Productivity Growth and Sustainability in Post-Green Revolution Agriculture in the Indian and Pakistan Punjabs

Irrigated agriculture in the Indo-Gangetic Plain of northern India and Pakistan is critical to the food security and livelihood of some 500 million people in an area that includes the largest concentration of poor people in the world. The agricultural sector, employing over 70% of the population and accounting for up to one-half of total GDP, is the key to income growth and poverty alleviation in this area. Beginning in the mid-1960s, most of the region has experienced rapid agricultural growth as a result of widespread adoption of Green Revolution technologies, accompanied by investment in irrigation and market infrastructure. The Green Revolution was spearheaded by the introduction of modern varieties of rice and wheat, the two major crops in the Indo-Gangetic Plain.

In recent years, there has been considerable concern regarding productivity growth and sustainability of these irrigated agricultural systems. Use of modern inputs is now high over much of the area, especially in the Indian and Pakistan Punjabs, and further intensification of input use is providing low returns at the margin (Byerlee, 1992). Intensification of cropping, especially in the wheat-rice cropping system which dominates much of the area, has also resulted in apparent degradation of the resource base in the form of salinization, over-exploitation of ground water, deterioration in soil physical and chemical properties, and pest and disease problems (Fujisaka, Harrington and Hobbs, 1994; Siddiq, 1994).

Despite this evidence, there is considerable controversy about the aggregate performance of the agricultural sector and the severity of resource degradation. There does seem to be clear evidence that the growth rates of yields of rice and wheat have slowed sharply over the past decade despite continuing increases in input use. However, good performance in non-food crops, such as cotton in Pakistan and oilseeds, fruits and vegetables in northwest India, and in livestock, may have offset the slowdown in foodgrains.

Crop yields are, however, only a measure of partial factor productivity. Total factor productivity (TFP) growth has been widely used as a measure of the overall performance of the agricultural sector. TFP compares an index of output changes with an index of input changes, so that the residual growth of productivity is attributed to technical progress, changes in the quality of inputs, and changes in the physical and economic environment. Over the longer run, experience from industrialized countries suggests that TFP in the agricultural sector should grow at 1.5-2% per year, with one-third to two-thirds of that due to investment in agricultural research and extension.

Recent estimates of TFP for agriculture in Pakistan and northwestern Indian provide conflicting conclusions. In Pakistan, two studies indicate negative TFP growth in the post-Green Revolution period, especially in the Punjab (Azam, Bloom and Evenson, 1991; Ali and Velasco, 1994). By contrast, a study by Khan (1994) concluded that TFP in Pakistan has grown sharply in the period 1980-92, at a rate of 2.1% annually, suggesting that the Pakistan agricultural sector has performed well in recent years. There is little recent evidence on aggregate TFP growth for northwest India although the few studies available generally indicate positive TFP growth (Kaur, 1991; Sidhu and Byerlee, 1992; Kumar and Rosegrant, 1994; Evenson, Pray and Rosegrant, 1999).

The above studies vary widely in the coverage of inputs and outputs, methods of valuing inputs, the index procedure used to estimate TFP, and the level of disaggregation. These differences make it difficult to draw meaningful comparisons and reach definite policy conclusions. Moreover, none of the studies have attempted to quantitatively disaggregate productivity growth into the effects of technical change, education and infrastructure, and especially, degradation in the quality of land and water resources.

Better estimates of productivity growth and sustainability under intensification are critically important to food security and poverty alleviation in the Indo-Gangetic plain, the breadbasket of northern India and Pakistan. This paper reports comparable estimates of TFP growth in the Indian and Pakistan Punjabs since the advent of the Green Revolution. Considerable resources were invested to collect data on individual crop and livestock products, inputs, and prices at the district-level in both states. To avoid the problem of aggregation across heterogeneous regions, TFP is estimated separately for specific agro-ecological zones defined in terms of cropping systems. The delineation of similar cropping systems across states enables direct comparisons of productivity trends and helps determine if productivity slowdown and/or environmental degradation is associated with particular cropping systems and ecologies. In Pakistan, where detailed data on resource quality were collected, productivity growth is econometrically decomposed into the effects of technology, resource degradation, human resources, and infrastructure.

There are several measures of TFP that use different rules for aggregating outputs and inputs (Alston, Norton, and Pardey, 1995). We use the chain-linked Tornqvist-Theil index since it provides an exact measure of technical change for the linear homogenous translog production function with Hicks-neutral technical change (Diewert, 1976). TFP is obtained by taking the difference between the growth rates of the aggregate output and input indices:

(1)

where QI_t is the aggregate output index, XI_t is the aggregate input index, and R_it and S_jt are the revenue share of output i and cost share of input j at time t, respectively.

For purposes of computing TFP, annual data on all inputs, outputs, and prices were collected at the district-level in the two Punjabs. Data were collected from statistical agencies and secondary sources for the period 1961-94 in India and 1966-94 in Pakistan. Crop and livestock products were aggregated into an output index using district-specific farm harvest prices for crops and market center-specific prices for livestock products. Input categories included land, labor, water, machinery, draught animals, fertilizer, and pesticide costs. To minimize aggregation bias in TFP, inputs of different qualities were valued by the price of each input-quality type. Land was divided into irrigated and unirrigated land, labor was disaggregated into skilled and unskilled labor based on the rural literacy rate in each district, water was divided into canal and tubewell water, and fertilizer into individual nutrient sources — nitrogen, phosphorous, and potassium.

Finally, comprehensive district-level data on soil quality (organic matter, phosphorous content, and soluble salts) and groundwater quality, based on soil and water testing, were assembled for the Pakistan Punjab during 1971-94. Disaggregated resource quality data for the Indian Punjab or for the pre-1971 years in Pakistan were not available.

District level data were aggregated to quantify TFP growth in terms of the dominant cropping pattern. The districts in India were divided into three cropping systems -- (1) wheat-rice, (2) wheat-cotton, and (3) wheat-maize, while there were four distinct systems in Pakistan; (1) wheat-rice, (2) wheat-cotton, (3) wheat-mixed summer crops (often maize or sugarcane), and (4) wheat-mungbean (or wheat-fallow).

Growth in TFP was analyzed for three periods corresponding to different phases of Green Revolution technical change (Byerlee, 1992). The Green Revolution period (1966-74) corresponds to the widespread adoption of input-responsive modern varieties of wheat and rice which led to a dramatic increase in production. This was followed by an input intensification period (1975-85 in India, and 1975-84 in Pakistan) when use of fertilizers and capital inputs increased rapidly, and a post-Green Revolution period (1986-94 and 1985-94 in India and Pakistan, respectively) when input use leveled off.

Table 1 describes the production record for the three major crops – wheat, rice, and cotton – in the two Punjabs. Technical change in agriculture in both states was triggered by the introduction of modern varieties (MVs) of wheat and rice during the mid-60s. MVs of wheat were adopted widely and rapidly in both states. Production of wheat during the Green Revolution period increased by over 7% annually in both states, with yield increases accounting for a little over half of that growth. Rice production also grew rapidly in both states, and especially the Indian Punjab. In the post-Green Revolution period, growth rates of yield have decreased to 2% per year for wheat, and become stagnant or negative for rice, fueling concerns for the sustainability of the irrigated wheat-rice cropping system.

While the overall trends in production of wheat and rice have been similar in both states, the gap between the yields in the two states has widened over time. Yields for wheat in India during the post-Green Revolution period were nearly double those in Pakistan, although they started at the same level at independence. Rice yields were also much higher in India, but this was partly due to specialization in low-yielding but high-valued Basmati rice in Pakistan. It is notable, however, that cotton yields in Pakistan were higher, both in level and growth rates.

These differences in production performance relate to differences in input use and cropping intensity (Table 2). Use of modern varieties stimulated rapid input intensification in both states. In the Indian Punjab, fertilizer use jumped from 33 kg of nutrient per hectare of cropped area in the Green Revolution to 156 kg per ha in the post-Green Revolution period. Labor use gradually declined, while mechanical power (tractors, harvesters, and threshers) increased from 4.3 hrs per ha in the first period to 41 hrs per ha in the most recent period. The same patterns were repeated across the border, in Pakistan. However, the use of fertilizer and machinery as well as cropping intensities were considerably lower than those in the Indian Punjab in all periods.

Finally, there are strong indications of degradation of the water and soil resource base in both states. The wheat-rice cropping system of the Indian Punjab has witnessed a steep decline in the water table while rising water levels have led to severe water logging in the wheat-cotton zone. Data from the Pakistan Punjab also confirm a serious problem of water logging and salinity, due in part to deterioration of tubewell water quality (reflected in a significant increase in residual carbonate and electroconductivity of groundwater in Figure 1). Soil quality in terms of available soil organic matter and phosphorus has also deteriorated, particularly in the wheat-rice zone (see Figure 1).

3.2 State-Level Trends in Total Factor Productivity

Table 3 summarizes trends in output, input, and TFP in the Indian and Pakistan Punjabs at the state-level. Overall, the Indian Punjab has had a better record of growth, with a 5% annual rate of growth in output, and a 1.9% rate of growth in TFP. In contrast, output in Pakistan grew at 3.2% per year, with TFP gains accounting for 1.5% of that growth.

In both states, there were sharp differences in TFP growth over time. Contrary to most views, TFP increased little in the Green Revolution, even as output growth was most rapid during this period. Surprisingly, TFP gains accelerated during the input intensification period, after the adoption of modern varieties was essentially complete.

There are three possible reasons for the time lag between the adoption of modern varieties and the realization of TFP gains. First, empirical evidence from areas of Asia that experienced rapid change following the Green Revolution suggests that in the initial years of adoption, technical inefficiency was fairly high (about 30%) but may have decreased with time as farmers learned about the technology. High levels of technical inefficiency could be traced mainly to deficiencies in information and technical skills (Ali and Byerlee, 1991). Both factors are likely to have been serious in the two Punjabs where a single generation of farmers, with low levels of education, switched from a largely traditional agriculture to complex multiple cropping systems produced with high levels of modern inputs.

Decrease in technical inefficiency a few years after the adoption of modern varieties may be attributed to ‘learning-by-doing’ as farmers gained experience using the new technologies and to increase in human capital as levels of schooling increased in both states. Indeed, evidence from India suggests that Green Revolution technological change directly increased the returns to education, spurring greater private investment in schooling particularly in states like the Punjab that grew most rapidly during that period (Foster and Rosenzweig, 1996).

A second reason for the time lag between technology adoption and realization of TFP gains is that adoption of modern varieties was accompanied by significant capital accumulation, particularly tubewells, during the Green Revolution. Indivisibility of investment, in combination with a preponderance of small farms as in the Indian Punjab, is likely to have led to the under-utilization of tubewells at least in the short-run. Under-utilization of quasi-fixed inputs leads to the underestimation of TFP gains during the period of adoption when investment costs are incurred, but TFP gains become observable as excess capacity is absorbed (Berndt and Fuss, 1986).

Third, low TFP growth during the Green Revolution relates to limitations of the conventional method of productivity measurement when technical change is biased towards saving one or more factors (Murgai, 2000). When technical change is biased, it is impossible to separately identify the contribution of technical change from that of factor accumulation since part of the contribution of technical change is captured in changes in factor shares which are used to aggregate inputs. In the case of the land and labor-saving Green Revolution technologies, conventional TFP estimates underestimate the contribution of technical change to growth, particularly during the Green Revolution.

The post-Green Revolution period witnessed a reversal in productivity performance between the two states. While TFP growth slowed slightly from 1.8% to 1.5% during this period in India, the rate of TFP growth rose sharply to 2.9% in Pakistan. The increase in Pakistan is partly due to a large increase in cotton yields (from 267 kg/ha during the intensification period to 601 kg/ha in the post-Green Revolution), mungbean production, and gains in the livestock sector. The strong performance of cotton and mungbean in Pakistan was due to the introduction of modern varieties of both crops and a sharp increase in pesticide use, although the latter has proven technically and environmentally unsustainable in recent years. By contrast, irrigated cotton in India was a much less important crop than in Pakistan (Chanmugam, 1994; Kurosaki, 1999). Until the recent reforms in 1994, quantity restrictions on cotton kept its domestic price below world prices, and non-tariff barriers restricted cotton imports as well (World Bank, 1996). Varietal research for irrigated India consequently lagged, although India has been very successful with rainfed cotton for central India. In addition, because of price policies favoring production for the domestic market, the Indian Punjab produced short staple cotton for local consumption and Pakistani cotton varieties (long staple for export) were not adapted to Indian conditions.

Comparing growth performance across cropping systems shows that there was considerable diversity in experiences within both states (Table 4). The wheat-rice systems that were major beneficiaries of the modern semi-dwarf varieties for both crops had the lowest rates of TFP growth. In the Indian and Pakistan wheat-rice zones, TFP grew at 1.4% per annum and 0.1% per annum, respectively. These results confirm widespread concerns that continuous double cropping of cereals, especially rice and wheat which require very different soil and water management practices, is an unsustainable cropping pattern (Pingali and Rosegrant, 1993; Byerlee and Siddiq, 1994; Cassman and Pingali, 1995; Ali, 1996). Deterioration in soil and water quality, discussed earlier, seems to be especially serious in this system, as evidenced by indicators of soil and water quality disaggregated by system in Pakistan (Ali and Byerlee, 2000). The better performance of this system in the Indian Punjab reflects in part the focus on early maturing coarse rice for local consumption, as well as concerted efforts to arrest resource degradation (e.g., through widespread use of gypsum to combat secondary salinity from tubewells).

By contrast, the wheat-cotton systems in both states had much higher rates of TFP growth. Better rural infrastructure, especially canal irrigation, and higher literacy may account for this difference. However, pesticide use in this system is high, and has led to environmental and health damages that are not accounted for in the TFP estimates.

TFP growth was relatively high also in the wheat-mungbean system of Pakistan. Production growth rate in this system was maintained in the post-Green Revolution period following the release of early-maturing mungbean varieties that allowed the adoption of a more sustainable cereal-legume rotation. In the wheat-mungbean zone, yield growth rates for wheat were among the highest in the state.

Together, temporal and spatial differences in productivity growth highlight the potential role of technological change, infrastructure and human capital, and resource degradation in determining TFP growth. Indeed, previous studies have related TFP growth in the Indian and Pakistan agricultural sectors to technical change (tied explicitly or implicitly to research investments), extension systems, infrastructure investments, human capital endowments, and policy reform (e.g., Rosegrant and Evenson, 1992; Kumar and Mruthyunjaya, 1992; Fan and Hazell, 1997 cited in Pingali and Heisey, 1999). However, there is little evidence of the quantitative impact of resource degradation on productivity growth.

Part of the problem has been the difficulty of agreeing on how to measure the impact of resource degradation on TFP. Some have argued that TFP measurement should incorporate changes in resource quality and externalities (Herdt and Lynam, 1992), while others measure TFP only with conventional inputs and consider resource stocks as a technical constraint that influences trends in TFP (Squires, 1992). We prefer Squire’s approach in TFP estimation for three reasons. First, attempts to account for market failure and social prices in estimates of TFP violate the theoretical basis underlying those estimates (Byerlee and Murgai, 2000). Second, in practice, it is difficult to value changes in resource quality, even where these changes can be physically quantified. Finally, in the medium term covered by this paper, farmers may not be able to observe resource degradation and therefore it is exogenous to decision making rather than endogenous.

Keeping these considerations in mind, in order to assess concerns of resource degradation in detail, we estimated a cost function relating costs of production to human and physical infrastructure development, technological change, and resource quality.ⁱ The adult literacy rate was used to capture the effect of changes in labor quality. The inverse of the distance of a village from the nearest metal road was used to quantify the effect of improvement in physical infrastructure. The effect of technological change was proxied by two variables (i) the proportion of area sown to modern wheat varieties, and (ii) cropping intensity. Cropping intensity is a proxy for adoption of MVs in summer crops (for which varietal adoption data were not available), since one of the major impacts of MVs of cotton, Basmati rice, and to a lesser extent, mungbean, was to shorten the growing period, and thus to reduce the conflict between harvesting of the summer crop and planting of the following wheat crop. The effect of resource degradation was estimated through measures of soil and water quality (soil phosphorus, soil organic matter, total soil soluble salts, and water electroconductivity). Region-specific time-trend dummies were included to capture the remaining unspecified effects of technological change, resource degradation, or change in resource productivity not included in the function.

Productivity trends in the crop sector in Pakistan were econometrically decomposed into the effects of technological change, improvements in human resources and infrastructure, and natural resource degradation by multiplying the negative of the coefficient in the cost function with the system-level rate of per annum change (in percentage) in each variable included in the cost function.

Technological change and improvements in human and physical infrastructure combined, produced a growth of 0.94% per annum with each accounting for about half the total. Resource degradation in aggregate lowered growth by 0.53% per annum (see Table 5).ⁱⁱ The combined effect of technological change, resource degradation, and improvement in human and physical infrastructure was negative in the wheat-rice system (overall increase in unit cost), while the contribution of technological change was highest in the wheat-cotton and wheat-mungbean systems. These results confirm the pattern found in the TFP estimates.

Soil and water degradation reduced productivity in all regions, highlighting the role of natural resource variables on productivity. In the wheat-rice system, resource degradation more than cancelled the productivity-enhancing contributions of technological change, education, and infrastructure. The unspecified “other factors” captured by coefficients on the regional time-trend variable also reduced productivity quite strongly in all but the wheat-cotton system. Among other omitted variables, these factors include resource degradation factors that were not measured, such as the development of pest complexes due to inappropriate use of pesticides and monocropping of cereals. More research is needed to identify the management practices causing such a decline. As massive public investment to control water logging and salinity is not included in the cost function which relates to private costs and returns only, the effect of these other factors is probably under-estimated.
Policy Discussion

Motivated by concerns of food security following the food aid crisis of the mid-60s, governments in both India and Pakistan concentrated resources in irrigated areas like the Punjabs that held promise for the greatest increases in crop yields (Sims, 1993). The subsequent gap in agricultural performance and sources of growth between the two states, despite quite similar agro-ecological potential, seems to relate to differences in non-price policies towards agriculture. Price incentives, a major policy instrument in both countries, were fairly similar.

At the core of the price policies was a strategy of massive subsidization of fertilizer, credit, power, and irrigation inputs. In India, Gulati and Sharma (1995) estimate that subsidies on the four major inputs grew at 9% to 12% per annum in real terms between 1981 and 1993, accounting for between 2.2% to 2.7% of GDP by the end of that period. Subsidies were an important element of public spending in Pakistan as well, but with a stronger bias toward large farmers (Sims, 1986; Sims, 1993). Inputs subsidies were maintained well beyond an initial period when they might be economically justified to overcome farmers’ risk aversion and learning by doing. Once established they proved to be politically difficult to remove. In the Indian Punjab, where small and medium farmers dominate the electorate, subsidies became even more entrenched.

At the same time, however, output prices on basic food grains were taxed to maintain low food prices through controls on food marketing that maintained grain prices below world prices, combined with overvalued exchange rates and tariff protection of the non-agricultural sector. Despite the high input subsidies, there was a large net transfer of resources out of agriculture. Effective protection rates for food grains over much of this period averaged between -33% and -50% in both states (Faruqee, 1995; Gulati and Kelley, 1999).

Empirical evidence on agricultural supply response and input elasticities in the Asian context suggests that it is differences in non-price incentives that are the more likely determinants of differential performance between the two states. Investments in rural infrastructure, human capital, and research and extension have been shown to play a dominant role in influencing supply and productivity growth (Binswanger et. al., 1993; Fan, et. al., 1999; Rao, 1989).

Based on comparisons of public expenditures in agriculture in the two states and the importance of education and infrastructure in growth evident from the cost function analysis, India would be expected to have more rapid input use and productivity growth. Pakistan’s share of public resources allocated to agriculture has been low relative to India (Choudhry and Faruqee, 1995). This difference in public expenditure is reflected in the disparities in public investment and quality of rural infrastructure, human capital, research and extension. Massive public investment in rural infrastructure in India meant that by the mid-1980s all villages in rural Punjab were electrified, the density of the road network was well above the Indian average and that of the Pakistan Punjab, and over 90% of cropped area was irrigated (Fan et. al., 1999). In Pakistan, investment in education and rural infrastructure has been comparatively much lower (Mujahid-Mukhtar, 1991; Rosegrant and Evenson, 1992; Faruqee, 1995). India also has a relatively better developed network of agricultural research centers and universities, especially in the northwest (Nagy and Quddus, 1998; Mruthyunjaya and Ranjitha, 1998).

Given the relative importance of non-price incentives for growth, the trend of expanding subsidies at the expense of productivity-enhancing investments is a cause of concern for the long-term sustainability of growth (Fan et. al., 1999). In Pakistan, the total share of public resources allocated to agriculture has been declining, and irrigation related expenditures have been particularly impaired with a 4% annual rate of decrease (Ahmad and Kutcher, 1992). Continuous under-investment in operational and management costs has seriously reduced the efficiency of the irrigation system (Chaudhry and Ali, 1989). Regular breaches, excessive seepage, and limited water supplies for the tail reaches of distributaries are some of the most common problems. Research spending also fell in real terms in the 1990s and accounts for a falling share of the agricultural budget (Choudhry and Faruqee, 1995). Across the border, even though the share of public resources devoted to agriculture has steadily risen since the early-1980s in India, agricultural subsidies have increased three times faster than other expenditures (World Bank, 1996).

Apart from crowding-out productivity enhancing expenditures, input subsidies have also been a major cause of over-capitalization, inefficient use of inputs, and a shift in cropping patterns towards water and fertilizer-intensive crops, thus contributing to soil degradation, salinity problems, and over-exploitation of groundwater in India (Joshi and Tyagi, 1994; Vaidyanathan, 2000). In Pakistan, Ahmad and Kutcher (1992) found that the subsidy on canal water prices has led to inefficient use of water, and contributed to the waterlogging and salinity problem. In addition, the flat rate structure of water and electricity prices, along with a subsidy on tubewell drilling without regulation on the number of tubewells together distorted the efficient use of water (Johnson, 1989). Resource degradation in itself is not a reason for policy intervention if it is internalized in producer decision making. In this case, distorted policies have led private and social costs to diverge.

Removal of price distortions in the form of input subsidies will be a major step toward encouraging more sustainable systems. Removal of subsidies would also free resources for high priority public investments in rural infrastructure, education and research and extension, that would encourage both higher productivity growth and more sustainable systems. Since the early to mid-1990s, both Pakistan and India have initiated steps to reduce price distortions in the agricultural sector but this has not as yet been accompanied by an increase in public investments in agriculture. ⁱⁱⁱ

Arresting resource degradation will require a concerted effort on several other fronts as well. First, the shortage of public investment funds has led to considerable under-investment in establishing new drainage systems that are central to resolving the waterlogging and salinity problems. It may be argued that irrigation is a private good and should gradually be left to private markets but drainage is a public good and will remain so, at least, beyond the farm-field where most of the drainage investments are needed (World Bank, 1994). On the irrigation front, new institutional structures are now being piloted in Pakistan, including the devolution of water management to farmers’ organizations, and the establishment of public utilities to operate and price water further up the system (Bandaragoda and Firdousi, 1992).

Second, research systems have been oriented toward developing technologies based on packages of modern inputs with little emphasis on practices that can arrest resource degradation. In irrigated agriculture, the move to more input efficient and environmentally- friendly practices will require considerable location-specific research in areas such as integrated pest and nutrient management, diversifying rotations to include legumes, and use of practices such as conservation tillage. At the same time, many such practices are information intensive and will require much greater efforts at information dissemination and extension.

Finally, part of the difficulty with addressing soil and water management problems in irrigated agriculture in India and Pakistan is the large number of institutions with overlapping mandates, and the general lack of coordination among them. For example, in the Pakistan Punjab alone, there are nearly a dozen institutions working on salinity. Although there is a lot of information and awareness of the land and water problems, much of the information and policy making is institutionally dispersed (John Mellor Associates, 1994). An important issue is therefore the establishment of a central agency to provide regular and up to date information on the state of land and water resources in irrigated areas.

This paper addresses the critical issue of long-term productivity and sustainability of irrigated agriculture in the Indian and Pakistan Punjabs. The results of this study confirm those of other observers who have noted much higher and more rapid growth of yields in the Indian Punjab compared to the Pakistan Punjab, at least for food crops. Only in cotton, Pakistan’s major export, has the Pakistan Punjab performed significantly better. However, the results suggest that most of the more rapid agricultural growth in the Indian Punjab can be attributed to more rapid input growth. Overall productivity growth in the Indian Punjab was higher but not by a large margin.

The gap in input use and agricultural performance between the two states, despite quite similar agro-ecological potential, seems to relate to differences in non-price policies that encouraged much faster growth in input use in India than Pakistan. While investment, both public and private, plays a central role in productivity growth, timing also matters. In both Punjabs, there was a considerable lag between investment in infrastructure and Green Revolution inputs, and the realization of productivity growth. In part this seems to relate to learning by doing and investment in human capital which have led to improvements in technical efficiency, and in part, to the better utilization of lumpy capital investments over time, especially tubewell capacity. For policy-makers, it suggests that a long-term commitment is needed to realize complementarities between investment in technologies and supporting infrastructure.

The results in this paper raise serious concerns about long term sustainability of these intensive irrigated systems due to resource degradation. This study, at least for Pakistan, provides the first quantitative evidence of the impact of resource degradation. Degradation is estimated to reduce productivity growth by one third overall, and in the case of wheat-rice to practically cancel the effect of technical change. These results, combined with the stagnation of cereal output in recent years underline the urgency of measures to arrest the problem and maintain the Punjab’s most valuable asset—its irrigated land base.

Ahmad, M. and G.P Kutcher. 1992. "Irrigation Planning with Environmental Considerations: A Case Study of Pakistan’s Indus Basin." Technical Paper No. 166, World Bank, Washington, D.C. Processed.

Ali, M. 1996. “Quantifying the Socioeconomic Determinants of Sustainable Crop Production: An Application to Wheat Cultivation in the Tarai of Nepal.” Agricultural Economics 14: 45-60.

Ali, M. and D. Byerlee. 1991. “Economic Efficiency of Small Farmers in a Changing World: A Survey of Recent Evidence.” Journal of International Development 3(1): 1-27.

Ali, M. and D. Byerlee. 2000. “Productivity Growth and Resource Degradation in Pakistan’s Punjab: A Decomposition Analysis.” Policy Economics Working Paper, World Bank (forthcoming).

Ali, M. and L.E. Velasco. 1994. "Intensification-Induced Resource Degradation: The Crop Production Sector in Pakistan.” Working Paper. International Rice Research Institute, Manila. Processed.

Alston, J.M., G.W. Norton, and P.G. Pardey. 1995. Science Under Scarcity: Principles and practice for Agricultural Research Evaluation and Priority Setting. Ithaca and London: Cornell University Press.

Antle, J. M. and McGuckin, T. 1993. "Technological innovation, agricultural productivity, and environmental quality." In: G.A. Carlson, D. Zilberman, and J.A. Miranowski, eds., Agricultural and Environmental Resource Economics. New York, NY: Oxford University Press.

Azam, Q.T., Bloom E.A. and R.E. Evenson. 1991. “Agricultural Research Productivity in Pakistan.” Islamabad: Pakistan Agricultural Research Council and New Haven: Yale University.

Bandaragoda, D.J. and G.R. Firdousi. 1992. "Institutional Factors Affecting Irrigation Performance in Pakistan: Research and Policy Priorities." International Irrigation Management Institute, Colombo, Sri Lanka.

Berndt, E.R. and M.A. Fuss. 1986. “Productivity Measurement with Adjustments for Variation in Capacity Utilization and Other Forms of Temporary Equilibrium.” Journal of Econometrics 33: 7-29.

Binswanger, H.P., S.R. Khandker, and M.R. Rosenzweig. 1993. “How Infrastructure and Financial Institutions affect Agricultural Output and Investment in India.” Journal of Development Economics 31: 337-366.

Byerlee, D. 1992. “Technical Change, Productivity, and Sustainability in Irrigated Cropping Systems of South Asia: Emerging Issues in the Post-Green Revolution Era.” Journal of International Development 4: 477-496.

Byerlee, D. and A. Siddiq. 1994. “Has the Green Revolution been Sustained? The Quantitative Impacts of the Seed-Fertilizer Technology in Pakistan Revisited.” World Development 22(9): 1345-1361.

Byerlee, D. and R. Murgai. 2000. “Sense and Sustainability Revisited: The Limits of Total Factor Productivity Measures of System Sustainability.” Agricultural Economics (forthcoming).

Cassman, K.G. and P.L. Pingali. 1993. “Extrapolating Trends from Long-Term Experiments to Farmers Fields: The Case of Irrigated Rice Systems in Asia.” In: V. Barnett, R. Payne, and R. Steiner (eds.) Agricultural Sustainability: Economic, Environmental and Statistical Considerations. New York: John Wiley and Sons Ltd.

Chanmugam, N. 1994. "Factors Influencing the Gap in Crop Productivity between the Indian and Pakistani Punjabs." M.Sc. Thesis, Cornell University.

Chaudhry, A. M. and M. Ali, 1989. Measuring Benefits to Operation and Maintenance Expenditure in the Canal Irrigation System of Pakistan: A Simulation Analysis, Agricultural Economics 3, 1989: 199-212.

Choudhry, Y. and R. Faruqee, 1995. “Pakistan: Public Expenditure in Agriculture”, Report No IDP-153, World Bank, Washington, DC.

Diewert, W.E. 1976. “Exact and Superlative Index Numbers.” Journal of Econometrics 4: 115-146.

Evenson, R.E., C.E. Pray, and M.W. Rosegrant. 1999. "Agricultural Research and Productivity Growth in India." Research Report 109, International Food Policy Research Institute.

Fan, S., P. Hazell, and S. Thorat. 1999. "Government Spending, Growth and Poverty: An Analysis of Interlinkages in Rural India." Research Report No. 110. International Food Policy Research Institute, Washington DC.

Faruqee, R., 1995. "Structural and Policy Reforms for Agricultural Growth: The Case of Pakistan." Agricultural and Natural Resources Division, South Asia Dept, World Bank, Washington.

Foster, A.D. and M.R. Rosenzweig. 1996. "Technical Change and Human-Capital Returns and Investments: Evidence from the Green Revolution." American Economic Review 86(4): 931-953.

Fujisaka, S., L. Harrington, and P. Hobbs. 1994. “Rice-Wheat in South Asia: Systems and Long-Term Priorities Established Through Diagnostic Research” Agricultural Systems 46: 169-187.

Gulati, A. and A. Sharma. 1995. “Subsidy Syndrome in Indian Agriculture.” Economic and Political Weekly 30(39): A93-A102.

Gulati, A., and T. Kelley, 1999. Trade Liberalization and Indian Agriculture. Oxford University Press, New Delhi.

Herdt, R. W., and J.K. Lynam. 1992. “Sustainable Development and the Changing Needs of International Agricultural Research.” in Lee, D. R., S. Kearl, N. Uphoff (eds), Assessing the Importance of International Agricultural Research for Sustainable Development, Ithaca: Cornell University, 1992.

John Mellor Associates and Asianics Agro-Development International. 1994. Institutional Reforms to Accelerate Irrigated Agriculture. Government of Pakistan and World Bank.

Johnson, R. 1989. "Private Tubewell Development in Pakistan’s Punjab: Review of Past Public Programs: Policies and Relevant Research." International Irrigation Management Institute, Colombo, Sri Lanka.

Joshi, P.K. and N.K. Tyagi. 1994. “Salt-Affected and Waterlogged Soils in India: A Review.” Joint Study of the Indian Council of Agricultural Research and the International Food Policy Research Institute.

Kaur, P. 1991. Returns to Investment on Agricultural Research and Extension. Ph.D. thesis, Punjab Agricultural University, Ludhiana.

Khan, M.H. 1994. “The Structural Adjustment Process and Agricultural Change in Pakistan in the 1980s and 1990s.” Draft paper, Simon Fraser University, Vancouver.

Kumar, P. and M. Rosegrant. 1994. “Productivity and Sources of Growth for Rice in India.” Economic and Political Weekly 29(53):183-188.

Kumar, P. and Mruthyunjaya. 1992. “Measurement Analysis of Total Factor Productivity Growth in Wheat.” Indian Journal of Agricultural Economics 47(3): 451-458.

Kurosaki, T. 1999. "Agriculture in India and Pakistan, 1900-95: Productivity and Crop Mix," Economic and Political Weekly December 1999: A160-A168

Mruthyunjaya and P. Ranjitha. 1998. “The Indian Agricultural Research System: Structure, Current Policy Issues and Future Orientation.” World Development 26: 1089-1102.

Mujahid-Mukhtar, E. 1991. “The History of Agricultural Development and the Role of Policy: A Comparison of the Two Punjabs (1849-1981).” The Indian Economic Journal 38: 39-61.

Murgai, R. 1997. “Diversity in Economic Growth and Technical Change: A Districtwise Disaggregation of the Punjab and Haryana Growth Experience.” Ph.D. Dissertation, University of California, Berkeley.

Murgai, R. 2000. “The Green Revolution and the Productivity Paradox: Evidence from the Indian Punjab.” Paper presented at the 24^th International Conference of Agricultural Economists, Berlin, 13-18^th August.

Nagy, J.G. and M.A. Quddus. 1998. "National agricultural commodity research priorities for Pakistan." Agricultural Economics 19: 327-340.

Pingali, P.L. and P.W. Heisey. 1999. “Cereal Crop Productivity in Developing Countries: Past Trends and Future Prospects.” CIMMYT Economics Paper 99-03, Mexico D.F.: CIMMYT.

Pingali, P.L. and M.W. Rosegrant. 1993. “Confronting the Environmental Consequences of the Green Revolution in Asia.” In Proceedings of the 1993 American Agricultural Economics Association International Pre-Conferences on “Post-Green Revolution Agricultural Development Strategies in the Third World: What Next?.

Rao, M.R. 1989. "Agricultural Supply Response: A Survey." Agricultural Economics 3:1-22.

Ray, S.C. 1982. "A Translog Cost Function Analysis of U.S. Agriculture." American Journal of Agricultural Economics 64:490-498.

Rosegrant, M. W. and R. E. Evenson. 1992. “Agricultural Productivity and Sources of Growth in South Asia,” American Journal of Agricultural Economics 74(3): 757-761.

Siddiq, A. 1994. “Sustainability of the Indus Basin: Impact of Tertiary Salinity/Sodicity on Wheat Productivity, Damage Assessment and Future Public Policy.” Ph.D. Dissertation, Dept. of Agricultural Economics, University of Illinois, Urbana-Champaign.

Sidhu, D.S. and D. Byerlee. 1992. “Technical Change and Wheat Productivity in the Indian Punjab in the Post-Green Revolution Period.” CIMMYT Economics Working Paper 92-02. Mexico, D.F.

Sims, H. 1986. “The State and Agricultural Productivity: Continuity versus Change in the Indian and Pakistan Punjabs.” Asian Survey, 26(4): 483-500.

Sims, H. 1993. “Economic Imperatives, Political Risks and Modes of Action: Agricultural Policy Implementation in India and Pakistan.” Journal of Developing Areas 27: 145-162.

Squires, D. E. 1992. “Productivity Measurement in Common Property Resource Industries: An Application to the Specific Coast Trawl Fisheries.” RAND Journal of Economics 23(2): 221-236.

Vaidyanathan, A. 2000. “India’s Agricultural Development Policy.” Economic and Political Weekly 35(20): 1735-1741.

World Bank. 1994. "Pakistan Irrigation and Drainage: Issues and Options." Report No. 11884-PAK, Washington D.C.

World Bank. 1996. “India Country Economic Memorandum – Five Years of Stabilization and Reform: The Challenges Ahead.” Report No. 15882-IN, Washington D.C.

Rinku Murgai is an economist in the Development Economics Research Group of the World Bank, Mubarik Ali is an agricultural economist at the Asian Vegetable Research and Development Center, and Derek Byerlee is a lead economist in the Rural Development Department of the World Bank.