Previous PageTable Of ContentsNext Page

Agricultural production and productivity in developing countries


In the second half of the twentieth century, farmers were presented with an unprecedented increase in the demand for food. In the first half of the century, the world's population grew by 960 million people. In the second half, the increase was
3 690 million people. The population of the developing countries as a whole increased from 1.8 billion to 4.7 billion during the second half of the century, an increase of 260 percent. In addition, per caput income, another factor that affects food demand, also grew in many developing countries over the period.

Farmers were presented with this demand challenge at a time when much of the land suited for cultivation was already used for crop production. In 1950, farmers in many countries were farming land intensively, with significant levels of irrigation and multiple cropping. In most areas it was not feasible to meet the demand challenge simply by expanding cultivated area. (Some regions, however, did have the potential for cropland expansion, e.g. part of Africa and the Cerrado region in Brazil.)

Figure 21

Figure 21 shows cereal production, imports and food aid per caput for the periods since 1961. While these data do not cover all categories of food or agricultural production, they do reflect the main trends of recent decades in five developing and four developed country regions.

These data are reported in per caput terms, allowing comparison by both region and time period. It may be noted that the apparent per caput consumption of cereal grains is highest in developed countries. This reflects high rates of cereals fed to livestock (and the fact that feed conversion to livestock product rates are low).1 For developing regions, where cereals are mostly consumed as foodgrains, per caput consumption is lower. Sub-Saharan Africa has the lowest levels because root crops are important in Africa. In South Asia, per caput consumption is also low, reflecting the fact that low levels of cereal grains are fed to livestock

Over the periods depicted, cereal production per caput rose significantly in Latin America, South Asia and the Far East. It declined in sub-Saharan Africa and the Near East from 1961 to 1981 but has risen in subsequent periods (note that the largest increase in population in these regions occurred after 1981).

Such a per caput production performance can be characterized as both extraordinary and uneven. Extraordinary in view of the massive increase in population; uneven because it has not been uniform over all regions and countries. This section examines the production record and the developments in productivity that lie behind it, both at the aggregate or global level and at the local or national level.

Agricultural production has increased enormously but unevenly over the past 50 years.


Economists and historians have offered several different perspectives on the agricultural growth process. These include:

Figure 21 (continued)

Apart from the Malthusian-resources perspective, these perspectives are not mutually exclusive. The Malthusian-resources perspective, however, focuses on the economic growth process when institutions and human resources are unchanging and technology invention and diffusion are not taking place.

The Malthusian-resources perspective draws attention to both population (and workforce) growth and available land and water resources. With abundant land and water resources, the ratio of population (workforce) to resources need not decline as the population grows, since more land can be brought under cultivation. However, as land (and water) frontiers are closed, the ratio of population to resources will rise and bring about a decline in per caput production. The Malthusian-resources perspective leads to a policy emphasis on population growth reduction. It does not formally recognize the demographic burden/gift.

The "demographic gift" allows a country to increase investment and savings while its workforce grows.

The demographic burden/gift effect is based on differences in the growth rates of population and of the workforce. When growth rates rise, as they did in virtually all developing countries during the 1940s and 1950s, population growth exceeds workforce growth for a number of years simply because of the fact that children and young people do not become workers until they reach a certain age. This creates a consumption burden. Conversely, when population growth rates decline, as they have in most developing countries since the 1950s and 1960s, a demographic consumption gift occurs as growth in the workforce exceeds growth in population. Most developing countries have experienced (at different times) a burden-gift cycle since 1950. The demographic gift is important, even when the Malthusian-resources perspective is valid.

An extension of the Malthusian-resource model recognizes that population growth or population density may actually stimulate investments in institutional and technological change, thus providing its own "antidote" to the diminishing returns feature of the model.

The institutional change, human capital, best practice and adaptive invention perspectives all depart from the Malthusian-resources perspective by introducing dynamics that enable producers to produce more with the same amount of resources (labour, land, etc.). That is, they introduce productivity change (see Box 21 for the arithmetic of agricultural productivity). Each of these perspectives is associated with the development of what here will be referred to as technological capital (TC), which represents a country's capacity to implement, adapt and develop productivity-enhancing technology.

The institutional change perspective addresses inefficiencies associated with transaction costs and imperfect markets. Infrastructure investments reduce transport and other costs, and can reduce transaction costs as well. Investments in institutions (credit institutions and legal systems) have been important for agricultural economies. Improved institutions and infrastructure provide a source of growth in per caput food production even in Malthusian-type economies, where few or no changes take place in the technology actually available to farmers.

The human capital perspective stresses that farm management skills and farm production skills (farmer human capital) can be improved through investment in training (schooling) programmes, experience and agricultural extension programmes. Investments in farming human capital can thus produce growth in per caput food production.

The best practice perspective focuses on the fact that, at any given time, farmers may not yet have tested and adopted existing technology that would reduce costs and produce growth because of failures in the information and demonstration systems available to them. Investments in agricultural extension systems will then produce growth in food production per caput by bringing farmers closer to best practice technology use.

Finally, the adaptive invention perspective emphasizes that agricultural technology is location-specific to a considerable degree. Biological processes are sensitive to soil, climate and even economic conditions. Natural "Darwinian" evolutionary change produced a rich diversity of species, resulting in natural differences in plant and animal life in each ecological niche. Farmers only partially overcame this niche phenomenon when they selected the landraces (farmers' varieties) that today constitute the genetic resource stock utilized by modern plant (and animal) breeders as they search for varietal (and breed) improvements. Modern plant breeders must also respect soil and climate factors and tailor varietal improvements to regions or niches. This means that technology that is valuable in one location may not be valuable in another. It also means that targeted invention (plant breeding) programmes can produce growth in per caput food production.

The Malthusian-resource perspective has, however, been linked to the other perspectives in a number of economic studies that address the relationships between population growth (relative to resources) and the policies and investments inherent in the institutional change, human capital, best practice and adaptive invention perspectives. One linkage is through population-induced changes and investments.3 Studies of this linkage usually focus on the demographic burden. The second linkage is through the complementary relationship between the demographic gift and institutional changes and investment.4

Box 21


Crop production (P) can be expressed as area (A) times production per unit area or yield (Y):

P = A x Y

The rate of growth in P (GP) is simply the sum of the rate of growth in area (GA) and the rate of growth in yield (GY):

GP = GA + GY

Production of crops (or livestock) (P) can also be expressed as a function of inputs, such as area (A), workers (W), machine services (M) and fertilizers (F):

P = F (A, W, M, F)

The rate of growth in production (GP) can then be expressed
as the cost-share weighted sum of the rate of growth in the production inputs plus a residual term measuring growth in total factor productivity (TFP) (GTFP):


This expression effectively defines TFP, which is the ratio of production (P) to an aggregated factor index (I). The growth
in I is:



TFP growth is, therefore, the difference between actual production growth (GP) and the production growth that would have occurred (GI ) if farmers had not changed the technology of production or their own efficiency behaviour. Growth in production can be achieved through increased factor use or through a more efficient use of production factors. It is the latter that is expressed by the TFP growth.


It must be emphasized that even the institutional change perspective is linked to investments, particularly investments in the production of public goods.5 The proper role of government in a market economy is to devise and administer institutions (legal systems, regulations, competition policy) that provide incentives for efficient private (farm) production while, at the same time, investing in the provision of public goods where appropriate. In practice, governments in many developing countries have often intervened in markets in inappropriate ways and have invested in state-owned production enterprises that have been inefficient. In recent decades, reform movements have been undertaken in many countries to privatize inefficient state-owned enterprises and to eliminate marketing boards and other cumbersome regulatory agencies. Some of these reform movements, however, have not fully appreciated the historical role of public goods in agriculture in all economies. Public sector investment in rural schools, agricultural extension and applied agricultural research has been vital to agricultural development in every economy in the world. Institutional reform without investment in these public goods does not produce economic growth in the agricultural sector. Growth is not produced by passive "let the markets work" policies that do not include critical public investment programmes.

Agricultural sector growth requires investments from the public sector.

The interrelationships between these public good investments are illustrated in Figure 22, which depicts schematically the process of improving agricultural productivity through technological progress, represented by five subsequent levels of technological capital (TC). For each TC level, four crop yield levels are shown for a given location. These are: (A) actual farm yield in this location; (BP) best practice yields, i.e. the profit-maximizing yields that are achieved when farmers use the best practices and technology suitable for this location; (RP) research potential yields, i.e. the yields that could be best practice yields if an applied adaptive research programme targeted to this location were in place; (SP) science potential yield, i.e. the yield level that could be the best practice yield if the applied adaptive research programme were supported by international and national "pre-invention science" programmes.

Associated with these yields, three "gaps" can be defined:

The process of improving agricultural productivity is associated with the progressive reduction of each of these gaps - starting with the extension gap, moving on to the research gap and then to the science gap - as the country's capacity improves for adopting and developing improved technologies, represented by the gradual movement from TC I to TC V.

In Figure 22, TC I is a level where little extension, research or science is being undertaken and where the research that is producing technology for other regions is not producing technology for the TC I region. Farmers' schooling levels are low, markets are poor and infrastructure is lacking. The extension gap is large in this stage, and thus investment in extension and infrastructure has a potentially high pay-off, even though there are few effective research programmes to raise best practice yields. Through extension programmes, the extension gap can be reduced to a fraction of its original size, represented by the transition to TC II. In order to move on from TC II to TC III, the economy must rely on the closing of the next gap, the research gap. This depends on the forging of a direct link between research and extension, where extension programmes are responsible for extending the results of applied adaptive research programmes to farmers. The further move from TC III to TC IV is associated with raising the research potential yield (RP), as the applied adaptive research programme is supported by international and national pre-invention science programmes. Further progress, i.e. to TC V, where the science potential yield can also be increased, depends on increasingly effective pre-invention science, research and extension programming.

Figure 22

Although most of Africa is at the level of TC II, some African countries have not yet made the transition from TC I and a few have achieved the transition to TC III, where research systems are producing significant flows of new technology suited to farmers in many regions. This is in contrast to the situation in both South and Southeast Asia where, by the mid-1960s, many economies were already at TC II. Since then, green revolution technology in rice, wheat, maize and other crops has enabled these economies to make the transition to TC III, and today, in many Asian and Latin American countries, there is the potential to reach TC IV.

It is possible that research systems in other regions can raise best practice yields in TC I economies before they have made the transition to TC II. In practice, most research gains have been realized in economies that have already achieved TC II or III market, infrastructure and skill levels. In some cases this has been induced by the development (often in international centres) of genetic resources and methods that increase the research potential (RP) yield levels. In some African countries, RP yield levels may be quite low because of limited genetic resources and difficult disease and insect problems, so the research gap is actually quite small. When this is the case, stimulus in the form of improvements in science (closing of the science gap) may be required to achieve better research performance.

Figure 23

Figure 23 provides further insights into the location specificity and pre-invention science issues. The figure comes from the first major economic study of the economics of agricultural technology, written by Zvi Griliches.6 It illustrates the adoption (in percentages of area) of hybrid maize by farmers in a number of states in the United States. Griliches noted that the technique of hybridization is actually the invention of a method of invention, i.e. it is a pre-invention science discovery. The actual inventions are the hybrid maize varieties targeted to regional niches, but the method itself is a product of pre-invention science (the modern counterpart of this is the development of biotechnology methods). Although the method has some location specificity, it is far less location-specific than the actual inventions. Thus, the inventions (hybrid maize varieties) suited to Iowa were not valuable (hence not adopted) in Alabama. It was only when breeding programme capacity in Alabama was improved and hybrid maize varieties had been developed for local conditions that the technology was made available to Alabama. Similarly, hybrid maize technology was not available to the Philippines or India until research stations were built in those countries. As will be seen below, this "Alabama effect" has been operative in all developing countries. Countries that do not have the capacity to adapt inventions have gained little from technology produced outside their own boundaries.

Productivity gains from technology occur in developing countries that have the capacity to adapt inventions from outside their own boundaries.


Productivity growth entails achieving increased production per unit of resource used to produce goods and services. Two types of indicator have been used to measure increases in productivity: partial factor productivity (PFP) indexes and total factor productivity (TFP) indexes. These indexes enable comparisons among regions (and countries) and time periods.

Measures of partial factor productivity

Figure 24

PFP measures are ratios of output to a single production factor. The most widely used PFP for the general economy is the labour productivity index, production (P) per worker (W) or P/W. Production can be measured in terms of a single product, in which case it can be measured in quantity units, or as an aggregate of products, meaning prices must be used to aggregate products into a constant price value. For the purposes of comparison over time, the prices used must be real or constant prices.

Yields have grown faster in developing countries than in developed countries but have not become more variable.

The most widely used PFP index for agriculture is production per unit of land, or crop yield. This index has been in standard usage for centuries and allows comparisons among locations and among time periods. Comparisons between locations must take soil and climate differences into account. Yield changes in a given location over time are widely used as indicators of improved economic efficiency.

Figure 24 portrays the yields of major crop groups in developed and developing countries for each decade from the 1950s until the 1990s. This Figure shows that, while crop yields are typically higher in developed countries, the rate of change in yields has been higher in developing countries.

Figure 25

A concern that is often expressed regarding yield increases is that the variability of yields seems to have increased as yields have increased. Increased yields are thus generally associated with increased farm income variability. To clarify this issue, the coefficient of variation7 is included in Figure 24. Coefficients of variation show no trends, and it appears that higher yields are not associated with higher variability in yields or production.

The chief limitation of PFP measures is that they can be affected by changes in factors other than the factor in the index. For example, increases in fertilizer applications per unit of land will increase yields, so it could not be inferred that an increase in crop yields is the result of genetic improvement or of reductions in transaction costs unless some control for other factor use (either statistically or experimentally) is utilized in the analysis. In spite of this limitation, yield indexes are very useful indicators of productivity change, as they are generally available for specific crops, enabling comparisons by crop.

Measures of total factor productivity (for aggregate agriculture)

Total factor productivity (TFP) measures are sometimes used to compare productivity in different locations, but they are chiefly suited to comparing productivity gains over time. They differ from PFP measures in that they are expressed as the ratio of production to a cost-share weighted index of different factors of production.

That is, TFP is the ratio of an index of production to an index of aggregate inputs. The rate of growth in TFP is the rate of growth in production minus the rate of growth in the aggregate input index. This corresponds to the difference between actual production growth and the production growth that would have occurred under conditions in which no productivity change occurred. (See Box 21, p. 248, for an arithmetic presentation of the concept.) Changes in TFP can also be interpreted as changes in the cost of producing a unit of the product, holding input prices constant.

A number of studies have measured TFP changes in different countries, but these are generally difficult to compare because studies and countries follow different practices regarding adjustments for changes in the quality of production factors and have different available data. However, data available from FAO sources allow the calculation of TFP growth rates for the 1961-1996 period for 89 developing countries, based on seven factors of production (see Box 22 for details of the calculations).

The results from these calculations are displayed in Figure 25 by region. Of the 89 countries, 14 had negative TFP growth rates over the 1961-1996 period. Of these, 11 were in sub-Saharan Africa and three in the Caribbean region. Given the crudity of the measurement, these negative TFP growth rates may be owing to error (only six were more than minus 0.5 percent). However, it is possible that some "real" TFP decreases have occurred where land degradation has taken place (discussed with the other unfavourable factors on p. 282). Interestingly, the region with the highest mean TFP growth is the Near East and North Africa. Perhaps the chief feature of these roughly estimated TFP growth rates is that they display a high degree of dispersion and variability from region to region.

Box 22


FAO AGROSTAT data allow the calculation of the TFP growth rate for 1961-1996 based on seven factors of production. First, rates of growth for production and for each factor are statistically estimated. Second, factor shares are estimated for three periods (1961-1976, 1971-1986 and 1981-1996) using factor-output ratios and estimates of factor shares from independent studies for India and Brazil. The arithmetic of the calculations is explained in the following.
Growth in TFP (GTFP) is defined as growth in production (GP) minus growth in production inputs (GI) (see also Box 21, p. 248):
Growth in production can be estimated for three periods (1961-1976, 1971-1986 and 1981-1996) from FAO AGROSTAT data as the coefficient b in a regression for each country.
log (P) = a + b year
Growth in inputs is defined as:
GA is growth in cropped area, SA is the cost share;
GI is growth in irrigated area, SI is the cost share;
GW is growth in agricultural workers, SW is the cost share;
GAN is growth in work animals, SAN is the cost share;
GT is growth in number of tractors, ST is the cost share;
GH is growth in number of harvesters, SH is the cost share;
GF is growth in fertilizer and SF is the cost share.
Growth rates for each factor were estimated for the three periods using the same procedure as for production.
Factor shares were estimated by calibration to TFP studies in India for several periods and in Brazil for several periods.

Sources: Factor shares for India: R.E. Evenson and Y. Kislev. 1975. Agricultural research and productivity. New Haven, Connecticut, USA, Yale University; R.E. Evenson and M.W. Rosegrant. 1995. Total factor productivity and sources of long-term growth in Indian agriculture. EPTD Discussion Paper No. 7, Washington, DC, IFPRI. Factor shares for Brazil: A.F.D. Avila and R.E. Evenson. 1998. Total return productivity growth in Brazilian agriculture and the role of Brazilian agricultural research. Economia Applicada.

Figure 26


The indicators discussed above are designed only to measure productivity growth. Additional methods are required to identify the sources of productivity changes. Four bodies of evidence can be used to gain insights into the sources, i.e. the investments, policy actions and institutions that produce productivity growth in agriculture. The simplest approach to identify sources is area-yield accounting. Area-yield accounting studies enable production growth to be separated into area and yield components. A more complex form of accounting are product-factor quality accounting studies. A third body of studies approaches the problem from a return to investment perspective. Finally, correlations between TFP growth and indicators of technological capital levels by country can also shed light on the question of sources of productivity growth.

Area-yield accounting

Area-yield accounting is based on the simple fact that growth in production can be separated into two components: growth in area and growth in yield.

Improvement in genetically determined plant performance, in particular, results in yield growth. Yield growth measures have been used as indicators of the contribution of the green revolution associated with improved (semi-dwarf high-yielding) varieties of wheat and rice in the late 1960s in South and Southeast Asia. These improved varieties are credited with increased yield and production and with saving millions of poor families from the dismal prospect of the Malthusian-resources development perspective.

Improvements in developing countries' agricultural yields did not stop with the green revolution but continue to grow.

Figure 26 (continued)

Figure 26 displays area-yield production accounting by decade for major crops in developed and developing countries. Briefly the comparisons show:

Studies of returns to investment in agricultural productivity

An alternative approach to TFP source identification is to measure the growth contribution of investments in production growth. Two different types of method have been applied: project evaluation methods and statistical methods.

Project evaluation methods. This approach attempts to measure benefits in the form of growth components that can be attributed to investments in agricultural research programmes, agricultural extension programmes, farmer training programmes and infrastructure. A number of studies attempting to measure the benefits attributable to agricultural research and extension programmes have been completed utilizing project evaluation methods. These allow the internal rate of return (IRR) of the project to be identified, among other indicators. The IRR is the rate of discount for which the present value8 of the benefits is equal to the present value of the costs. It can be interpreted as the rate of return or interest rate realized on investments in the programme, evaluated over a long period of time.

Table 16



of IRRs

Percent distribution

Median IRR


















By region:




















Latin America


















Applied research









By region:




















Latin America


















Pre-invention science









Private sector R&D









Ex ante research









Statistical studies of returns to investments. A second body of return to investment in agriculture and agricultural extension is based on statistical estimates of coefficients in PFP and TFP decomposition studies (see Box 23 for technical details). The basic idea of these studies is to identify, statistically, the contributions to growth in PFP or TFP with variables that are based on investments. Research programmes typically affect productivity with a time lag, which results from time lags between expenditure and discovery and between discovery and diffusion. Studies have estimated this time lag to be between five and ten years from spending to full productivity impact.

Decomposition studies can then compute an estimated benefits stream associated with a one-unit investment in a given period (or over several periods). This enables the analyst to calculate a marginal internal rate of return to investment which can be interpreted as the returns to public investment in the programme, where benefits are measured as the total benefits to producers and consumers.

Box 23


Statistical decomposition methods require that variables for a region and time period of the following general types be defined:

GTFP= a + b1Res + b2Ext + b3Sch + b4Inf


GTFP is a measure of TFP (when PFP measures are used, input prices should be used as independent variables);

Res, Ext, Sch, Inf are variables corresponding to research, extension, schooling and infrastructure services respectively.

The method enables the analyst to identify TFP growth contributions statistically with variables that are based on investments. Each of the explanatory variables is designed to reflect TFP services for the unit of observation. These services have both time and spatial dimensions which should be estimated and included in the statistical design.

For example, the research programmes servicing a region in the period t to t+1 will be based on investments made before period t. Research programmes typically affect productivity with a time lag, not only from expenditure to discovery, but also from discovery to diffusion. Studies have estimated this lag to be five to ten years for full TFP impact. Research service variables are constructed accordingly as time-weighted cumulations of prior investment.

Spatial dimensions must also be addressed because a region may benefit, not only from the research station associated with the region, but also from research findings from other stations and from private firms. This requires estimation of spill-in weights (although few studies have actually estimated these).

With time and spatial weights and estimates of b1, b2, etc., decomposition studies that compute an estimated benefits stream associated with a unit of investment in a given period can cover several periods. This enables the analyst to calculate a marginal internal rate of return to investment.

The estimates of b1, b2, etc., combined with changes in the Res, Ext, etc. variables, can also be used to attribute portions of TFP growth to each investment (see Table 17, p. 267).

It should be noted that TFP growth needs to be measured consistently as regards factor quality treatment. It is inappropriate to mix "raw" TFP estimates with quality adjustment estimates.

Studies on returns to investments (project methods and statistical methods). IRRs from studies of returns to extension and research based on both project evaluation and statistical methods are summarized in Table 16.9 While most studies on extension used statistical methods, studies on applied research used both methods. Distributions of IRRs for a number of study categories are presented. Two features characterize virtually every category. The first is that the IRRs are high - 74 percent of the extension IRRs and 82 percent of the research IRRs exceed 20 percent. The second feature is that the IRRs have a broad range of estimates.

Given the breadth of the range of IRRs in each category, it is difficult to draw strong conclusions regarding differences in means between categories. It can be noted, however, that the categories with the greatest proportions exceeding 40 percent are pre-invention science, private sector research and development, rice research, and fruits and vegetables research. A slightly higher proportion of research studies exceed 40 percent (59 percent) than is the case for extension studies (51 percent).

Regional distributions vary, with studies of both research and extension in Africa having lower proportions exceeding 40 percent than in other regions. Asian research IRRs are especially high.

There does not appear to be a time trend in the IRRs reported. Studies for later periods show IRRs similar to those in studies of earlier periods.10 This body of evidence indicates that both research and extension programmes are high pay-off investments for taxpayers.

Many studies have made comparisons between improved and modern varieties and traditional varieties (see Box 24). Plant and animal breeders have contributed to productivity growth by developing plant cultivars and animal breeds (and individual animals) that are more productive and less susceptible to damage by pests and pathogens (see Box 25).

Returns to public investments in agricultural research and extension are very high.

Box 24


The green revolution, popularized in the media in the late 1960s and early 1970s, associated productivity gains with high-yielding or modern crop varieties of wheat and maize. There is little question that improved varieties were important and that they were a catalysing force in productivity improvement. The images created by the popular press, however, were misleading in many respects. They left the impression that the green revolution was created by the exceptional insights of a small group of scientists in IARCs, and that the modern varieties produced as a consequence were confined to rice and wheat varieties developed in the late 1960s and diffused during the 1970s.

In reality, there were many breeding programmes in developing countries for a number of crops before the IARCs were established. These programmes developed many rice and wheat varieties for tropical and subtropical climate regions in developing countries. In rice breeding, for example, programmes to bring Japonica (temperate) genetic resources into Indica (tropical) rice cultivars had already achieved success in the 1960s. Progress in bringing temperate zone wheat genetic material into subtropical wheat cultivars had also been made in Mexico (led by a Rockefeller Foundation programme that was a predecessor of CIMMYT).

The IARCs brought more concentrated and directed attention to the plant breeding challenges presented by tropical and subtropical production conditions. They facilitated the exchange of genetic resources and of advanced breeding lines used as parents in national agricultural research systems (NARS) breeding programmes. Access to genetic materials was improved by the establishment of IARC gene banks and international nursery systems. IARC programmes complemented NARS breeding programmes and stimulated their expansion.

Recent studies of varietal production and release show that the rate at which new varieties were being developed increased during the 1960s and 1970s in several crops and that this rate of varietal production has been maintained since then. In rice, for example, more than 2 000 modern varieties have been released in more than 100 breeding programmes. New traits (disease resistance, drought tolerance, etc.) have enabled the expansion of high-yielding varieties into more production environments over time.

Many of the studies of returns to research are based on the contribution model by improved crop varieties. The development of improved crop varieties appears to be the key to moving from TC II to TC III in the technological capital formulation. A recent study of rice research concluded that roughly one half to two thirds of the productivity gains produced by rice research programmes was attributable to genetic improvements.

The International Agricultural Evaluation Group (associated with the Consultative Group on International Agricultural Research [CGIAR]) has undertaken a recent study of crop genetic improvement in developing countries. The study reports the following findings:

Crop breeding programmes in IARCs complement the crop breeding programmes in NARS by providing advanced breeding lines and varieties that have been extensively used as parent materials by NARS.

The rate of development of new crop varieties rose for wheat, rice and potatoes from the 1960s until the 1980s and remained constant in the 1990s. For other IARC crops, the rate of new varietal development continued to increase throughout the 1990s.

The complementarity between IARC breeding programmes and NARS breeding programmes has stimulated higher levels of investment in NARS breeding programmes.

New crop varieties have continued to make significant impacts on crop productivity as recently released improved varieties replace earlier improved varieties. Different rates of diffusion of improved crop varieties have been observed in different crops and regions, however. Some regions are disadvantaged because their soil and climate conditions are such that new crop varieties have little or no impact. This is a factor in the unevenness of productivity change.

Returns to investment in plant breeding programmes have been very high.

Accounting for contributions to productivity growth. The statistical studies summarized in Table 16 can also be used to account for PFP or TFP growth. Each variable (research, extension, etc.) in the statistical model has an estimated coefficient (or coefficients). These coefficients indicate how much PFP or TFP growth is associated with a given change in the same variable. Thus, it is possible to compute the contribution to growth produced by each source over a particular period.

Table 17 reports such growth accounting calculations based on studies for the United States, India and Brazil. For India, the contribution of high-yielding varieties is measured; these varieties were a major source of growth in the 1970s. The accounting structure used identifies several sources of TFP growth, including research programmes, extension, schooling and market developments. Such accounting, it should be noted, does not indicate that each source of growth is independent of other sources. In fact, the sources complement one another.

Table 17



US agriculture

Brazilian agriculture

Indian agriculture
















Annual TFP growth









Proportion resulting from: -Public sector research









(High-yielding varieties)









-Industrial R&D









-Agricultural extension









-Farmers' schooling









-Government programmes
























Sources: United States data: W.E. Huffman and R.E. Evenson. 1993. Science for agriculture, p. 212, Table 7.10. Ames, Iowa, USA, Iowa University; Brazil data: A.F.D. Avila and R.E. Evenson. 1998. Total return productivity growth in Brazilian agriculture and the role of Brazilian agricultural research. Economia Applicada, Table 13; India data: R.E. Evenson, C.E. Proy and M.W. Rosegrant. 1999. Agricultural productivity growth in India. IFPRI Report No. 109. Washington, DC, IFPRI.

Empirical evidence on the role of technological capital

In Figure 22, p. 251 productivity gaps were conceptually defined in terms of technological capital (TC) classes. Box 26,
p. 273 provides an empirical methodology for classifying countries into different TC classes based on eight TC indicators. Using this classification method, it was possible to place 89 developing countries into one of four TC classes in each of three different periods (1961-1976, 1971-1986, 1981-1996). Most countries achieved an improvement in TC class over the three periods, as indicated in Box 27, p. 274 where countries are identified in terms of their technological capital in each period.

Figure 27 displays distributions of TFP growth rates by TC change classes. First consider the 21 countries that have not advanced beyond TC II (i.e. 112 and 222). These countries effectively inherited little (TC I) or poor (TC II) technological capital and have achieved little improvement over the 1961-1996 period. Seventeen of the 21 countries are in sub-Saharan Africa. The sample of countries does not include several countries that have probably remained at TC I over the period, notably Somalia, the Congo and Ethiopia. Dispersion in TFP growth for this group is high with seven negative TFP growth rates, possibly reflecting soil depletion, but probably also problems of social stability, including civil strife. Many of these countries have limited capacity to provide basic services for their populations, none are industrialized. Five countries in the group had TFP growth rates of over 1 percent, but the group mean was only 0.2 percent. It would be reasonable to infer that these countries are still governed by Malthusian conditions.

Box 25


Prior to the advent of modern biotechnology methods, which enable genetic material from one species to be transformed into another species, breeders were constrained to searching for better genetic combinations within the existing cultivated crop and livestock species. For most of these species, rich within-species diversity is available to breeders in the form of landraces in crops and breeds in livestock. The diversity was created by farmers over previous centuries as they selected new types to suit new conditions in new locations and for expanding populations.

This legacy of farmer-selected biodiversity plus mutants and wild or weedy related species is of great value to breeders today. Many of the productivity gains in agriculture can be attributed to genetic improvements. Ex situ gene bank collections of genetic resources have been developed for all major crop species. These collections hold a high proportion of potentially collectable material, for which a policy of exchange has been maintained.

To date, several studies of genetic resource value have been made. Most work has been done on rice.1 These studies conclude that genetic resources have high economic value and that further collection, evaluation and prebreeding (to identify parental values) are economically justified. The studies also conclude that, as breeding moves into the biotechnology era, traditional genetic resources will be enhanced in value.

1 R.E. Evenson and D. Gollin. 1997. Genetic resources, international organizations and improvement in rice varieties. Economic Development and Cultural Change, 45: 471-500.

The second panel in Figure 27 displays TFP growth rates for 24 countries that moved from TC II in the first period to TC III in the second or third period (14 of these countries are in sub-Saharan Africa). This group of countries has an average TFP growth of 0.6 percent. Eighteen of them showed positive TFP growth, with three countries in the 1 to 2 percent range. This means that the average country in this group is realizing approximately enough TFP growth to prevent production per caput from declining, and a few countries are realizing good economic growth rates.

The third cluster of 29 countries were at TC III in the first and second periods. Twelve of them moved to TC IV in the third period. The mean TFP growth for this group was 1.53 percent, a very respectable rate. Only one country had negative TFP growth and nine had more than 2 percent TFP growth per year.

The fourth group of 14 countries includes countries that either inherited TC IV levels or had moved into them by the second period. This group includes China, India and Brazil. It has performed extremely well, with a mean TFP growth of 2.3 percent and four countries achieving more than 3 percent. This set of countries has R&D capacity in industry.

While these TFP measures are crude and the TC classes somewhat arbitrary, the TFP-TC correlations tell a powerful story.11 The story, simply put, is that countries in TC I realize little or no TFP growth. Moving to TC II, where basic government institutions are in place and some agricultural research capacity exists, modest growth is achieved. Countries with TC III capabilities can realize high rates of TFP growth and have significant agricultural research and extension systems. Countries with TC IV capabilities can realize super growth in agriculture. Since TC IV encompasses R&D capabilities in the agricultural supply sector, part of this super TFP growth is a spillover from industrial TFP growth.


Productivity changes and resource degradation

The sustainable development movement of the past two decades draws attention to the possibility that productivity gains in many countries may have been realized at the expense of degrading resources. It is often not recognized, however, that resource upgrading, either through investment in drainage and irrigation or through farming practices (crop rotation, fertilizing, liming), also occurs.

It is important to note that the TFP measures reported in the previous section, in principle, do take account of net degradation or net upgrading of the resource base. It is quite plausible that countries with low TFP growth rates are experiencing net degradation, and that countries with high TFP growth rates are experiencing net upgrading. To the extent that this is the case, resource degradation is inversely related to technological capacity and per caput income.

The literature on resource degradation rates identifies several problem areas, including salinization, soil erosion and intensification.

Salinization problems have been severe in some locations (e.g. Pakistan). In some cases, better management would have prevented these problems. In other cases, management may not have prevented the problem and, with hindsight, it can be concluded that irrigation should have been controlled.

Soil erosion occurs continuously; some areas benefit from erosion, others lose. Erosion on some soils translates into reduced natural productivity. In others this is not the case. Soil erosion can be, and in many countries has been, controlled and managed, in particular where households have secure property rights. (Crosson12 provides a useful survey of soil erosion and productivity effects.)

Box 26


Developing countries were classified according to technology capital (TC) classes in three different periods (1961-1976, 1971-1986, 1981-1996). Four TC classes were defined, based on eight indicators as shown below. The criteria ensure that countries are included in only one TC class in each period. Most countries achieved TC improvements in recent decades (see Box 27).






 Adult male illiteracy





 Proportion of labour  force in industry





 Foreign direct investment/GDP

Little or none



0.25% or more

 R&D in manufacturing firms/manufacturing value added





 Royalties and  licence fees paid





 Royalties and  licence fees received





 Agricultural research investment intensity

Low <0.25% of agricultural production

Moderate 0.25-0.5% of agricultural production

High =0.25% of agricultural production

High =0.5% of agricultural production

Intellectual property rights



Weak protection

Moderate protection

1 Indicator data are drawn from the World Bank Development Indicators database.

Intensification, i.e. the use of fertilizer and other chemicals and improved varieties to achieve high yields, is seen by some as damaging to land productivity. There is concern that such agriculture in the tropics may have negative environmental effects that have not been realized in high-yielding agriculture in the temperate zones. The experience with high-yielding agriculture in temperate zones is that, with proper management, even poor natural soils can be improved. This requires investments, and most soils in tropical regions, especially in Africa, have not benefited from these investments. The majority of soils in Africa have large unexploited potential for improvement through investment.

Box 27


Over recent decades most countries have changed their TC levels through investments and institutional development. Below countries are listed according to their classification by TC levels in each of the three periods 1961-1976, 1976-1986 and 1986-1996 (e.g. 112 indicates TC I in the first two periods and TC II in the third period).
Most of the countries in the 111, 112 and 222 categories inherited TC I levels in the 1950s; the countries in the 222 classification had upgraded their TC rating to TC II by about 1970, but have remained there since. Only a few countries have achieved more than one TC upgrade over the 40 years covered. More than half have achieved one upgrade.

COUNTRY CLASSIFICATIONS (1961-1976, 1976-1986, 1986-1996)












Burkina Faso


Dominican  Rep.







Côte d'Ivoire















Costa Rica



Lao People's Dem. Rep.






Korea,  Rep.






Islamic Rep.

El Salvador






















Guinea- Bissau




Libyan Arab Jam.

Saudi Arabia







Sierra Leone









Sri Lanka








Tanzania, United Rep.










Viet Nam


























Syrian Arab  Rep.





Trinidad and Tobago




Productivity changes and income distribution

The benefit from productivity gains can be distributed in different ways among producers and consumers and across locations, according to circumstances. Productivity gains lower costs, which leads to increased supply. In a small economy open to international trade, domestic prices are determined by international market prices, so prices do not fall when production costs fall. In this case, consumers do not gain from productivity change and producers capture all the benefits. For an economy that is closed to international trade (or for non-traded goods), on the other hand, prices will fall when productivity gains lower costs. This produces gains to consumers, while individual producers will gain or lose according to whether their average costs declined by more or less than prices declined.

Figure 27

Cost declines may vary among producers, and farmers may differ in adoption patterns. Early adopters will then have cost declines before they accrue to late adopters. Credit-constrained farmers may have slower cost declines than farmers with good access to credit. These factors appear to have been of importance in developing countries, but extension and infrastructure programmes have had a levelling effect on them.

The most important factor affecting cost declines in an economy, however, is the nature of biotechnology and its interaction with soil and climate conditions. The first modern green revolution rice varieties were adopted on only about 30 percent of India's irrigated and rainfed rice land. Even though they were very productive under certain conditions (e.g. with good water control) their advantage was reduced or lost under other soil, climate and pest/pathogen conditions. Several generations of breeding for host plant resistance to insect pests and diseases and for host plant tolerance to abiotic stresses (e.g. drought stress) produced rice varieties that are now planted on roughly 90 percent of India's irrigated/rainfed rice area. But this high-yielding varietal technology still remains unavailable to many farmers in upland, deepwater and other unfavourable production conditions for rice.

Technology-induced declines in production costs may benefit only some farmers and may harm farmers who do not have access to the technology.

This favourable-unfavourable production environment situation has been an important element in most developing economies. It creates serious inequities and conditions where improved technology benefits all consumers and some producers, but actually harms producers who do not have access to cost-reducing technology for reasons of location. There are some remedies for this situation, although they will not necessarily create equal access to productivity gains. The most important one is to develop research and experiment station systems designed to serve all regions. This can then be supported by extension, schooling and infrastructure investments. Plant breeding programmes, as noted in the Indian example, can and do tailor genetic improvements to local conditions. Another remedy is the movement of people. A recent study by IRRI13 concluded that labour mobility enabled workers to avoid the unfavourable environment effects on wages.


Economic studies of productivity gains in agriculture show considerable variation in productivity performance across countries and periods; but patterns do exist. Studies of productivity change (and the TFP-TC correlations reported above) are consistent in showing that technological capital is critical in determining productivity performance. Technological capital is cumulated over long periods of time. Developing countries differ greatly as regards the capacity inherited from colonial regimes in the early part of the second half of the twentieth century, and also in their investment in upgraded capacity over the half-century.

From the agricultural development models discussed earlier in this chapter, the following generalizations appear to be supported by the experience of the past several decades.

The Malthusian-resource model could be considered valid for a subset of developing countries. Those economies that did not invest in a minimal level of TC realized few or no productivity gains and experienced declining per caput incomes unless they had abundant land resources. Of the 21 developing countries failing to move beyond TC II over the past half-century, none appear to have attempted to solve their problems with the classic Malthusian remedy - population control. Had they done so without investing in technological capacity, it appears that little income growth would have been achieved. Some studies nevertheless conclude that high population densities stimulate investments in technological capital.

Some of the TC II countries attempted to achieve productivity growth by stressing the best practice model. This model calls for investment in agricultural extension as the lead TC strategy and has had rather limited success. When countries have moderate levels of literacy, extension "mining" (see Figure 22, p. 251) to improve farmer productivity by using existing technology more effectively produces only limited growth for a limited period of time.

It is the adaptive invention model that enables countries to achieve high productivity growth. This means building a research capacity. It also means building a capacity to train researchers. With the development of IARCs and their support of TC III national agricultural research programmes, high rates of productivity growth have been realized. As the programmes are complemented by institutional investments in markets and infrastructure, their effectiveness increases.

However, super productivity growth in agriculture takes place only when productivity growth is being realized in the industrial sector. This occurs in TC IV countries and, when it occurs, there can be massive poverty reduction. There is little or no evidence for leaping from TC I to IV - countries do not achieve TC IV without having first achieved III. This indicates that the agricultural sector is an important catalyst for TC IV development.

Real prices of most cereal grains have dropped by about 50 percent over the past 50 years.


The production and productivity performance of the agricultural sector in developing countries over recent decades was both extraordinary and uneven; extraordinary because of the magnitude of the production and productivity gains achieved, and uneven because production and productivity gains were realized at different rates in different regions and countries. The extraordinary quality of the production performance in today's relatively open and globalized economy has been a factor in the realization of low food prices. In fact, the real prices of most cereal grains in 1999 were roughly half their 1950 levels.

As for investments, the experience of the past half-century has been very different for investments in true public goods production for agriculture (research, schooling and extension) and for investments in most state-owned enterprises. Investments in public good production have yielded high returns to taxpayers, while investments in most state-owned enterprises have not. Governments in developing countries and development agencies have not always been able to distinguish between productive and essential public investments, and unproductive and non-essential public investments where the private sector is the efficient form of economic organization.

There are many lessons from the events of the past half- century. Experiences differ by country, period and commodity, but certain patterns stand out and are of relevance in the context of future planning. The following are some of the general lessons that can be drawn.

Productivity gains require investments. It simply is not the case that countries can achieve productivity gains via technological spillovers ("spill ins") without investing in technological capital.

Investment can be achieved via two means. The first is to provide an institutional and policy environment that provides incentives for private sector investment, including foreign direct investment. The second is to engage in public investment and, in many cases, public sector control of activities that the private sector does not invest in (and cannot be encouraged to do so). This balancing of public and private investment is not easy. It requires careful evaluation and planning.

Population control programmes based on the Malthusian-resource model will not, in and of themselves, produce real welfare improvement. They must be complemented with investments in technological capital.14

Most inventions in developing countries are adaptive inventions, i.e. adaptations of inventions made in developed countries. In public sector agricultural research programmes, IARCs facilitate adaptive inventions, as do private multinational firms. Developing countries can exploit international productivity sources if these are accessible to them and if they have invested in domestic capacity.

There is a sort of technological sequencing within productivity improvement. Genetic improvement appears to be central in that it provides a complementary relationship with extension and agronomy, pathology, entomology, economics and other research fields. Once an economy reaches an advanced level of technology capacity, the effectiveness of extension and management advice depends on investment in research.

The full effects of productivity improvement are realized through broader economic changes. Agricultural productivity gains may not accrue exclusively, or even primarily, to farm producers. Markets distribute the gains to both producers and consumers and enable the general population to benefit.

As developing countries seek to move to TC IV, they must first progress through I, II and III. For most countries in TC I and II, the agricultural sector is the dominant sector of the economy and obviously critical to development. The technological capital investments required in agriculture to move to TC III are investments in public sector agricultural research and extension.

Projections suggest that productivity growth in agriculture will continue for the next 25 years.


As the twenty-first century begins, there are both favourable and unfavourable factors which will play a role in agricultural performance. Several recent studies by the World Bank, FAO and the International Food Policy Research Institute (IFPRI) have projected global production trade and prices for agricultural products. All agree that for the next 25 years or so the favourable factors will outweigh the unfavourable factors and that per caput food production will increase sufficiently to prevent food prices from rising. In fact, all three models project declining real food prices. Thus the extraordinary nature of the food and agriculture sector performance is likely to continue for some time. (These models take into account the fact that very new technology will become effective in the near future.) The unevenness of food and agricultural performance is, unfortunately, also likely to continue.

Favourable factors

Virtually all developing countries experienced high rates of population growth in the 1950s and 1960s. Declines in death rates, especially of infants, had occurred in the 1940s and 1950s producing a population boom. This also gave rise to a population burden, because populations grew faster than the workforce (it takes 15 years for the newly born to become workers) i.e. the dependency ratio rose.

Partly for policy reasons (family planning programmes, rural health progress, etc.) and partly for economic reasons, families began to reduce fertility rates and thus enter the second phase of demographic transitions. The fertility rate began to decline in different countries at different times, with Taiwan Province of China, Singapore, Hong Kong and the Republic of Korea - the "Asian Tigers" - leading the way. By the 1970s, most Latin American countries and other countries in Asia were also experiencing fertility declines and, by the 1990s, they had spread to virtually all developing countries. For developed countries the fertility decline was truly astounding; by 1995 almost all developed countries were below replacement fertility levels.

There is still considerable population momentum associated with the fact that more children in the last generation means more mothers in this generation, but the largest global population increments occurred a few years ago and further increments will become smaller each year. As fertility declines and population growth slows, a demographic gift in the form of a falling dependency ratio takes place because the workforce grows more rapidly than the population. This gift is very favourable for agriculture, since workers are vital to agricultural production.

At the beginning of this half-century, the technological capital of developing countries was very limited. Only a few had achieved productive research capacity (TC III) in 1950. By 1990, most developing countries had technology capital levels of TC III or IV.

In 1950, the CGIAR system of IARCs had not yet been developed. By the 1990s, several IARCs had produced important research and technological findings, many modern high-yielding variety programmes were being developed by IARCs, and the exchange of genetic resources, including advanced breeding lines, was facilitated by the IARC gene bank and international nursery systems. This made NARS more productive.

The development of IARC-NARS programmes is still not complete, however. Progress in increasing the technological capacity of countries with limited capacity has not been particularly good in the past decade or so, as international support for the programmes has diminished. Nonetheless, the existing capacity and the development of improved varieties, agronomic practices (integrated pest management), etc. provide technological momentum that will ensure further productivity growth probably at something like the growth record of the 1990s.


The green revolution High-yielding varieties of rice are
now planted in 90 percent of India's irrigated and rainfed rice


The biological sciences that provide the scientific foundation for the agricultural sciences have also made extraordinary progress in recent decades. Fundamental scientific discoveries in the basic fields of science are occurring at unprecedented rates and it is no exaggeration to suggest that a scientific revolution has been under way in recent years. This scientific revolution has stimulated a technological revolution in the form of biotechnology, which is still in its formative stages and has accumulated a broad array of critics. The most aggressive investors and developers of biotechnology products for agriculture have been private companies. Transgenic products are now widely used in a number of developed countries, and stronger intellectual property right (IPRs) are critical to private investors in these countries.

The public sector research system in developed countries is also responding to pressures created by advances in science, by strengthened IPRs and by the rapidly growing private sector research and development activities. The response is seen in graduate study programmes, research project selection and design and a research system culture in which free exchange of scientific information is possibly being curtailed. Public research systems see both threats and potential in the biological science revolution.

To date, the response of agricultural research systems in developing countries has been very slow, and developing country access to biotechnology products from the private sector will be more difficult than in developed countries. At present, the TC IV developing countries are realizing some of the benefits of biotechnology, but the TC III countries (and certainly the TC I and II countries) have yet to build capacity to benefit from this scientific advance.

Productivity experience over recent decades clearly shows that the agricultural sector benefits from robust growth and development in the industrial sector in both developed and developing countries. The agricultural sectors in developed countries are undergoing structural change (farm size, specialization, contracting, etc.) as a result of industrialization, but structural change has been less important in developing countries. Industrial growth provides improved factors of production to agriculture. It also improves the functioning of labour markets and provides income growth which stimulates improved market efficiency. Industrialization in developing countries, notably in East and Southeast Asia, has proceeded rapidly. The crisis of the late 1990s in Southeast Asia appears to be subsiding and the next decades are likely to be decades of rapid industrialization.

Unfavourable factors

Land degradation has probably occurred at a substantial rate in some countries over recent decades. Land improvement has as well. In almost all countries, good management of lands leads to net improvements. It will be the case, however, that not all countries will have good management in future decades. This is especially important for countries where the long process of developing practices to raise soil productivity is in its early stages. Land degradation and land improvement take place in all countries, but in countries with high TC levels and supporting institutions, land tends to be managed more efficiently.

In developed countries, cropland expansion has virtually ceased (and actually reversed in most cases). With investments in drainage and soil conservation measures, land is more productive than it was half a century ago. In developing countries with low levels of technological capital, poor institutional conditions and high rates of population growth, cropland expansion and fallow period reduction are occurring. Land improving investments are, however, also occurring - especially irrigation investments. It is possible that some of the negative TFP performances realized in some countries reflect significant land degradation.

Water is scarce in some regions, abundant in others. Irrigation systems have expanded over the past half-century in most developing countries, but it is probably the case that irrigation investment opportunities have been exhausted in many regions. Furthermore, irrigation system management has not been ideal in many regions. But water scarcity is similar to land scarcity; productivity gains, especially genetic improvement gains, that enable more production per unit of land also enable more production per unit of water.

When referring to resource constraints, however, it is important to emphasize that technological capital and supporting institutions can alleviate such constraints. Resource scarcity and degradation are, therefore, most important in the poorest countries.

A lack of capital and favourable institutions in poor countries impedes solutions to resource degradation problems.

Most TC I and II developing countries have been losing ground to developed countries for the past half-century. These countries face difficult political and policy environments, both domestically and internationally. The biotechnology revolution will almost certainly mean that they will lose further ground, given today's political and policy climate and the failure of these countries to make the investments required to achieve productivity growth.

Those developing countries that have achieved TC III status have generally gained ground on developed countries over most of the past half-century. They have reduced their technological lag through adaptive invention programmes and have been aided by IARCs. For the past decade, however, they have been losing ground as biotechnology inventions at the frontier have been made and put into practice. The political hostility to biotechnology has contributed to this by inhibiting the IPR reforms and other policies important to gaining access to this technology. It has also contributed by inhibiting investment in training required to modernize the agricultural sciences in both NARS and IARCs.

Many developing countries that have achieved TC IV status, however, do have the institutional backing to enable them to benefit from advances in biological sciences and the associated biotechnology inventions and are less likely to lose ground to the developed countries because of delayed implementation of these advances.


1 For the developed countries overall, total feed use in 1997 corresponded to 60 percent of domestic supply of cereals, as against only 21 percent for the developing countries.

2 Thomas Robert Malthus (1766-1834), in his Essay on the principle of population as it affects the future improvement of society, discussed the interaction between an exponentially growing population and a linearly growing natural resource base which, without restraints on population growth, would lead to continued pressure on living standards.

3 J. Simon. 1977. The economics of population growth. Princeton, New Jersey, USA, Princeton University; E. Boserup. 1981. Population and technological change: a study of long-term trends. Chicago, Illinois, USA, University of Chicago Press.

4 D.E. Bloom and J.G. Williamson. 1998. World Bank Economic Review, 12: 419-456.

5 Public goods are goods that cannot be withheld from single individuals without being withheld from everybody (classic examples are national defence, police protection, street lighting). For this reason they cannot be provided by private entrepreneurs (who would not be able to impose a payment for the public good on its beneficiaries, and would thus have no incentive to provide it) and must, therefore, generally be provided by the public sector.

6 Z. Griliches. 1957. Hybrid corn: an exploration in the economics of technological change. Econometrica, 25: 501-522.

7 The standard deviation divided by the mean for each decade, the coefficient of variation is a statistical indicator of the degree to which the various observations in a sample are dispersed around its mean. The smaller the coefficient of variation, the closer the observations on the whole are to the mean; the larger the coefficient of variation, the more they are dispersed around the mean value of the sample.

8 The present value is the value today of a future sum or flow of money. It is calculated by discounting, from the future sum or flows, an interest rate equivalent to the interest at which the sum could have been invested.

9 R.E. Evenson. 1999. Economic impact studies of agricultural research and extension. New Haven, Connecticut, USA, Yale University (mimeo).

10 Ibid.

11 Note 27, however, that the TC classes are actually quite firm, regardless of the indicators on which they are based. Fewer or alternative indicators and indicator criteria would not change the membership in the clusters in Figure 27 very much.

12 P. Crosson. 1995. Soil erosion and its on-farm productivity consequences: what do we know? Resources for the Future Discussion Paper No. 95/29; and P. Crosson. 1997. Will erosion threaten agricultural productivity? Environment, 39: 4-9, 29-31.

13 C.C. David and K. Otsuka, eds. 1994. Modern rice technology and income distribution in Asia. Los Baños, the Philippines, IRRI.

14 Boserup, op. cit., note 3.

Previous PageTop Of PageNext Page