Previous - Next
7.5 Lessons and policy implications: short-run adjustments and long-term strategies
Policies for the short and medium term
The implication of the analysis in the previous sections is that a "proper" macroeconomic environment (price stability, a competitive exchange rate, an interest rate that reflects the balance between credit demand and savings) can greatly contribute to improving the conditions for agricultural growth. The main conclusion from reviewing past experience is that piecemeal solutions cannot work as crises are usually the result of not one but many policies, often working at cross-purposes. A consistent set of policies is necessary to avoid crises or reverse them. For instance, demand contraction without devaluation in the presence of inflexible prices and wages, may cause a crisis, in countries suffering from reversals in foreign capital inflows. On the other hand, real devaluations are not sustainable without proper monetary and fiscal policies. The case for a wide flexibility in the exchange rate policy is gaining ground. However, devaluation, even when necessary, cannot substitute for the need to achieve, through productivity increases, long-term improvement in competitiveness. Despite the fact that devaluation may give a boost to agriculture by restoring the price incentives of tradables, and may provide short-term relief in the balance of payments, long-term performance of the agricultural sector and the economy is contingent upon real productivity gains.
Box 7.2 Determinants of success of policy reforms: the case of South Asia and sub-Saharan Africa
The variety of experiences with (apparently similar) policy reforms warrants a closer look at the socioeconomic and/or political factors determining consistency in reform implementation and success in achieving growth. In a paper written for FAO (Subramaniam et al., 1994) a comparison was made between a "typical" poor Asian and a "typical" poor African country (hereafter referred to as Asia and Africa) to show how differences in the economic and institutional structures may determine the nature of adjustments needed as well as reform outcomes. The major results are summarized below:
1. In Africa reforms were more often initiated not by the countries themselves but imposed by lending agencies. They were general and uniform rather than country specific, and often they ran contrary to the interests of the ruling coalitions. Reforms in Asia were often initiated by the countries themselves, were more sensitive to the local needs and conditions, invited less opposition and were politically more sustainable. Therefore their implementation was more consistent.
2. Short-run political sustainability was hampered by the frequency and magnitude of the external shocks to which the African economies were subjected, and the ensuing economic hardships that prevented the compensation of the "losers" from the reforms. The inability to provide compensation hindered consensus.
3. Despite the fact that the "post adjustment" or "redefined" role of the state was similar for regions (providing public goods, dealing with externalities, reducing transactions costs, assuming social functions) several differences exist: The required reduction in the role of the state was smaller in Asia than in Africa, since in the latter region the pre-reform state intervention was more extensive. As a result of these differences, agencies supplying public goods were less disrupted by the reforms in Asia than in Africa.
4. In agriculture, private marketing networks and indigenous rural institutions were weaker and more suppressed in Africa where the mistrust for the private sector was greater than in Asia. The implication is that in the "typical" Asian country the private sector was able to step in and assume functions previously in the domain of the state.
5. Growth in the typical African country is hampered by the undeveloped state of financial markets. The weaker fiscal system makes the African economy more dependent on commodity taxes for collection of fiscal revenues. The relative inability of the African state to collect revenues contrasts with the finding that the marginal impact of additional infrastructure in terms of higher agricultural productivity is larger in Africa than in Asia.
The question whether, in principle, a system of fixed or flexible exchange rates is superior is not the right one to ask. The real issue is whether monetary and fiscal policies are supportive to the exchange rate system chosen. The reason why flexible exchange rates may be preferable derives from the fact that fiscal and monetary policies to support a fixed exchange rate are almost untenable in practice, especially in the presence of external shocks. For countries constrained in terms of their ability to adjust the value of their currencies, a fiscal policy that avoids overvaluation and prices and wages that are flexible to respond to disequilibria constitute the cornerstones of a supporting macroeconomic stability. The recent decision to devalue the CFA and CF francs (the currencies of the 14 African countries of the franc zone described in IMF Survey, 24 January 1994) demonstrates the difficulties in adhering to the "iron laws" of fiscal and monetary policies needed to prevent real exchange rate over valuation under a "nominal anchor" system, and in the face of external shocks. Namely, while the fixed exchange rate system was a source of stability, low inflation and growth until the mid-1980s, the decision to maintain a fixed parity in the face of reductions in the terms of trade and an appreciating French franc meant that international competitiveness needed to be restored through containment of production costs, and/or with the use of trade measures. Failure to contain costs through internal adjustments caused a collapse of exports, investment and growth and a financial crisis in a number of these countries.
Experiences with adjustments undertaken in the wake of the economic crisis show that several countries used trade restrictions and foreign exchange rationing to deal with balance of payments crises. As allocation of restricted imports and rationed foreign exchange often reflects the ability of groups or sectors to exercise political power, agriculture has mostly lost out in the process.
Within a framework of macroeconomic reforms as described above, funding for agricultural infrastructure will be conditioned by the constraints imposed on fiscal expansion under the reforms. As was mentioned previously, in the short run and until more efficient income tax collection mechanisms are developed, taxation, inter alia, of agricultural exports and imports, may be unavoidable. If such direct taxation of agriculture is to be imposed (or continued) then it is more important that indirect taxation of agriculture in the forms of overvalued exchange rates, etc., be removed. Uniform taxation or border protection although sub-optimal from an efficiency standpoint may be the best way given the implementation difficulties and the possibilities for rent seeking associated with non-uniform measures. Converting quantitative restrictions to tariffs will add to revenue collection. Several developing countries have already made such shifts.
Other sources of funding also need to be sought. Savings from phasing out various unproductive/uneconomic subsidies and from the reform or privatization of state enterprises and parastatals can be used to finance public goods as well. In several countries additional funding can be provided by reallocating overall expenditures among expenditure lines, e.g. from recurrent expenditures to capital spending as the efficiency of the public sector improves.
In addition to an "enabling" macroeconomic framework, liberalization of the agricultural sector will be ineffective unless balanced reforms are undertaken across economic sectors. Promoting efficiency reforms in agriculture by lowering import controls, tariffs or subsidies, but maintaining protection in other sectors, puts agriculture at a distinct disadvantage. Lowering half of the distortions in the economy is different from lowering all distortions by half.
The international prices of some commodities (e.g. cereals, dairy products, sugar) are lower and often more volatile than would be the case in the absence of trade-distorting policies of developed countries. However, in some cases these very policies contributed to enhance security of supplies in world markets, particularly of cereals. This is because they caused stocks held by these same countries to be higher than they would have been in the absence of such policies (O'Brien, 1994; see also Chapter 2).
International market prices represent the actual opportunity cost of resources facing developing countries and could be used, in principle, as the basis for calculating relative protection or taxation of commodities, i.e. as an anchor for calculating distortions. However, such prices are sometimes abnormally low and opaque as a result of heavy, irregular and often not transparent export subsidies. The issue, therefore, is whether such prices should be allowed to penetrate unhindered the agricultural economies of the importing countries because they can disrupt the orderly development of domestic production. The gains from the lower import prices can be offset in varying degrees, or more than offset, if producer incentives are subject to such erratic and artificial signals from the international markets. It is for this reason that several importing developing countries have adopted border measures to filter such signals, e.g. through the adoption of "price bands" for import prices. Such price bands are meant to allow longer term price trends to penetrate the domestic economy, but not the erratic and temporary price movements caused by certain types of export subsidization, both open and hidden ones.
In conclusion, in considering the role of international market prices as opportunity costs influencing incentives to domestic producers, allowance should be made for the higher risks involved in such use of distorted international prices. They are not only lower but also more volatile than they would be without these policy distortions. They are also less transparent (e.g. when export subsidies are not fully reflected in quoted prices) and they may be less durable than price trends generated by market forces, because the policies that lowered them in the first place may be changed at short notice.
Arguments have been advanced in favour of positive protection of (especially food) crops, mainly for improving food security (often wrongly associated with food self-sufficiency) as well as for supporting rural employment and incomes. Some of those arguments in favour of increased food self-sufficiency can be valid also on efficiency grounds if they correct market failures or other distortions. Likewise, if food production has multiplier effects that private producers fail or are unable to take into account in their decision-making process, or are unable to act upon due to, e.g. lack of credit, some public intervention may become necessary to maximize net social welfare (Matthews, 1989; this issue is discussed below). When such policy objectives are pursued, the instruments by which food production is to be promoted need careful consideration, so that they attack the root of the problem. Thus, interventions to increase productivity in the smallholder sector through non price-based measures (such as supportive credit organizations and cooperatives, facilitating technology adoption, etc.) may be more efficient or more cost effective than interventions which affect the structure of price incentives. The possible broader implications of those productivity-increasing interventions for agricultural and overall growth are examined in the next section.
Increased volatility of both import bills and export receipts may lead to occasional national food security crises in low-income, food-deficit countries. In these conditions a greater degree of food self-sufficiency than would be suggested by pure efficiency considerations may be justified to the extent that it reduces the risk element associated with world market price volatility and to the extent that it is difficult to implement alternative policies to cope with this risk, e.g. maintenance of foreign exchange reserves targeted for food imports, use of variable tariffs to compensate for the effects of artificially and erratically volatile world prices due, inter alia, to trade-distorting policies, or resort to world futures or option markets.
As was mentioned previously, deviations of domestic from international prices may be justified to the extent that international prices are the result of unsustainable policies of food exporters (rather than technological improvements), implying a high risk that they will be eventually reversed. This is important, especially if decisions are to be taken on investment projects in the food sector. The rate of return on infrastructural projects of long duration should take into account "sustainable" levels of world prices. Several other arguments may be advanced in favour of increased self-sufficiency over and above what would be under strict efficiency criteria. Such arguments may reflect social and political values and realities in different countries, their validity should be examined on a case by case basis, keeping in mind the income losses or gains associated with such choices as well as whether or not they can be sustained.
From short-term policies to long-term strategies: the role of agriculture in economic development
The trend towards fiscal prudence and market-oriented policy reforms in developing (but also developed) countries is likely to continue. The particulars of reform (timing, sequencing, speed, etc.) may change as lessons are learnt and more data are gathered on performance of countries undergoing reforms. It is now clear that more time has to be allowed than was initially thought, for the reforms to produce results, especially in low-income countries where institutional and infrastructural deficiencies prevail. In such countries the response of agriculture to macroeconomic reforms has been mixed.
In the future, reform programmes will likely continue to pay more attention to alleviating infrastructural bottlenecks, promoting research and extension, providing funding for environmental conservation and for direct measures to alleviate poverty in its different aspects (food security, health, etc.), in addition to restoring equilibria in external and internal balances and in relative prices.
Most of what has been discussed so far relates to short- and medium-term policies. The purpose of such policies is to create the conditions for efficient allocation of productive resources among sectors and within agriculture. Short term policy reforms aimed at correcting economic imbalances and alleviating associated crises, while necessary, may not be sufficient for the resumption of growth. Long-term growth strategies involve going beyond short-term policies, and are based on the identification of those sectors where public investment funding will be given priority.
The severe economic costs incurred as a result of the neglect of agriculture in developing countries in the context of import substitution/industrialization, prompted a closer look at agriculture as a driving force for overall economic development, especially for countries with large rural populations, and a high share of agriculture in GDP, exports and employment. Thus, agriculture-based development strategies have gained new momentum.
Agriculture-based development strategies go beyond the adoption of short term measures that eliminate discriminatory policies against agriculture. They suggest an active support of the agricultural sector in terms of massive public investment allocations, pointing out the broad benefits of such a strategy in terms of improvements in food security and nutrition, an egalitarian income distribution, employment generation and, more importantly, stimulation of overall growth through production (supply) and income (demand) linkages between agriculture and the rest of the economy.
Mellor and Johnston (1984) proposed an "interrelated" strategy which emphasizes the benefits resulting from increases in the productivity of smallholder (mainly food) producers through the development of infrastructure and the dissemination of appropriate labour intensive technologies. Such a strategy would improve the nutritional status of the population by both increasing food production and (more importantly) improving access to food through the secondary effects of increased agricultural production on employment and incomes. Increased smallholder production and incomes have spillover effects on the rural non-farm and urban manufacturing sectors through increased demand for agricultural implements, and consumer goods that are income elastic. Both of those non-agricultural sub-sectors constitute presumably labour-intensive activities and their growth further increases employment. Increased non-farm employment and incomes absorb the increased food production without a collapse in prices. Surplus to industry is transferred through cheaper food and savings. Increasing productivity and incomes of small farms, and the resulting increased farm and non-farm employment, result in improved income distribution. Thus the strategy can generate overall economic and social gains.
A similar development strategy, emphasizing the primacy of the agricultural sector in the development process, has been proposed by Singer (1979) and empirically tested by Adelman (1984) and especially by Vogel (1994). As in the case of the "interrelated" strategy, the proposed agricultural-demand-led industrialization (ADLI) strategy is based on the strong demand linkages between agriculture (especially food production by small- and medium-level farmers) and the non-agricultural sectors. Emphasizing smallholder food production results in increased derived demand for (mainly traditional) intermediate inputs which can be satisfied domestically, in contrast to export agriculture's demand for modern intermediate inputs a large part of which "leaks" into the international economy.
Furthermore, increases in incomes of smallholder producers result in increased demand for domestically produced basic consumer goods, thus providing a stimulus to domestic light manufacturing and further increasing employment. The ADLI strategy is proposed as an alternative to past trade based strategies emphasizing import substitution or export promotion policies. Empirical testing of such strategies for their superiority in enhancing overall growth is subject to many caveats, including those emanating from the correct, or otherwise, representation of the specificities of the agricultural sector in the relevant models, e.g. resource and technology constraints or the functioning of international commodity markets (see Alexandratos, 1992).
Under the "linkages" criterion, the allocation of public investment priority to agriculture is justified when it increases GDP (in the sector itself and, through the demand linkages, in the rest of the economy - principally in the sectors producing inputs and consumer goods for agriculture) by an amount exceeding that obtainable from alternative allocations of public investment. The most comprehensive empirical test of this question to date has been performed by Vogel (1994), using social accounting matrices from 27 countries to derive agricultural and non-agricultural multipliers. The results of the study confirm agriculture's strong backward links to non-agricultural production activities at low levels of development. Also, at low levels of development the rural household income multiplier (i.e. the effect of increases in rural household income on the demand for non-agricultural consumption goods) is the dominant one, while during the development process, the backward agricultural input output multiplier becomes dominant, i.e. the demand for agricultural inputs, including agricultural services and credit, is expanding. The findings indicate that the rural household multiplier declines with development as one would expect.
Whether for correcting years of neglect and discrimination or as a result of an "active" strategy that places agriculture at the leading sector in the development process, agriculture is and will probably continue to be the focus of the development efforts in developing countries, especially low-income ones. The case for an agriculture-led development strategy seems strong for several countries, but care should be exercised not to make sweeping generalizations. The different "models" or strategies giving priority to agriculture were developed in the context of more-or-less specific regional environments, stages of development, and/or global economic climates. For instance the interrelated strategy is based largely on evidence from the successes of the green revolution in Asia. Several countries for which an agriculture-based development strategy seems to have worked (for example Taiwan (province of China), Thailand, Malaysia) are countries that were also promoting an export-led growth strategy. Therefore, it is difficult to determine which element (pro-agriculture or pro-export) was at the basis of the growth successes in these countries.
ADLI strategies, looking at demand linkages, assume that supply elasticities in all sectors are infinite, so no disruptive large price increases take place. This means that success of such demand-based strategies hinges on large supply elasticities which in many countries and/or sectors may not exist and may have to be "created" in the long run through the elimination of various bottlenecks. In a situation where supply elasticities are low, the objectives of the strategy are not met no matter how large the magnitude of the multipliers.
Problems will also be created where there is an asymmetry in the response of the agricultural and manufacturing sectors to demands from the other sector. For instance the stimulation of too rapid an increase in food production without a concomitant increase in employment and incomes in the rural economy (and given inadequate export structures or high transfer costs) may result in collapse of food prices in producing areas unless the excess is removed from the market or area. Alternatively, government stockholding may be needed to keep prices from crashing (Mellor, 1986, reports such a case in India). Such delicate intersectoral equilibria are difficult for most governments to manage and it is almost impossible to plan them ahead of time. In the same fashion, if the supply elasticity of non-agricultural goods is small, then an increase in demand for those goods will cause inflationary pressures or "leakages into imports". ADLI "strategists" suggest measures, including infant industry protection, to counter the problem. As already noted, such policies may result in chronic protection and inefficiencies. The above-mentioned "interrelated" strategy supports a liberal open trading regime even if that means that part of the increased incomes would be spent on imports.
A strategy (e.g. a scheme to allocate public investment resources or recurring development expenditure among sectors) should consider several options taking into account the long-term comparative advantage of the country. Although the "linkages" criterion may be considered, this should not be the only one. Public investment should also be directed where it can attract private investment activity. The distinction between "export promoting" and "domestic market expanding" strategies may be misleading. Even for countries that experience declines or stagnation in their markets for traditional exports, an inward-looking strategy may not be advisable. The potential for non-traditional agricultural exports, for agriculture-based products with higher level of processing, etc., should also be explored.
In general, increasing smallholder food production although desirable due to its possible multiple growth and social effects, should not be pursued at any cost to overall social and economic efficiency, i.e. even when strong comparative advantage exists in other agricultural (or non-agricultural) activities. If such sectors are neglected, the overall social welfare can be reduced substantially. In case an alternative strategy is followed, the "second round" effects of an agriculture/food-based development strategy (such as improvements in nutrition or rural poverty alleviation) may be more efficiently achieved by direct interventions. Although there is little doubt that in most low-income developing countries agriculture can be the most important sector in terms of potential returns (both social and economic) to public investment, caution should be exercised in general in picking "winners". The successful experience of East Asian countries in picking winner industries may not be feasible for a number of countries.
Agricultural growth and the non-agricultural rural sector
Another dimension of the agriculture-based development strategy has to do with the implications of such a strategy for growth through agriculture's links to rural non-farm activities. Non-farm employment constitutes an important share of total rural employment especially where communities are organized in small rural towns. The share varies among developing countries and country regions depending on the concentration mode of the population. Thus, shares are higher in Latin America and Asia where the rural population is more concentrated in rural towns, than in Africa where rural populations are more dispersed in small rural settlements (for more extensive discussion see Haggblade et al., 1989, 1991; Hazell and Haggblade, 1993).
Non-farm rural activities offer employment and income opportunities to the rural poor- landless or near landless-and women, the latter being generally a significant share of total rural non-farm employment (food preparation and processing, tailoring, etc.). As in the case of links between the farm and overall economy, strong production and consumption linkages exist between the farm and non-farm rural economies. Growth in the rural non-farm economy will have positive effects on poverty reduction and income distribution, and will stem the rural to urban migration. According to the available empirical results, the average agricultural income multiplier is around 1.6, i.e. each additional unit of agricultural income generates additional income of 0.6 units in the non farm rural sector. The multiplier seems to vary by developing country and region, being in general lower in Africa than in Asia and Latin America.
Research has demonstrated the importance of rural towns in strengthening the linkages between the farm and rural non-farm sectors and for increasing the multiplier effects of agricultural growth (Hazell and Haggblade, 1993). Thus, for Africa, the lower multiplier effects can be partly explained by the very low concentration of population mainly in small rural settlements and the difficulties in connecting population concentration points. This result is in accordance with what has been presented in Box 7.1, i.e. the importance of rehabilitating and expanding the transport network in Africa for increasing agricultural growth and facilitating the realization of its "secondary" effects.
From a policy standpoint, the implication is that apart from policies that promote growth in the agricultural sector, policies to strengthen the ability of the rural non-farm sector in responding to the demands of a growing farm economy have to be considered. In addition to increasing infrastructural facilities and human resource capacities in the rural areas and small towns, a clear, enabling legislative framework for private small business activity is needed including legislation concerning labour markets. Donor support should consider promoting activities such as training in small business skills (accounting, etc.), technical assistance in workshop facilities and the creation of an efficient credit system in the rural areas.
Support of non-rural farm activities also addresses the important issue of export diversification of developing countries to products with a higher share of value added. A move away from their dependence on raw agricultural commodities, in addition to the growth implications mentioned above, reduces the vulnerability to the wide swings of international commodity prices. Developed countries can contribute to this effort by removing discrimination to imports from developing countries with a high value added share, e.g. through tariffs escalating with the degree of processing.
The above results are relevant in rethinking the process of economic transformation of developing countries from basically agrarian to basically industrial economies. The growth potential of other sectors such as rural commerce and rural services has to be considered along with (or instead of) manufacturing.
Concluding remarks: agriculture-based strategies and the role of the state
In considering the role of agriculture in development, a strategy cannot and should not ignore the fact that economic development leads ultimately to a reduced role of agriculture in the overall economy in the long run. The debate on agriculture-based development strategies does recognize the need for such transformation, but rejects the idea of agriculture as simply a resource reservoir. Instead, these strategies view the transfer of resources to non-agriculture as a result of surplus generation in agriculture as opposed to surplus extraction. In other words, this transfer is effected as a result of growth in agricultural productivity which, in combination with the fact that demand for agricultural products grows less fast than that for other products, makes it economically advantageous for resources to be increasingly directed to other sectors as development advances. The role of the state is crucial in these strategies in promoting technological improvements in agriculture (including human capital) to increase productivity, and promote institutions that facilitate the smooth functioning of markets for inputs and outputs.
In this connection, it is noted that more recent thinking on economic growth tends to underplay the importance of physical capital in economic development in favour of a broader definition of capital, including the stock of knowledge and human capital. As knowledge generated in a sector may have positive spillover effects on other sectors, public investment in research and development is crucial, as is the promotion of industries that produce knowledge. For agriculture, the positive effects of human capital, learning, and research and development have long been demonstrated (see Chapters 4 and 10) and have been embodied in the literature on agricultural development since the 1960s, so this evolution in thinking contributes little that is new to the analysis of the determinants of agricultural growth.
Other reasons for which the new growth theories are of limited use for the understanding of agriculture in economic development are: (a) they deal with one sector models (at least in their recent form); and (b) they have been largely concerned with long-run steady states. Descriptions of steady states may be of limited usefulness when describing agricultural growth and the changing role of agriculture in economic development, especially during the agricultural transformation.
In summary, although there are strong arguments in favour of priority to agriculture in development strategies in well-identified situations, there are no ready-made universal answers of general applicability. Research results on the effectiveness of public investment at the country or regional level are difficult to generalize into a single strategic approach to agricultural development. The dynamic long-term effects of government interventions and public investment activities have not been assessed due also to the lack of detailed data of public expenditure going to agriculture as opposed to rural areas in general, and on the amount of private investment that they can generate, including investment financed from farm savings and own-labour investment not giving rise to observable financial flows (Timmer, 1991).
It is also risky to make generalizations regarding the appropriate "mix" of public and private participation. The tasks generally accepted to be appropriate for the public sector, as discussed in this chapter, are neither well determined for each country nor may they be considered to be immutable over time. The view that development depends to a large part on organization, good governance, structure of institutions (including people's participation in the decision-making process) and political consensus is gaining more and more ground. These characteristics differ among countries and over time within the same country as institutions and organizational structures evolve. While the general proposition that governments are not generally well equipped for direct interventions in the production and distribution systems is now widely accepted, it is equally well recognized that they must become more effective in the provision of public goods, etc. The extent to which they can do so will go hand in hand with improvements in their organizational/managerial abilities.
Development failures in the past were to a large extent due to attempts by governments, sometimes encouraged by the "donor" community, to do too much, in their effort to balance the (often contrasting) interests of too many groups, by circumventing or supplanting market activity. The relevant experience indicates that greater reliance than in the past should be placed on markets and private agents, and that the conditions that cause markets to fail (information deficiencies, lack of clear property rights, lack of a clear and stable legal framework) should be corrected. Within a market-oriented framework, there is plenty for governments to do: establish and enforce rules and regulations for the smooth functioning of markets; establish and enforce property rights, especially land tenure rights; provide public goods and correct externalities; establish quality standards for food; carry out specific interventions to alleviate poverty and improve food security and nutrition.
NOTES
1. Such arguments were the cornerstone of writings by Soviet authors such as Preobrazhensky (Sah and Stiglitz, 1984).
2. Labour is in "surplus" in a particular sector (e.g. a traditional sector) when it can be transferred out of that sector and into another sector (e.g. the modern sector) at the prevailing wage without loss of output in the former. The concept presupposes that no additional capital is used to substitute for labour in the traditional sector. This concept, developed by A. Lewis, became one of the most influential contributions in development economics.
3. Lewis himself did not identify the backward sector with agriculture, and the capitalist sector with industry. Plantation agriculture, for instance, was part of the advanced sector. Interestingly enough, the Lewis model on surplus labour was used later to support a "pro-agriculture" development strategy (see Johnston and Mellor, 1961). In fact, as early as 1953, in his advice to the government of Ghana, Lewis emphasized stagnant agricultural productivity as the main obstacle to industrialization.
4. The "secular decline" argument developed simultaneously by Prebisch and Singer was based on a combination of assertions and data analysis results. The question of a secular decline in primary commodity prices, although generally accepted today, has a number of different dimensions. Research undertaken by FAO shows that "... on a purely qualitative basis the study confirms the original observation by Prebisch and others that the international barter terms of trade of primary commodities have a tendency to deteriorate". But it cautions that the measured tendency to decline is: (a) small in size; (b) statistically significant at the lowest confidence level; (c) in most cases reversing itself given a sufficiently large number of years; and (d) erratic in sign and size if considered over a small number of years. Stronger tendencies to decline can be detected though over smaller periods (Scandizzo and Diakosavvas, 1987). The possibility that the estimated trend in real commodity prices is biased downwards is raised in a recent World Bank study (World Bank, 1994a: 14). Namely, if the manufactures price index is adjusted for quality improvements, then the real commodity price index will be increased by about half a percentage point per year between 1900 and 1992. The same index would be increased by a further 0.1 points p.a. if the deflator is only for the price of manufactures exported from the developed to the developing countries; and by another 0.4 points if adjustments are made in this latter index for quality changes "within product categories". The sum of the three components means that the long term trend in real commodity prices may have been underestimated by as much as one percentage point a year.
5. For a detailed discussion and review on the reasons behind the rise in interest for "pro-agricultural" strategies see Staatz and Eicher (1990).
6. On the role of liquidity, see Rausser et at (1986).
7. Tradable commodities are those whose prices are, by and large, determined in international markets.
8. The above analysis is not meant to assign blame to lenders or borrowers. The interest is to extract useful lessons on the policy dimensions of the crisis, i.e. the extent that it depended on policies under the control of countries.
9. Detailed data on a number of countries are provided in a five-volume World Bank study (Krueger et al., 1991). Results are summarized in Krueger et al. (1988), and in Schiff and Valdés (1992). It should be noted that the study contains only a small sample of developing countries from all regions, and thus the results should be interpreted bearing this in mind. See also Norton (1992).
10. A number of such schemes are presented in FAO's study on agricultural price policies (FAO, 1987).
11. Public investment projects expanded rapidly in the 1970s although many questions have been raised ex post about their economic rates of return. An ex-post evaluation of the World Bank's agricultural investments for six African countries showed that 36 percent (in value terms) had negative economic rates of return while 18 percent had rates of return lower than 10 percent (Lele and Myers, 1987).
12. For definitions and methods of measuring the rates of direct and indirect protection of agriculture see Krueger et al. 1988, 1991.
13. For further evidence on the intersectoral consequences of macroeconomic and trade policies see World Bank (1986).
14. The reverse could also be true. Governments used the profits made by parastatals or stabilization funds during favourable years to finance budget deficits. For evidence see Claassen and Salin (1991).
15. The extent to which these measures result in a net disincentive to producers of competing products depends on the total package of policies. The latter may include compensating influences from devaluation and/or adjustments of border measures (e.g. price bands) to filter the penetration of signals from non-permanent and volatile distortions in world market prices and conditions of trade (subsidized credit, etc.)
16. The countries considered to be "healthy adjusters" were able to reduce domestic imbalances by increasing savings rather than reducing investment, and external imbalances by increasing exports rather than by reducing imports.
17. Platteau (1993) performed statistical analysis of the growth performance differences reported in Binswanger (1989). He could not find statistically significant differences of growth rates in agricultural production between countries providing a "favourable price environment" to agriculture and those that did not.
18. In the quest for evaluation of the effects of adjustment on economic performance, several criteria have been used to classify countries as "adjusting" or "non adjusting" according to timing, intensity and consistency of adjustment efforts. Countries have even been put into "finer" categories (early intensive adjusters, early non-intensive, late intensive, etc.). The criteria used have been the number of structural adjustment loans received, or the period between agreement on and disbursement of a structural adjustment loan. It is unlikely that either measure constitutes a credible indicator. The "Third Report on Adjustment Lending" of the World Bank suggested a closer look at the record of reforms to evaluate the performance of adjustment programmes rather than relying on number and timing on adjustment loans. The reason is that (according to the World Bank): (a) mere existence of the programme does not guarantee reforms taking place; (b) reforms agreed are not fully implemented; and (c) reforms executed are often reversed (see World Bank, 1994b). Structural adjustment programmes have often been criticized as being too rigid, i.e. as ignoring the political realities and dynamics of the implementing countries. Thus, failure of implementation may at times be an endogenous feature of some of the programmes. It is shown elsewhere in this chapter that when governments have to cut expenditure, less "visible" (soft) items in the budgets are cut first. Thus, public investment programmes are reduced before cuts are made in wage bills of the civil service, or in military expenditure.
19. Lack of imported consumer goods in the rural areas has been identified as an impediment to price response by farmers. For instance, this phenomenon is believed to have been present in Tanzania in the 1970s. Non-availability of consumer goods, in either the official or black markets, pushes the marginal utility of additional income earnings close to zero. Thus, with higher prices, the same level of income can be achieved with more leisure, and the supply response may actually be negative (Bevan et al., 1987).
20. The lack of transparency in the formation of international prices under various open and hidden forms of export subsidies can be deduced from the following examples: quoted export (Fob) prices of wheat were in early 1994 $130 150/tonne (US Soft Red Winter No. 2) or around $120 (Argentina Trigo Pan). At the same time, Kenyan producers complained that "wheat is pouring into Mombasa at $90 and $100 a tonne-we can't compete with those prices" (Financial Times, 28 June 1994). China is reported to have purchased in January 1994, 815000 tonnes of US SRW wheat with an EEP (US Export Enhancement Program) bonus of US$65.6 per tonne, resulting in an implicit discounted price of US$93.
21. The strategy considers the "dualism" in capital allocations (i.e. unduly large allocations of capital to industry and the unproductive elements of the private sector) as a major obstacle to the efforts of countries to increase employment, combat rural poverty and malnutrition. It is thus defined as a development strategy with a central role given to food production by a unimodal smallholder-based agricultural sector.
22. The observed positive correlation between food production and food imports in many developing countries is used as a major argument in support of the strategy. The conclusion is that an increase in food production causes an increase in food imports due to its general positive effects on overall income, but this conclusion applies in varying degrees to the different countries and may not apply at all to some countries (see de Janvry et al., 1989).
23. The social accounting matrix (SAM) is an accounting framework that maps out the aggregate structural interrelationships between different sets of economic "agents" and traces the circular flows of incomes and expenditures on goods and services.
24. According to Mellor, it was the critical importance of trade expansion in an agriculture-based strategy that created the perception of those countries as examples of export-led growth rather than as examples of successful agriculture-based strategies (see Mellor, 1986).
25. Data show that rural non-farm employment (excluding earnings from seasonal and part-time activity) accounts for 19, 36 and 47 percent of rural employment for Africa, Asia and Latin America respectively, when rural towns are included. Respective income shares range from 25-30 percent for rural Africa to 30 40 percent for rural Asia and Latin America (including part-time and seasonal employment but excluding rural towns) (Hazell and Haggblade, 1993).
26. A third linkage (labour market linkage) between the rural farm and non-farm activities has also been investigated. Growth in agriculture may increase the rural wage rate and cause substitution of rural non-farm economic activity to shift from low skill, labour intensive to higher skill more capital intensive, high return ones (see Hossain, 1988).
27. Results from different regions vary around the average according to the type of agricultural activity but also the methodology used. Most studies use demand driven multipliers, and input output models which assume a fixed proportions technology and a perfectly elastic supply of non-tradable commodities. Thus the derived multipliers are probably overestimated. The study by Haggblade et al. (1991) constitutes an exception in which prices are endogenous, input substitution is allowed and the supply of non-tradables is not perfectly elastic. Using this method, consistently lower multipliers are obtained compared to the fixed-price models.
28. For a comprehensive presentation of the evolution of new growth theories, see Stamoulis (1993).
29. For a critical review of the role of agriculture in economic development in the context of various growth theories, see Stem (1994).