Liberalizing Indian agriculture : an agenda for reform

WPS 1 172 Pollcy Research WORKING PAPERS Trade Policy Policy Research Department The World Bank September 1993 WPS 1172 Liberalizing Indian Agriculture An Agenda for Reform Garry Pursell and Ashok Gulati India's incentive system heavily favors manufacturing and dis- criminates against agriculture. This proposed reform agenda would remove major policies that distort agricultural imports, exports, inputs, and domestic markets. It would protect low- income groups against necessary increases in food prices. PolicyRcar hWoingPpasdi_nateinthefindingsofwoknprognsandcncouragethcxchangcofidcasamongBankstaff and aBothes intsradindedvlppmwntissues.hepaps.distzibutedbythcRcsslrchAdvismyStaff,canythenamaoftheauthon.rdflect cnytdrviows.andghouldbeusedand dtcdacrdingly.Thefindings.intepeutions.andconclusionsatiheauthoown.lTheyshmdd na be sbuAted to the Wold Bank, its Board of Dirctoai, its management, or any of its mber countrics.  Polloy Research Trade Policy| WPS 1172 Research for this paper- ajoint product of the Trade Policy Division, Policy Research Department of the World Bank, and the National Council of Applied Economic Research in New Delhi - was carried out mainly by consultants in India. It is part of a long-term research program to quantify the impact on agriculture of India's trade and other incentive and regulatory policies. The paper was funded by the Bank's Rescarch Suppoit Budget under a dissemination grant (RPO 678-04). Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Dawn Ballantyne, room NIO-023, extension 37947 (44 pages, September 1993). in July 1991, India embarked on a program of * Initially, allowing the export only of high- econormic decontrol that greatly speeded the quality, high-priced varieties of such commodi- previously slow process of liberalizing trade and ties as cotton ani rice, to limit upward pressures domestic regulatory controls begun in 1978. But on domestic prices of lower-quality varieties, the focus of reform has been on manufacturing. which are important to consumption in low- Reformn has barely touched agriculture, which income Indian households. accounts for two-thirds of employment in India and about 30 percent of India's GDP. * Liberalizing fertilizer imports and deregulating domestic manufacturing and the Although some crops (notably oilseeds) distribution of fertilizers. receive heavy protection, the net effect of interventions to date is to heavily favor manufac- * Removing subsidies on irrigation, electric- turing over agriculture. In this agenda for reform, ity, and credit (and creating conditions to facili- Pursell and Gulati recommend: tate the trading of canal irrigation water rights). - Removing all quantitative export and import * Deregulating the wheat, rice, sugar, cotton, controls on agriculture, except for special and edible oil and oilseed industries, and abolish- treatment (such as export taxes) when Indian ing compulsory government acquisition at expol-t would be substantial enough to depress below-market prices of sugar, molasses, and worid prices (most likely with rice). milled rice. * Further reducing protection on manufactur- * Reforming the food security system to ing, rather than bringing protection for agriculr protect low-income groups from the increase in ture up to the same level. the general level of food prices required by the liberalization of agriculture. This would involve * As a transitional measure, consider:ng the better targeting of food subsidies and associated use of variable tariffs based on weighted aver- reforms of the public distribution system, or even ages of past intemational prices as a way to its eventual replacement b) a food stamp system. parlSy insulate domestic prices from extreme fluctuations in world prices. The Policy Research NVorking PaperSeries di&seminates the findings of work under way in the Bank. An objectiveof the series is to get these findings out quickly, even if presentations are less than fully polished. The fndings, interpretations, and conclusions in these papers do not necessarily represent official Bank policy. Produced by the Policy Research Dissemination Center  LIBERALISING INDIAN AGRICULTURE: AN AGENDA FOR REFORM by Garry Pursell and Ashok Gulati Garry Pursell is Principal Economist, Trade Policy Division, Policy Research Department at the World Bank. Ashok Gulati is Director, National Council of Applied Economic Research, New Delhi. We are grateful to Hans Binswanger for comments on an earlier version of this paper. The findings, interpretations and conclusions of this paper are the authors' own, and should be used and cited accordingly. They should not be attributed to the National Council of Applied Economic Research or to the World Bank, its Board of Directors, its management, or any of its member countries.  A-Note This paper is based on the final chapter of the monograph atled 'Trad Policy, Incentives and Resource Allocation in Indian Agriculture' by Ashok Gulati and Gury Purl (1993), which is being prepared for publication. AkreYadoQn CCI Cotton Corporation of India EPC Effect;ve Protection Coefficient ESC Effective Prote'nion Coefficient FCI Food Corporation of India PICC Fertiliser Industry Coordinating Committee GOI Government of India JRY Jawar Rojgar Yojna MFA Multi Fiber Arrangement MMTC Minerals and Metals Trading Corporation NAFED National Agricultural Cooperative Market Federadon NDDB National Dairy Development Board NPC Nominal Protection Coefficient O&M Operation & Maintenance PDS Public Distribution System QRs Quantitative restrictions RPS Retention Price Scheme SEB State Electricity Board SSI Small scale industry STC State Trading Corporation TMO Technology Mission on Oilseeds UP Uttar Pradesh Gloss=r Crore 10,000,000 Lakh 100,000 Quintal 100 kilos ,E,,xanoe Rate Financial year averages (April 1 to March 31): Rs/!US 1985/86 12.2 1986/87 12.8 1987/88 13.0 1988/89 14.5 1989/90 16.6 1990/91 17.9 1991/92 24.5 1992/93* 30.3 Jan. to Dec. 1992  CONTENTS Introduction I. The Existing Controls and their Impact on Cutivators' Incentives I. 1 External trade policies 1.2 Domestic regulatory policies I.3 Impact on the level and structure of incentives before the 1991 devaluation 1.4 The structre of incentives after the July 1991 devaluation H. Eolicy Reform: Objectives for the Medium or Term and Tacics for the Short Term ..1 Removing anri-agriculture bias and creating more neutral incentives within agriculture (i) Removing anti-agriculture bias (i) Creating more neutral incentives (iii) Short term tactics H.2 Special treatment of commodities in which the poteal Indian supply or demand may appreciably affect world prices in the long run 11.3 Dealing with unstable world prices and fluctuations in domestic producdon 1.4 Removing import controls, high tariffs and domestic reguatory controls fr m tradeable inputs used in agriculture 11.5 Removing large subsidies on non-traded inputs: canal irrigation, electricity and credit 11.6 Removing controls and distortions associated with the 'food security' complex (i) Food Stamps (ii) Reforming the PDS system 11.7 Removing other domestic regulatory controls and distortions m. Policy Reforms: Some General Suggestions IV. Likely Implications End Safety Nets  LIBERALISING INDIAN AGRICULTURE: AN AGENDA FOR REFORM Introduction Beginning in July 1991, the Indian govermnent embarked on a program of economic decontrol which greatly speeded up the slow process of economic liberalisation that had been under way since about 1978. The stabiiisation program adopted at the same time, and particularly the devaluation and floating of the Rupee, has had indirect effects on the whole economy including agriculture. But the focus of the specific reforms has been almost entirely on manufacturing, with the abolition of most of the stifling system of industrial licensing, the removal of import licensing from nearly all manufactured intermediate goods and capital goods, tariff reductions and the relaxation of the rules regulating foreign investment. By contrast, very little of the reform effort has so far been directed towards agriculture, even though the agricultural sector in India is quite large, accounting for two thirds of the employed population and about 30 percent of GDP. As we believe that extending the liberalisation process to Indian agriculture has the potential to greatdy increase the efficiency and the growth momentum of the economy (as has happened in China), and because there is a widening domestic public debate on the subject in Indian academic journals and in the media, we have thought it worthwhile to set out our ideas in the form of a reform agenda. To this end, the paper first briefly summarises previous research on the nature and degree of govermnent interventions in agricultural markets, and what impact these interventions had on the overall incentives for agricultural production and on the relative effective incentives for the cultivation of different crops during the 1980s (section 1). It also suggests what has probably happened to these incentives as a result of recent policy changes, particularly the devaluation of July 1991 and the associated and continuing liberalising reforms which have so far principally affected manufacturing. Section It delineates the broad direction in which agricultural policy reforms ,5ed to move in the medium to long run so that the system becomes more transparent and promotes both allocative and technical efficiency in the use of scarce resources. In this section we also attempt to chalk out details of a strategy that we feel might be politically feasible in the current environment. Based on the experience of other countries and India's experience so far with policy reforms affecting manufacturing, Section Im makes some suggestions of a general nature on the tactics of reform which we hope might be useful for policy makers . And finally, in section IV, we discuss some of the likely implications of this reform agenda, and suggest what could be done to minimize its possible adverse effects on the poor.  L The Exsting Controls and their bripact on CultIvaors Jncendlvs IL1 All except a few agricultural imports and exports are subject to non-tariff controls of one kind or another, including import and export licensing, canaisation' in which only one specified parastatal is allowed to import or export the commodity, and the use of minimum export prices.' On the import side the only exception is pultses, which are imported by private traders over a low tariff (at present 10%). Agriclwture was not included In the trade Ilberalisatlon measures taken during 1991 and 1992, except fo. the reiaxation of some export controls which in most cases left other controls on the same commodities in place. At the end of 1992 about 60 agricultural and livestock products were subject to some form of export control, as well as about 46 manufactred products, most of which were processed primary commoxiities. in April 1993 a fi-ther range of products was removed from this list, but most products which are actually exported or which have export potential either remain on the list and are subject to various kinds of -. ort control, or were removed from the list but are now subject to ad hoc export controls to be arnounced in public notices. The 1991/92 reforms reduced the share of inrnationally tradeable GDP subject to some form of quantitative import restriction from about 93 per cent at the end of 1990 to about 75 per cent in May 1992. But practically all of this change was in m, for which the share of value added subject to QRs fell from 90 per cent to about 46 per cent. By contrast the share of agricultural and livestock GDP subject to QRs barely changed, from 94 per ceut before the reforms to 93 per cent in May 1992.2 .2 Domestic Regguaeor Police For the most part domestic trade in agricultural commodities within India is not physically restricted. Most products are traded and transported nationally. Ihere is excellent and up to date information on prices nation wide, and private markets operate remarkably efficienly considering the very considerable communication, transport (notably the numerous municipal road tax- "octroi"- points ) and other handicaps they face. Nevertheless, there are a number of physical constrains on the free movement of agricuteural commodities and regulatory and other interventions which seriously distort domesdc markets. There was no liberalisation of these controls during 1991 and 1992 when regulatory controls over manufactring were significantdy reduced. These include: - The periodic physical botding up of the wheat surpluses in the north west (Punjab, Haryana and western Uttar Pradesh) in order to allow the Food Corporation of India ( PCI) to procure its requirements at the official procurement price. -The "levy 1.ice" compulsory acquisition ) systems for rice and sugar which are used to obtain the estimated govetrnme 'eguirements for the public distribution system ( PDS)P and for buffer stocking, an,' which s%v ..Iy distort these two markes. 2  -The operations of the PCI and the PDS and the associated regulatory controls implemented by the Department of Food and the Deparctmw of Civil Supplies. These distort the normal regional and seasonal variations of commodity prices, prevent or constrain efficient private trading operations, and distort production decisions by farmers P-r .gards toth timing and location. In particular, because of procurement prices which are the sane throughout the year, very large grain deliveries by farmers in the surplus north west are concentrated in a highly wasteful manner into just a few weeks. There is also a great deal of waste, inefficiency and overemployment and large scale rent seeking (e.g. about a third of all the wheat, rice and sugar, and over half the edible oils are estimated to be diverted from the PDS'. -Periowic controls on the movement of groundquts and groundnut oil out of Gujarat. -Some state level controls, e.g., monopsonistic purchases of rice by the government in the Thanjavur district of Tamil Nadu and of cotton by the Maharashtra Federation (often referred to as monopoly procurement schemes). -The pervasive controls on the operations of private traders, including (i) the general ban (with a few minor exceptions) on futures trading (ii) inventory control1s (iii) credit controls. -Discrimination by Indian Railways against private traders in favor of parastatals such as FCI. -The regulatory and other activities of the various commodity-specific boards or other government organisations, e.g., the Cotton Corporation of India, the Ministry of Textiles, the National Dairy Development Board, the Jute Corporation of India, the Tea, Coffee, and Tobacco Boards etc. -Extremely detailed and highly distortive regulation of the sugar industry, both by the central government and by the governments of the main sugar producing states . -Price and other regulatory c-rols over the processing of primary commodities which seriously inhibit the efficiency of the 'modem' sectors of these industries while allowing the continued existence and/or further development of inefficient, high cost small-scale processors which are free of all controls and taxes. Examples: sugar mill controls and khandsari units in UP; price controls on cotton ginning; controls over edible oil processing; controls (the levy system) on rice milling, etc. -Subsidies to farmer cooperative tradig and processing organisations which make it difficult or impossible for private traders or processors to compete. Examples: sugar milling in Maharashtra, subsidised edible. oil mills supportod by the National Dairy Development Board ( NDDB). Many "cooperatives" (e.g., in UP ) are in practice state government controlled and managed, and highly politicized. L2 hIact on the Level nd Structure of Incentives Before the 1991 Devaluation6 Before 1991, on the basis of measured nominal protection (i.e., comparing domestic and world prices), in the aggregate Indian agriculture was heavily discriminated against relative to 3  manufacturing. This is shown in Fig. 1, which graphs estin-tes of the weighted average nominal protection coefficients for agriculture as a whole fo; the 25 years 1964/65 to 1989/90; estimates of the weighted average NPCs of manufacturing for th3 17 years 1970/71 to 1987/88; and the relative NPCs of agriculture for the 1970/71 to 1987/88 perik4 obtained by dividing the coefficients of agriculture by those of manufacturing. I On average, during the 17 years for which the comparison has been made, the protection level for agriculture was about half the protection level for manufacturing. The gap widknew in the early 1970s as vg icultural protection declined steeply while manufacturing protection increased, and during the re' . of the 1970s the anti-agriculture bias was particularly pronounced, with the average agriculture NPC only about a third of the average NPC for manufacturing. During the 1980s until 1987/88 there was a pronounced trend in the opposite. direction, with the average NPC of agriculture rising steadJy and a substantial decline in manufacturing protection. The decline in manufacturing nominal protection presumably reflects the easing of import controls and the liberalisation of domestic industrial licensing and other controls on manufacturing during this period, as well as the slight real appreciation of the Rupee between 1978/79 and 1983/84. Even so, in 1987/88 a big gap remained, with a nominal protection rate of just above zero for ag.iculture and almost 50 percent for manufacturing. After this the average NPC of agriculture declined along with the Rupee devaluation which accelerated after 1988, and an increasing trend in world commodity prices. A corresponding post 1987/88 series for manufacturing is not available, but in all likelihood, owing to the Rupee devaluation, manufacturing prices would have also declined relative to Rupee denominated world prices. Whether, compared with 1987/88, the anti-agriculture bias of the system would have increased or declined would largely depend on the speed and extent of the upward movement of domestic agricultural prices compared to the upward movement of domestic manufactured goods prices in response to the devaluing Rupee. 4  FIG 1: AGGREGATE NOMINAL PROTECTION COEFFICIENTS FOR AGRICULTURE AND MANYUFACTURINNG 2 - 1.8 MANUFACTURING 1.6 - 1.4- 1.2- O F RICULJ~RCULURE 0.8 0.6 0.4- RELATIVE (Agr/Manuf) 1965 1970 1975 1980 1985 1990 Nominal Protection Coefficient , value of output in domestic prices value of output in reference prices Note: References prices are world prices adjusted for port costs and domestic transport and marketing costs to the point at which the Indian products do or would compete internationally. The aggregation for agriculture treats all the aggregated commodities as import substitutes i.e. it is made on the importable hypothesis.  This finding of a marked and continuing anti-agricultural bias in the incentive system holds up after allowing for the effects of protection on the cost of tradeable inputs used in agriculture, the subsidies to agriculture's principal non-tradeable inputs. and the exemption of agricultural activities from corporation and Income taxes. Aa shown in Table I: Table I Indicators of Incentives to Agrliculture and Manufa~tuxD Oumit a (adj. for a. tI. ocanp.) Agriculture 0.88 0.97 0.86 0.97 to 0.90 (1980/81 to 86/87) 1.07* Manufacturing 1.42 1.44 1.34 n.a. 1.41 (1986/87) Ratio Ag/Manuf. 0.62 0.67 0.64 n.a. 0.64 source, Gulati and Pursell (1993). Notes: NPC-Nomlnia protection coefficient EPC=Effective protection coefficient BSC=Effective subsidy coefficient 7 The ESC for agriculture ranges from about 0.97 to about 1.07 depending upon the definition of the canal irrigation subsidy i.e., whether the cost of canal irrigation includes operating and maintenance (O&M) expenses only or the eatmated annualised capital cost of all irrigation schemes, as the cost to be recovered from the farmers. -Allowing for the protection of tradeable inputs reduces anti-agricultural bias (in this case measured by the ratio of the aWegate effective protection coefficient of agriculture to the aggregate effective protection coefficient of manfacurl by a negligible amount. For agriculture, the principal tradeable inputs are fertilisers, farm achinery, seeds and pesticides. On average, agriculture obtained its internationally tradeable inputs at less than world prices. This was entirely due to fertliser, which in most years was supplied to farmers at prices well below the border price plus estimated delivery costs to the farm. The nominal protection of farm machinery was low to moderate by Indian standards (there is a relatively efficient domestic tractor and farm machinery industry) but nominal protection of pestcides was high. -Subsidies to non traded agricultural inputs viz. canal irrigation, electricity, and credit are substantial and are reflected in an aggregate effective subsidy coefficient (ranging from 0.97 to 1.07, S  depending upon the definition of tho Irrigation subsidy adopted) which is well above the aregate effective protection coefficient (0.86). Even so, without allowing for any subsidies for manufacturing and simply comparing the agriculture ESC above with the EPC estimsie for manufacturing, the ratio (0.72 to 0.80) stll indicates substantial anti-gricultural blas. Ithe comparison would be less favorable for agriculture if allowance were made for the subsidy squivalent of the various forms of government support for the large number of "sick' manufactrin5 firms, and for loss-making public enterprises which can only continue to operate with such support even though they may not be offitially defined as sick. -The exemption of agriculture ftom income and corporate taxation increases aggregate incentives for agriculture relative to aggregate incentives for mu ng to a very minor extent by comparison with the anti-agricultural bias resulting from trade policies. During the 1980s, the excess of the free trade exchange rate over the official rate increased from about 30 percent at the beginning of the decade tc about 40 to 50 percent towards the ed. The increasing prczium reflected the substantial increzv in impsot duty rates over the period and the growing trade deficie. The discrimination against agriculture iro; the overvalued exchange rate was therefore substantial, a finding which is consistent with the findings of the Krueger-Schiff-Valdes country studies. 10 Within agriculture, there were large differences in net incentives between crops. At the official exchange rate, as measured by the effective subsidy indicator (which allows for the protection of output, tradeable inputs and subsidies on non traded inputs) the Gulati-PurseU et al research classifies the main crops broadly as follows: NGegatiyeincntive Rice Cotton Zero or low incentives Wheat Coarse grains Pulses Tobacco High incentives Oitseeds incl. coconut/copra Rubber Sugarcane The following crops not included in the research probably have negative or low incentives Coffee Tea Cocoa Jute Spices Fruit and vegetables 6  There are no quantitative incentive estimates for livestock and fishing but net incentives are probably low. (Wool is imported without restriction over a 10 per cent tariff; there are export controls on the exports of hides and skins. On the other hand the import of meat and dairy products is banned, except for small regulated imports of powdered milk). Because if international and domestic trans.nort and marketing costs, the measured incentive level of a crop can be considerably affected according to whether it is treated as an importable or an exportable. For example, the classification above is based on wheat as an import substitute, but in the 1980s the net incentive to wheat production was quite high if wheat is considered as an exportable. Treating wheat and rice as exportables raises the aggregated incentive level for agriculture as a whole, but even with this adjustment it is still well below aggregate incentives for manufacturing . As noted, these classifications are in relation to world prices converted at the official exchange rate. If the incentives are considered in relation to average manufacturing incentives or to the estimated free trade exchange rate, the crops with zero to low positive incentives are strongly discriminated against, and the net incentives to the oilseeds, sugar and rubber are lowered. The incentives for oilseeds and sugar in most years were nevertheless high, even by comparison with some of the most highly protected manufacturing industries. IA. the Strcture of Incentives after the July 1991 Devaluation For this period, there is io comprehensive empirical work on manufacturing protection and only some limited nominal protection estimates for agriculture. But it is unlikely that the basic structure of relative incentives for agriculture will have changed very much. The overall anti-agriculture bias is certainly still in place for the following reasons: -Measured manufacturing protection will have come down owing to (i) the removal of QRs on most manufactured intermediate goods and most capital goods (ii) the reduction in the maxium tariff to 110 % in 1992 and to 85 % in 1993 (iii) the large number of effectively non traded manufactured products resulting from continuing import bans (e.g., cn all consumer goods) or from redundant tariffs. The domestic prices of the latter are delinked from world prices and most of them have probably not risen by the full amount of the devaluation. But in agriculture the domestic prices of wheat, rice and coarse grains have also risen by much less than the devaluation, so that the average price level of agriculture has also fallen relative to world prices. Other agriculture prices (e.g. oilseeds, pulses, cotton) have moved up more or less in line with the devaluation, but the main graim dominate agricultural GDP. The big dispersions of incentives within agriculture are basically unchanged and may have Increased in some respects. Following the devaluation up to about February 1992, border prices of wheat and rice were about double domestic prices. For the first time in many years market prices of wheat in Punjab were consistently well below (about 40 percent) estimated export parity prices, i.e., estimated fob prices minus transport costs to Bombay from Punjab minus Bombay port costs. The domestic prices of common rice were even lower than this in relation to export parity prices. Punjab prices remain 7  well below export parity prices despite substantial increases in the government's procurement prices for the 1992/93 season. This situation can only be maintained by the continuing export controls. On the other hand domestic edible oil prices rose substantially and remained at moie or less the 1980s level (two or three times as high) in relation to import prices. If maintained, this disparity will increase the substitution of oilseeds for wheat and other grains and maintain about the same pull of resources from pulses and other crops whose prices have risen more or less in line with the devaluation. Relative to border prices, the cost of traded inputs for agriculture has probably declined slightly. Fertiliser prices have increased , but overall by not as much as the devaluation (see later discussion). The costs of farm machinery, pesticides and minor tradeable inputs have also probably gone up less than the devaluation. The non tradeable Input subsidies to agriculture will have declined In real teems but are still substantial and highly distortionary. As long as the devaluation remains a real one, i.e., is not erased by increases in the prices of non tradeables, the subsidies on the non tradeables will represent less in terms of the world prices of the agricultural commodities. However, this effect may be offset to some extent by increases in the default rate on agricultural loans resulting from a politically motivated govermment program in 1990 which waived repayments of agricultural loans. The overall level of input subsidies is also probably higher in budgetary terms. IL Policy Reform: Objectives for the Medium or Long Term and Tactics for the Short TOm This section suggests objectives for the medium or longer run, and some ideas on what might be feasible start in the present environment. The next section makes some general suggestions on the tactics of reform. II.1 Removing anti-agriculture bias and creating more neutral incentives within agriculture 0) Removing anti- agricultural bias The main instrument for this should be the continued reduction of protection to manufacturing. The government has stated that it intends to continue removing QRs applied to manufactured goods i.e to manufactured consumer goods, since most intermediate and capital goods are already freed from import licensing. In 1992 it made a small beginning by allowing certain exporters to use part of their foreign exchange earnings to import a number of specified consumer goods, and in 1993 the "baggage allowance " for Indians reurning from abroad was relaxed. It has also announced that tariffs (present maximum 85 percent) will be reduced to about 20 or 30 percent in two to three years. (Ihis was originally stated as the target for the maximum tariff, but since then 8  there appears to have been some backslding in that the latest _ refer to this as the target for average tariffs). The reduction of manufctrig protection should be accompanied by whatever devaluations of the real exchange rate are needed to keep the trade deficit under control. It would be a major milstake - and In any case Impractcal to attempt to offset the high protection of manufacturing with high protection for agricultur Agricultural tariffs should be zero or at least kept down to a maximum of say, 10%. But for the above process to improve relative incentives for agriculture, it will be essentlal to allow the devalued exchange rate to feed through to the prices of agricultural commodities. The best way for this to occur naturally and smoothly is to remove the quantitative controls on agricultural imports and exports and marketing board and other ineventions which affect domestic prices, so that domestic prices are linked directly to world prices. The removal of agricultural QRs and of prohibitive import tariffs wIll involve: - Abolishing the present export controls, including canalisation (i.e parastatal export monopolies), export licensing and quotas, minimum export prices etc. This could be done more or less immediately without much difficulty. A clean sweep should be made of all these export restrictions, with the onus on those who want to phase some of them out more gradually to make a convincing case. A case for some form of more permanent special treatment (export taxes ?) could perhaps be made for common rice, sugar, tea and jute. There may also be a case for partial insuation of the domestic prices of wheat, rice and sugar from the larger fluctuations in international prices, at least as a transitional measure. These last two points are discussed below. - Abolishing import canalisation (i.e., parastatal import monopolies) and other quantitative import controls and setting zero or low import duties. These reforms will have to he accompanied by far reaching reforms of the activities of the parastatal organisations and marketing boards (e.g. FCI, the State Trading Corporation, the Cotton Corporation, the Rubber Board etc) and of domestic reguatory policies, and are likely to be difficult and politically highly sensitive. A major task in designing a feasible reform strategy will be deciding how, in what order, and over what time period this should be done. In liberalising import and export controls and reducing tariffs, the considerable potential for trade with India's neighbours should not be forgotten. Recently India has susrted importing short and medium staple cotton from Pakistan. Provided political problems can be overcome, there is a very considerable potential for greatly expanded trade in agricultural products with all the neighbours, especially Pakistan, Bangladesh and Sri Lanka. (O) Creating more neutral incentives within agricultre This involves: - Removing export controls which depress domestic prices below export prices. Particularly important: the controls on the export of cotton, wheat and common rice. - Removing import controls and setting zero or low tariffs on highly protected commodities. Particularly important: edible oils and oilseeds, sugar and rubber. -Removing regulatory and other domestic controls which would otherwise impede the transmission of world price signals to firmers. 9  These reforms should apply to the commodities in their pr zessed and unprocessed forms. nernational trade is often maiy in the processed commoditiei, e.g., edible oils and oilmeals versus oilseeds; sugar and molasses versus sugarcane. When both the primary commodity and the processed versions are Internationally traded, liberalising one wil require liberalising the others. For example, the removal of export controls from cotton wiUl need to be accompanied by the removal of export quotas from cotton yarn (except the country-specific quotas required by the MFA). (l) Short term tactics Both the above objectives L.e, removing anti-agriculture bias and creating more neutral incentives within agriculture, can be achieved by remsving impediments to Lnports and exports and thus linking domestic prices to world prices. But in view of the long past history of controls on international trade and the need to take account of the likely reactions of politically powerful groups and to avoid exposing low income groups to sudden adverse price changes, in some cases this may have to be done gradually in order to ensure a smooth transition. As regrds exports, the prevailing philosophy has been to treat the export markets of most agricultural commodities as 'residual' markets, i.e., exports are only allowed if the country has a surplus after meeting domestic needs, even if domestic prices are much lower than export prices and the exports would be both privately and economically profitable. A classic example of this philosophy in action is cotton, which has been subjected to export quotas and minimum export prices that are haltingly announced against a background of intense lobbying and bargaining between the textdle industry, the handloom industry, and the cotton growers and traders. This results in the sporadic appearance of Indian cotton in expor markets, basicaUy in years when there is an overall surplus in the domestic market. The uncerinty which this has created for both Indian exporters and foreign importers has discouraged them from investing in long term marketing facilities and relationships and has contributed to Indian cotton varieties being exported at substantial discounts from equivalent varieties exported from other countries. Furthermore, by periodically depressing the domestic prices of the various varieties below world prices in arbitrary ways, it has contributed to an unpredictable and inefficient structure of effective protection for cotton yarn production and has led to the wasteful use of high quality long staple cottons in the production of low and medium count yarns for sale in the domestic market. The philosophy ilustrated by the treatment of cotton exports clearly has to change if Indian agriculture is to be integrated into global markets. Policies should cease disciminating against exportables such as coucn so that India can emerge as a regular and reliable exporter of those commodities in which it has a comparative advantage. For most agricultural products it should be possible to immediately remove all the remaining export controls, including minimum export prices.u A mumber of these are in any case basically importable with domestic prices exceeding world prices (e.g. oilseeds and edible oils) so that the abolidon of export controls would have little overall impact on these industries or on consumers. Nevertheless it is possible that the controls may prevent exports in some circumstances e.g. temporary surpluses in areas where transport costs and delivery times would make it more profitable to supply neighbouring countries rather than more distant parts of India. In the case of pulses, export controls are prevening exports of particar varieties of processed and unprocessed pulses for which there is a substal foreign demand (particularly among Indian communities in the Middle East and elsewhere) even though the substantal unrestricted Indian imports of different types of pulses from Turkey, Australia and other countries can assure adequate supplies to the domestic market. However for the reasons given above, it may be better to move more slowly with some especially "sensidvew commodities. We have in mind in particular rice, wheat, cotton, and sugar, 10  and there may be others. For rice, wheat and cotton we would suggest a 'selective variety approach" in which, as a transitional measure, only the export of particular higher valued varieties which are not generally consumed by lower income consumers would be allowed initially. In the case of rice, this has already been implemented for some time by allowing exports of basmati rice, and in early 1993 quota restrictions were removed from exports of superfine long grain rice. This could be extended to fine varieties. In the case of wheat, the same principal could be applied by initially allowing exports of the durum varieties. In the case of cotton, export controls initially could be removed from longer staple cottons, say beyond a staple length of 30mm. In order to moderate the upward pull that would be exerted on the prices of the lower quality varieties of these products, or at least to provide a ceiling, imports of lower quality varieties could be allowed. In the case of sugar, controls could initially be removed from exports of sugar in raw form (brown sugar), for which there is so far little domestic demand, and whose production should be encouraged. We recognise the administrative problems that this 'selective variety approach' may pose, and therefore suggest it only as a short term transition towards fully opening up the exports of these commodities. In addition to removing explicit export controls, other regulations and controls should be investigated and if necessary changed if it turns out that they are impeding exports. There is probably considerable potential, in particular, for greatly expanded exports of unprocessed and processed fruit and vegetables, cut flowers, and fresh and processed seafood. The traditional export industries subject to commodity board interventions- tobacco, tea, coffee, spices (particularly chillies and black pepper)- should also be looked at from this perspective. As an example, the Coffee Board keeps domestic prices below export prices by a mechanism involving compulsory purchase of all coffee production and the allocation of the acquired coffee to separate auctions for domestic and export sales. In 1993 the system was liberalised to some extent by allowing up to 30 percent of plantation production to be sold directly in the internal market. It is difficult to see the rationale for these interventions, since India, with a world market share of about 2 to 3 percent, has no significant power to influence world prices, and the international coffee agreement which set export quotas, has been inoperative since 1989. Furthermore, coffee is hardly an imrportant item of consumption by low income Indian consumers.12 If quantitative export controls are ever needed again in the future as a result of new international agreements, the most straightforward way to implement them would be to auction export licences corresponding to the agreed quantity of exports. In some cases controls on the export of the commodity in its primary or raw form are imposed in order to indirectly subsidise its export in processed/packaged form. In order to avoid an excessively abrupt removal of such subsidies, in some cases the export controls could be temporarily replaced by an export tax. But the aim should be to remove these export taxes over a reasonably short period, so that the form in which the products are exported are not artificially distorted.'3 In other cases the export controls may have been introduced in the hope that the net return to the industry could be increased by restricting Indian supply and raising export prices. The potential for this kind of benefit is often vastly overestimated, in particular by failure to predict the speed and extent of the development of alternative sources of supply in the world market, and similar underestimates of the extent of substitution away from the product by consumers. Allowance is also seldom made for the fact that the overvaluation of the exchange rate inherent in the general system of import controls and tariffs is already restricting export supply even when there are no explicit export restrictions or taxes. Finally, even if there are benefits from export restriction over and above the restriction due to the overvalued exchange rate, account has to be taken of the economic costs of the administration of the restrictions and the rent seeking which is likely to accompany them, especially if they consist of export licensing or some other discretionary controls, rather than export 11  taxes. For these r-asons, we believe that it will seldom be possible to make a convincing case for the terms-of-trade argument for export taxes or restrictions: a few possible exceptions (rice, tea, sugar and jute) for which some kind of special treatment such as export taxes might be justified, are discussed below. On the imports front, the extent of the Rupee devaluation over the past three years has meant that at present, if they were imported, the landed costs of most major commodities would considerably exceed domestic prices. This provides an excellent opportunity for decanalising and otherwise removing non-tariff controls from imports of products such as wheat, rice, coarse cereals and cotton, without disturbing domestic markets to any significant extent. The present extent of the devaluation would also facilitate the removal of quantitative controls from imports of rubber and sugar, which on average have been highly protected in the past. The general principle would be to replace the physical controls with tariffs. In most cases these should be zero from the beginning, but in others they could start at higher levels and could be reduced over time, preferably according to a pre-announced timetable. If need be, particularly if there are signs of deliberate dumping by some supplier for the very short run (as may happen in the case of sugar), temporary anti-dumping duties could be introduced to safeguard the interests of the domestic producers. Care needs to be taken, however, that anti-dumping measures do not simply become a means for reintroducing arbitrary, ad hoc and lobby- prone protection by the back door"4. As a transitional measure, to alleviate concerns about excessive fluctuations in world prices, consideration could also be given to variable tariff schemes which link domestic prices to a moving average of world prices (see discussion below). In freeing up imports, there is need for special caution in dealing with edible oils. Most of these have protection levels equivalent to two or three times world prices, and even more in the case of copra and coconut oil.1' Suddenly opening up edible oil imports over low tariffs would seriously damage the large investments that have gone into this sector over the past three or four years as a consequence of the 'success' of the Tecrhnology Mission on Oilseeds." Nevertheless, the message should be conveyed in no uncertain terms to the edible oil industry as also to the oilseed producers that given the level of technology and of prices in world markets, it is not advantageous for India to increase its production through price hikes culminating in area shifts. Edible oil imports should be decanalised and subject to whatever is the current maximum import duty rate (presently 85 percent) in the first year, to be reduced to say 60, 40 , 20 and 10 or zero percent in subsequent years. As a transitional measure, it has also been suggested that India could arrange a short term bilateral contract (for say three years) with a country like Malaysia for the import of palm oil against exports of rice. The main argument for this is that it could help sell the initial stages of the required reduction in oilseed protection, by providing an easily visible counterbalancing benefit. But there are many dangers in this type of strategy, and we believe the potential costs outweigh the benefits. If a process of non -discriminatory import liberalisation is at all feasible, it would be much preferable even if the rate of tariff reduction is relatively slow.'7 Xj3 Special treatment of commodities in which the potential Indian supplv or demand may aDnreciably affect world orices in the long run. While sudden changes in Indian imports (e.g., in STC's imports of edible oils) or exports may cause temporary blips in world prices in the short run, the only major agricultural commodities in which India probably has some significant long run market power are rice, sugar, tea and jute. Some kind of special treatment (e.g., export taxes or import duties) may be justified in these cases, bearing in mind that - as noted above- exports are already taxed by the overvaluation of the exchange 12  rate resulting from import protection, mainly to nu . Although India and Bangladesh dominate world production of Jute, their individual and combined market power is probably limited owing to substitute synthetics. The main efficiency gains would be from opening up trade in both raw jute and jute products across the border with Bangladesh. High quality basmati rice is exported, but some limited exports of common (non basmati) rice began to be allowed for the first time only in 1991. In 1993 larger exports of superfine rice were permitted subject to a minimum export price. Ihese liberalisation measures should be continued and expanded. Doing so will not fully raise domestic prices to the present level of world prices, since world prices will be depressed by Indian exports on any substantial scale. Furthermore, export markets demand high quality rice with low broken percentages which substitutes imperfectly for the predominant poorer quality rice (mainly about 20% broken) sold in domestic markets. The optimum export tax and the optimum level of exports wil depend on this price depressing effect (i.e., the elasticity of excess demand), the extent of the exchange rate overvaluation, and the level of the non traded subsidies (irrigation, electricity and credit) to rice growing. Substantial export earnings (say $1 billion for 3 miUlion tons?) should be easily achievable in a short time with a moderate increase in domestic prices (say 10% to 20% ?). Apart from these major commodities, Indian exports constitute a fairly large share of the world market In the case of cardoman, black pepper and processed cashew nuts. For all these, the supply from other countries is likely to be quite elastic. Unless a convincing case can be made, we would not recommend any measures such as export taxes that would restrict supply more than the restriction already brought about by the overvaluation of the exchange rate. f, Dealing with unstable world prices and fluctuations in domestic production Until now, the insulation of the domestic markets from world market conditions and the government's buffer stocidng policies in the case of foodgrains, have meant that the domestic prices of most agricultural commodities have been considerably more stable than international prices. An exceptions is cotton, for which during the 1980s domestic prices have actually been less stable in important respects than world prices, in large measure because of the erratic application of export controls. Other exceptions are pulses and wool, whose domestic prices have moved broadly in line with import prices after the freeing of imports in the early 1980s. Freeing imports and exports of wheat and rice and the other commodities from canalisation, licensing and other controls will mean that domestic prices will move up and down with international prices. At the same time, the monsoons and other conditions wiUl contnue to cause fluctuations in domestic production. Is this a problem, and If so, how should it be dealt with? In the long run, there are strong arguments for allowing domestic prices to be directly linked to international prices. Most importantly, this means that production and consumption decisions wiUl constandy take account of Ildia's comparative advantage without the lags and other disturbances in the price signals which result from trade-intervening measures. Provided households below the poverty line can be protected from fluctuing prices by an efficient targeted system of food subsidies (see later discussion) the remaining consumers (say 70% of households) should be able to adjust without difficulty to the somewhat greater movements in food prices knat would occur. Secondly, as regards fluctuations in domestic production, India's present self-sufficiency in grain despite the strong anti-agriculural bias of the bintve system suggests that a more neutral structure of incentives could lead India to become a permanent, relatively large grain exporter. Despite continuing growth of the very large population, land scarcity and environmental problems, this also 13  seems plausible because of the very large productivity differences between grain production in the north-west and elsewhere. In such a scenDrio, in which there would be substantial exports of wheat but in which rice exports would be taxed in order to take account of the narrownes of the world rice market, domestic prices will be determined by export prices (minus the export tax in the case of rice) and all or most of the impact of poor monsoons would automatically be absorbed by declining exports. If imports are needed, the ceiling for domestic prices would be cif prices plus transport and other costs to the point at which imported and domestic grains compete. As an indication of the scope that this would allow for weather related price flucations in wheat and rice, between 1985 and 1987 the estimated cif-fob margin (as a percentage of the cif price) averaged about 5% in the case of rice and 17% in the case of wheat. Domestic port, transport and markeedng costs videned these margins considerably, however. For example, for rice, the average estimated pre-fob price at Calcutta (i.e. the price before loading onto ships) was about 20% lower than the estimated landed cost of imported rice. The price difference in the Punjab (the main surplus area for rice) was about 25%. The corresponding differences for wheat were 36% a- the port (assumed to be Bombay) and 43% in Punjab. 1' Although allowing domestic wheat and rice prices to fluctuate with world prices as described above should be the long term objective, we recognise that it may not be desirable or politically feasible to move to such a system in the short run, especially in view of the deficiencies of the present safety net for poor households and the time probably required to reform it. In that event the governmeat could consider a system of variable import and export taxes and subsidies which would be based on a moving average of past inerational prices. If properly implemented such a system can smooth out the effects of short and medium term fluctuations in world prices on domestic prices while ensuring that they move with world prices in the longer run. At the same time no discretionary, quantitative controls on trade are required and imports and exports can be made by private traders without restriction, subject only to the payment of the current variable export or import tax, or receipt of the current variable import or export subsidy. Ihere would be no need for a government buffer stock: inventories would be held by private agents, including private firms in other countries with an interest in supplying or buying from India. Based on their assessment of the probability of a bad monsoon, it would also pay private traders to hold inventories over from one harvest season to the next, to take advantage of any potential movement of domestic prices upward from the fob levels towards or to cif levels. If the monsoon in fact turns out to be poor, like a government buffer stock this wiUl reduce the required volume of imports. There is a lot of experience with moving-average price band schemes in other countries which should be studied carefully.'9 From our reading of this experience, our preliminary suggestions for India are that: -The government should be wiling to commit itself to pay out import subsidies when current international prices go above the calculated moving average target price. Failure to do this and the consequent failure to protect consumers against high international prices has been a major deficiency of other schemes. A symmetric scheme which both taxes and subsidises imports and exports is similar to a stabilisation fimd which could hedge its risk in intnonal commodity ftres markets.3 - The period covered by the moving average should not be too long. Say weekly avere prices for the previous four or so years? 14  -The prices used in the moving average formula should preferably be taken from some international souwce (e.g., Chicago Board of Trade prices for wheat), not from records of Indian import or export prices which the firms affected by the application of the formula can influence. The moving average price would then be adjusted in a transparent way for transport and other costs to give the desired price band; for example the difference between L-e calculated moving average cif import price and the calculated moving average fob export price. The formula is used to calculate the percentage difference between the moving average price and the transport-cost adjusted current price, also calculated from the price at the international market center. If the current price exceeds the upper level of the band, this is announced as the a valorem export tax rate for exports from India for the coming period (say a week?) and as the ad valorem import subsidy rate for the same period. If the adjusted current price lies within the band, there are no export taxes or subsidies. If the adjusted current price is less than the lower level of the band, the percentage difference becomes the ad valorem export subsidy and import duty respectively. - The formula and the data on which it based are made public so that participants in the trade can accurately predict the duty or subsidy rate which will be applied to their imports or exports in the near term and hedge their risks in futures markets when planning further ahead. For obvious reasons, it will also be vital to stick to a formula once it is agreed and to resist pressures to make ad hoc changes. -Moving average schemes in other countries have usually been applied to imported commodities only and have included a band around the moving average (say plus or minus 10 percent) within which import prices are allowed to vary before attracting the special import tax. (As noted above, probably reflecting the greater bargaining power of farm lobbies compared to the lobbying power of consumers, import subsidies have not been paid in practice). In India, there is a concern that the potential fluctuation of domestic prices between fob and cif limits is already excessive (owing to unpredictable monsoons and both rice and wheat shifting from being exportables to being importable from year to year). For this reason, such a scheme could start without any additional scope for price variation other than the fob/cif gap. Margins for further variation could be introduced if the commodities become firmly established as exportables or as importable despite substantial variations in weather conditions. Alternatively, if the potential for domestic price fluctuations resulting from the fob/cif gap is considered to be excessive (as for wheat?) the allowable price band could be narrowed. Given the objective of reducing the exposure of Inaian producers and consumers to fluctuations in world prices, moving- average price band schemes have important advantages over government buffer stock schemes which involve the maintenance of quantitative controls over imports and exports, and most likely the maintenance of controls over private traders to prevent them from accumulating or running down stocks and offsetting the activities of the buffer stock organisation. But like buffer stock arrangements, the schemes also introduce distortions of various kinds, both directly into the markets of the products to which the schemes apply and indirectly into markets of substitutes. For example, an export tax on wheat required by the scheme's formula may prevent or limit wheat exports even though domestic production costs may be lower than world prices Again, if wheat exports are taxed under the scheme but coarse grains are not covered by it, coarse grain might be exported even though the economic return from exporting wheat would be higher. There are also potential complications with downstream products. For example, if wheat 15  imports are taxed but no such special import tax is applied to flour or flour products, domestic producers of the latter are likely to be squeezed. Effects of these kinds are also likely to lead to pressures which have the potential to greatly complicate and undermine the transparency of the schemes and to generate a lot of lobbying. For these reasons there is much to be said for treating the schemes as transitional devices and phasing them out over some defined period. This could be done, for example, by pre-announced increases in the price bands to the point where intervention would only occur in the event of extreme peaks or troughs in world commodity prices of the order of magnitude of those experienced ac the time of the first oil- price shock. I[5 Removing imnort controls. high tariffs and domestic regulatory controls from tradeable inputs used in agriculture The most important things to do here concern fertilisers. Except for some recently deregulated phosphatic and potassic fertilisers, the industry is controlled by a combination of an import monopoly by MMTC, a fixed single nationwide wholesale price for each fertiliser, a cost-plus pricing system for each individual fertiliser plant, and large subsidies provided by the central government. The pervasive distortions on both the production and distribution sides of the fertiliser industry are well known and well documented. Our analysis of the subsidies, based on comparisons of domestic and international prices during the 1980s, revealed that roughly half of the budgeted subsidy supported high cost fertiliser producers (Gulati, 1990; Gulati and Kalra, 1992). Consequently the entire burden of the removal of fertiliser subsidies would not fall on farmers: a significant part would require the rationalisation of fertiliser production. The key element in price control is the Retention Price Scheme (RPS), under which a normative cost- plus pricing formula is applied to each individual plant. Reflecting differing feedstock prices, technologies, locations and operating efficiencies, there are large differences in retention prices as between fertiliser plants. The situation is further distorted by the fact that there is no systematic link between the controlled prices of the various feedstocks (naphtha, fuel oil, natural gas and coal) and their international prices. Given these distortions and the absence of import competition, the location of fertiliser plants is also suboptimal. On top of all this, the Joint Parliamentary Committee on Fertiliser Pricing (GOI, 1992) noted that the cost figures supplied by various fertiliser factories were seldom cross checked by the Fertiliser Industry Coordination Committee (FICC), which administers the RPS. This indicates the existence of considerable scope for exaggerating the normative cost both directly and by understating capacity. In July 1991 a start was made on moving away from the pervasive regulation of the industry, by increasing controlled wholesale selling prices by 40 percent, decontrolling the prices of low analysis nitrogenous fertilisers, and setting a ceiling on the subsidy for Single Super Phosphate (SSP). This was a sensible attempt to begin ieversing a trend of increasing subsidies under which nominal farmer prices had been virtually unchanged for ten years and the total budgetary subsidy had grown to more than half of central government spending on agriculture and to about one percent of GDP. 5ubsequent policy changes have been halting and contradictory, however, and in some important ways have worsened resource allocation. They are recounted below in order to illustrate the importance of obtaining agreement on clear objectives for policy reform, especially when politically powerfil groups are significantly affected. In summary: August 1991 . Previous increase in controlled fertiliser prices of 40 percent rolled back to 30 percent , and small and marginal farmers exempted from the increase altogether. March 1992. Rock phosphate and sulphur imports (inpats for phosphate fertiliser plants) decanalised (i.e.the import monopoly of MMTC was removed and private imports allowed). 16  Auggst 142. Controlled prices of urea (accoundng for about half of all fertlliser sales) reduced by a further 10 percent. Price controls reintroduced on low analysis nitrogenous frtlisers. Prices of phosphatic and potassic fertlisers decontrolled. September 199-. Farmer subsidy of Rs 1000/ton introduced for di-ammonium phosphate (DAP) and MOP (muriate of potash): equivalent to price reduction of about 13 percent. DAP imports decanalised i.e. MMT C import monopoly removed and private imports allowed. Controlled naphtha and fuel oil feedstock prices to fertliser plants increased. Jue 122a. DAP subsidy of Rs 1000/ton contiued for domestically produced DAP but discontinued for imported DAP. New subsidy depending on phosphate content introduced for domestically produced complex phosphatic fertilisers but not for imports of the same ferdlisers. The effect of these changes was to keep the controlled farmer prices of nitrogenous fertilisers (of which by far the most important is urea) well below border prices while contimiing to pay large subsidies on imported fertilisers and to the nitrogenous fertliser manufacturers equivalent to the difference between these prices and their production costs. By contrast, when the prices but not the imports of phosphatic and potassic fertlisers were initially decontrolled, their prices suddenly doubled or more than doubled, and in the case of DAP they went from about 20 percent below international prices to about 50 percent above. Ihe freeing of DAP inmports then brought DAP prices down to about the cif level plus port and domestic transport and marketing costs, but this was reported to have caused the closure of 8 out of 11 DAP m a plants. Attempts were made to respond to the resulting political pressures, first by a general fixed subsidy paid on both imperted and domestic fertilsers which aimed to reduce the prices paid by famers, but subsequently by limiting the subsidies to domestic fertlisers with the aim of protecting the local manufcturers and enabling them to reopen. Apart from generating a great deal of uncertnty for all market participants, especially farmers, the net effect of these changes was to send a signal to farmers to increase urea consumption but to cut down on the consumption of DAP and MOP. Already the relative consumption ratios were not favourab!e from an agronomic point of view, and these price signals made the situation even worse. As expected, during the rabi season of the 1992-93 crop, while the sales of DAP dropped by about 30 per cent, and of MOP by about 50 per cent (compared to the previous rabi crop), urea sales went up by more than 20 per cent and in some areas it was being sold at a premium. Halting and contradictory atempts of this kind to liberaise the production and distribution of fertilisers should be avoided. As an example of a possible approach, in the 1993-94 season, the controlled farmer price of urea could be raised by about 15 to 20 per cent, while the subsidy on DAP and MOP could be reduced to, say, Rs 600 per tonne. Ihe subsidy on nitrogenous fertlisers (urea) could be limited to a maximum of Ra 1000 per tonne. Taking account of projected future world urea prices, the government could then announce a package which would include increases of future urea prices for farmers, limits on the per ton subsidy to industry, and the liberalisation of ferdliser imports. Changes such as these would help with dismantling the retention price scheme in due course and with the other far reaching reforms which are needed. In summary, the longer term reform objectives in the fertliser sector should be to: -Decanalise ferdliser imports, i.e., remove the MMTC import monopoly and allow private imports; 17  -Set low to moderate tariffs (say 10 % or 15% ?); -Abolish the retention price scheme for fertUiser plants; -Abolish the fixed subsidised domestic fertiliser price to farmers and the uniform naiionwide pricing which goes with it; -Abolish the fertiliser subsidy as regards both farmers and producers; -Adjust feedstock prices for fertiiser plants to reflect their opportunity costs. India probably has little long run market power in nitrogeneous ferilisers, but conceivably could have some in phosphatic fertilisers. Special treatment of some kind (an above average import tax?) might be justified for the latter. Imports of seeds, pesticides, farm machinery, plastic piping and other farm Inputs are subject to import licensing &/c; high tariffs (for example, tractor imports are effectively blocked by a tariff of 80 %). A beginning on freeing up these controls was made in April 1993 by allowing agro-industries which ernort 50 percent or more of their output ("Export Oriented Units") to import their inputs duty free, and to import capital equipment at concessional import duty rates. This initiative should be extended and all agricultural and agro-industrial inputs and to agricultural machinery, which should be freed from non-tariff controls and subject to low to moderate import tariffs. The prices of the domestic producers of these products in many cases are well below duty - inclusive import prices, so the overall cost of these inputs to farmers would probably not be reduced by much. Nevertheless, such a reform will make all kinds of technologies available which at present are not found in India, and will shake up the domestic manufacturers.2' In 1991, tractors , combine harvesters and rice transplanters were included in the list of products for which there is now automatic approval of foreign technology agreements and of foreign equity of up to 51% .However most agricultural implements and other farm inputs such as plastic piping and sheeting are reserved for production by small scale firms'. By preventing small firms from growing and larger firms from competing, small scale industry (SSI) reservation adversely affects the quality, technological level and marketing of these inputs. SSI reservation creates similar problems for the efficiency of agricultural processing industries such as rice milling, cotton ginning and oilseed crushing. It also prevents direct investment by foreign firms in the production of the reserved products. While SSI reservation is a general problem affecting the whole manufacturing sector in India, it has a particularly marked negative impact on the efficiency of farming and of the agro processing industries. For this reason , special attention should be paid to removing agricultural and agro-processing inputs from the small scale industry reservation lists. At the same time, other regulatory impediments (e.g. excessive red tape and delays in obtaining enviromnental clearances, registering land etc) to competition and to direct investment by foreign firms in the agricultural input industries, should also be removed. The overall combined impact of these changes will be to increase the average cost of tradeable inputs to farmers, because the freeing up of fertiliser prices and the phasing out of fertiliser subsidies will dominate reductions in the prices of the other inputs. On the other hand better quality inputs embodying later technologies and more varieties of inputs will become available, and also better and more efficient distribution, provided inventory and other controls on private traders are 18  eliminated. As with the removal of irrigation, electricity and credit subsidies, to help defuse farmer opposition, the fertiliser reforms should be accompanied by the actions discussed previously to remove the discrimination against agriculture on the output side. IL6 Removing large subsidies on non traded inputs: canal irrigation, electricity and credit The serious distortions resulting from subsidised charges for these services are well known. Reform involves major and far reaching changes in the organisation and operations of the irrigation commands, state electricity boards, and the banking system. As with fertilisers, instituting prices for these services which reflect opportunity costs will be more acceptable if it is seen by farmers to be accompanied by measures which increase relative agricultural incentives on the output side, as also increased efficiency in the delivery of these inputs. But it will still be politically difficult in regions or for crops (e.g., rice in some parts of the north wvst) in which the size of the input price increases may outweigh any feasible or desirable increases in output prices. Likewise, it will be politically difficult for those portions of highly protected crops which are irrigated (notably oilseeds and sugarca'ie in most years) which should face declining outpu: prices despite the general increase in the price level of all or most other crops. Substitution into the more profitable crops (to the extent that it is possible) will frequently not offset the loss of the large economic rents inherent in the present system of input subsidies. As regards the subsidy on canal waters, the situation is extremely serious. But since irrigation is a state subject, and the form of subsidy is somewhat different, it does not create much 'noise' in the central budget or in the corridors of the Ministry of Finance, as does the fertiliser subsidy. The nation has spent more than Rs 600 billion ($US 36 billion) at 1988-89 prices on canal networks during the last forty years, adding an irrigation potential of more than 22 mitlion hectares (Gulati, 1993). Today, the direct recovery from farmers towards the cost of canal waters is only a small fraction of operational and maintenance expenses, not to speak of capital costs. The low cost recovery (in most states the rates have not been revised during the last 10 years or so) is starving state exchequers and irrigation departments. As a result, minimum essential repairs remain neglected and in many cases the continued existence of the systems is at stake. In 1972 the Irrigation Commission recommended that the price of canal waters should account for about 5 per cent of the gross revenue of farmers in the case of foodgrains while for cash crops it should be near 12 per cent. The present reality, however, is that water charges probably average only around one percent of the gross revenue of irrigated farms.' Recently, an Expert Committee on Pricing of Irrigation Waters (GOI, 1992)24 examined the financial position of the irrigation sector in gri-at detail, and recommended an increase of more than six times in water charges collections (from existing levels of Rs 50 per hectare to Rs 310 per hectare) through a two part tariff structure. According to the Committee's calculations,this would amount to about 6 per cent of the gross revenue of an average farm .It would, however, ;over operation and maintenance (O&M) costs and 1 per cent of capital costs (calculated at historical prices without taking care of the gestation lag factor). In addition, the Committee recommended that the irrigation commands should limit themselves to wholesale distribution of water with volumetric pricing to farmer groups who would be responsible for the subsequent water distribution and management of the system over areas of up to about 500 hectares. While we strongly agree that such reforms -if implementable -would constitute major improvements on the present situation, we wonder why the Committee has suggested that only one per cent of capital costs should be recovered. That might be a strategic compromise, but we feel that basic principles should not be 19  relegated to the background, and that fall recovery of the relevant capital costs ( after allowing for urban beneficiaries and public good externalities aue as flood control etc) should be the long run target. Secondly. we suggest that experiments sbould begin to make project authorities/irrigation departments financially and operationally autonomous. This could be combined with initiatives to make the farmers ca-owners of the irrigation systems by issuing 'water bonds' to the tune of -say - five per cent of tae equity of the system. This could be made somewhat compulsory in the sense that water would le supplied on a priority basis (or only) to those who 'own' the system through equity participation. This would help recover some part of the capital cost, and also contribute to farmers feeling that the irrigation system which supplies them also in part belongs to them, thereby inducing them to take greater interest in its mnanagement. The institution of arrangements under which farmers pay for the cost of canal water delivered to them will reduce the wasteful use of water, contribute to better water allocation within irrigation commands, provide funds for improved operation and maintenance 2, and somewhat reduce the rent seeking activity generated by the present administrative methods by which water is allocated in most systems. However, on its own, this reform cannot efficiently allocate water between different users, and mechanisms need to be found by which the supply of water to farmers and other users is responsive to the value it has to them. By far the most promising method for achieving this aim would be to create conditions which would allow the existence of efficient markets in tradeable water rights A i avincing case for tradeable water rights is made, and the extensive literature on the subject surveyed,in a recent paper by Rosegrant and Binswanger (1993). If rights to the delivery of water can be freely bought and sold, farmers with new crops or in new areas will be able to obtain water provided they are willing to pay more than its value to existing users, and established users will take account of its sale value in deciding on what and how much to produce. In this way there is great potential for mitigating some of the pervasive problems of Indian irrigation commands, for example the 'tail ender' problem where farmers at the top ends of the canal systems obtain ample water to cultivate water intensive crops such as rice and sugar cane, while down-canal farmers are starved of water even though its marginal value to them may greatly exceed its value to the up-canal users. Ideally, the creation of a market for water should be accompanied by reforms which charge users for the marginal cost of delivery i.e marginal operation and maintenance costs for deliveries within established networks, and marginal operation and delivery costs plus incremental capital costs if new investment is needed. In this way farmers (and non-farm users) will be obliged to take these costs into account in making their trades. However, the institution of tradeable water rights will lead to a very substantially improved - if not fully optimal - allocation of water, even In the absence of proper recovery of marginal delivery costs. In India, as e'sewhere, it may be politically extremely difficult, or even impossible, to fully recover these costs, since farmers strenuously resist increasese in water charges, which amounts to the expropriation of economic rents built into land values. While we believe that strong and continuing attempts should nevertheless be made, at the same time the government should push the reforms needed to establish water markets e.g. the conditions and institutions required for contract monitoring and enforcement, reliable water delivery and measurement, mechanisms for internalising or otherwise taking account of the interests of third parties etce The electricity subsidy to the rural sector has already crossed Rs 40 billion ( $US 1.3 billion ) per annum. The pricing of electricity for rural areas is one of the major reasons that most state electricity boards (SEBs) are in the red, and it is becoming increasingly difficult to sustain this financial burden. With the rapid energisation of indian agriculture, coupled with the existence of a flat rate tariff for electric purnpsets in most states, this subsidy increased especially rapidly during the 1980s. The flat rate system means that the marginal cost of additional electricity use falls to almost zero, which provides an incentive to go for water heavy crops based on groundwater reserves 20  even in areas of low rainfall. The widespread emergence of paddy in the Punjab-Haryana belt is a case In point. While the annual rainfall of this region is about 60 cms, irrigated paddy requires more than 200 cms, although its consumptive use is less. This requirement for irrigation water is primarily being met through groundwater tube wells running on electricity. In a country where electricity is a very scarce resource, and its opportunity cost is not below Rs 2/kwh7, agriculture gets this resource at throw- away prices. In Pmjab, for example ,the average revenue from farmers is less than 7 paiselkwh against an average state wide cost of generation and distribution of more than 110 paise/kwh. The story is not very different in other states (Gulati and Katula, 1992). Tamil Nadu in fact supplies power to its farmers totally free. Populist measures of this kind are clearly incompatible with the efficient use of this scarce resource.8 We have three suggestions for reforms in the provision of power to the rural sector. First, the state electricity boards should be made more accountable to consumers as regards their costs of operation and generation. Their costs should be scrutinised by agencies which incorporate representatives of consumer groups, including farmers. The structure of electricity tariffs and tariff increases should be debated in public and the relevant cost and demand data should be made publicly available. This would help induce the SEBs to economise on their costs of generation and distribution. Second, the flat rate system for pumpsets should be replaced by volumetric pricing by installing meters. The early argument that the cost of installing meters and administering volumetric pricing would not justify the benefits is no longer valid (if it ever was) iz view of the manifold increase in the level of electricity consumption by Indian agriculture in recent years. Third, the distribution of electricity should be increasingly transferred to the private sector on an attractive commission basis, especially the distribution to agriculture. Some farmers' cooperatives might be particularly suited to this task. While the government has stated that electricity generation is now open to the private sector,' higher priority should be given to private sector participation in electricity distribution and in the collection of dues. The ability to subsidise farmers through rural credit is an ace card with the politicians. It is therefore not surprising that there has been pervasive political interference culminating in the extremely damaging loan waiver scheme which was a direct outcome of the 1989 elections. At present the farming sector is starved of funds, and the whole process of rural lending is in jeopardy. The annual subsidy to farmers through concessional rates of interest and bad debts is in excess of Rs 30 billion ($US 1 billion), the amount depending upon the definition used to measure it (Katula and Gulati, 1992). Although the required reforms of rural credit involve far reaching and difficult reforms of the whole banking and financial system3s, in present circumstances we feel that three reforms could be carried out without generating a great deal of opposition. First, the concessions on rates of interest for rural loans should be reduced and then abolished, while increasing the availability of credit. Secondly, efforts should be made to evolve group lending in rural areas, where members of the group act as sureties for each other. In case of default by any member, the entire group may be sued and banned from further loans. Thirdly, defaults with a two to three year bistory should be treated severely under present laws, perhaps through special tribunals. Unless financial reforms along these lines are initiated, the process of recyc!ing deposits, loans and recoveries will continue to be disrupted and an inflexible rural credit system is likely to slow down the response of agriculture to the kinds of trade policy, regulatory and other changes that are urgently needed. As emphasised later , the required reforms of iertiliser pricing and distribution, canal irrigation, electricity and rural credit are likely to be polid,ally extremely difficult. To increase the chance that they will be implemented, it is important th;* they are perceived to be part of a reform 21  package in which farmers gain from reforms on the output side which on average involve real price increas for agricultural products, while paying higher prices for these inputs. If the output and Input reforms are not tied together, and efforts to remove the input subsidies are delayed, the gains from the output price reforms are liely to be pocketed by the farmers with little or no action on the input side. I. Removi controls and distorions associated with the 'food securit' conmlex This is the source of the largest and most pervasive distortions and inefficiencies in domestic agricultural markets, including the two-price systems for major commoditdes both at procurement oevy prices etc.) and distribution, and the associated rent seeking. The Essenial Commodides Act is also an important source of distortions and rent seeking.'1 Apart from the powerful groups with vested interests in the system as it now funcdons, the main stumbling blocks to reform are the need to retain some way of permanently reaching low income and deprived groups and dealing promptly with drought and other emergencies, both local and nationwide. Various approaches are possible. The most far reaching would rely on food stamps: a second less radical reform would keep the fair price shops but drastically change other aspects of the present system. (i) Food Mns. The main elements of a reform based on food stamps would be: -Food stamps would be issued to low income households based on income/wealth criteria and used to buy from private retalers according to the type of food stamp system adopted. They could be administered by the states and reimbursed (according to a variety of formulas) by the central government. The food stamp system would replace the PDS, which would be abolished. -The central government would be responsible for policies (for example, vaiable export and import duties and subsidies related to a moving average of world prices, as discussed previously, or, failing that, a buffer stock system) aimed a! preventing excessive peaks and troughs in domestic prices. However, it would be important not to inhibit normal seasonal and regional price variations which reflect carrying, marketing and transport costs. Insofar as overall stabiisation continues to involve direct goveroment interventions, purchases and sales would only be made in major wholesale markets. Storage and transport of buffer/emergency stocks could be subcontracted out to the private sector. -The periodic imposition of physical and other controls on traders to prevent the movement of grain out of the surplus north west region would cease. -As at present, the Centre in combination with the states would intervene in regions affected by drought or other emergencies, by emergency work programs (including food-for-work prcSrams), sales in local wholesale markelt, etc. Some of FCI's storage facilities in drought prone areas could be retained for this purpose, although again these fumctions could also be subcontracted out to the private sector. -FCI and NAFED would get out of the business of physically handling gris and other prmary commodites. Taking delivery, storage and arranging shipment would all be done by the private sector. 22  -The levy price systems for rice, sugar and molasses would be abolished. -Lnports and exports of wheat, coarse grains, and edible oils would be freely made by the private sector without restriction. Exports and imports of rice and sugar would be freely made by the private sector, but subject to special tariff/export tax treatment as mentioned earlier. (ii) Reforming the PDS systm If the reform program were based (at least in its initial stages) on keeping the fair price shops and ration cards rather than on a food stamp system, it would still be possible to do this while keeping most of the reforms mentioned above. One approach would be to decentralise procurement for the fair price shops and for emergency stocks down to some regional level. That is, a regional organisation with storage capability (preferably subcontracted out to the private sector) would take bids for the delivery of the grains needed for a network of fair price shops for which it would be responsible. In order to support low income targeting the grains purchased would be at the low end of the quality spectrum. They could be purchased from anywhere (locally, from the surplus north west areas, imported), from anyone, and delivered to whatever time schedule corresponded best with the demand from the regional FPS network. The organisation would be reimbursed for the difference between its purchases and other costs and its receipts from the FPS network sales. Emergency stocks and expenses would be accounted for and subsidised separately. Incentives would have to be established for the management of the organisation to optimise its purchases in the light of the demand. In order to concentrate the benefits on low income households, it would be highly desirable to restrict the issue of ration cards on the basis of whatever information is available on income and wealth (as is now being done in seven states). This would have the further advantage of reducing leakage from the system back to the open market and perhaps to the regional procurement organisation. If this or any other version of the present PDS, edible oils and sugar should be removed from the system, which would handle only rice, wheat and coarse grains (and perhaps gur/khandasari instead of sugar). PDS wheat and rice should be at the low end of the quality spectrum in order introduce some measure of self targeting. Edible oils should be removed from the PDS for two reasons. First, PDS prices have consistently been maintained and remain at approximately double border prices: liberalising edible oil imports (see discussion below) will make much cheaper edible oil available to everyone, including low income consumers not reached by the PDS. Secondly, more than half of the edible oil allocated to the PDS is diverted (much of it in bulk to edible oil refineries well before reaching the fair price shops), and only about one fifth actually reaches consumers in the bottom 40 per cent of the income d&-tribution. Sugar should be removed for similar reasons. Firstly, it is a small share of the budgets of low income households. Secondly, removing it and abolishing the levy on sugar mills will bring down free market prices substantiaily, since the levy share of total sugar sales is normally high (currently 45%). Thirdly, on average domestic sugar prices have been well above world prices and can be expected to decline with trade liberalisation, even though sugar is one of the commodities which qualifies for special treatment of some kind owing to the narrowness of the world market in relation to ldian demand and supply. Fourthly, about a third of the sugar supplied to PDS is estimated to be diverted. Fifth, it has been suggested that in some regions gur and low quality 23  khandsari could be distributed through fair price shops in place of sugar, on the argument that this would automatically target the consunption to low income groups. Finally, the removal of sugar is likely to reduce the attractiveness of the PDS to middle and higher income households and thus facilitate targeting. HI Removine Other Domestic Regulatory Controls and Distortions: These can be roughly classified under four main headings, although there is considerable overlapping among th'. first three. (i) Renoving controls on markets, traders and processors and subsidies to cooperatives: These affect the markets for all agricultural products, although their application is often commodity specific. Thcy are the responsibility of the Central government, but there may be some additional state controls and subsidies. Reform would involve: -Abolishing the Essential Commodities Act; -Abolishing the general ban on futures trading; -Abolishing inventory controls; -Abolishing selective credit controls on inventory financing; -Removing the discrimination of Indian Railways in favor of shipments by parastatals; -Treating farmer cooperatives on an equal footing with the private sector, i.e., removing their preferential access to subsidised credit, their preferential tax treatment, their exemption from various regulatory rules applied to private firms, and direct subsidies. A general problem will be that the inventory and credit controls are perceived as technuques for preventing the private sector from offsetting inventory accumulation or decumulation by parastatals such as FCI, NAFED, the Cotto l Corporation of India, etc. (i) Abolishing state-implemented movement controls: These include: -The Maharashtra monopoly procurement scheme for cotton; - The isolation of Thanjavur district in Tamil Nadu for rice procurement; -Gujarat's periodic movement controls on groundnuts and groundnut oil. It is possible that other state-implemented movement controls exist. Because of the obvious local political sensitivity, removal of the controls would probably have to be accompanied by some offsetting benefits (e.g. would a substanial long term improvement in cotton prices resulting from open trade in cotton be sufficient to offset the perceived benefits of the Maharashtra cotton scheme?) (ill) Removing commodity-speciflc controls: There are large numbers of regulatory controls (which affect farming, marketing, distribution and processing) implemented by commodity boards and by central and state government departes, 24  which reduce economic efficiency in the industries to which they are applied. For example, the Tobacco Board attempts to set annual production quotas for each one of more than 10,000 individual growers of Virginia tobacco. Controls of this kind, which are clearly inefficient, largely unenlforceable or both, should be identified and abolished. At the same time the roles of each of the commodity boards or of other Intervening agencies should be assessed and specific reform progrms (which may involve their abolition or substantial changes in their functions) should be developed. By way of illustration, we make some suggestions below for removing domestic controls from from five major commodities -wheat, rice, sugar, cotton, and oilseeds . (a)As regards ykhn, informal movement restrictions have often been Imposed on the surplus states of Punjab, Haryana and on western Uttar Pradesh so that the govemmont's requirements for the PDS and for buffer stocks can be purchased at the official procurement price. In his February 1993 budget speech, the Finance Minister announced a welcome general policy change, under which there would be no further administrative restrictions on movements of agricultural products within the country. It should be made clear that the new policy wiUl also mean that the government will no longer pressure private traders to shun the primary grain markets, as was done in the marketing season of 1992.-2 The government should also announce and commit itself to support prices (covering say the bulk line Raid out costs' of the farmes), while procurement should be done at market prices In competition with the private trade in the open market. Farmers should have the right to sel to anyone offering better prices. This is important becuse informal controls on either the movement of wheat or on the participation of the private trade in the market, lead to aU sorts of corruption within each state and at state borders, and undercut the support of this politically Important farmer group for any general program of agricultural reforms. Furthermore, the uncertiny and transaction costs involved reduce the attractiveness of wheat production and contribute to farmers switching to other crops that do not face the same movement controls, such as oilseeds.'3 (b)ln the case of ni, rice millers in the three major surplus states are at present subject to a levy (i.e. compulsory acquisition at fixed prices ) of 75 per cent of their production in Haryana and Punjab and 50 per cent in Andhra Pradesh.' A visit to these rice mills easily reveals how they try to evade this levy, and how they succeed in avoiding the minimum quality conrol and supply the poorest quality rice to the procurement agencies, all at the cost of the exchequer. The economic rents in this system are largely approprated by the millers and by the inspectors and other employees of the procurement agencies. The levy system should be eliminated e.g. in Punjab and Haryana, by reducing the levy percentage from 75% to 50% in the first year, to 25% in the second year, and fully witdrawing it in the third year. As long as the PDS remains the chief means of providing food to low income groups ,to meet its procurement targets, the government should invite tenders from rice mis and procure from the lowest bidders. If the PDS is decentralized as suggested previously, these bids could be for delivery to the location served by the regional or state agency requirig the rice. At the same time ,as for wheat, the governt should provide a support price for paddy based on the bulk line paid out costs of the farmers. With the market determining the prices of milled rice,there would be an incentive for millers to upgrade their techology and to reduce the breakage ratio. At present, the signals are in fiat in the opposite direction. Many millers also install small inefficient hullers, which are exeempt from the levy, and in recent years the number of such hullers has increased in the northern belt. Furthermore, rice milling is reserved for the smaU scale sector, which deters larger modemnsing investments. It should be removed from this reserved list. 25  Theso measures would increase the prices the mills would be willing to pay for superior rice varieties, and thereby would increase the incentive of farmers to produce them. (c)In the case of ag, the established mills have to deliver a fixed proportion (currently 4S per cent) of their production to the government at a fixed levy price which generaly does not cover their full production costs. They are expected to compensate this from open market sales of sugar. But even In the open market, sales of sugar are regulated by a system of releases which allocates a centrally determined sales quota to each one of the approximately 386 sugar mills each month. In addition, they are subject to controlled minimum prices for their purchases of sugar cane and a variety of other controls, including compulsory crushing quotas and mandatory crushing periods duriog seasons of excess cane production. Furthermore molasses (the principal by-product of sugar milling) is under a 100 percent levy at prices which are generally less than a quarter of open market prices. By contrast with established mills,new sugar mills and expansions in the capacity of ecisting mills are exempt from the sugar levy for periods of from 5 to 10 years, depending on where they located. Together with the other controls, this artificial incentive to the establishment of new mills and new capacity is an important reason for persistent excess capacity and widespread "industrial sickness" (i.e. bankrupt and loss making mills) in this industrym Given the poor financial condition of many sugar mills3a, and the large scale evasion of the molasses levy, we suggest that molasses should be decorolled immediately 38. It is difficult to think of a justification for subsidising the consumers of alcohol at the expense of consumers of sugar. Next, the sugar levy should be removed. This could be done by reducing it in the first year from 45 percent to say 20 percent, and withdrawing it totally in the following year. For reasons given earlier, it would be better to remove sugar from the PDS. If the government wishes to continue subsidising sugar consumers through the PDS, it should be procured by competitive tendering from the sugar mills, as recommended for rice. Storage could also be arranged by competitive tendering in which the mills would doubtless participate in order to use the storage capacity they have built to store government owned sugar under the present systemY Elsewhere (see Bhide and Gulati, 1992), we have als argued for delicensing of the sugar industry. Among other things, this would help to circumvent the problems created by the fact that the states often set much higher minimnum prices for sugar cane than the minimum prices recommended by the Commission on Agricultural Costs and Prices and announced by the central government. (d)As regards , monopsonistic purchases by the Maharashtra State Cotton Markedng Federation (often referred to as monopoly procurement of cotton in Maharashtra) distorts the national cotton market. It often leads to "smuggling" of cotton between Maharashtra and adjoiing states whenever the prices in these states are higher or lower than the buying prices of the Maharashtra Federation. The Cotton Corporation of India already provides a nationwide set of floor prices for cotton as insurance against any drastic collapse of market prices, and it is difficult to see the rationale for the continued monopoly procurement operations of the Federation. A second urgently needed reform is the removal of sate govenmment controls over ginning margins, which is a serious impediment to the badly needed m saon of this indutry. (e) As regards edible oils and oilseeds , the suggestions made earlier for trade policy reforms, in particular removing STCY' iamport monopoly of edible oils and progressively loweing Import tariffs,would be incompatible with the price maitnce and buffer stocking scheme at present managed by the National Dairy Development Board. This scheme should be abolished. In July 1991 the processing of oilseeds was liberalised in a significant way by the fact that the vanaspati industry and the solvent extraction idustry were among the many industries freed from industrW 26  licensing. However two major reforms are still needed. Firstly, the detailed regulatory controls which the Ministry of Food and Civil Supplies applies to the vanaspati industry should be removed. These Include informal price controls, controls on the processes which can be used, controls on the kinds and quantities of crude oil inputs, and a complicated system of differential excise tax rebates aimed at encouraging the use of oils from ricebran and minor oilseeds. Secondly, the reservation of oilseed crushing for small scale industry should be abolished, as it creates an artificial barrier between activities which in other countries are predominantly carried on by integrated firms. It is also a deterrent to direct investment by foreign firms in the oilseed processing industry, which the government's general liberalisation of the foreign equity and technology rules (also in 1991) was intended to encourage. Especially for the above mentioned five commoditlfs, which account for more than half of India's gross cropped area and value of crop output, we feel that the time has come to allow and to promote futures trading. rutures trading had been bannad for many years on the argument that it encourages speculation, and during years of acute shortaje, exploits consumers. Conditions have dramatically changed since these arguments had some popular appeal. While some varieties of cotton have been opened up for futures trading lately, after a gap of more than 25 years, these reforms for cotton should be broadened and extended to other commodities. Futures markets have an important role to play in stabiising commodity markets. Their existence will be particularly important for domestic food industries such as oilseed processing to be internationally competitive, since they are critical for dealing with risk and uncertainty in the face of constantly fluctuating prices and fine margins between the various processing stages. (iv) Removing agriculture's exemption from income tax: Under the Indian constitution, income tax on agricultural incomes is a state subject, but only seven states actually levy such a tax, and the revenue from it is very low to negligible. For the purposes of the cental government income tax, agricultural income is supposed to be combined with non- agricultural income in determining marginal tax rates on non- agricultural income. However, as a result of the way this provision is worded the reported additional tax collections are negligible, and the provision has not effectively prevented large scale evasion of income taxes by individuals who arrange their affairs to show that most of their income is from agriculture (Gupta, 1991). As shown previously, this de facto exemption of agricultural incomes from taxation does not amount to much in the aggregate when compared to the trade related measures and the non-traded subsidies affecting agricultural incentives. Nevertheless, it distorts choices between agricultural and non agricultural activities, and the extent of the distortion will grow as income taxes become more important sources of government revenue, as is normally the case in the course of economic development. Since the overall effects of the reforms we are suggesting will be to significantly improve real farm incomes, we believe that this would be an appropriate time for the states to reconsider their present policies and to introduce taxes on farm incomes at somewhere about the levels of the central government income taxes on non-farm incomes. Because the central exemption levels ensure that only relatively high incomes are subject to any tax, this would not affect marginal and small farmers. At the same time, it would make a badly needed contribution to state government revenues and would remove an important avenue for the evasion of the central government income tax. It could be made more palatable to farmers if it were used to help finance increased state expenditure on rural infrastructure. 27  IIL Policy Reforms : Some General Suggestions The broad objectives of reforms in the medium or long term, and the tactics for the short run suggested above, are comprehensive and ambitious. In order to get reform under way, experience from Indian manufacturing and agriculture, and liberalisation episodes in other countries suggest that the following would be useful: - Include at an early stage a "demonstration liberalisation", i.e., the liberalisation of an industry which is likely to have an early and readily apparent favorable, positive impact as regards output and employment. - As far as possible, combine reforms with a contractionary impact with reforms which are likely to be expansionary, either with regard to the same crops, or crops to which the adversely affected farmers can switch. - Do not allow the reform process to become hung up on attempts to retain price stability over time and territorial price uniformity. Commodity prices are inherentdy unstable: to have any chance of implementing substantive liberalising reforms with an appropriate role for private intermediaries and processors, this needs to be recognised. -Do not build up overoptimistic expectations that the reforms will necessarily lead immediately to a noticeable increase in the growth of the farm sector and in rural incomes and employment. Although some farming activities and regions are likely to gain in the short run, others are likely to lose or at least face a difficult period of adjustment. It may take several years before higher growth in the whole farm sector becomes apparent, and even that may be delayed or even prevented altogether by a variety of factors. It is better to be realistic about possible future problems and prospects. Disappointed, overoptimistic popular expectations can easily lead to policy reversals. - Studies of the various kinds will be needed , but it is vital that they be clearly focussed on the issues at hand in the reform process. Broad, unfocused studies can muddy the issues, drag on for too long, and just serve as an excuse for delaying effective action. As regards a "demonstration liberalisation n", the Indian debate on many of the issues is highly ideological. For historical and other reasons there is in particular a deep seated distrust of private traders and processors and private markets in general, not only in the ministries and parastatals but amongst economists who work on agricultural subjects. The resulting interventions and regulatory controls emasculate or distort the operations of the private sector and expand the scope for rent seeking and black economy activities, which reinforces the prevailing conviction that it is inherently deficient as well as corrupt. In order to help break down this self fulfilling process of distrust and control, it would be helpful at an early stage to have a successful 'demonstration liberalisation". The evident and widely recognised success of the deregulation of the cement industry in the early 1980s greatly helped in mobilising support for the broader deregulation of manufacuring which came later. In our opinion the cotton industry would be an excellent candidate for such a "demonstration liberalisation' for the rest of agriculture, although a start should be made on other fronts at the same time. The case for an early concentration on cotton is that, because of India's comparative 28  advantage in the production of longer staple , labor ntnsve -ottons, the removal of export controls (including minimum export prices) is likely to lead to a substantial Increase in the export of these varieties and to a corresponding increases in production amd employment . Furthermore , unlike cotton yarn and fabrics, the international market for cotton is reatively open and in particular is wt restricted by the multi-fiber arrangement which creates difficulties for exports of textiles to developed country markets. Simultaneously with the removal of export controls, imports should also be allowed without import licensing or other controls and with a zero tariff. While domestic prices are well below intenational prices as at present, there will be no or few imports, but allowing imports without any restriction will ensure that the textile industry is not injured as a result of future shortages due to drought or other events, especially shortages of coarser cottons in the production of which India has less of an advantage. As noted earlier, as a by -product, by aligning domestic and world prices ,the opening of cotton exports and imports would also contribute to greater economic efficiency in the textile industry. As well as these trade reforms, a policy reform package for the industry would also include the domestic regulatory reforms referred to earlier, namely: -Removal of the export quota allocations to the CCI, the Maharashtra Federation and similar organisatons and the creation of conditions for eff-ctve export marketing by the private sector. The present regulatory functions of the Ministry of Textiles over the cotton industry would be abolished. -Removal of the general export controls on cotton yarn (except for the controls required by the multi-fiber arrangement ). -General permission for futures trading, which in the case of the cotton industry could be rapidly and effectively implemented. -Removal of inventory controls from cotton traders and textile firms. - Reviews of the roles of CCI and the Maharashtra Federation and ensuing reforms which enure that they do not impede or distort trade in cotton. -Removal of the state ginning margin controls and the creation of conditions which would make investment in the modernisation of cotton ginning attractive to private industry. We believe that this reform package would encourage new investment and increased competition at all stages in the production and distribution of cotton, and in particular would improve the quality of Indian cottons and its markedtng both domestically and in export markets. There are obvious advantages in reform packages which include ezansionary refornm which would partly or fully offset contractona reforms. A few examples: -Reductions in the non-traded subsidies (canal irrigation, electricity and credit) should be accompanied by increases in the selling prices of grains and cotton, which in turn would follow from the removal of export controls and from general trade and domestic deregulation in the markets for these products. In particular, abolition of the compulsory levy prices for sugar and rice would benefit these farmers in a clear way which would be apparent to them and everyone else. This change could be included in a reform package involving increased charges for inputs. - Decontrol of grains and cotton, leading to higher prices, could help offset decontrol and reduced protection for oilseeds and sugar. -In the south, decontrol and higher prices for rice, spices coffee and tea could help offset lower protecdon and reduced prices for rubber and cocomut/copra. As regards price stablity and tertorlal price uniformity, seasonal variations and differences which reflect local conditions and transport costs are necessary for the efficient timing and location of production. With the agricultural economy open to international trade, local prices 29  should also reflect these t opporunies. It may be economically efficient, for example, to Import a commodity during part of the year and export it during another part of the year (especially in regions bordeing neighbouring countries). Likewise, it may be economically efficient to import a commodity In one part of the country while exporting the same commodity from another location. Efficient resource allocation may also require substandal price variations from year to year and over longer periods. Excessive preoccupation with stability over time and with geographical uniformity could make any substauive liberalisation of the present controls very difficult to achieve. In this regard, a few general commentB are worth malking. First, as is well known, price stability is not the same as income stability for fiamers. Secondly, insulation from or only partial exposure to world markets does not of itself guarantee stable prices: for example, in certain respects domestic cotton prices have been less stable than they would have been if they had been direcdy linked to world cotton prices. Thirdly, Indian farmers in this and other industries (e.g., pulse farmers, whose prices have been principally determined by import prices since the early 1980s) have lived with unstable prices without disastrous consequences, even though regulatory controls have gready inhibited the extent to which they themselves &/or private intermediaries have been able to deal with the price risk. Fourthly, as a reform strategy, farmers (including farmers producing the major grains) may be willing to trade off less price stability for a higher general level of prices. As regards the need to avoid building up overoptimistic expcatlons, even if a thoroughgoing reform program were implemented, it is important to recognise that the supply response of agriculture as a whole is likely to be quite low in the short run, even if the longer run impact is substantial. Whereas increases in incentives for individual crops can elicit large increases in production in the course of a season as farmers switch from other crops , the aggregate short and even medium term (say over three or four years) response of the whole farming sector is limited by the supply of agricultural land and by the time required for new on- farm, off-farm and especially infrastructure investment to take place.4