicator value. Planned N is obtained by regressing N on the lagged value of itself.

According to the theoretical approach adopted here, the policy-makers respond to the availability of foreign aid by reallocating money to the various budgetary categories. Although the model assumes bounded rationality, the reallocation itself is in response to additional amounts of foreign aid and is therefore in keeping with allocation at the margin. My major concern here is to examine the allocation of finance to development and non-development expenditures. An additional area of interest is the impact of aid on domestic revenue raising.

The econometrics and other technical details related to the estimation of aid impacts on the MIT economies have been described elsewhere. Here the empirical results are summarized for Indonesia, and where relevant for Malaysia and Thailand.

The data set comprises of Foreign Aid to Malaysia, Indonesia and Thailand from 1970-1996. This is the period when all three economies took off. This is also the period — for Indonesia in particular — when the New Order Government under Soeharto undertook successive development efforts in Indonesia. In what follows, the Indonesian case study is described in somewhat greater detail than those done for Malaysia and Thailand. The results are broadly similar for all three countries. In addition to the aid data, the annual fiscal statistics on revenues and expenditures were also collected both from Indonesian and non-Indonesian sources. Among Indonesian sources are the documents of BPS (the central bureau of statistics) and Bank of Indonesia (annual reports). Indonesia Source Book from the National Development Information Office also served as a source of information. After reconciling the statistics from various sources, all the data were converted to constant Rupiahs at 1980 purchasing power parity prices.

For the purpose of estimating and interpreting the model correctly, the fact mentioned earlier that the policymakers work with actual budgetary data and not with theoretical entities we have in the model becomes relevant. A translation between the two modes is necessary. Fortunately for our purpose, however, the Indonesian budgetary categories do correspond to Development and Non-development expenditures to a large extent. All the published categories such as Agriculture and Irrigation, Industry, Mining and Energy, Transportation, and Communications, Public Works and Transmigration, Education, Health and Family Planning can be used directly, Local and Regional Development and expenditures also occur as a separate category. There is a large "other" or residual category. After discussion with the Indonesian scholars and officials, it was decided that part of this "catch-all" category, in fact, caught some "non-development expenditures." It was estimated to be between 25% and 40%. After further discussions and checking (a very time-consuming process) with the Ministry of Finance and BPS officials an estimate of linkage to non-development expenditures was arrived at for each year between 1970 and 1996.

On the revenue side Development Funds including Project Aid are clearly marked off from the other items. The flow from income, value added tax, excise and import tax receipts constitute the major sources of government tax revenues. The tax-collection system was standardized and modernized as a result of the post-1983 reform program. Corporate and personal income taxes are now set at the top marginal rate of 35 percent on annual incomes above Rp. 50 millions. Tax revenues have risen in recent years. However, a large part of revenues has traditionally come from the oil and gas sectors. In 1987 Indonesia was the lowest taxed nation in Southeast Asia with a tax to GDP ratio of 9.1 percent. By 1990, the ratio rose to 12.5 percent.

How has foreign aid influenced the fiscal behavior of the Indonesian policy makers?

The results of the empirical exercise for Indonesia are given in Table 8. The structural equations in section 2 contain parameters r _R, r _B and r _M

by way of constraints (3) and (4). These three parameters show the nondevelopment expenditure responses to an increase in domestic revenues, bilateral aid, and multilateral aid respectively. In the table, estimates for these three parameters together with some others are shown for the eight different models describing eight different policy-maker types ranging from fiscally liberal to fiscally conservative, statist to non-statist and developmentalist to non-developmentalist as shown in Table 9. After some general observations, I have chosen to discuss two cases in detail for illustrative purposes. Others can be interpreted following a similar approach.

Looking across the rows in Table 8, it is striking that for both developmentalist and non-developmentalist types of policy-makers both types of

Table 8

Table 9

aid matter; however, bilateral aid seems to have had a greater impact than the multilateral aid in almost every case on development expenditures. According to the classification adopted here Types I-IV are the non-developmental policy-makers and Types V-VIII are the developmental ones. It is also interesting to see the difference between the two types. The co-efficients (with varying degrees of significance) r _B vary between .6123 and .7393 for models I-IV. That means that in the presence of bilateral aid approximately 26 to 39 percent of this aid goes to development expenditure on the margin if the policy-maker is non-developmental. On the other hand, from models V-VIII, the corresponding

percentage of aid going to development expenditures is between 67 and 53 percent. For models I to IV, r _M varies between .6454 and .8219. For models V

to VIII, the range is between .4581 and .7235. Thus in terms of influencing development expenditures in Indonesia Rupiah for Rupiah bilateral aid has been more successful than the multilateral aid. In addition to revealing the influence of bilateral aid, the above co-efficients also indicate that the type of the policy-maker really can make a difference. This is also true in terms of financing development expenditures out of domestic revenue. For a non-developmental policy-maker r _R varies between .7235 and .8221. Rather dismally, this implies that between 72 and 82 percent of domestic revenues may go to non-development expenditures in the presence of aid.

What kind of policy-makers did make the decisions in Indonesia regarding development? This is a particularly fascinating question, but is hard to answer in a definitive fashion. The "best guess" one can make must use a great deal of reliable institutional history. In case of Indonesia, this is largely unavailable. The books and articles written on this subject deal at best with particular episodes. On the whole, however, a picture of at least partial commitment to genuine development objective emerges. This is also consistent with my own visits to Indonesia and extensive investigations with the Indonesian and non-Indonesian academics and development practitioners on the subject.

I am also able to offer some econometric evidence to corroborate the above characterization. In Table 8, the last column presents the value of the Akaike Information Criterion (AIC) for each of the eight models. AIC is a model selection criterion that can be applied to any model that can be estimated by the maximum likelihood method. One simply minimizes (2LogL)/n + 2k/n where k=the number of parameters in the likelihood function L and n is the number of observations. Particularly for a non-linear model the AIC is a convenient econometric discriminator among different model specifications. It would seem that by this criterion at least type VII policy-maker model may be the most appropriate one for Indonesia during the period of observation. This means that both developmental and statist concerns dominated the real fiscal agenda during this period. This too, seems to be consistent with the institutional studies and my own informed observations.

Let us consider then the type VII policy-maker first. According to the typology in Table 9 this is further a fiscally liberal policy-maker. All the r 's are positive and significant at .05 level. In the presence of foreign aid almost 50% of the additional revenue goes to non-development expenditures. For bilateral foreign aid the percentage going to development expenditures is 67% whereas 54% of aid from all other sources is spent for non-developmental purposes. Thus, a straightforward interpretation would have been to claim the superiority of bilateral aid over other aid in this case. However, some caution is required. We do not know if the presence of aid pulls some money out of the domestic revenue to non-development purposes. It is reasonable to suspect that for some categories of aid (for both generally Japanese and other aid) this may be partially the case. Under these circumstances, if the substitution effect is not too high, (i.e. aid doesn't replace completely development expenditures that would have been financed out of domestic revenues) only then there is an incremental effect of aid on development expenditures. Under this scenario, bilateral aid would seem to be more effective Rupiah for Rupiah than other aid. I show next that in case of Indonesia this may be a reasonable conclusion.

The ratios of the parameters from the loss function (the a 's) can be readily interpreted by referring to the structural equations. In the simultaneous equations framework, given the specific objective function and constraints, the ratios of a 's (e.g. a _D/a _R or a _N/a _R) indicate how to explain the changes in domestic revenue in the presence of foreign aid. For the type VII policy-maker both a _D/a _R and a _N/a _R are significantly different from zero. The interpretation of the first of these coefficients is as follows: in the presence of foreign aid any increase in development expenditures reduces the domestic revenue raising effort. The quantitative magnitude is given in a non-linear fashion by the product of this coefficient and (1-r _R). However raising the target for development expenditures even with aid coming in will lead to an increase in R. The coefficient a _N/a _R also gives an estimate of (partial) impact of non-development expenditures on R. In this case, an increase in non-development expenditures also leads to an increase in R. Also, this magnitude is further increased by the magnitude of R*. Thus, bureaucratic or political decision to increase R* will lead to an increase in revenues as well. We may call the above description the aid-dependent revenue effect.

If aid-dependent revenue effect is positive, then the presence of aid actually increases domestic revenue. In the case of Indonesia for model VII type of policymaker this will be true. Let us now turn to the model which has the least AIC value among the rest; this is model VI. As can be seen from Table 2 this is the developmental, non-statist and fiscally conservative type.

Looking across the row under the headings for the various parameters the contrast is indeed quite reassuring empirically. More than 52% of the domestic revenue goes towards development expenditures even in the presence of foreign aid. The coefficient is significant both statistically and economically. Out of bilateral aid, again in a statistically significant sense, about 61% goes to development expenditures. Of the other aid receipts about 55% goes to development expenditures. Thus, a major hypothesis of this study is verified: the more developmental the orientation of the policy-maker, the more foreign aid influences spending in the direction of development. It also corroborates the earlier finding that bilateral aid performed well in general.

Turning now to the other coefficients a _D/a _R and a _N/a _R have absolute values of .0213 and .0235 and both are statistically significant. Looking at the revenue equation for this type of policy-maker it is possible to see that the negativity of a _D/a _R (estimated) implies that revenue increases as indicator levels of development expenditures increase although the rate of increase is quite slow. This is consistent with a developmentalist but fiscally conservative preference. Aid finances development expenditures more than domestic revenue raising efforts. In the absence of aid, such expenditures may drop dramatically. Non-development expenditures also lead to an increase in revenue raising. This is consistent with a balancing the budget fiscal conservatism. It also suggests that foreign aid is only marginally diverted to non-development expenditures when finance is needed. It is more likely that domestic revenues are increased more than proportionately to cover these non-development expenditures.

From the discussion of the two cases, it would seem that developmental, statist Indonesian policymaking environment contributed to the salutary effects of foreign aid. Whether the policymakers were fiscally conservative or liberal may not have made that much difference. If we go by the evidence of budget deficits model VII would indeed seem to be the right model and my earlier observations would be strengthened.

These results are very much at variance with the received wisdom on the effect of foreign aid on public expenditures. Aid may have been more effective in Indonesia because of the links with infrastructure investment. It may also be the case that the microlevel projects are more successfully managed through technical cooperation. There is some evidence for this (Khan, forthcoming).

These results also confirm Howard Pack and Janet Pack's findings with regards to fungibility of aid in Indonesia (Pack and Pack, 1990). Their study did not separate out bilateral aid. But they find "that in the largest categories, aid is spent for which it is given." They also find as in some of the models discussed above that taxes are raised in the presence of aid. In the present study, the suggest that a drive for independence leads to the raising of taxes. It could also be the case that some aid flow requires matching funds (Cashel-Cordo and Craig, 1986; Booth, 1988).

In case of Malaysia and Thailand the results are broadly similar to Indonesia. The parameter values are given in Tables 10 and 11. In both the

countries aid led to more development expenditures. Like the Indonesian case the flow of aid seems to be also associated with an increased governmental effort to raise tax revenues as well. Also, bilateral aid flows seem to perform better than their multilateral counterpart.

The effectiveness of bilateral aid flows seems to be connected largely to Japanese bilateral aid. This has been demonstrated in a series of papers (Khan [1995a, b, c]). The main reason appears to be the infrastructure projects financed by Japanese aid. Since Japanese bilateral aid is a dominant share of all bilateral aid in these economies, these results are perhaps not altogether surprising.

However, in considering the effectiveness of aid — both multilateral and bilateral — we need to go beyond infrastructure, or even public investment in general. Here certain weaknesses emerge in the MIT economies. These weaknesses have largely to do with the shortfalls in human resource development in these countries. While a detailed discussion of this is beyond the scope of this chapter, Figures 8 and 9 illustrate part of the problem. Figure 8

Table 10

Table 11

Figure 8

figure 9

is a regression of gross secondary enrollment rates against per capita income of selected Asian economies. Both Malaysia and Thailand fall below the expected rates of secondary enrollment given their GNP/capita while Indonesia is only slightly above the trend. The quality of education also leave a great deal to be desired. Figure 9 shows the status of research and development capabilities. Here, surprisingly the MIT economies lag behind poorer Asian countries such as Pakistan and India. Although it is difficult to establish direct causality between aid and human resource and technology development from this data alone, it would seem that aid has either not been used, or it has not been used effectively to promote human resource development and innovation. In light of the state strategies of these countries to develop further through exports and given the vulnerability of their current export-mix to lower wage economies such as PRC, it seems doubtful that these economies will be able to diversify their exports towards higher value added activities with such poor human resource and technological base.

Another aspect of foreign assistance which is directly connected to multilateral policy-based lending is the structural adjustments undertaken by the MIT economies in the 1980s. Thailand and Indonesia adopted SAPs early in 1980s (1982/83). Indonesia’s second adjustment period followed in 1985-86. Malaysia undertook privatization policy in 1983/84, and adjustment and liberalization in 1986 after the severe 1985-86 recession. At the time the assessment of these reforms by both IMF-World Bank and domestic experts (a few dissidents aside) was highly positive. However, with the recent financial debacles considerable skepticism has developed with respect to the liberalization of financial sector in particular.

Part of the structural adjustment for Thailand in particular also involved reorganization and realignment of functions of development institutions such as the NESDB. In case of Indonesia, the key role of BAPPENAS in aid management was attenuated somewhat, although it continued to play a powerful role along with the Ministry of Finance and other functional ministries in the Indonesian government. Perhaps Malaysia was the only country where there was far less a gap between the rhetoric of shared growth (partly through external assistance) and the reality through the New Economic Policy.

To the extent aid has been successful in these economies, this has been largely the result of a conscious allocation of aid through the government’s budget towards development expenditures. In a stable macroeconomic environment, with the help of supporting governmental structures, aid could be used to finance development projects that may not otherwise have been financed. Yet the results discussed here suggest that perhaps both multilateral and bilateral aid could be better utilized especially for human resource development.

4.Lessons for Africa

In a recently published book on Asia and Africa, the editors explain:

Although in my view many African countries — the poor, highly indebted ones in particular — are facing initial conditions (internal and external) that are much less favorable than the editors of the above volume claim, I do subscribe to the view that there is much the African countries can learn from the experience of the Southeast Asian countries. However, in the area of foreign aid in particular, the lessons are both positive and negative.

On the positive side the most important lesson is to have some factors complementary to foreign aid in place so that foreign aid can lead to the greatest possible development impact. As many of these factors may not be present in large quantities, both foreign aid and domestic resources may need to be mobilized first to create some of these. A kind of "strategic bunching" may be necessary where infrastructure, human capital and technology adaptation may all need to be undertaken at once so that a critical mass of FACE is created. Further injection of aid then can be expected to lead to a greater impact than the one produced initially.

The above strategy also calls for not evaluating the initial lack of immediate impact of aid on certain macro variables including growth rate too harshly. Instead, we must ask: are complementary factors to aid being created at a sufficiently rapid pace? If not, then the effort to build a FACE should be intensified.

It also follows then that human resource may indeed be a crucial policy area. Here the lesson to be learned from the MIT experience is largely negative. Instead of following a haphazard and undirected path, the African countries should plan carefully for their human resource needs. A part of the foreign aid receipts should be earmarked for HRD programs and projects.

Just like in MIT economies in the 80s trade — exports in particular — should accompany aid. Without markets in developed countries, aid-dependence will become a chronic disease. As Khan (1997a) has advocated with respect to U.S. policy towards Africa, the possibilities for regional integration along the ASEAN model may also be considered:

In Africa foreign aid needs also to be directed toward helping the small producers through providing funds and expertise. Some aid should be used to foster human development through the African NGOs. Such NGOs should also be given access to the global communications networks. By interacting with progressive groups outside and allocating funds to projects, at the grassroots level such NGOs can play a crucial role in future African development.

5.Conclusions

Perhaps the central fact of African development is the failure of the Sub-Saharan African (SSA) countries to grow. At the beginning of 1990s per capita GDP was almost at the same level as during the early 70s. This failure to grow has naturally meant a stagnation — indeed a deterioration — in the standard of living of the majority of the people in SSA. These economies have also been extremely vulnerable to terms of trade shocks and environmental deterioration.

In addition to economic stagnation and environmental decay SSA has also been subject to persistent macroeconomic imbalances. These imbalances have included on seemingly ineradicable malaise of a perpetual balance of payments deficits that tends towards a steady deterioration. In fact, SSA’s current account deficit during 1973-1992 averaged 23 percent of export earnings — a far higher proportion than any other region.

It is clear that an appropriate use of foreign aid must include redressing the macroeconomic imbalances. But at the same time as this is done and growth is promoted, environmental decay must be checked as well.

Given the dismal development record, it is not surprising that Africa falls short of being able to provide its people with adequate resources to have even the basic capabilities and functionings. (The interested reader is referred to the apendix for an extended discussion of capabilities and functionings). Moreover adding insult to injury, the virulent spread of acquired immunodeficiency syndrome (AIDs) threatens to undermine the slight progress that has been made in such areas as life expectancy and basic health care provision.

It is in this desperate context that external assistance becomes much more crucial to Africa than it was for Southeast Asia. Because the extent of capabilities failure is so alarming in Africa, there is both economic scope and a strong moral case for direct intervention to improve the basic capabilities. By preventing hunger and malnutrition, providing people with basic health care and spending on creating a literate, educated population the foundation for long-run growth can be laid. However, an egalitarian distribution of the benefits of aid — in the sense of equalizing capabilities — is a sine qua non for this to happen. Thus the top heavy aid bureaucracy is not an appropriate way to distribute aid to Africa. A much more grass roots based — perhaps primarily through reputable international and domestic NGOs — and accountable institutional setting is necessary. Justice for Africa demands urgently this type of response both from the donors and the African recipient countries. Only this altered manner of giving and receiving aid for capability-enhancement in Africa will lead to a real difference between the past and the future performance of foreign aid.

APPENDIX

Capabilities and Functionings

If we define development as the creation of capabilities of individuals to lead their lives in certain ways then it is possible to argue for a positive relationship between certain types of aid and development

Capabilities can be construed as general powers of human body and mind that can be acquired, maintained, nurtured and developed. They can also (under circumstances such as malnutrition or severe confinement) be diminished and even completely lost. I have emphasized elsewhere the irreducibly social (not merely biological) character of these human capabilities. Sen himself emphasizes ‘a certain sort of possibility or opportunity for functioning’ (Crocker in Nussbaum, 1995, p. 162).

In order to assess the future of African development a capabilities perspective we need to go further and try to describe more concretely what at least some of the basic capabilities may be. David Crocker has given an admirable summary of both Nussbaum’s and Sen’s approach to capabilities in a recent essay. Mainly relying on Nussbaum but also on other sources he has compiled a list that is worth reproducing here:

The social capability approach outlined above emphasizes the positive freedom to choose a good life. Clearly then, such an approach to development must necessarily entail democratic freedoms as well. It is impossible to elaborate on all the aspects of both well-being and agency freedoms that such a concept of development must encompass. But in case of Africa even the short list given above makes it clear how far the continent is from providing its people with the necessities for life with freedom and dignity.

Bibliography

Bacha, Edmar L., 1984, "Growth with Limited Supplies of Foreign Exchange: A Reappraisal of the Two-Gap Model" in Moshe Syrquin, Lance Taylor, and Larry Westphal (eds.) Economic Structure and Performance: Essays in Honor of Hollis B. Chenery. New York: Academic Press.

-----, 1990, "A Three-Gap Model of Foreign Transfers and the GDP Growth Rate in Developing Countries." Journal of Development Economics, 32: 279-196.

Bank of Indonesia, Annual Reports.

-----, 1991, The Development Frontier: Essays in Applied Economics, London: Harvester/Wheatsheaf.

Bauer, Peter T. 1984, Reality and Rhetoric: studies in the economics of development. London: Weidenfeld & Nicolson.

Biro Pusat Statistik (BPS), Jakarta, National Income Accounts, various years; Social Accounting Matrices, various years.

Binh, Tran-Nam and Mark McGillivray, 1993, "Foreign Aid, Taxes and Public Investment: A Comment," Journal of Development Economics 41, 173-176.

Booth, Anne, 1988, "Survey of Recent Developments," Bulletin of Indonesian Economic Studies, Vol.24, No.1, pp.1-26.

Browne, Stephen, Foreign Aid in Practice, New York: New York University Press, 1990, pp.193-95.

Bulletin of Indonesian Economic Studies, Economic Surveys, various years, Canberra:Australia.

Carneiro, Dionisio, and Rogerio Werneck (1987). "External Debt, Economic Growth, adn Fiscal Adjustment." Rio de Janeiro: Departamento de Economia, Pontificia Universidade Catolica.

Cashel-Cordo, Peter and Steven G. Craig, 1992, Donor Preferences and Recipient Fiscal Behavior: A Simultaneous Analysis of Foreign Aid Fungibility and the Flypaper Effect, Manuscript, Canisius College, Buffalo, NY 14208.

Chenery, Hollis B., and Alan M. Strout, 1966, "Foreign Assistance and Economic Development", American Economic Review, 56:149-179.

Cline, William R., 1979, "Resource Transfers to Developing Countries: Issues and Trends" in W.R. Cline (ed). Policy Alternatives for A New International Economic Order. New York: Praeger.

Economic Commission for Latin America and the Carribean (ECLAC), 1990, Latin America and the Carribean: Options to Reduce the Debt Burden. Santiago

Europa World Yearbook (London: Europa Publications Ltd.), various years.

Fanelli, Jose Maria, Roberto Frankel, and Carlos Winograd, 1987, "Argentina" Stabilization and Adjustment Policies and Programmes Country Study No. 12. Helsinki: WIDER.

Frey, Bruno and Friedrich Schneider, 1986, "Competing Models of International Lending Activity," Journal of Development Economics 20, 224-245.

Gang, Ira N. and Haider Ali Khan, 1989, "Modelling Foreign Aid and Development Expenditures," paper presented at AEA conference, Atlanta.

-----, 1990, "Some Determinants of Foreign Aid to India, 1960-1986," World Development 18, 431-442.

-----, 1991, "Foreign Aid, Taxes and Public Investment," Journal of Development Economics, 34, 355-369.

-----, 1993, "Reply to Tran-Nam Binh and Mark McGilliray, "Foreign Aid, Taxes and Public Investment: A Comment," Journal of Development Economics 41, 177-178.

-----, 1994, "Foreign Aid and Development Expenditures: Does the Policymaker Make Any Difference?" unpublished paper.

The Government of Japan: Ministry of Foreign Affairs, Waga Kuni No Seifu Kaihatsu Enjo, 1990, Tokyo.

Griffin, Keith B., 1970, "Foreign Capital, Domestic Savings, and Economic Development." Bulletin of the Oxford University Institute of Statistics, 32: 99-112.

Helleiner, G. K., 1986, "Balance of Payments Experiences and Growth Prospects in Developing Countries: A Synthesis." World Development, 14: 877-908.

Heller, Peter S., 1975, "A Model of Public Fiscal Behavior in Developing Countries: Aid, Investment and Taxation," American Economic Review 65, 429-445.

Holt, C. C., 1962, "Linear Decision Rules for Economic Stabilization and Growth," Quarterly Journal of Economics 56, 20-45.

The International Bank for Reconstruction and Development (World Bank), Bangladesh: Recent Economic Developments and Medium Term Prospects, Vol. 1, March 1986; The World Development Report, 1986, 1987.

Islam, Shafiqul(ed.), Yen for Development, New York, Council on Foreign Relations Press, 1991.

Johansen, Leif, 1960, A Multi-Sectoral Study of Economic Growth. Amsterdam: North-Holland.

Keuning, Steven J., 1984a, "Distributive Aspects of Indonesian Agriculture: With Respect to Small-holders Land and Farm Food Crops," Institute of Social Studies Working Paper Series No. 2.

-----, 1984b, "The Distribution of Profits in Indonesia," Institute of Social Studies Working Paper Series No. 3.

-----, 1984c, "Modelling the Indonesian Social Accounting Matrix," Institute of Social Studies Working Paper Series No. 1.

-----, 1998, Technology, Development and Democracy, Aldersho & U.K.: Edward Elgar.

Khan, Haider Ali, 1997a, African Debt and Sustainable Development: Policies for Partnership with Africa, New York: Phelps-Stokes Fund.

-----1997b, "Japanese Foreign Aid, Development Expenditures and Taxation in Thailand", Unpublished Paper.

-----, 1997c, "Japanese Foreign Aid Development Expenditures and Taxation in Malaysia, Unpublished Paper.

-----, 1997d, "How Effective is Japanese Aid? Econometric Results from a Bounded Rationality Model for Indonesia", Unpublished Paper.

-----, 1995a, "Does the Policy-Maker Make a Difference?" paper presented at AEA/ASSA meetings, Washington D.C., January 1995.

-----, 1995b, "Does Japan's Aid Work?" unpublished paper, University of Denver.

-----, 1995c, "The Impact of Japanese Aid in a Bounded Rationality Model: The Indonesian Case", Unpublished Paper.

-----, 1995d, "Economic Justice and Democracy", Unpublished Paper and E. Thorbecke, 1988, Macroeconomic Effects of Choice and Diffusion of Technology: The Indonesian Case, Aldershot, U.K.: Gower Publishing Co.

Khan, Haider Ali and Elichi Hoshino, 1992, "Impact of Foreign Aid on the Fiscal Behavior of LDC Governments," World Development 20, 1481-1488.

Kumssa, Asfaw and Haider Ali Khan, 1995. "Foreign Aid and the Fiscal Behavior of the Recipient Government: The Case of Kenya." Regional Development Studies, Vol. 2, winter, pp. 157-173.

Lensink, R. and P.A.G. van Bergeijk, 1991, ‘Official Finance Requirements in the 1990s, World Development, vol. 19, no. 5, pp. 497-510.

Levy, Victor, 1988, ‘Aid and Growth in Sub-Saharan Africa: The Recent Experience’, European Economic Review, Vol. 32, Nov., pp. 1777-95.

-----, 1987b, ‘Does Concessionary Aid Lead to Higher Inverstment Rates in Low-Income Countries?, Review of Economics and Statistics, Vol. 69, pp.152-56.

Lindauer, David and Michel Roemer, 1994, Asia and Africa: Legacies and Opportunities for Development, San Francisco: ICS Press.

Lipton, M. and John Toye, 1990, Does Aid Work in India? A Country Study of the Impact of Official Development Assistance, Routledge.

Maizels, Alfred and Machiko K. Nissanke, 1984, "Motivations for Aid to Developing Countries," World Development 12, 879-900.

McCawley, Peter, 1983 "Survey of Recent Development, Bulletin of Indonesian Economic Studies, April 1983.

-----, 1981, "The Growth of the Industrial Sector", in Anne Booth and Peter McCawley, (Eds.), The Indonesian Economy During the Soehart Era, Oxford University Press, Oxford, 1981.

Mosley, Paul, 1987, Overseas Aid: Its Defence and Reform, Brighton: Wheatsheaf.

d:\pats\aid1.doc

Return to the previous page

Return to the UNU Homepage