India - Effective incentives in India's agriculture : cotton, groundnuts, wheat and rice

Policy, Planning, and Research WORKING PAPERS Trade Policy New Delhi Resident Mission and Country Economics Department The World Bank January 1990 WPS 332 Effective Incentives in India's Agriculture Cotton, Groundnuts, Wheat, and Rice Ashok Gulati with James Hanson and Garry Pursell A policy that moves prices closer to free trade levels would shift resources from groundnuts (or oilseeds) into cotton, rice, and wheat - crops for which India has more of a comparative advantage and would earn more in foreign exchange. This AMLL would also allow agriculture to compete with industry f investment rupees. ', > The Policy. Planning, and Research Comnplex distributes PPR Working Papers to disseminate the findings of work in progress and to enourage the exchange of ideas among Bank staff and all others interested in development issues. These papers carry the names of the authors, reflect only their views, and should be used and cited accordingly. The findings, interpretations, and conclusions are the authors' own.They should not be attributed to the World Bank, its Board of Directors, its management, orany of its membLrcourtries.  Plc,Planning, and Research Trade Policy This paper -a product of the Country Operations Division, New Delhi Resident Mission, and the Trade Policy Division, Country Economics Department- is part of a larger effort in PPR to quantify incentives in agriculture and to analyze the effects of the trade policies of developing countries on the allocation of resources bctween different crops and between agriculture and industry. Copies are available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Ghislaine Bayard, room N10- 021, extension 38004 (138 pages with tables). Detailed estimates of effective subsidy coeffi- Investment programs aimed at increasing cients for four crops - cotton, wheat, rice, and cotton, rice, and wheat production appear to groundnuts - yielded the following conclusions have high economic rates of return - higher at about agricultural incentives in India (among the margin than investment in industry. others): Andhra's cotton is a financially a'id economi- cally profitable export even at current exchange Wheat, rice, and especially cotton have been rates. Investment in rice and wheat would yield disprotected (in effect, taxed) in the 1980s. high economic rates of return as import substi- Groundnuts have been heavily protected, en- tutes, particularly in -areas where subsidies are couraging a different allocation of resources low. Their financial and economic profitability than under free trade. as exports is, however, more doubtful. The evi- dence on these major crops suggests that in India The incentive framework generates static (as in other developing countries) agriculture as efficiency losses and net foreign exchange a whole is underpriced, which produces a bias losses, particularly in areas where the crops toward the industrial sector. compete directly for resources. A policy that moved prices closer to free trade levels would Subsidies in Indian agriculture are substan- shift resources from groundnuts into crops that tial - about 10 percent of value added for would eam more in foreign exchange. This groundnuts and cotton, 25 percent for wheat, and increased output, particularly of cotton, could be 35 percent for rice - and should be considered used to purchase edible oils - with a net gain in explicitly in evaluating incentives for agricul- foreign exchange. tural investment and production. The PPR Working Paper Series disseminates the findings of work under way in the Bank's Policy, Planning, and Research | Complex. An objective of the series is to get these findings out quickly, even if presentations are Icss th-n fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent official policy of the Bank. Produced at the PPR Disscmination Center  Table of Contents Page No. Executive Summary i Chapter 1 An Overview of the Incentive Structure in Indian Agriculture 1 A. Objectives and Methodology 1 B. Estimation of Protection Coefficients and Results 5 C. Implications and Concluding Remarks 11 Chapter 2 India: Effective Incentives for Agriculture: The Case of Wheat and Rice 18 Introduction 18 A. Nominal Protection Coefficients (NPC) of output 20 B. Effective Protection Coefficients of Output (EPCs) 25 C. Effective Subsidy Coefficients of Output 28 D. Concluding Observations and Future Outlook 36 Annexures 41 Chapter 3 India: Effective Incentives for Agriculture: The Case of Cotton Introduction 56 A. A Brief Introduction to the Indian Cotton Economy 57 B. Nominal Protection Coefficients of Seed-Cotton (Kapas) 61 C. Effective Protection Coefficients 63 D. Effective Subsidy Coefficients 65  E. Concluding Remarks and Future Outlook 67 Annex 71 Annexures 75 Chapter 4 India: Effective Incentives for Agriculture: The Case of Groundnut 104 Introduction 104 A. The Groundnut Economy: A Brief Sketch 105 B. Nominal Protection Coefficients of Groundnut 110 C. Effective Protection Coefficients of Groundnut 112 D. Effective Subsidy Coefficients of Groundnut 112 E. Concluding Remarks and Future Outlook 114 Annexures 119 References 134  List of Tables Page No. Chapter 1 1.1 Crop and Region-Specific Dispersion of Effective Incentives in Indian Agriculture (Average 1980-81 to 1986-87) 7 1.2 Temporal Behaviour of Crop-Specific Effective Incentives and Prices in Indian Agriculture. 9 Chapter 2 2.1 Nominal Protection Coefficients of Wheat 23 2.2 Nominal Protection Coefficients of Rice 24 2.3 Effective Protection Coefficients of Wheat 28 2.4 Effective Protection Coefficients of Rice 29 2.5 Subsidies on Non-Tradeable Inputs of Wheat 31 2.6 Subsidies on Non-Tradeable Inputs of Rice 32 2.7 Effective Subsidy Coefficients of Wheat 34 2.8 Effective Subsidy Coefficients of Rice 35 2.9 Regional Dispersion of Effective Incentives for Wheat & Rice Cultivators in India (Average 1980-81 to 1986-87) 37 2.10 Temporal Behaviour of Protection Coefficients of Wheat and Rice (Weighted Average) 38 Annexures 41 Chapter 3 3.1 Nominal Protection Coefficients of Seed-Cotton 63 3.2 Effective Protection of Coefficients of Seed-Cotton 64 3.3 Subsidies on Non-tradeable Inputs of Seed-Cotton 66 3.4 Effective Subsidy Coefficients of Seed-Cotton 67  3.5 Regional Dispersion of Effective Incentives for Cotton Cultivators (Average 1980-81 to 1986-87) 69 3.6 Temporal Behaviour of Protection Coefficients 70 Annexures 71 Chapter 4 4.1 Nominal Protection Coefficients of Groundnut 111 4.2 Effective Protection Coefficients of Groundnut 113 4.3 Subsidies on Non-tradeable Inputs of Groundnut 115 4.4 Effective Subsidy Coefficients of Groundnut 116 4.5 Temporal Behaviour of Protection Coefficients (weigh.ted Average) and International Process 117 4.6 Regional Dispersion of Effective Incentives for Groundnut Cultivators 118 Annexures 119  Executive Summary 1. Government intervention in Indian agriculture is pervasive. It includes support/procurement price policies, explicit and implicit subsidies on inputs that now represent a significant fraction (over 2.5%) of GDP, control over the links between external and domestic markets, research and extension programs, and crop specific programs. Assessments of these wide ranging and sometimes contradictory policies have usually been based on analyses of their impact on the net income per hectare from the crop in question. Sometimes consideration also is given to the opportunity cost of policy-induced substitutions between one crop and another. However, almost all of these analyses rely on domestic prices for their evaluations. This implicitly autarkic approach neglects the possiblities of gains from trade and additional foreign exchange earnings. 2. This study and other recent World Bank analyses of agriculture (Krueger, Schiff, and Valdes) take a different tack -- explicitly considering the possiblities of international trade in evaluating the impact of incentives. Under this approach, incentives for different crops are measured by the ratios between domestic and international prices (appropriately adjusted for input costs, subsidies and the s^arcity value of foreign exchange). ComrAodities with ratios in excess of one are protected by the policy regime, commodities with ratios below one are disprotected (in effect taxed) by the policy regime, relative to the situation that would prevail under free trade. 3. Specifically, this study estimates three standard coefficients (ratios): a) the nominal protection coefficient (NPC) -- the ratio of domestic to international border prices (the reference prices) b) the effective protection coefficient (EPC) -- the ratio of value added measured at domestic prices (the domestic price less the per unit domestic cost of tradeable inputs) to value added at international border prices (the invernational price less the per unit cost of tradeable inputs at international prices) c) the Effective Subsidy Coefficient (ESC) -- the effective protection coefficient adjusted for the subsidies and taxes on non- tradeables The EPC measures the protection and subsidies accorded to the land, labor, and capital and non-tradeable inputs used i- che production of a commodity, while the NPC measures the protection accorded to the commodity, including all the inputs that go into its production. The importance of this distinction can be seen by noting that a high NPC provides little incentive to use resources in the production of a commodity, if the NPC of the inputs is also high. Hence, the EPC is a better measure than the NPC of the incentive for resources to be used to produce a commodity. Similarly, if non-traded inputs such as credit and electricity are subsidized, then there is greater incentive to apply inputs of land, labor and capital than without subsidy. Consequently, the ESC is a better a.aasure of incentives than the EPC.  ii 4. The study also includes estimates of the nominal protection coefficient adjusted for a premium on foreign exchange (ANPC) and some discussion of how such a premium would affect the EPCs and ESCs. Inclusion of a notional premium on foreign exchange is used to reflect the necessity of a compensatav change in the exchange rate were the economy to move to a free ;rade ei.!ronment,as well as the possiblity that the current account deficit is not sustainable in a given year and hence the currency is overvaled even given the trade and exchange control regime (See for example Krueger, Skiff, and Valdes). With the inclusion of such a premium, the estimated coefficients become good proxies for cost benefit indicators such as Domestic Resource Cost (DRC) ( See Pursell and Roger and Scandizzo and Bruce). As this estimate is only indicative, the adjustment is formally made only on the NPC. 5. This study covers four commodities: wheat and rice, India's iost important cereals; cotton, an important cash crop used in the country's large textile industry; and groundnut as a representative of the oilseeds crop complex. Together, these four crops account for roughly 45 per cent of gross cropped area and above 50 per cent of gross value of crop output. Wheat, rice and cotton experienced technological breakthroughs during late 1960s and 1970s in certain regions, while groundnut yields have lagged behind. Presently, efforts are being made to stimulate oilseed production under the Technology Mission on Oilseeds launched by Government of India in 1986 and a buffer stock/price support scheme was introduced in 1989. 6. Table 1 summarizes the estimates for the protecion coefficients for the four commodities for the 1980s, in addition it contains estimates of comparable domestic and world prices. These coefficients are weighted averages of coefficients calculated for the principal producing states; the state coefficients differing in prices, transport costs to market (which affects the 'border' reference price) and degree of subsidies. The coefficients have been calculated under two assumptions: that che crops compete with imports and that they are exported to a particular market. The difference is basically that the international border or reference price is computed by adding on international shipping costs to the price at the foreign market under the importable hypothesis and by deducting international shipment costs from the price at the point of foreign sale in the case under the exportable hypothesis. These two different treatments of transport costs allow the analyst to see whether the Indian production is competitive with imports given the protection accorded by transport costs and what subsidy, if any, would be required to export the crop. 7. The major results can be summarized as follows: a) Cotton, and to a lesser extent wheat and rice have been disprotected or in effect taxed during the 1980s, relative to the prices that would have prevailed with free imports (the importable hypothesis). This is shown by protection coefficients that generally are less than one. b) Groundnuts have received substantial protection -- their protection coefficients exceed one and on average, are 1.8 to 2.6 times those for the other crops. In fact the average EPC, 1.56, is above the average effective protection coefficient for industry of 1.4 (World Bank 1987).  tii Sabic le 1: SooraI Behaviour of Crom-SoeofA Eet rngentives nd Prices in indian Arioulture Hypothesis /Crop/ProtectLori Coefficient 1980-81 1981-82 1982-83 1983-84 1984-85 1985-86 1986-87 Averase --------------------------------------------------------_----__--------------__----------------------- nmsortcblL ieothes is WIEAT NPCs 0.72 0.73 0.84 0.84 0.76 0.76 0.91 0.80 EPC 0.67 0.68 0.80 0.80 0.72 0.72 0.89 0.75 ESCs 0.76 0.85 0.99 0.99 0.90 0.90 1.12 0.93 ANPCs 0.57 0.58 0.67 0.67 0.60 0.60 0.73 0.63 Dometic Prico/a (RaIq) 117 130 142 151 152 1!7 162 World Pria-lb (Rslq) 127 148 152 159 170 173 145. Price Ratio 0.92 0.88 0.93 0.95 0.89 0.91 1.12 0.94 RICE NPCs 0.45 0.51 0.75 0.70 0.72 0.78 0.80 0.67 LPC-a 0.43 0.48 0.73 0.68 0.70 0.78 0.78 0.65 ESCs 0.58 0.65 0.99 0.90 0.96 1.01 1.06 0.d8 ANPCS 0.36 0.40 0.60 0.56 0.57 0.63 0.64 0.54 Domestic PeiceIc (RaIq) 130 164 174 189 196 203 2C8 World Price/d (Rs/q) 353 353 255 291 293 274 274 Price Ratio 0.42 0.46 0.68 0.65 0.67 0.74 0.76 0.63 COTTON NPCs 0.77 0.94 0.82 0.68 0.73 0.86 0.83 0.80 EPCs 0.64 0.82 0.71 0.58 0.62 0.74 0.69 0.69 E8Cm 0.71 0.90 0.79 0.64 0.67 0.82 0.75 0.75 ANPCa 0.60 0.72 0.64 0.53 0.57 0.66 0.64 0.62 Domastic Pricele (Rs/q) 895 1118 1118 1176 1206 1250 1264 WorLd Pricalf (RsIq) 1629 1415 1594 1949 1-'7 1333 1734 Price Ratio 0.55 0.79 0.70 0.60 0.70 0.94 0.73 0.72 GROUtNDNUS NPCs 1.06 1.37 1.66 1.41 1.40 1.53 2.05 1.50 EPCs 1.09 1.44 1.74 1.47 1.47 1.58 2.13 1.56 ESCS 1.20 1.55 1.93 1.58 1.60 1.76 2.32 1.71 ANPCs 0.84 1.08 1.31 1.11 1.11 1.21 1.63 1.18 Domstic Prlels (RaIq) 294 386 421 450 486 500 528 World Price/h (RsIq) 382 S39 389 396 497 445 409 Price Ratio 0.77 0.72 1.0W 1.14 0.98 1.12 1.29 1.01  ivr (contd.) TabLe 1.2: Tpmnorsl Behaviour of Croy-Soacifto Effectlve rncentives and Prices In Indian AtgMcultury Hypothesis/Crop/Protection CoeffLeient 1980-81 1981-82 1982-83 1983-84 1984-85 1985-86 1986-87 Averasge EhoortabLe Hypothesis WHEAT NPCs 1.15 1.09 1.24 1.29 1.27 1.33 1.99 1.34 EPCs 1.29 1.16 1.35 1.47 13Ss 1.71 3.44 1.71 ESCs 1.57 1.35 1.57 1.73 1.81 2.00 4.07 2.01 ANPCs 0.87 0.83 0.93 0.96 0.93 0.97 1.38 0.98 RICE NPCs 0.50 0.60 0.94 0.89 0.94 1.09 1.16 0.87 EPC& 0.47 0.57 0.93 0.88 0.95 1.13 1.19 0.87 ESCs 0.65 0.77 1.26 1.21 1.30 1.60 1.68 1.21 ANPCs 0.39 0.46 0.70 0.68 0.72 0.83 0.87 0.66 cOTTON NMPC 0.89 1.13 0.92 0.74 0.83 1.01 0.93 0.92 EPCs 0.78 1.04 0.80 0.62 0.71 0.89 0.79 0.80 ESCs 0.84 1.13 0.88 0.69 0.78 0.99 0.86 0.88 ANPCs 0.70 0.87 0.72 0.58 0.64 0.76 0.72 0.71 GROUNDNUT NPCs 1.25 1.72 2.03 1.69 1.76 2.03 2.87 1.91 EPCs 1.44 2.09 2.44 1.90 2.07 2.36 3.45 2.26 ESCs 1.57 2.24 2.69 2.04 2.31 2.63 3.76 2.47 ANPCe 0.99 1.34 1.57 1.32 1.37 1.56 2.20 1.48 Notes: la Domestic price of wheat is approximated by Its procurement price for FAQ. lb World price ls of US Hard Red WLnter No. 2 with ordlnary protein, fob US gulf (at official exchanse rate) for ApriL to June quarter. /c Domestic prlce of rice is estLmted as procurement price of paddy dlvided by 0.7, whlch Is paddy-rLce conversion factor for Indian 'Common' rice. id World -rice of rice Is of Thai vhite (Miled) 51 Broken, fob Bangkok (at official rate) for October to January. /a Domestic price of cotton Is approximated by procurement price of kapa. (J-34/ 414F/H-777 variety) divided by 0.34. whlch Is kapas-llnt conversion ratlo. If World Price of cotton is that of cotton outlook index 'A'. cif Liverpool. Ig Domestic prlce of groundnut is its procurement price (Ln cermw of earnels). /h World price of groundnut Is of KerneLs of sny origLa, cLf Europe (Rorterdam).  v c) EPCs for cotton,wheat and rice are below the corresponding NPCs (under the importable hypothesis). This reflects the disprotection of these crops (NPCs less than one) combined with the protection of the industrial inputs (pesticides, machinery, fertilizer) used in their production. On the other hand groundnuts have a higher EPC than NPC, because the protection of groundnuts exceeds the protection of its inputs. d) Estimated net subsidies are large, about 10% of value added in the case of groundnuts and cotton, about 25% in the case of wheat and about 35% in the case of rice. AssuAming these large subsidies are distributed evenly across all producers would imply that they offset some of the disprotection -- this is shown by the EPSs that are greater than th-e EPCs and quite close to one (e.g. wheat 93, rice .88, cotton .75 under importable hypothesis). However, 'If this assumption is not corret .hen, as discussed below, some producers of these crops are in effect bea.rh taxed (as shown by the EPCs) while others are receiving large benefits. e) Cotton appears to be an efficient export (protection coefficients less than one under te exportable hypothesis.) Rice and, to a much lesser extent wheat, would be efficient exports provided a 25% premium were accorded foreign exchange, *1s shown by the ANPCS. Groundnut, on the other hand, would need a foreigr. exchange premium of 40% or more to be aii efficient export. f) The degree of disprotection for cotton, wheat and rice has been reduced over the l980s. This largely reflects falling world prices, that have not been offset even by the depreciations of the rupee against the dollar, and not any major rise in domestic prices. In fact, domestic prices rose 5-6% p.a. over the period in question, about the same as the wholesale price index in general. Protection on groundnut also has risen over the period, but this results both from falling world prices and domestic policies that have increased domestic prices about 11% p,.a. during the period. g) The calculations of protection coefficients across states (not shown) indicate at least two significant spatial variations: 1) Punjab rice has a much higher NPC, EPC, and ESC than the other states, reflecting higher prices and, to a lesser extent, the irrigation subsidy; 2) Maharashtra cotton (and to a lesser extent Punjab cotton) received much higher protection and Andhra cotton much lower protection than the average for cotton, reflecting differences in procurement price policy during the period and, to a lesser extent, irrigation subsidies. h) Of the two main crops, rice is _omewhat more disprotected (taxed more heavily) than wheat, compared 1^ the prices that would prevail under free trade. Moreover, the .iverage .i-antive to rice is brought up by the high protection accorded Punjab rice. Hence, rice and wheat producers in the Punjab and Haryana, and wheat producers elsewhere, receive more incentives than rice producers in the rest of the country.  vi 8. These results have a number of implications: a) The disprotection accorded cotton, rice and wheat and the high protection accorded groundnuts suggests that the incentive framework for agriculture hi. stimulated the allocation of resources to groundnuts and discouraged resource use in the other three crors0, compared to the allocation of resources that would prevail under a free trade regime. b) The difLerence in effective incentives also suggests the possibility that the incentive framework has generated static efficiency losses &nd ret foreign excharRe lusses, particalarly in areas like Gujarat and Andhra Pradesh, where thu crops compete directly for resources. A price policy that moved relative prices closer to the free trade levels would shift resources from groundnuts into the other crops, with a net Increase in value product measured at international prices. This increased oucput, particularly if it were cotton, could be sold in international markets in order to buy edible oils, with a net gain in foreign exchange. On the other hand, higher prices for edible oils would tend to reduce the acreage and production of internationally competitive crops. Hence, programs stimulating groundnut production through higher prices would have to have accord a high implicit benefit for self-sufficiency in order to be justifiable. c) Investment programs aimed e- increasing cotton, rice, and wheat production appear to have high economic rates of return. All three of these crops have protection coefficients which, adjusted for an exchange rate premium, are below unity and, as mentioned above, such coefficients are good proxies for cost-benefit indicators such as domestic resource costs (DRCs). By comparison, for example, the average effective protection coefficent for industry is about 1.4. Hence, these estimates suggest that at the margin investment in the three crops would yield a higher economic rate of return than investment in industry. Andhra cotton in particular appears to be exportable even at current exchange tates. The results suggest that investment in rice and wheat would yield high economic rates of return up to the point where they become exports -- i.e. they are efficient import substitutes. This is particularly true in areas where the subsidies are currently low; in the Punjab and Haryana, where effective subsidies to rice are high, the economic returns to further investments in rice production would be lower than elsewhere (See e below). An exchange rate premium of about 25% would mean investments in rice for export markets would also yield a satisfactory economic rate of return, provided world prices rise somiewhat and provided the expansion takes place in areas where subsidies to rice are low. Also, the absolute size of rice exports might be limited by the narrowness of the international rice market. d) Investment programs aimed at increasing groundnut production are likely to yield lower rates of return than cotton, rice, and wheat. Protection coefficients exceed those in Indian industry, suggesting that fairly high protection has been needed to make the present level of resouce allocation to groundnut profitable. e) Subsidies in agriculture are fairly large and need to be considered explicitly in evaluating the incentives accorded to production of the various crops and the feasiblity of investment programs. The subsidy on canal irrigation -- basically the difference between the sum of annualized capital costs and operating costs less actual revenues -- accounts for  vii about 70% of total subsidies estimated here. (The fertilizer subsidy looms large in the government accounts, but much of it goes to the producers of fertilizers, to ensure that their production is profitable at current prices, rather than substantially lowering the domestic farmgate price of fertilizer below world prices. Hence the fertilizer subsidy is less important to agriculture than one would expect from looking solely at the budget. It also should be noted that the fertilizer subsidy is included in the EPC, as it is a subsidy on traded inputs, as well as in the ESC -- the other subsidies appear only in the ESC.) The small size of the other subsidies suggests that they could be reduced substantially without requiring much of an offsetting increase in procurement prices (See also below para 9 b ). On the other hand, the large subsidy on irrigated crops suggests that further expansion in already irrigated areas may not be as efficient as w3uld seem from simply looking at net incomes per hectare, or even prices, and shifting to less water intensive crops would be warranted on efficiency grounds. There also would appear to be room for increasing water charges, which would provide an incentive for this shift to less water intensive crops. f) The prevalence of protection coefficients below unity (under the importable hypothesis) supports Schultz's thesis that agriculture in developing countries is underpriced and that there is a bias toward the industrial sector, which in India has an average effective proteciton coefficient of about 1.4. The results are also similar to those reported by Krueger, Schiff and Valdes, particularly in cotton where the controls on exports have delinked the domestic from the international market and depressed the price below international levels. At the same time, imported food products (edible oils of which groundnut is a major component) receive substantial direct protection. 9. The study also raise:s a number of important issues that deserve further study: a) Why are certain agricultural nroducts competitive and others uncompetitive? On one level, the answer is that it is not products, but marginal production that is competitive or uncompetitive. The delinking of the domestic and international market has allowed domestic and international prices to diverge and tended to encourage (discourage) additional resource use in crops where prices have risen above (fallen below) international levels. If domestic and foreign markets were to be linked, then domestic and foreign prices would become equal, and producers of all crops would be competitive internationally. However, the current resource allocation pattern would change, with resources being drawn out of the currently protected sectors and into the currently disprotected sectors. On another level, one can ask what policiies and investments would lead to a more internationally competitive agricultural sector. As just noted, closer linking of prices to international levels is one way to make the sector more efficient in a static sense. In terms of growth and investment, the protection coefficients, adjusted for an exchange rate premium, are good proxies for cost benefit indicators, such as domestic resource costs. Hence they are indicators fo relative investment productivities (measured at international prices). It also should be noted that argument for high farmgate prices stimulating development of  viii high yield varieties (a variant of the infant industry argument for protection), is probably invalid in the Indian context, given the limited impact of market signals on India's research bureauacrac. Finally, more investigation is warranted on the role of transport, marketing and processing costs in limiting India's agricultural export potential -- for example, not only are transport, processing an marketing costs high for Indian groundnuts, the procedures contribute to the high moisture content of the kernels, which in turn increases the risk of aflatoxin that limits Indian groundnuts and groundnut cake exports to the EEC. b) The impact of subsidies: Estimated subsidies are equivalent to a large percentage of agricultural GDP (16-17% see Gulati 1988a) but their influence on agricultural prices may not be as large as generally thought. Subsidies on irrigation, electricity, and credit lower the costs only of those farmers who have access to canal water, power lines and bank loans. If farmers that do not have such access account for a large fraction of the supply, then their costs determine agricultural prices. In this case, the main impact of the subsidies will be on income distribution -- through the high profits they generate for the recipients - rather than on prices. This suggests that further investigation is necessary to determine the impact of subsidies on prices and incentives and on income distribution. c) The potential impact of increased Indian trade on world prices and incentives: Large increases in Indian agricultural exports or imports might affect international prices, owing to the narrowness of some agricultural markets. This is not an argumert for ignoring the potential exports, even if the international price falls somewhat the foreign exchange income earned from additoional exports would certainly be positive. However, additional analysis is necessary to see what markets, if any, might be subject to this limitation and how this might affect the protection coefficients and what are the appropriate Indian policies.  CHAPTER 1 AN OVERVIEW OF THE INCENTIVE STRUCTURE IN INDIAN AGRICULTURE A. Obiectives and Methodology 1.01 Government intervention in Indian agriculture is pervasive. The Government tinkers with the market mechanism through its price support/ procurement policy on the one hand, and its subsidization of the major agricultural inputs on the other. The Government's price policy played a role in speeding the adoption of high yielding varieties (HYVs) of wheat and rice in the so-called Green Revolution and it has helped to provide greater certainty for farmers in terms of the prices they can expect to receive. Subsidies have encouraged the use of modern inputs. Estimated subsidies, including irrigation charges that do not cover fully costs and non-payment of water charges, low interest rates and non-payment of credits, low electricity rates and non-payment of electricity charges, and the difference between international and domestic prices of fertilizers, are now equal to about 16% - 17% of the agricultural GDP (higher in the case of wheat and rice) and, correspondingly, about 2.5% of total GDP. Other Government policies have delinked internal and external markets by imposing quantity and/or price con_rols over exports/imports of agricultural commodities and by "canalizing" the sale of agricultural exports and imports through public corporations. The Government also extends support to agriculture through research and extension programs. Besides such broad interventions, the Government has frequently launched crop-specific programs with more limited objectives. 1.02 The simultaneous implementation of such wide ranging, and sometimes divergent, policies impinges upon farmers' incentives in many complicated and sometimes conflicting ways. Analysts of the impact of Government policies typically have examined their effect on a summary variable -- net income per hectare in a given crop -- in assessing whether the policies in question have stimulated additional output. More sophisticated analyses also consider the opportunity cost of the policy, in terms of the policy-induced substitution between one crop and another. The inclusion of opportunity costs clearly represents an improvement over the simple calculation of net income per hectare. However, both approaches rely on domestic prices in their evaluations. ThuL, they implicitly adopt an autarkic view of the agricultural sector and thereby neglect the possiblities of gains from international trade and chlaiges in net foreign exchange earnings. 1.03 Another approach, taken in this study and in recent World Bank work, (Krueger, Schiff and Valdes) is to analyze how agricultural policies affect the differential between domestic prices (adjusted for subsidies and input costs) and international prices. In other words, are certain crops, and agriculture in general, being protected or disprotected (in effect taxed) by the maintenance of prices that diverge from those that would prevail under free international trade? To see the importance of this question, suppose that policies combine to keep the domestic price of  one commodity above the world price (protecting that crop) and prices of a second commodity that could be produced on the samoe land, below world prices (disprotecting or in effect taxing that crop). Then, from an efficiency standpoint, too much land, labor and capital are being used to produce the first commodity and too little are being used to produce the second. Resources and foreign exchange could be saved by adjusting prices to stimulate a shift of resources from production of the first commodity to the second. Demand could be satisfied and foreign exchange earned on balance by exchanging the increased production of the second crop for imports of the first in international markets. 1.04 The differentials between domestic and international reference prices are measured in this study by three standard ratios, referring to comparisons at three levels of increasing complexity -- nominal protection coefficients (NPCs), effective protection coefficients (EPCs), and effective subsidy coefficients (ESCs). These coefficients are defined below but basically the NPC indicates the incentive to pruduce a commodity, the EPC indicates the incentive to the use of resources (land, labor and capita'.) in producing the commodity after deducting the cost of tradeable intermediate inputs, and the ESC adjusts the EPC to take into account taxes and subsidies on non traded inputs (e.g. irrigation, electricity); all relative to what would exist under free trade wi.hout taxes and subsidies. 1.05 In calculating the international competiveness of a commodity, transport costs can make a potentially enormous difference. In terms of competing with imports, international transport costs provide a degree of protection for domestic producers. In contrast, exporting means that the domestic producer's price must be low enough to make the product competitive in foreign markets, including transport costs to the market. A simple example will make clear how important this difference is: If international transport costs are $10 per ton and the international price of a good at the foreign point of sale was $100, then domestic producers could compete effectively with imports at the border, providing they produce profitably at a price of $110. However, in order to export, domestic producers would need to be able to produce profitably at a price of $90. Thus, the competitive border price is about 22% higher than the competitive export price. This means that investments in production could easily earn a high economic rate of return as long as domestic production is substituting for imports, but might become unprofitable once it became necessary to capture an export market. Moreover, domestic transport costs can provide additional "natural protection" to production in inland areas. 1.06 To cover this issue of transport costs, the NPCs, EPCs, and EPSs have been calculated under two hypotheses: a) the crop in question is imported and thus competes at the domestic port with imports including their transport cost (the importable hypothesis) and b) the crop in question is exported and thus competes at a foreign port including transport costs (the exportable hypothesis). Thus the two different treatments of transport costs allow the analyst to see whether the Indian producer is competitive with imports given the protection accorded by transport costs and what, if any, subsidy would be required to export the crop. Adjustments are also made for domestic transport costs, as explained in the detailed crop studies.  -3- 1.07 In additicn to NPCs. EPCs, and EPSs, this study also comments on the size of the price differentials under the Assumption that foreign exchange is accorded a notional premium of 25% over the going exchange rate. The inclusion of such a premium is assumed to cover not only the possibility of currency overvaluation under the current trade and payments regime, but also the compensatory change in the exchange rate that would be necessary if all protection were to be eliminated. With the inclusion of such a premium, the coefficients become good ?roxies for cost benefit indicators such as Domestic Resource Cosc (DRC). Finally, the coefficients also are estimated for specific regions and crop varieties to provide some idea of the differences in incentives between regions. 1.08 The Nominal Protection Coefficient (NPC) is defined as the ratio of the domestic price to the world reference price of the commodity under consideration. Symbolically, (1) NPC - p /P where, NPC - Nominal Protection Coefficient PD - Domestic Price of the commodity in question at the farmgate P - Reference Price of the commodity in question, i.e. what the farmer would have received in the case of free trade. 1.09 In this study, the domestic price is approximated either by what the cultivators of the relevant commodity receive- or what the Government announces as its support/procurement price; the world reference price is derived from the international price, adjusted for transport cost (both foreign and domestic), and marketing and trading margins, including any processing necessary to make the domestic commodity equivalent to the internationally tvaded commodity. 1.10 If the NPC is greater (less) than one, then the commodity is protected (disprotected or in effect taxed), compared to the situation what would prevail under free trade. Of course, one needs to compare NPCs across crops and industries and make an adjustment for the premium on foreign exchange to get a better idea of whether the policy regime provides more or less incentives for production of a given crop than under free trade. For example, if the average NPC was 1.4%, then crops with NPCs of 1.4% or more would be receiving an incentive vis a vis other import substitutes; crops with lower NPCs would be receiving an incentive vis-a-vis exportables and, perhaps, non-tradeables. 1See Garry Pursell and Neil Roger and Pasquale L. Scandizzo and Colin Bruce, for further details.  -4- 1.11 The Effective Protection Coefficient (EPC) is defined as the ratio of value added at domestic prices to value added2 at world reference prices. Symbolically, (2) EPC = VADIVAR = (pD_SUm(aij PDj))/(PR_Sum(aij pRj) = (PD_TID)/ ((PD/NPC)_(TIDNPCi) where VAD = Value Added in Domestic Prices VAR = Value Added measured at International Reference Prices pD = Domestic Price of the good in question (the ith good) pR * International reference price of the good in question. aij = Quantity of the jth input used to produce the ith good pDj = Domestic Price of the jth good pR. = international Price of the jth good. TID = Sum (aij PD.) NPC = Nominal Pro ection coefficient of the good in question NPCi= Weighted average of the Nominal Protection coefficients of the tradeable inputs into good i, with the weights equal to the value shares of the inputs in the reference price. Thus, a prerequisite for estimating EPCs is a detailed knowledge of the input structure of the commodity under consideration and the nominal protection not only on the output but also on its traded (tradeable) inputs. 1.12 Again, a coefficient greater than (less than) one indicates protection (disprotection or, in effect, taxation) compared to the free trade. Effective protection (EPC) will be greater than (less than) the nominal protection (NPC) to the extent that tradeable inputs into the production process have a lower (higher) NPC than the product. Thus the EPC refers to the protection accorded to the land, labor and capital (the value added) used in the production of the commodity in question, while the NPC reiers to the protection accorded a commodity, including all the inputs that go into its production. The NPC overstates (understates) the incentive to apply resources to production if the inputs have higher (lower) NPCs than the final product. Hence, the EPC is a better measure of the incentives to produce a commodity than the NPC. 1.13 Finaily, the Effective Subsidy Coefficient (ESC) essentially adds net subsidies (subsidies minus taxes) on non-traded (non-tradeable) inputs in the numerator of EPC and divides by value added at world reference 2 Value added as used here refers to the difference between the output price and the per unit value of all traded (tradeable) inputs that go to produce one unit of output. Theoretically, the non-traded inputs should be decomposed into their tradeable and primary, non-traded components, and different treatments accorded to both, but in practice such refined treatments are precluded by lack of data.  -5- prices. Symbolically. (3) ESC - (VAD+NS)/VAR 'here NS - Subsidies (net of taxes) on non-traded inputs, and VAD and VAR were defined previously. Net subsidies on traded inputs are already taken into account in their prices and thus are already included in the EPC calculation. The ESC goes a step beyond the EPC in measuring incentives to the use of resources in a crop or a sector, because it takes into account not only protection and subsidies on traded (tradeable) inputs, but also net subsidies on non-traded (non-tradeable) inputs. 1.14 The three estimates of protection coefficients defined above help (1) to measure the extent of divergence (distortion) between domestic and international prices as a result of the regulatory policies on international trade and domestic markets; (2) to measure the level and differences in 'effective incentives' for cultivators of different crops in different regions. This in turn can throw some light on issues related to equity in the distribution of incentives across regions, as also assist in identifying the causes that lead to faster growth of some crops in certain regions than in others; (3) to measure comparative advantage and thus degree of competetiveness of various crops in different regions, considering foreign trade as a transformation frontier. This would be valuable information assisting in allocating existing and investment resources more efficiently, and evolving a rational/desirable cropping pattern; (4) to test the hypothesis of underpricing of agriculture in developing countries (Schultz, 1978) and to measure the extent of bias against agriculture and in favour of industry. B. Estimation of Protection Coefficients and Results 115 Four comaodities are covered in this study: wheat and rice, India's most important cereals; cotton, an important cash crop used in the country's large textile industry; and groundnut as a representative of the oilseeds crop complex. Together, these four crops account for roughly 45 per cent of gross cropped area and above 50 per cent of gross value of crop output. Wheat, rice and cotton experienced technological breakthroughs during late 1960s and 1970s in certain regions, while groundnut has lagged behind. Presently, efforts are being made to stimulate oilseed production under the Technology Mission on Oilseeds launched by Government of India in 1986 and a buffer stock/price support scheme vas introduced in 1989.  - 6 - 1.16 Tables 1.1 and 1.2 suimarize the estimates of the protection coefficients for each of these crops for the 1980s (1980-81 to 1986-87), for the major producing states and c5op varieties, under both the importable and exportable hypotheses . Details on the individual crops are found in Chapters 2 (wheat and rice), 3 (cotton) and 4 (groundnuts) and the references cited there. 1.17 Starting first with the incentives for crops as import substitutes (the importable hypothesis), the crop-specific NPCs4 for the period 1980/81 to 1986/87 were lowest in case of rice (0.67) and highest for groundnut (1.50) (See Table 1.1). Wheat and cotton fall in between, with both having NPCs of 0.80. Across states, the lowest NPC is for Andhra Pradesh cotton (MCU-5 variety, NPC-0.63), followed by rice in Bihar and Orissa (NPC-0.65). In sum, domestic prices of rice, wheat, and cotton were all well below comparable world prices, indicating disprotection or effective taxation of these commodities compared to border prices. In contrast, groundnut was highly protected. The relative incentive provided for groundnuts by the policy regime and the delinking from international markets is more than 123% higher than for rice. 1.18 Turning to effective protection (under the importable hypothesis), the EPCs for wheat, rice, and cotton all were less than the coresponding NPCs. This is because tradeable inputs used in these crops were protected, while the crops themselves were disprotected. The difference between the EPC and the NPCs is greatest in the case of cotton (13.75%) and smallest in the case of rice (2.98%). The large difference between the EPC and NPC in cotton, especially that of Gujarat (by 31.5%), reflects the large share of tradeable inputs, primarily pesticides, in the production of cotton, and the fact that NPC of pesticides is much higher than NPC of cotton. On the other hand, the protection on groundnuts was 3 In the case of wheat, the coefficients are estimated separately for four states - Uttar Pradesh, Punjab, Haryana and Madhya Pradesh. Fair Average Quality (FAQ) of domestic wheat, as categorised by Food Corporation of India, is compared with US Hard Red Winter No. 2 with ordinary protein. In the case of rice, the coefficients are calculated for Andhra Pradesh from the south, Bihar and Orissa from the east, Punjab and Uttar Pradesh from the north and Madhya Pradesh from the central region. Indian "Common" (FAQ) rice is compared with Thai (Milled) white, 5 per cent Broken. The protection coefficients for Cotton are worked out for Maharashtra, Gujarat, Punjab and Andhra Pradesh. In case of Maharashtra Hybrid-4 variety is compared with its like Mexican cotton; Gujarat's Shankar-4/6 is compared with California (SM l 1/8"), Punjab's J-34/320F is compared with Orleans/Texas (1") and Andhra MCU-5 with Giza 67/69/81. In case of groundnut, protection estimates are for Gujarat, Andhra Pradesh and Tamil Nadu. 4 The crop specific NPCs were calculated as a weighted average of the state NPCs, with value of the crops in the states used as the weights.  -7- Table 1.1: CrOD and Reienon-Soecifig Disuersion of Effective incentives in Indian Agriculture (Average 1980-81 to 1986-87) Indeax of ESCs Crop Variety (Welghted --------------------------- Hypothesis/Crop/State NPCs EPCs ESCs Average-100) ANPCz Domestic International ---------------..--------------------------------------------__--------------__--------------------------------------------- Importable 'Jypothesis WHEAT Haryana 0.8% 0.79 1.03 110.75 0.66 PAQ Hard Red WLnter No. 2 Madhya Pradesh 0.75 0.73 0.96 103.22 0.60 -do- with ordinary protein (USA) Punjab 0.85 0.80 0.93 100.00 0.66 -do- -do- Uttar Pradesh 0.77 0.73 0.91 97.85 0.62 -do- -do- Weighted Average 0.80 0.75 0.93 100.00 0.63 -do- -do- RICE Andhra Pradesh 0.69 0.66 0.88 100.00 0.55 Common Thai White (Milled) 5X Broke Bihar 0.65 0.64 0.86 97.72 0.52 -do- -do- Madhya Pradesh 0.67 0.65 0.85 96.59 0.53 -do- -do- Orissa 0.65 0.63 0.84 95.45 0.52 -do- -do- Punjab 0.74 0.72 1.01 114.77 0.58 -do- -do- Uttar Pradesh 0.66 0.64 0.85 96.59 0.52 -do- -do- Weighted Average 0.67 0.65 0.88 100.00 0.54 -do- -do- COTTON Maharashtra 0.96 0.93 0.98 103.67 0.74 H-4 Mexican Gujarat 0.89 0.61 0.67 89.33 0.68 5-4/6 California (SM 1 1/8-) Punjab 0.83 0.74 0.86 114.67 0.o4 J-341320F Orleans/Texas (1") Andhra Pradesh 0.63 0.54 0.55 73.33 0.50 MCU-S Giaa-67169181 Weighted Average 0.80 0.69 0.75 100.00 0.62 GROUNDNUT Gujarat 1.47 1.59 1.70 99.41 1.17 Andhra Pradesh 1.50 1.54 1.72 100.58 1.18 Tamil Nadu 1.53 1.55 1.73 101.17 1.21 Weighted Average 1.50 1.56 1.71 100.00 1.18 Exportable Hypothesis WHEAT Punjab 1.34 1.71 2.01 0.98 FAQ Hard Red Winter No. 2 (USA) RICE Punjab 0.87 0.87 1.21 0.66 Common Thai White (Milled) 5S broke COTTON Maharashtra 1.13 1.12 1.17 132.95 0.86 H-4 Mexican Gujarat 1.10 0.74 0.81 92.04 0.78 S-4/6 California (SM 1 1/8") Punjab 0.98 0.89 1.04 118.18 0.75 J-34/320F Orleans/Texas (1") Andhra Pradesh 0.67 0.56 0.58 65.91 0.53 MCU-5 GLza-67169/81 Weighted Average 0.92 0.80 0.88 100.00 0.71 GROUNDNUTS Gujarat 1.87 2.48 2.60 105.26 1.45 Andhra Pradesh 1.91 2.15 2.42 97.97 1.48 Tamil Nadu 1.95 2.13 2.38 96.36 1.51 Weighted Average 1.91 2.26 2.47 100.00 1.48