Table 1, Table 2, and Table 3 present development data for some 73 developing countries. The bulk of the data derives from World Bank sources. Table 1 contains data on population (1985), per capita income in 1976 and 1986, growth rates in per capita income for 1960-1976 and 1965-1986, and the percentage of the labor force in agriculture in 1960 and 1975. The latter provides a rough indicator of the extent of structural transformation (transition from agriculture to manufacturing) that has occurred within a given economy after 1960. National income figures for 1986 show a wide range of per capita incomes, from $120 (Ethiopia) to $2920 (Venezuela). There is also a considerable range of growth rates in per capita incomes over these countries, from -2.6% (Uganda) to 6.7% (Republic of Korea). Significantly, all but two of the negative growth rates occur in sub-Saharan Africa. Table 1 also reports preliminary results of the United Nations International Comparisons Program (ICP), a program that attempts to adjust national GNP data by country consumption good prices. This index constructs a purchasing power parity (PPP) conversion factor that makes it more plausible to make cross-country welfare comparisons (World Development Report 1989, table 30 and technical notes). The ICP value represents a percentage of United States per capita GNP for 1985 (about $17,000). This value can be used to construct a poverty budget, permitting an estimate of the proportion of the population of a given country falling below the poverty line.
Table 2 presents a compilation of available inequality measures for a subset of these 73 countries. Two measures are provided: the Gini coefficient and the ratio of income shares of the top quintile to the bottom two quintiles of income. The two measures are defined differently, but show a reasonably high level of correlation in the small number of cases in which data is available for both (R2=.73). Income inequality data are both more difficult to come by and less reliable than national income data; these data have been drawn from World Bank compilations and several published review articles (Ahluwalia 1974, Fields 1980). Here too there is a wide range of country conditions. Available Gini data range from .378 (Republic of Korea) to .605 (Brazil); while quintile ratio values range from 1.57 (Korea, 1971-75) to 9.87 (Colombia, 1961-65). For our purposes it would be desirable to have reliable and comprehensive time-series data on income inequalities, in order to reach conclusions about the direction of change in this variable. However, limitations of available data preclude this. Multiple estimates of inequalities during different five-year periods are available for only a few countries, and these generally derive from different studies embodying somewhat different methodologies. So it is not possible to draw strong conclusions about the direction of change in income inequalities from this data set. It is suggestive, however, that the average income ratios for the five periods surveyed show a strongly rising trend.
Table 3 presents data on several indicators of population welfare for the 73-country set. This table provides an estimate of the percentage of the population below the poverty line in 1969 and 1985, longevity data for 1960 and 1975, infant mortality data for the same years, and primary school enrollment percentages for the same years. In the final column is an estimate of the Physical Quality of Life Index (PQLI) for each country, based on data from the 1970s. Several of these variables require some further comment. The percentage of the population falling below the poverty level is problematic because it is necessary to take account of variations across countries in the cost of living; a uniform dollar figure will not suffice. These data reflect adjustment of country income date by a purchasing power parity inflator reflecting the variations in the cost of subsistence goods and non-tradables across countries. ICP data provide the basis for such comparisons, but these data are available only for a minority of countries. Column 2 of table 3 represents my preliminary attempt to compute the poverty percentage based on a hypothetical income distribution (low, middle, and high), a poverty budget corresponding to the 45th percentile of India's income distribution, and ICP-corrected per capita incomes for several countries. (See figure 2.) These computed values are speculative, but they are plausible estimates. The model on which this computation is based can be improved by making use of actual income distribution data for each country; but this data is available in only a few cases.
The PQLI indicator is designed to provide a composite index of the quality of life in developing countries. It reflects average performance on variables such as infant mortality, literacy, and longevity. As was true concerning income inequalities, it would be desirable to have time-series data on poverty and quality of life, in order to see whether economic development has generally tended to reduce the former and increase the latter. However, once again it emerges that this time-series data is unavailable, so it is necessary to base conclusions on the direction of change in poverty and quality of life on other grounds.
Finally, table 4 presents a crude aggregation of all of these variables by region. This aggregation does not weight variables by population; instead, it represents simple averages of the country data by region. Even this level of aggregation is illuminating, however. It shows that there is a good deal of inter-regional diversity in each of the variables under consideration-per capita income, poverty, and inequality. The greatest income inequalities are found in Latin America, Central America, and the Caribbean. However, these regions also fall on the high end of the per capita income-leaving open the possibility that the condition of the poor may be higher in these regions than in more equal Asian economies. And in fact the regional aggregate figures show that the highest percentage of poverty in 1969 is found in South and Southeast Asia. (It is noteworthy, however, that sub-Saharan Africa and South Asia switch relative positions depending on whether we consider GNP or ICP-adjusted GNP.) The aggregate data also suggest a rough correlation between PQLI and per capita income; South Asia has the second lowest per capita income and the second lowest PQLI (after sub-Saharan Africa), whereas Latin America has the highest per capita income and PQLI. This impression is born out by a simple rank-order correlation; the rankings imposed by these two variables on the ten regions produce a Spearman's rho rank correlation of .76. Poverty percentage and PQLI are somewhat more loosely correlated, with a Spearman's rho value of -.53.
Regression analysis of possible correlations among various pairs of variables provides some interesting results. First, there is a strong negative linear correlation between the poverty percentage and per capita income (R2=.58)-that is, economies with higher per capita income tend to have lower poverty percentages. Second, there is a moderate positive correlation between income inequalities and per capita incomes (R2=.25)-that is, income inequalities tend to rise as per capita incomes rise. There is a moderate negative logarithmic relationship between PQLI and the poverty percentage (R2=.25)-that is, PQLI falls as the poverty percentage rises, an intuitively plausible finding. And there is a moderate correlation between PQLI and per capita income (1986) (R2=.49). This reflects the fact that some very low-income countries-e.g. Sri Lanka-have achieved high PQLI, whereas some middle-income countries-e.g. Algeria, Jordan, or Tunisia-still have low PQLI.
When we turn to possible correlations between the poverty ratio and measures of inequality, we find that there is a moderate negative correspondence between these variables: economies with higher inequalities tend to have lower poverty ratios. There is a moderate negative linear relationship between the poverty ratio and the Gini coefficient-suggesting that higher inequalities correspond to lower absolute levels of poverty (R2=.23). However, this finding is based on only nine data points, and is probably not significant. When we turn to a possible relationship between poverty and the income inequality ratio we find the same general pattern; there is a weak reciprocal correspondence between the poverty percentage and the income ratio (R2=.13). This correlation is based on 28 data points. These correlations might seem to suggest that rising inequalities are favorable for reducing poverty; however, when we recall that there is a positive relation ship between per capita income and income inequalities (i.e., societies with higher per capita incomes tend to have greater income inequalities), it is reasonable to hypothesize that these correlations represent a form of collateral causation: falling poverty percentages and rising inequalities are both collateral effects of rising per capita incomes.
A final set of variables of interest are the poverty percentage, PQLI, and the percentage of the labor force in agriculture. These variables allow us to track the consequences of the process of structural transformation on poverty and physical quality of life. Here we may take the agricultural work force as the independent variable and speculate that more agricultural economies will have higher poverty percentages and lower PQLI. This hypothesis is born out by moderate correlations between the two pairs of variables. There is a positive logarithmic relationship between the poverty percentage and the agricultural work force: societies with higher percentages of agricultural workers tend to have higher poverty percentages as well (R2=.17). And there is a negative linear relationship between PQLI and the agricultural work force (R2=.50). This means that countries with higher percentages of labor in agriculture tend to have lower PQLI. Finally, inequalities tend weakly to increase as the percentage of the workforce in agriculture falls; but this trend also probably derives from the stronger association between rising per capita incomes and rising inequalities.
The country data set serves chiefly to provide a general picture of the state of development in the less-developed world at the end of the 1980s. In order to improve this picture, however, it is necessary to take a closer look at several countries. Below we will turn to a comparative study of six Asian countries with a variety of development experiences over this period.
I turn now to a brief review of several Asian country development experiences with regard to the issues of growth, inequalities, and poverty. Asia remains the home of the most extensive of the world's poverty: African poverty is deeper but less extensive, and Latin American poverty is on the whole both less extensive and less deep. Indonesia and India are home to the majority of the world's poor; so the development experience in South and Southeast Asia is critical for the prospects of poverty alleviation. At the same time, Asia is the locus of two of the most dramatic development experiences in the past forty years: the peasant revolution in China, on the one hand, and the breakthrough of other East Asian economies (Taiwan, South Korea, and Japan), on the other.
Korea is, of course, one of the success stories of post-war economic development. Its GNP per capita jumped from $670 in 1976 (1976 dollars) to $2370 in 1986 (1986 dollars). Moreover, the benefits of development were broadly distributed over all sectors of the Korean economy-urban and rural, industrial and agricultural. Income inequalities have been consistently low by international standards-a gini coefficient of .378, and an average income ratio of 2.13-substantially below mean LDC measures of inequality. What were the chief characteristics of Korean development?
First, of course, is the spectacular success of Korea's industrial development strategy. Korea has pursued a strategy of export-led growth, with a rapid buildup of industrial production and foreign trade. Manufactured exports increased from .2% of total exports in 1954 to .3% in 1960, 19.7% in 1965, and 39.2% in 1976 (Mason et al:137). Moreover, the pattern of industrial growth tended to be labor-absorbing, with the result that demand for labor grew at a respectable rate from the mid 1960s onward. Industrial employment expanded at an annual rate of about 7% during the period of 1963-76 (Mason et al:109), while agricultural employment expanded at only 1% during this period. This pattern facilitated the process of structural transformation from a largely agricultural workforce in the early 1960s to a largely industrial workforce in the late 1970s; agriculture's share of the labor force declined from 63% in 1963 to 45% in 1976.
Important as this process of structural transformation was in Korea's economic development, agricultural development played a central role as well. For the bulk of Korea's population was rural until the mid-1970s; so the benefits of economic growth could only reach this part of the population through rural development. Korea's experience of rural development is framed by the occurrence of a major land reform in the 1940s and 1950s. This land reform created the basis for a substantially more equal pattern of asset-ownership in the rural economy, leading to a narrower range of income inequality as well. Moreover, this early redistribution of productive assets in the countryside influenced the pattern of economic development more broadly. "A lower level of rural inequality in turn had a substantial influence on much else that occurred in the Korean agricultural sector. The pace with which new techniques spread from one farm or region to another, to take only one important example, was undoubtedly accelerated by the absence of extreme inequality" (Mason et al:210). The land reform was quite extensive; 60% of the arable land was owned by landlords in the 1930s, and virtually all of this land was distributed to tenants and landless during the 1940s. The land reform program was established by the U.S. Military Government in 1945; it was politically feasible because many of these landlords were Japanese, and the Korean landlords were discredited after the war by their pattern of collaboration with the Japanese (237). The land reform program involved compensation to landlords, but at a very low level; with the result that the productive assets of the landlord class were virtually wiped out. Mason et al provide a reconstruction of the effects of land reform on household income, which shows that household income almost doubled between 1933 and 1975 (239).
It is often argued that agricultural development must precede industrial growth; but the Korean experience contradicts this assumption. In this case rising demand for agricultural outputs (food and inputs for manufacturing) stimulated the growth of the rural economy. "In a basic sense, therefore, it was rising urban and export that `caused' a large part of the increase in farm demand" (213). Moreover, the robust demand for labor in manufacturing soon gave an impetus to raising farm productivity (because of the induced increase in the cost of rural labor; 220). From the 1960s forward there is a rising trend in real wages for both factory and farm workers (220). This led to a round of agricultural modernization, including mechanization of farm work; which in turn increased the productivity of the farm economy.
The Korean state provided only modest assistance for agricultural development during the 1960s, but this pattern began to change in the 1970s. The government provided some support for rural credit for poor farmers, though demand for credit greatly exceeded supply (233). And in the late 1960s the government created a series of grain price policies that improved the terms of trade for agriculture.
In this case, then, we have growth with equity. Mason et al write, "One of the most striking features of Korean economic development since 1945 is that development has been achieved without requiring or causing a highly unequal distribution of income" (408). And this high degree of equity depends on two things: an effective land reform that leveled out access to assets in the farm economy before the period of rapid economic growth, and a pattern of industrialization that was heavily labor-absorbing, creating strong demand for labor and stimulating rising wages.
Turn now to a second important example: India. A. K. Sen's (1989) recent review of India's development experience is a useful point of departure. First, the successes: India has succeeded in developing a massive industrial capacity; it has experienced a fairly consistent rate of per capita growth of about 2 percent a year; and it has successfully introduced Green Revolution farming on a massive scale, making India food self-sufficient. Its failures, however, are daunting; most importantly, it has made little or no progress in reducing the mass and depth of poverty since Independence, and inequalities-particularly rural inequalities-remain high, by Asian standards. In spite of a national development rhetoric that gives priority to poverty alleviation, the condition of the poor (largely rural) has changed very little over a forty-year period. Atul Kohli summarizes the situation in these terms:
Three decades of planned economic development have failed to improve the living conditions of India's poor. This persistence of poverty is clearly manifest in the continuance of low percapita income. It is nevertheless clear by now that higher growth rates, and therefore higher percapita income, are not sufficient to improve the lot of the poor. New wealth has not "trickled down." The solutions to the problem of India's poverty will thus not emerge from higher rates of economic growth alone; if they emerge at all, they are likely to involve conscious state intervention aimed at reconciling growth with distribution. (1987:1)
India's industrial development has been largely led by a model of state planning that combines Soviet-style encouragement of heavy industry and state intervention with a market economy. Indian industrial policy emphasized national planning mechanisms to encourage a high degree of industrial autarchy in key production industries-steel, heavy equipment, etc. (a policy of seizing the "commanding heights" of industry)-and a relatively low degree of emphasis on export production and consumer goods. (Each of these features has changed somewhat in the 1980s as the result of a series of liberalization reforms initiated by the government of Rajiv Gandhi.)
Central to India's agricultural development has been the adoption and diffusion of Green Revolution technologies. During the 1960s and 1970s India absorbed the new varieties of wheat and rice, along with the chemical and water technologies needed for NV cultivation (about 15.5 million hectares of each in 1977; Griffin 1989:148). The results were dramatic; India's grain production increased sharply, and farm incomes rose as well. Food output rose from 72 million tons in 1965-66 to 108 million tons in 1970-71 (Varshney 1989:83).
How has Indian development affected poverty and inequality? In brief, there has been little or no positive effect on either poverty or inequality. William Murdoch emphasizes the magnitude of continuing rural inequalities in India, writing "In parts of India the income of the large landowner from the farm alone is typically 100 times that of the farmworker and the family farmer" (1980:160), and he reports that real rural income in Bihar actual declined during the 1960s; "between 1961 and 1971 the proportion of the rural population below the poverty line increased from half (21 million) to 59 percent (30 million)" (242). The Rudolphs report data showing that the extent of landless ness and the land poor has remained virtually constant in India between 1954 and 1972 (1987:336). Griffin reports that wages rates in Maharashtra in the late 1980s are actually lower than in the mid-1970s (1989:138). And Gillis et al conclude that "the most recent analyses suggest that the proportion of people who are poor has changed little since the 1950s" (1987:91).
Given its successes in both industrial and agricultural development, why did India fail so completely in reducing poverty and increasing the absolute incomes of the poorest 40% of its population? Atul Kohli's The State and Poverty in India (1987) is a comparative study of three Indian states-West Bengal, Karnataka, and Uttar Pradesh-that attempts to answer this question. He argues that poverty alleviation requires positive policy efforts on the part of the state; the normal workings of a market system do not inevitably or commonly lead to improvement of the condition of the poor. However, some states in India have done better than others in poverty alleviation. What are the social and political factors that influence the welfare of the poor in the process of third-world economic development? Kohli finds that the welfare of the poor is not correlated with the overall prosperity of a state. Instead, the critical variable is the type of regime in power during the process of economic development: regimes formed by strong, competent political parties of the left succeed in tilting the process of development towards poverty alleviation, whereas weak regimes and regimes dominated by the propertied classes have a poor record of performance in poverty reform. Poverty reform requires a political regime that has both the will and the means to implement it, and a regime that is relatively autonomous from the political reach of economic elites. The Communist Party, Marxist (CPM) in West Bengal succeeded in bringing tangible benefits to the poor through poverty reforms including tenancy reform, rural credit programs, and rural employment schemes. CPM is a leftist party with a coherent redistributivist ideology, competent party organization extending down to the village level, and effective leadership. The Urs regime in Karnataka also possessed a redistributivist ideology, but lacked effective political organization and had a fragmented leadership; its efforts at poverty reform were not successful. And the Janata party in Uttar Pradesh was dominated by the rural landowning class and lacked the will to implement poverty reforms. Kohli explains the presence or absence of poverty alleviation in a state, then, as the result of the presence or absence of a regime which has both the will and the means to implement poverty reform.
Kohli finds that West Bengal, though the poorest of the three states considered, has made a perceptible impact on rural poverty in its decade in office. (About 65 percent of the rural population lives in conditions of absolute poverty; 118). The CPM is strongly committed to a political program favoring the interests of the rural poor; it has effective administrative capacity through which to implement its program; and it has an electoral strategy that permits it to retain office while implementing these policies. What, then, are the poverty-oriented policies that have been adopted in West Bengal? Kohli singles out three main policy efforts described as operation barga: tenancy reform, programs for small farmers (credit for sharecroppers); and employment and wage schemes for landless laborers (117).
First, there is a program of tenancy reform. Sharecropping is the primary form of land tenure in West Bengal (96% of overall tenancy). CPM chose to pursue tenancy reform rather than land redistribution because it judged that the political risks of the former (determined landlord opposition) were too great. The party has made a determined effort to register sharecroppers (to improve their legal position vis-a-vis landlords); the chief aim of this program is to increase tenure security. Kohli does not assert that the tenancy reform program took as a goal the reduction of the share taken as rent, but his survey data shows that the rent share prior to this program was almost universally 50%, whereas after the program 66% of tenants paid 25% and 32% paid 50% (130).
Second, CPM put into place several support programs for sharecroppers and smallholders, including particularly state-funded credit. The regime undertook to facilitate low-interest loans to smallholders and sharecroppers (133). It also attempted to induce commercial lenders to extend credit to poor farmers. These programs brought credit to 430,000 small farmers-a small fraction of the total population, but a significant number nonetheless.
Finally, Kohli describes a package of wage and employment schemes created by the CPM regime for landless agricultural workers. Landless agricultural workers can depend on 3-4 months employment annually; the CPM regime undertook to improve both employment opportunities and wage levels for this group. It created the National Rural Employment Program (NREP) to offer employment to landless workers in the off-season at a daily wage about 25% higher than the rural average. This program provided about 25 days employment to about one-third of all landless families in West Bengal (137)-thus increasing annual income for these families by about 25%. (It appears that this labor was expended on local public works projects.) CPM has also attempted to facilitate unionization among unskilled workers to raise wage rates, with inconclusive results.
A final feature of Indian development is the major land reform that was effected in the state of Kerala. Kerala, though one of India's poorest states, has had a relatively spectacular record on indicators of the welfare of the poor: high education, low infant mortality, and high longevity. Ronald Herring describes this experience in Land to the Tiller. In brief, a commu nist government (CPI) enacted a package of land reform legislation that put a low ceiling on land holdings and effectively redistributed the surplus to the land-poor and landless. This eventually involved about 40% of the total arable land in Kerala (Herring 1989). This policy redistributed income dramatically to tenants. But only a small minority of Kerala's rural poor benefitted; in a study of one region Herring finds that only 12% of landless workers received land, and these generally in amounts of less than an acre (1989). In this recent article Herring suggests that the progressive efforts of the communist government have stalled as a result of the shifting of interests created by the land reform: smallholders no longer have a material interest in supporting programs that favor the interests of those even poorer than them.
The Indian case, then, represents a pattern of development in which both industrialization and agricultural improvement have worked reasonably well, and yet in which the condition of the poor has been little improved. And a persuasive case can be made that the explanation of the latter fact has a great deal to do with the lack of institutional reforms improving the position of the poor (access to land and education) and a capital-intensive industrial strategy; and these in turn can be explained as a consequence of the absence of a political regime committed to poverty reduction.
China's development experience is substantially different from either of these. [1] It is not possible to review the whole complicated story here, but several central themes emerge, and the overall record is mixed. On the one hand, the economy has shown respectable rates of growth, poverty alleviation, reduction of inequalities, suppressed population growth, and high rates of savings. Seen from this perspective, China represents a strong model for other developing countries. On the other hand, China's economy during this period shows some crucial flaws as well. Growth has not been based on rising productivity but rather extensive expenditure of capital and labor. Much of this expenditure has been of low efficiency, producing products of poor quality and diversity. The central planning process has produced some of the same problems of allocative inefficiency to be found in the Soviet system. Rural incomes witnessed little improvement until the post-Mao reforms. Urban-rural inequalities have remained significant (though they have declined since 1978). And tumultuous political events (GLF, Cultural Revolution, and the democracy movement of the last few years) have disrupted the economy and the process of economic planning. Let us look briefly at some of the most salient characteristics of this development experience.
As the CCP pursued and consolidated power in the late 1940s and early 1950s, a program of land reform redistributed land and farm capital from landlords and rich peasants to poor peasants. The lands of landlords, and to a lesser extent rich peasants, were confiscated and redistributed to poor peasants. According to Robert Dernberger, "By the end of 1952 all rented land (about 40 percent of China's cultivated area) had been redistributed to poor and landless peasants" (Dernberger 1982). However, the "land to the tillers" program was only the beginning of the agrarian reform. The next step was the creation of "Mutual Aid Teams" (MATs)-small groups of households which were encouraged to exchange draught animals, labor, and tools (Shue 1980:145). MATs were designed to build upon traditional modes of cooperation in Chinese village life; but they were also intended to begin to establish a basis for more extensive cooperation and collective ownership in the future. MATs were confronted with several administrative tasks almost immediately: assigning work points and compensation for contributions of draft animals and tools, and coordinating the expenditure of labor and other resources efficiently. This package of rural reforms had the potential for dramatically improving the performance of agriculture, particularly when supplemented by rural credit, marketing coops, and the like. For under these circumstances small farmers have both the incentive and the capacity to increase output and productivity.
The next major step in the process of Chinese agrarian reform was the creation of cooperatives for marketing (Supply and Marketing Co-op) and credit (Credit Co-op) (Shue 1980:196 ff.). The chief function of marketing and credit cooperatives, however, was not so much to coordinate production as it was to alter the economic environment within which farming took place and to discourage the reemergence of capitalism in the countryside. By controlling markets and access to credit, the state was in a stronger position to prevent the concentration of wealth that might otherwise have occurred.
The next stage in this process was the creation of Agricultural Production Co-ops, which were designed to directly organize the production process at the local level. Up to this point cultivation took place within the altered circumstances of private ownership and rent that were established by the land reform laws; the evolution of production cooperatives, by contrast, was designed to lead to full collective ownership and management of land and capital equipment. One goal of the Agricultural Production Cooperative was to encourage rich peasants to invest their surplus in capital available to the Cooperative, thus increasing the productivity of local farming.
Elementary cooperatives involved a larger scale of cooperation than MATs, but they continued to work through private ownership and compensation. Each member made his labor, capital, and land available to the cooperative, to be used jointly; but the owners of these resources were to receive compensation in proportion to their contributions. Thus each member potentially received both work points and "rent" on the land and capital he provided to the cooperative. According to Shue, the average size of an elementary cooperative was between 27 and 32 households (291). Advanced co-ops went one step further, in that all capital goods were to be turned over to the cooperative, with some small compensation to the owner. Income was based solely on labor contribution (Dernberger 1982:72).
The transition to production cooperatives was initially propelled through gradual and voluntary means; in 1955, however, the government took the decision to collectivize agriculture immediately by law. The most immediate goal of collectivization was to rationalize the production process through economies of scale. Land was to be pooled and farmed on a larger basis; labor was to be allocated more efficiently; collective goods (dams, reservoirs, ditches, roads, etc.) could be provided using surplus co-op labor; and so on. This led in a short time to the formation of very large brigades, collectives, and communes. It led also, during the Great Leap Forward, to a massive crisis in agriculture over several harvests culminating in a famine in which perhaps 30 million deaths occurred.
The post-Mao reforms in agriculture involved several major changes in policy. First, the household replaced the production team and other collectivized units as the basic production unit. Through the family responsibility system farmers were given longterm contracts for parcels of land and were given wide authority to make production decisions. This increased authority permitted farmers to specialize in high-value crops and crops well-suited to their factor endowments. And second, market institutions and price reform were reintroduced into the rural economy. Real prices of agricultural products rose sharply for the first time in several decades-producing corresponding incentives to increase production and cut costs. The net result was a sharp increase in output and productivity in the rural economy; and these gains flowed to rising rural incomes to an extent unprecedented in China's development experience. [2] Grain output increased almost 5% a year between 1978-1984. And non-grain output rose even more rapidly. And these increases have been reflected in rising rural incomes as well, with rural incomes rising from 134 yuan in 1978 to 355 yuan in 1984 (Lardy 1985:17).
Indonesia is often described as a model of market-driven development that reconciles growth and equity goals. It is held that a combination of appropriate government macro- and micro-economic policies, oil revenues, and industrial expansion have led to the outcome that the Indonesian economy has expanded at a rate of about 4.5 percent, while improving its poverty performance over the past 10 years. Peter Timmer and other development economists argue that the data in Indonesia support the notion that growth leads (or can lead) to a reduction of poverty. Moreover, this process has taken place under inauspicious political circumstances-the New Order regime (1966-present) has few social commitments to the rural poor and no appetite for land reform or other agrarian reforms.
Many observers suggest that Indonesian development in the past decade has had two important features: industrial development has led to an expansion in the demand for labor, leading to a rise in real wages in both industrial and agricultural sectors; and agricultural development has enhanced the productivity of rice cultivation, again with the effect of increasing incomes to participants in the rural economy. A recent study by the Stanford Food Research Institute finds that in the case of Indonesia, poor rural people are displaced in the rural economy, but move more or less easily into the bottom rungs of the urban economy with a net and immediate increase in income (Stanford Food Research Institute 1988). And they find that rural household incomes have tended to rise across the board in the agricultural sector, as productivi ty increases improved farm income and rising demand for labor improved wage income. In particular, Roz Naylor computes data suggesting a positive trend in farm labor real wages since 1980 (chap. 5, p. 25) (with a dramatic slowdown in growth in 1985-87). "Policies designed to stimulate rice production have enhanced farmers' incomes and have generated a significant amount of employment for unskilled labor in rural areas" (SRI, chapter 5, p. 1). Small-scale rural enterprises provided much of the non-farm demand for labor in the off-season (15). The report concludes, "The existence of a broadly competitive rural labor market in Indonesia implies that adjustments in rice production or in labor saving innovations in rice farming can occur smoothly as long as the demand for labor in non-agricultural sectors of the economy remains strong" (36).
An important component of Indonesia's development story is the Indonesian government's agricultural policies, including particularly its price policies and its financial support for irrigation and modernization. Indonesia's price policies turned towards agriculture after Sukarno through a series of increases in the floor price for rice (Timmer 1989:30). Higher prices stimulate further production, technical innovation, and higher farm incomes.
The question to be asked in the context of the issues raised in this paper is this: what is special about Indonesia? For we have seen that modernization directed at maximizing growth rates does not generally have positive effects on poverty and the incomes of the poor; so what special features of post-1960s Indonesia produced this seemingly more benevolent outcome? Moreover, we do not have in this case a regime that is primarily rooted in the material interests of the poor (as was the case with the communist regimes in Kerala and West Bengal); instead, Indonesia's New Order government has no inherent interest in poverty alleviation or attenuation of rural inequalities.
One possible answer to this question is that the current view of Indonesia is overly rosy. For when we turn to village studies of rural change in Java, we get a rather different picture: an increase in land inequalities as larger farmers benefited more fully from Green Revolution technologies, an accelerated social differentiation among large farmers, smallholders, and landless workers, and a shrinking demand for farm labor as larger farmers introduced labor-saving innovations (Hüsken and White 1989:236).
Hüsken and White analyze New Order agricultural strategies in terms of the rural social tensions engendered by unequal access to land in Java. After suppressing the Communist insurgency in the mid-1960s, the New Order regime sought to consolidate its power in the countryside. The state made strenuous efforts to created administrative forms capable of penetrating local society-replacing both traditional village arrangements and the interest-group organizations of the early 1960s (250). The New Order regime attempted to increase rice production through a combination of quotas and subsidies imposed on rice cultivators. This program included subsidies on fertilizer prices, subsidized agricultural credit, state purchases of paddy, and subsidy of irrigation water (253); the result was a rough doubling of rice output on a fixed arable land. And Hüsken and White argue that the preponderance of this state support for agriculture-and ensuing income benefits-flowed to the richest 1/3 of Javanese farmers, who produce by their estimate about 75% of Java's rice crop, since the richest 10-20% of farmers control 70-80% of farmland. And they find that "the share of output received by hired laborers in the form of wages has declined proportionately to the much more rapid growth of the farmers' net income from crop production" (254). On the basis of a detailed seven-village study in Java White and Wiradi find that income shares increased substantially more for labor-hiring farmers than hired laborers during the 1970s-though each group showed increases in almost every case (White and Wiradi 275). "These data therefore indicate a growing divide between `farmers' on the one hand, and `hired laborers' on the other, in terms of their relative ability to command incomes from paddy production" (274). Moreover, White and Wiradi find that the demand for labor per hectare actually fell in each village sampled (285). But they find that wage rates for virtually all agricultural tasks rose during the period 1971-1981-a fact they explain as the consequence of the increasingly rigid scheduling requirements in modern-variety cultivation. (Though overall demand for labor has fallen, the demand at peak periods has risen; 288.)
In short, the picture that we get from village-level studies of Java is one of rising inequalities in land ownership and incomes and slowly rising real incomes to small farmers and landless workers-a finding much less rosy than that of the SRI team.
If Indonesia is a possible success story, the Philippines is the reverse. Its record of development since 1950 shows a combination of urban industrial growth, enrichment of the wealthiest families, Green Revolution innovation in agriculture, and stagnating or falling real incomes in rural areas. The Huk rebellion (1946-early 1950s) mobilized small peasants in rice and sugar areas in Luzon. And a weak land reform was undertaken, first in 1953 without effect, and then again under Marcos in 1971. The Marcos land reform redistributed some land, with massive evasion by land owners; it also increased tenure security for remaining tenants (Fegan 1989:134). But land inequalities remain extreme, and farm wage incomes have fallen substantially during the 1970s and 1980s.
The Philippines was one of the first and most extensive adopters of Green Revolution technologies in the 1960s. Innovations developed by the IRRI (new seed varieties, chemical fertilizers, and mechanized cultivation) were rapidly adopted; by 1980 about 78% of rice land was planted in high-yielding varieties (Griffin 1989:153; Barker et al 1985). State grain price policies created a bias against agriculture, however, that made rice farming increasingly unprofitable during the 1970s (Fegan 1989:138), and mechanical innovations in rice cultivation have put pressure on unskilled farm labor markets. "The prospects for landless laborers seem bleak: they will be largely displaced from the rice production cycle by machines and chemicals. IRRI innovations, in the context of government policies that make capital artificially cheap, have marginalized landless agricultural workers" (Fegan 1989:138). Export crop cultivation on large plantations represents the most profitable form of agriculture in many regions, and is dominated by large land holders and foreign corporations.
Philippine development, moreover, has bifurcated sharply by sector. Industrial development has proceeded much more rapidly than agricultural development (Fields 1980:219), and the benefits of growth have been narrowly concentrated in upper income groups. ""Despite a tripling of the national product and a doubling of national product per capita, mean family incomes grew by less than 1% per year" (Fields 220). And, even more disturbingly, the real income flowing to the poorest 20 percent (chiefly rural) in fact fell by more than 10% between 1961 and 1971 (Fields 222). Fields argues that this extreme instance of growth with poverty is explained by the character of the Philippine regime: "successive regimes in the Philippines did not take direct measures to spread the benefits of growth" (224-25), and in the absence of such measures, the least-well-off received virtually nothing.
Turn now to another example from Southeast Asia: the effects of the Green Revolution in the rice-growing regions of Malaysia. James Scott provides a careful survey of the development process in Malaysia in Weapons of the Weak. The chief innovations were these: a government-financed irrigation project making possible double-cropping; the advent of MV rice strains; and the introduction of machine harvesting, replacing hired labor. Scott considers as relevant parameters the distribution of landholdings, the forms of land tenure in use, the availability of credit, the political parties on the scene, and the state's interests in development. His chief finding is that double cropping and irrigation substantially increased revenues in the Muda region, and that these increases were very unequally distributed. Much of the increase flowed to the small circle of managerial farmers, credit institutions, and outside capitalists who provided equipment, fertilizers, and transport. Finally, Scott finds that the lowest stratum-perhaps 40%-has been substantially marginalized in the village economy. Landlessness has increased sharply, as managerial farms absorb peasant plots; a substantial part of the rural population is now altogether cut off from access to land. And mechanized harvesting substantially decreases the demand for wage labor. This group is dependent on wage labor, either on the managerial farms or through migration to the cities. The income flowing to this group is more unstable than the subsistence generated by peasant farming; and with fluctuating consumption goods prices, it may or may not suffice to purchase the levels of food and other necessities this group produced for itself before development. Finally, the state and the urban sector benefits substantially: the increased revenues created by high-yield rice cultivation generate profits and tax revenues which can be directed towards urban development.
Scott draws this rather gloomy conclusion:
The gulf separating the large, capitalist farmers who market most of the region's rice and the mass of small peasants is now nearly an abyss, with the added (and related) humiliation that the former need seldom even hire the latter to help grow their crops. Taking 1966 as a point of comparison, it is still the case that a majority of Muda's households are more prosperous than before. It is also the case that the distribution of income has worsened appreciably and that a substantial minority-per haps 35-40 percent-have been left behind with very low incomes which, if they are not worse than a decade ago, are not appreciably better. Given the limited absorptive capacity of the wider economy, given the loss of wages to machines, and given the small plots cultivated by the poor strata, there is little likelihood that anything short of land reform could reverse their fortunes. (Scott 1985:81)
This example well illustrates the problems of distribution and equity that are unavoidable in the process of rural development. The process described here is one route to "modernization of agriculture," in that it involves substitution of new seed varieties for old, new technologies for traditional technologies, integration into the global economy, and leads to a sharp increase in the productivity of agriculture. Malaysia is in effect one of the great successes of the Green Revolution. At the same time it creates a sharp division between winners and losers: peasants and the rural poor largely lose income, security, and autonomy; while rural elites, urban elites, and the state gain through the increased revenues generated by the modern farming sector.
[1] For a valuable survey of this historical experience of development see Carl Riskin, China's Political Economy: The Quest for Development since 1949 (1987). It should be noted that Tom Rawski argues to the opposite conclusion. He holds that growth and investment were occurring in Republican China in the 1930s as a result of normal market-driven economic processes.
[2] This summary reflects Lardy 1985:1-10.