India: Demand and Supply Prospects for Agriculture SWP500 World Bank Staff Working Paper No. 500 October 1981 Prepared by: James Q. Harrison South Asia Programs Department Jon A. Hitchings Treasurer's Department John W. Wall South Asia Programs Department Copyright ® 1981 The World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. ws and interpretations in this document are those of the authors 44-01 0220 PUBI )uld not be attributed to the World Bank, to its affiliated HG I ations, or to any individual acting in their behalf. Feathery Ja8ies 38811.5 IN 234 .W57 W671 no . 500 The views and interpretations in this document are those of the authors and should not be attributed to the World Bank, to its affiliated organizations, or to any individual acting in their behalf. WORLD BANK Staff Working Paper No. 500 INDIA: PAPERS ON DEMAND AND SUPPLY PROSPECTS FOR AGRICULTURE October 1981 Since India's foodgrain situation began to improve after the mid-1970s, the World Bank's economic work on Indian agriculture has concentrated on the implications of this development both for foodgrains and for other major agricultural commodities. This volume contains several papers that report on work accomplished so far. Jon Hitchings' paper projects demand for major agriculture commodities through the year 2000 based on consumption expenditure data from the 1973/74 National Sample Survey along w4th estimates of future population and income growth, rates of urbanization and trends in the distribution of income. The analysis reveals the differential effects of long-run income growth, and other factors, on demand for various crops. The individual commodity papers were prepared to analyze the long-run supply prospects and to compare these with the projected demand. The foodgrain paper, prepared by James Harrison and John Wall, raises the distinct possibility of foodgrain self-sufficiency and even a potential for an eventual exportable surplus. The vegetable oil paper, by John Wall, is less optimistic and projects a persistent domestic shortage of vegetable oils. This underlines the need for greater efforts by the agricultural support institutions to stimulate oilseed production and for an incentive pricing policy. The sugar paper, by James Harrison, analyzes the sugar cycle and stresses the disruptive effects of very large fluctuations around the production trend, which is essentially adequate to meet domestic requirements and provide for some exports. The utility of maintaining a buffer stock of sugar to stabilize year-to-year sugar supply is discussed. prepared by: James Q. Harrison South Asia Programs Department Jon A. Hitchings Treasurer's Department John W. Wall South Asia Programs Department Copyright0 1981 The World Bank 1818 H Street, N.W. Washington, D.C. 20433 U.S.A. INDIA PAPERS ON DEMAND AND SUPPLY PROSPECTS FOR AGRICULTURE Table of Contents Page Number Part I: DEMAND PROJECTIONS FOR INDIA . . . . . . . . . 1 Jon A. Hitchings Part II: THE FOODGRAIN ECONOMY. . . . . . . . . . . . . 51 James Q. Harrison and John W. Wall Part III: THE VEGETABLE OIL ECONOMY . . . . . . . . . . 67 John W. Wall Part IV: THE SUGAR ECONOMY . . . . . . . . . . . . . . 109 James Q. Harrison PART I DEMAND PROJECTIONS FOR INDIA Jon A. Hitchings June 1981 1 DEMAND PROJECTIONS FOR INDIA TABLE OF CONTENTS INTRODUCTION Page I. Data and Methodology .................................. A. Population Projections ............................ 5 B. Baseline Consumption .............................. 7 C. Estimation Approach ....... ............ 10 II. Elasticities and Projection Results ................... 12 A. Comparisons of Expenditure Elasticities . . 12 B. Quantity Projections .. 15 III. Income Redistribution . . . 21 A. Gini Ratio Estimation and Comparisons ........22 B. Effects of Income Redistribution on Demand Projections 28 IV. Sensitivity Analysis ...... ............................ 31 A. Population Growth and Urbanization . . 31 B. Expenditure Growth and Redistribution ...... ....... 35 V. Conclusions ........................................... 41 APPENDICES I. Demand Projections with Moderate Redistribution ....... 44 II. Gini Ratio Estimation Program ......................... 45 III. Sensitivity Analysis Data .. 46 IV. Foodgrain Equivalents .. ...... . .... ............. ....... 47 V. List of Consumption Groups and Items ....... .. ......... 48 REFERENCES ........ 49 - 3 - LIST OF TABLES AND FIGURES Tables Table 1 Population and Urban Share Projections Table 2 Selection of Base Quantities for Projections Table 3 Expenditure Elasticity Estimates Table 4A, 4B Demand Projections Assuming 3.5% and 5.0% Expenditure Growth Table 5 Projected Annual Growth Rates of Demand for Selected Commodities Table 6 Comparisons of Demand Projections Table 7 Projections of Food Energy Demand Table 8 International Comparisons of Income Distribution Ordered by the Gini Ratio Table 9 Some Estimates of Gini Ratios for India Table 10 Gini Coefficients for Individual Consumption Expenditure Table 11 Future Expenditure Shares and Gini Ratios Given Redistribfi- tion Rates Table 12 Changes in Quantities Demanded Resulting from Expenditure Redistribution Table 13 Population and Urban Share Values for Sensitivity Analysis Table 14 Sensitivity of Projections for Year 2000 to Altered Population and Urban Share Assumptions Table 15 Sensitivity Comparisons Table 16 Sensitivity Analysis of Expenditure Growth Rates Table 17 Elasticities of Projections Table 18 Correspondence of Multiplicative Factors and Annual Gini Ratio Change Rates - 4 - Figures Figure 1 Distribution of Income at Different Levels of Per Capita GDP Figure 2 Projection Isoquants for Year 2000 - 5 - INTRODUCTION This paper reports a set of demand projections for agricultural commodities in India. The main objective was to assess the effects of income growth, population growth and other non-price variables on the future pattern of consumption of agricultural commodities in India. A second objective was the development of a projection model that would be compact, flexible and fully documented. The model should be capable of projecting demand for any number of commodities and future time periods in a single pass with a minimum of file manipulation. It should also readily accommodate altered assumptions regarding urbanization, expenditure growth, income redistribution and population growth. The resulting model is documented separately. 1/ The 28th Round of the National Sample Survey (1973/74) is the most recent large-scale expenditure survey available. The 32nd Round (1977/78) is also directed to household expenditures, but it has not yet been released. Restricted to this data base, the projections assume relative prices are constant, and model only the demand side of the market. Projections for 17 commodities at five year intervals, expenditure elasticities, and calorie demand per capita are presented and compared with other research. Changes in income distribution and consequent effects on demand received particular attention. 2/ Sensitivity analysis was performed on expenditure and population growth, expenditure distribution, and urbanization. As a final step in sensitivity analysis, estimates were made of the elasticities of future demand with respect to key assumptions. I. DATA AND METHODOLOGY A. Population Projections. Four types of data were required for the projections: population and urban share projections, the starting quantities of consumption, the sectoral distribution of initial consumption, and expenditure data from a household consumption survey. The population and urbanization forecasts that were used are given in Table 1. Historical figures for the urban share of the population show a growth rate of about 0.2% per year. 3/ Projections made by the UN Population Division/Urbanization show a general but bumpy continuation of 1/ "Documentation of a Demand Projection Model Prepared for India," Jon Hitchings, Division Paper for ASADB (March 25, 1981). Available from the India Division of the World Bank, Washington, D.C. 2/ Appendix II contains a program for estimating Gini coefficients. 3/ Census estimates for 1951, 1961 and 1971 are 15.9%, 18.0% and 19.9%. (UN Source). - 6- Table 1 Population and Urban Share Projections Population Urban Share (million) ----(%) 1973/74 595.6 20.6 1979/80 672.2 22.3 1984/85 744.2 24.3 1989/90 820.5 26.9 1994/95 897.7 28.1 1999/2000 973.6 30.0 Sources: For population, World Bank estimates. For urban share, Population Reference Bureau, which interpolated projections from U.N. Population Division/Urbanization. The urban share for 1999/2000 was somewhat arbitrarily lowered from a U.N. projection of 34.0% which implied an abnormally large increase in the last five years of the century. The population forecasts are also reported in Population Projections 1980-2000, World Bank, DEDHR, (July, 1980), p. 212. this trend and are used without modification by the World Bank and the Population Reference Bureau. Their figures were utilized in this analysis except that the estimate for 1999/2000 has been modified. 1/ The latest World Bank population projections were used in this study. However, provisional 1981 population census data have been released which show higher-than-expected rates of growth during the 1970s. Although the population forecasts have not yet been altered accordingly, the potential implications of the new census data for the demand projections are presented and discussed with other sensitivity analysis issues in Section IV. B. Baseline Consumption. Estimates of total baseline consumption for 1973/74, the year of the household expenditure survey, are compared in Table 2. The organization of the table follows the expenditure classifications of the NSS survey. A list of commodities in each category appears in Appendix V. World Bank estimates were used whenever available. Total consumption of cereals was allocated to categories such as wheat, maize, etc., using weights implicit in FAO data. Sweeteners were disaggregated using unpublished tabulations from the NSS survey. The quantity of clothing estimate is based on production minus non-fabric2 uses. 2/ However, using the standard industry conversion rate of 10 m2/kg, 0.722 million tons of "clothing" corresponds to 7.22 billion square meters of material which is only 6% below an independent industry estimate of 7.68 billion square meters. 3/ The industry estimate for woven textiles includes cotton, synthetic, and blended fabrics. Technically, the projections should be increased by a few percent to better represent expenditures on non-cotton textiles which were recorded in the survey, and which increased the expenditure elasticities. Using the lower base allows the projections to be interpreted more in terms of demand for cotton, but introduces a small downward bias. 1/ See the note on Table 1. 2/ In the expenditure survey, "clothing" includes bedding, upholstery, and other textiles. 3/ See footnote 14/, Table 2. -8- Table 2 lNDIA SELECTION OF BASE QUANTITIES FOR P'ROJECTIONS (Hiousehold Demand or Net Food Avai lability) 'ACGREGOR- FAO IBRD SOURCES NSS IMfPLIE1D SANDBRSON/gOY-/ GROSSIT8 SELECTED 1973-74 1972-74 1973-74 1973-74 31 1973-74 IACTOR- BASE __ _ _ _ _ . ------------------________…1,000,000 mt--… --------------------- ulses, Cereal Substitutes 11.082-/ 8.210 8 715/ 1.143 S,71 Edible Oils 2.90 2.748 2.6677/ 1.79 1.109 2.667 Meat/Fish/Eggs -_ 2.398 1.0 2.398 Vegetables 6.928 33,428 1.0 6.928 Fruits/Nuts 7.18212/ 16.571-/8 1.0 16.571 Sugar/Khandsari 11,621- 3.690 4 ,519-D 2.64 1.0 4,519 Gur/Other Sweeteners 6.429 7.3368/ 3.53 1.0 7.336 Spices 0.684 0,691 1.0 0.691 Beverages 0,305 0,318 1.0 0.318 Tobacco/Pan/Itltoxicants 0.46414/ --- 1.0 0,464 Clothing 0,722- --- 1.0 0.722 Milk and Products 24.2102/ 19.748 1.0 19.748 Rice 36.593-7/ 38.131 48,31 1.143 40.363- 2/ ~~~~~~~~~~~~~~~~~~~~~~~~11) Wheat 23 362- 21,283 28.12 1.143 22 536-. Maize 3.943,-L 3,848 5.20 1.143 4.0361- 3/3 1/ Sorghum/M'illet/Barley 19.596- 16,182 23.54 1.143 17,154-/ A31 Sugars 11,621 10,119 11.8555/ 6.17 All Cereals 83.495 79.444 84.09 - 105.17 96.81 84,09 __________. __________-_-… - … …1------- -1,000,000 …-------- Population 583.9 595.6 588.3 595.6 NOTES -/ c :rcaf±gure 1f 5.6/1 s-r - ,v-r-tpd ton 'u-r atr t1% oxtractton. -/59.956 for rice & wheat, allocated by NSS quantity weights. (These cereals were not dis- aggregated by MlacGreggor.) 3/23.539 for other cereals, allocated by NSS quantity weights. (maize was grouped with the other coarse cereals in the study.) A/Cashews, almonds, walnuts, groundnuts, coconuts, and fruit. 5/ 1973-74 average. Excludes Topioca -/Includes topioca, 7/IBRD, ASADB source, 8/ IBRD, ASADB source, 9/Fred H. Sanderson and Shyamal Roy, Food Trends and Prospects in India (Washington, D, C.: The Brook- ings Institution, 1979), Table 6.1. 10/The projections for pulses, cereals, and oils contain a grossing factor shown in this column to re- present nion-household demand. Seed, feed and waste account for about 12.5% of final demand for foodgrains: it/ 1/(1-.125) = 1.143. - The selected base quantities for cereals were found by allocating the IBRD total for cereals according 12/to proportions iap;iicit in FAO quantities. - Includes gur and other sweeteners. 12/The 'SS im'plied base figutres are inferred fre, data on nuantities consumed Per capita. Refers to residual of production weight of baled cotton mLinus seed, post-harvest loss, and exports. Ihis is about 6% lower than the estimate of 7.68 billion square meter: of cotton, non-cotton, and mixed fibre cloth, assuming ten squiare meters per kg. This estimate is from the altndboolk of Statistics o01 the Cottotn Te\tile [n_uJtrv, The Indidn Cotton Mills Federation, Bowbay (September 1, 1980), p. 35. Ihe 1974 e-stimated aggregate household consumption was adjusted for 1973/74 average availability (Table 19). _ -John Macgregor "Agricultural Dcmand Projections for India," The World Bank, ASADB (Draft Divisional Paper, 1979). - 9 - Fixed grossing factors for non-household demand (seed, feed, minimal industrial use, and loss) are shown for foodgrains and edible oils. Future demand for these commodities represents total demand, whereas projections for other items refer only to household demand. As a consistency check, baseline quantities for food can be compared with the population estimate in terms of calories per capita per day. The conversion from weight to calories utilized foodgrain equivalent factors shown in Appendix IV. The level of food availability implied by the selected baseline quantities is quite consistent with other estimates: 1,952 calories per capita per day (Table 7). Production indices in 1973/74 for foodgrains contracted somewhat, but were comparable with other below-average years, such as 1970/71 and 1976/77. Cereal production in calendar 1973 was 5% below the surrounding five-year average. However, an increase in net cereals imports in 1973, which partially offset the production shortfall, the reasonableness of the per capita calorie estimate, and the fact that FAQ base quantities which were adopted for some commodities were already three-year averages, led to a decision to dispense with more elaborate modifications of the base level of total consumption to adjust for starting period aberrations. Consumers may have been adjusting expenditures to rapid price increases in 1973/74. Price indices for major agricultural commodities rose rapidly in the early 1970's, particularly 1973-75, in contrast to the more stable pteice environment of the late 1960's. These considerations highlight a second area of potential sensitivity of the projections to the base year chosen. Set against these precautionary notes is the rather uniform set of expenditure elasticities obtained from various sources and time periods which are compared with estimates from these data in the following section. The similarity suggests typical patterns of demand may not have been too disrupted by these price movements. Weights for sectoral consumption are required to combine urban and rural demand into a national estimate. Quantity-based weights should be used to avoid price differentials and these were derived from the NSS survey for edible oils, sweeteners, and cereals. Value-based weights without price adjustment were used for the other commodities. A detailed exposition of the methodology adopted for projections is presented in a separate paper, available from the India Division of the World Bank, "Documentation of a Demand Projection Model Prepared for India." The principal features of the approach are outlined below. - 10 - C. Estimation Approach. Cross-sectional survey data from the NSS 28th round were used to estimate expenditure regressions and elasticities. 1/ Generalized least squares (GLS) regressions were run on the expenditure survey data using the ratio semilog inverse functional form. 2/ This form automatically satisfies the Engel aggregation condition of consumer demand theory since the dependent variables are budget shares. 3/ (If this condition were not met, the sum of marginal propensities to consume would not exhaust marginal expenditure which would introduce awkward inconsistencies into a set of demand projections). The functional form also does not impose an a priori relationship between the income level and the elasticity. One regression per commodity or commodity group was used to estimate the elasticity expression for all expenditure classes. The expression yields distinct elasticities for each expenditure class when it is evaluated using that group's mean total expenditure, and mean commodity-specific expenditure. Thus projections for each income level utilize separate elasticities, although they are derived from a single regression. GLS methods are needed to compensate for the heteroscedastic properties of grouped data having different numbers of observations per group. Persons per household times households per class formed the weights. Future demand was geared to the growth rate of expenditure in the economy, which can be selected at will during a computer projection run. From this growth rate, urban and rural per capita expenditure growth rates were derived for each projection period (base date to projection date) which are consistent with: a) the selected total rate of expenditure expansion; b) population growth and urbanization rates; 1/ National Sample Survey Organization, Tables on Consumer Expenditure, 28th Round, No. 240, Department of Statistics (1977), New Delhi. 2/ This form is Y/X = a + b ln X + c/X where Y is the commodity expendi- ture and X is the total expenditure. The expenditure elasticity is then e = (a + b + b lnX) X/Y. 3/ Price data would be necessary to check the other conditions, namely the negativity of the own substitution effect, the symmetry of cross-substitution effects, and homogeneity of degree zero for the system of demand equations. (The last condition implies that multi- plication of prices and income by a constant would leave demand pat- terns unchanged, i.e., there is no "money illusion" in consumption.) - 11 - c) historical rates of growth between the base year and the most recent observation--about 3.5% in real terms from 1973/74 to 1979/80. Table 3 of the documentation paper provides evidence that since 1961, there has been very little difference between the sectors in nominal per capita expenditure growth rates. Consequently, none is assumed here. An annual rate of change in the Gini ratio can also be selected during program execution. The effects of redistribution on demand are incornporated through a reallocation rule which satisfies the Gini ratio modification in such a way that changes in expenditure are proportionate to an income group's distance from income equality. Per capita demand growth for each commodity was projected for every income class, both sectors, and the future time periods of interest given the rate of redistribution and the derived per capita expenditure growth rate. The future periods chosen were five year intervals beginning in 1979/80 (to better coincide with the Plan period) and ending in 1999/2000. Rural and urban population growth factors were applied, the income classes were aggregated, and the sectors combined, using appropriate weights. Since demand growth is expressed as a multiplicative increase, the outcome can be multiplied by grossing factors for non-household demand (for foodgrains and edible oils) and by the base levels of consumption. The result is final demand in quantity terms. The adopted approach assumes that relative prices are constant through time, and equal for all consumers within same sector, either rural or urban. Non-household demand is neglected except for edible oils and foodgrains. Proportionality assumptions are made to infer non-household demand from household demand for these commodities. Individual household data such as occupational category or educational level of the head-of-household, caste/ethnic group, farm characteristics of farming households, etc., were not available, although they can illuminate consumption patterns. A final important assumption is that real-expenditure growth per capita is equal in the rural and urban sectors. This assumption at least holds in nominal terms since 1961. (Evidence to this effect is presented in the model documentation paper.) The programs in the model print out the expenditure elasticities, the starting and ending Gini ratios, expenditure shares held by population groups, projected demand for all commodities and time periods under the redistribution and expenditure growth assumptions, and aggregate demand for foodgrains and sweeteners. - 12 - II. ELASTICITIES AND PROJECTION RESULTS A. Comparisons of Expenditure Elasticities. The expenditure elasticities in this study satisfy the Engel aggregation condition for the rural and urban sectors (Table 3): the sum of expenditure elas5icities weighted by budget shares equals one. 1/ The' lowest adjusted R for the underlying generalized least squares reg essions was around 0.95 and most were 0.98 or above. However, the high R s are partly the consequence of using observations grouped by income class. 2/ The elasticities for a number of commodities are quite similar to previous estimates. Note, for example, comparisons with estimates from the National Commission on Agriculture (NCA) for pulses, meat/fish/eggs, sugar and khandsari, beverages, clothing, milk and maize. 3/ Although the expenditure elasticities for wheat and rice are higher than found by NCA, the estimates for foodgrains as a whole are consistent with other research. Desai's estimates of the rural and urban elasticities for foodgrains are 0.52 and 0.30 (not shown in Table 3) which, are close to the present estimates of 0.63 and 0.39. 4/ Mellor's national estimate is the same as the present urban figure. The national estimate of the expenditure elasticity for foodgrains is 0.59 in the present study, 5/ whereas the middle of the range reported by Scandizzo and Bruce for India is 0.60. The estimates they report use longitudinal data 1/ Meeting this condition is an attraction of using the ratio-semilog-inverse form in which budget shares are dependent vari- ables. The weighted expenditure elasticities equalled 1+ 0.001 in each case. 2/ This is an additional reason why the correlation coefficient is not particularly suited for choosing among functional forms even after appropriate weighting and econometric techniques have been applied. 3/ The "Other" category in Table 3 includes fuel and light, footwear, miscellaneous goods and services, rents, taxes, and durable goods. An elasticity for this category must be estimated to check the Engel aggregation condition. The projections for this category are expressed in terms of a multiplicative increase over the base level. 4/ B. M. Desai, "Analysis of Consumption Expenditure Patterns in India," Occasional Paper No. 54, Department of Agricultural Economics, Cornell University (August 1972), Table 3. Found from log-log-inverse regres- sions on 1963/64 NSS data. 5/ Found by combining the urban and rural estimates with population and expenditure weights. - 13 - TABLE 3 EXPENDITURE ELASTICITY ESTIMATES FOR INDIA Present Study 1/ NCA 2/ Mellor Various Rural Urban Rural Urban National 3/ National 4/ Pulses 0.86 0.81 0.85 0.66 - 0.32 Edible Oils 1.00 1.01 0.88 0.97 5/ 1.05 - Meat, Fish, Eggs 1.15 1.12 1.22 1.07 6/ 1.33 Vegetables 0.82 0.94 - - - Fruits, Nuts 1.36 1.54 - - - Sugar, Khandsari 1.51 1.06 1.65 1.11 1.10 7/ - Gur, Other Sweeteners 1.09 0.53 0.96 0.23 - Spices 0.67 0.50 - - - Beverages 1.20 1:43 1.29 1.33 6/ - - Tobacco, Pan, Int6x. 0.90 0.96 0.72 0.79 - - Clothing 1.82 1.68 1.86 1.64 8/ - - Milk and Products 1.73 1.43 1.46 1.30 1.60 - Rice 0.71 0.42 0.41 0.18 - 0.94 Wheat 1.01 0.55 0.67 0.37 - 1.06 Maize 0.03 -0.76 0.01 -0.47 -- Sorghum, Millet 0.07 -0.59 - - - All Other 1.11 1.28 - - - Foodgrains 0.63 0.39 - - 0.39 0.49-0.71 1/ Found from regressions of the ratio semilog inverse form on NSS 28th Round data (1973/74) weighted for households per income class and persons per household. Elasticities at the mean are reported, 2/ Report of the National Commission on Agriculture, Demand and Supply (1976), Appendix 10.2. 3/ John W. Mellor, "Agricultural Price Policy and Income Distribution in Low Income Countries", World Bank Staff Working Paper No. 214, September 1974. Estimated from NCAER "All-India Consumer Expenditure Survey, 1961L/65" using log-log-inverse functional forms. 4/ Various sources using longitudinal data compiled in "Methodologies for Measuring Price Intervention Effects", Pasquale L. Scandizzo and Colin Bruce, World Bank Staff Working Paper No. 394 (March 1980), p. 80. 5/ These elasticities are for vegetable oils. Elasticities for vanaspati are rural: 1.98, urban: 1.48, (reported by NCA). 6/ Budget shares were not reported, so these elasticities are simple averages for commodities (average of tea and coffee for beverages). 7/ Applies to all sweeteners. 8/ These elasticities are for mill-made cotton clothing, and are intermediate between handloom cotton clothing and khadi cotton clothing. - 14 - through the mid to late sixties. I/ The similarity of estimates from longitudinal data, cross-sectional data a decade earlier, and the current analysis, (Desai--1963/64, current--1973/74) indicates stability in the demand pattern for foodgrains and increases confidence in the projections. Several factors motivated re-estimating the elasticities despite their availability in other sources. Foremost among these is the fact that the elasticities derived here and used in the projections vary between expenditure classes, although they are only shown in Table 3 as they appear after evaluation at the grand mean. Moreover, while many functional forms have this property, the elasticities are usually forced to vary in the same direction across income groups, regardless of the commodity. The derivative of the elasticity with respect to income is negative a priori for certain estimation forms. In the regressions estimated in this study, the direction in which the elasticity changes depends on the commodity. This variation was required.since future demand for a given commodity was projected for each expenditure class, and marginal propensities to consume generally are not constant across income groups. Additional advantages of these estimates include more commodity disaggregation, weighting by persons per household and households per income group, and the consistency of Engel aggregation, as already mentioned. Thus the similarity of some elasticities at the overall mean to other estimates does not greatly detract from the fruitfulness of the exercise. Some of the notable features of this set of expenditure elasticities are: 1) the urban and rural preference for wheat and pulses compared with rice as total expenditure increases (a pattern supported by NCA estimates); 2/ 2) higher foodgrains elasticities in rural than in urban areas; 3) very low rural and negative urban elasticities for maize, sorghum, and millet; and 4) high elasticities for clothing, milk, milk products, fruit and nuts. These relationships among the elasticities foreshadow certain characteristics of the projections and sensitivity analysis, namely: 1/ See Footnote 4/, Table 3. 2/ This preference may have implications for the proportions of wheat and rice the public sector should hold for distribution. - 15 - 1) the increase in the proportion of wheat and pulses demanded out of foodgrains if expenditure growth is rapid (Table 16); 2) the apparent dampening effect of faster urbanization on the growth in foodgrains consumption (Table 14); 3) the somewhat reduced demand for coarse cereals given accelerated economic growth; and 4) the sensitivity of projections for clothing, milk, and milk products to expenditure growth rates. B. Quantity Projections. The projections for two rates of expenditure growth, assuming no income redistribution, are given in Tables 4A and 4B. The projections for various foodgrains and edible oils contain grossing factors listed above (Table 2) for non-household demand. Foodgrain demand would more than double by the end of the century under the lower rate of growth, and increase by 220% if 5% expenditure growth were experienced. 1/ Higher expenditure growth adds at least one percent to the growth rate of demand for sweeteners, edible oils, wheat and pulses through 1984/85 (Table 5). Although the growth in demand slackens somewhat after 1984/85 for these commodities, rice, and all foodgrains, it remains well above 3.0% for sugar/khandsari and edible oils even under the lower alternative. Pressures, therefore, may persist either to continue the importation of large volumes of edible oils, or to let prices rise sufficiently to induce a substantial supply response. 1/ These increases are relative to the 1973-74 base year. - 16 - TABLE 4A Demand Projections Assuming 3.5% Expenditure Growth FOR THE FOLLOWING PROJECTION TOTAL EXPENDITURE GROWTH ASSUMPTION IS 3.5 PERCENT GINI RATIO CHANGE RATE IS 0 PERCENT DEMAND IN MILLION METRIC TONS 73/74 79/80 84/85 89/90 94/95 99/00 PULSES 3/ 8.71 11.36 13.19 15.26 17.64 20.25 EDIBLE 01LS 2.67 3.36 3.99 4 75 5.59 6 56 MEAT, FISH. 6 EGGS 2.40 2 70 3.25 3.91 4 68 5.62 VEGETABLES 6.93 7.84 9.21 10.81 12 S9 14 62 FRUITS & NUTS 16.57 18.51 23.13 29 05 36.39 45 91 SUGAR & KHANDSARI 4.52 5.25 6.44 7.89 9.57 11.63 GUR & OTHER SUGARS 7.34 8.33 9.60 10.96 12.61 14.31 SPICES 0.69 0 77 0.89 1.01 1.15 1.30 BEVERAGES 0.32 0.34 0.43 0 56 0.70 0.90 TOBACCO. PAN & INTOX 0.46 0.51 0 60 0.71 0.83 0.98 CLOTHING 0.72 0.83 1.05 1.31 1.67 2.13 MILK 19.75 23 24 28 74 35.48 43.84 54.17 RICE 40.36 52.22 59.31 66 87 75.26 83.51 WHEAT 22.54 29.19 33.97 39 29 45.20 51 47 MAIZE 4.04 5.09 5.50 5 87 6 28 6.59 SORGHUM & MILLET 4/ 17.15 21 59 23.52 25.36 27.28 28 84 OTHER V/ 1.00 1.06 1.30 1.61 1.99 2.47 SWEETENERS 11.85 13.58 16.04 ,18.84 22.18 25.93 FOODGRAINS 92.80 119.46 135.49 152 65 171.66 190.66 TABLE 4B Demand Projections Assuming 5.0% Ex-enditure Growth FOR THE FOtLOWING PROJECTION TOTAL EXPENDITURE GROWTH ASSUMPTION IS 5 PERCENT GINI RATIO CHANGE RATE IS 0 PERCENT DEMAND IN MILLION METRIC TONS 73/74 79/80 4/84/85 89/90 94/95 99/00 PULSES 3/ 8.71 11.36 13.92 16.89 20.39 24.29 EDIBLE OILS 2.67 3 36 4.24 5.33 6.56 8 00 MEAT. FISH. & EGGS 2.40 2.70 3.51 4.55 5.86 7.52 VEGETABLES 6.93 7.84 9.71 11.97 14.55 17.55 FRUITS & NUTS 16.57 18.51 25.93 36.45 50.98 71 60 SUGAR 6 KHANDSARI 4.52 5.25 7.04 9 36 12.27 15.99 GUR & OTHER SUGARS 7.34 8 33 10.18 12.20 14 60 17.00 SPICES 0.69 0.77 0.92 1.08 1.26 1.45 BEVERAGES 0.32 0.34 0.49 0.70 0.99 1.42 TOBACCO. PAN & INTOX 0.46 0 Si 0.64 0.81 1 02 1 29 CLOTHING 0.72 0.83 1.21 1.74 2 52 3 62 MILK 19.75 23 24 32.12 43.94 59.81 80.82 RICE 40.36 52.22 61.30 70.72 80 34 88 54 WHEAT 22.54 29.19 35.70 42.91 50 68 58 60 MAIZE 4.04 5.09 5.48 5.78 6.08 6.20 SORGHUM & MILLET 4/ 17.15 21.59 '23.38 24 91 26 32 27 08 OTHER 1/ - t.00 1 06 1.44 1.98 2.71 3.73 SWEETENERS 11.85 13.58 17.22 21.57 26.87 32.99 FOOOGRAINS 92.80 119.46 139.78 161 21 183.82 204.71 1/ Multiplicative factors, not million metric tons. 2/ Assumes 3.5% expenditure growth through 1979-80. 3/ Includes cereal substitutes and grams. 4/ Includes barley and other coarse cereals. - 17 - TABLE 5 Projected Annual Growth Rates of Demand for Selected Commodities a/ 1979/80 to 1984/85 1984/85 to 1999/2000 Low High Low High …____________________(% )…______ ___________ Sugar & Khandsari 4.2 6.0 4.0 5.6 Gur & other sweeteners 2.9 4.1 2.7 3.5 Edible oils 3.5 4.8 3.4 4.3 Rice 2.6 3.3 2.3 2.5 Wheat 3.1 4.1 2.8 3.4 Pulses 3.0 4.1 3.0 3.8 Foodgrains 2.6 3.2 2.3 2.6 a/ The "Low" and "High" growth rates assume 3.5% and 5.0% total expenditure growth, respectively. The reasonableness of the projections can be examined through: 1) coinparisons with actual demand (availability) in 1979/80, which is the first projection period; 2) comparisons with other quantity projections; 3) converting food demand to calories per capita and comparing these figures with expected levels of consumption and other calorie projections. The actual availability of foodgrains for calendar year 1979 is estimated to be 114.11 million tons (103.44 million tons cereals and 10.66 million tons pulses) which can be compared with the projection for fiscal 1979/80 of 119.46 million tons (108.1 million tons cereals, 11.36 million tons pulses), which is about 5% higher. Foodgrain production in 1979/80 was 17% below the previous year's record level and availability in calendar year 1980 was sharply reduced. The projections would therefore further exceed availability if 1979 and 1980 were averaged. Nevertheless, it is to be expected that the decline in incomes and upward pressure on prices following a sharp decline in the production of foodgrains would have dampened demand beneath its forecast level and this undoubtedly explains part of the discrepancy. - 18 - Household consumption of sweeteners, averaged for 1978/79 and 1979/80, was 13.1 million tons whereas the projection was 13.6 million tons. 1/ The 1979/80 consumption of edible oils (including non-household consumption) was estimated to be 3.55 million tons compared with a projection of 3.36 million tons. 2/ The projections for these commodities are fairly close to observed and estimated consumption, particularly if allowances are made for the 1979/80 crop year. Projections for the year 2000 by the National Commission on Agriculture for rice, wheat, foodgrains, sugar and khandsari, edible oils and meat/fish/eggs are very much in line with those in the present study (Table 6). Estimates of demand in 1984/85 for rice from several sources lie in a narrow range. The present projections for sugar also closely match EPD's longitudinal estimates. Converting food demand to calories per capita allows a determination of whether the projected demand levels for all foods are reasonable in comparison with the population projections. First, the grossing factors were removed (since household consumption alone is relevant here), then quantities were converted to foodgrain equivalents with food energy weights, and finally the cross-commodity summation was expressed in calories per capita per day (Table 7). 3/ The resulting figure of 1,952 calories for 1973/74 is similar to estimates by NCA, The Brookings Institution, FAO, and the World Bank in its World Development Report. 4/ NCA's high projection for 2000 differs by only 22 calories per capita per day from the present analysis, and the low projections are eight calories apart. These differences are negligible in per capita terms. This is a remarkable concurrence considering that different consumption data, population projections, and methodologies were used. Sanderson and Roy's low projection (Brookings Institution Alternative C) 1/ Consumption estimate from U.S.D.A. preliminary figures and "Sugar Situation in India," GOI, (September, 1980 mimeo). Averaging compen- sated for the reduced cane production in 1979/80. Sweeteners do not contain adjustments for non-household demand. 2/ World Bank, estimate. 3/ The food energy weights and conversion factors are reported in Appen- dix IV. 4/ Some FAO base figures were adopted which makes this estimate less independent of the present study, but differences in population, foodgrains and some other crops remained which balanced out in terms of calories per capita. - 19 - 1ABILE 6 CoMp&risons of Demsi,d Projeetlons 1/ 1985 1990 2000 Low High Low High Low 111gb Rice -------- tillion tons----------- Present Study 59.31 61.30 66.87 73.72 83.51 88.54 NCA 59.89 64.69 - - 78.59 84.56 EPD 56.45 - 64.19 - - - USDA - 62.00 - - - - Iowa State 57.0 - - - - - Wheat Prcsent Study 33.97 35.70 39.29 42.91 51.47 58.60 NCA 33.83 38.16 - - 46.91 52.45 EPD - 43.02 - 53.61 - - USDA - 41.00 - - - - Other Cereals Present Study 29.02 34.28 - - 35.43 33.28 NCA 35.35 35.36 - - 43.19 42.90 Pulses Present Study 13.19 13.92 - - 20.25 24.29 NCA 16.95 20.27 - - 23.15 28.23 Foodgrains Present Study 135.49 139.78 152.65 161.25 190.66 204.71 IICA 146.03 158.25 - - 192.36 208.14 Brookings - - - 173.30 - - Sugar Present Study 2/ 5.09 5.56 6.23 7.39 - - EPD 5.39 - 6.30 - - - Sugar and Rhandsari Present Study 6.44 7.04 - - 11.63 15.99 NCA 6.55 8.62 - - 10.33 13.31 Edible Oils Preseuit Study 3.99 4.24 - 6.56 8.00 NCA 4.19 5.27 - - 6.45 8.05 Milk Present Study 28.74 32.12 - - 54.17 80.82 NCA 33.73 44.17 - - 49.36 64.40 Meat, Fish. Eggs Present Study 3.25 3.51 - - 5.62 7.52 NCA 3/ 3.95 5.16 - - 6.00 8.00 Notes: I The low and high projections reported for "Present Study" assum. 3.5% and 5.0% total expenditure growth, respectively, and no income redistribution. The NCA estimates presented for comparison are based on .onsumer demand, miltiplied by constant grcssirg factors given in Table 2. NCA's gross demand estimates for some tommodities actually use increasing factors for nan-hojsehold demand. 2/ Found from sugar and khandsari projections by assuming 34% of the cane crop was processed into sugar and 9Z into khandsart. 3/ A conversion factor of 48g/egg was used. Source": NCA - Report of the National Commission on Agriculture, PartIll, Demand and Suipply (1976). Table 10.7 (see Footnote 1). EPD - Econz,mic Analvsis and Projectiton, Depirtment, World Bank, USDA - U.S. Dcpartment of Agriculture, Anthony Rojko, et al, Alternative Futtires fir World Fvocd. 1985, Vol. I (1978), Table 28. Iowv State - Leroy Bla.keslee, Earl Heady, Charle3 Trani1ngham, World n,^l ProId, tion Deimnnd. anI Trade, Center for Agriculture and Pural i)eveluiment, Iowj State UTniver,.ity (1973), Ames. Rrookinpgq - The Bronkings l11,ritIlEtin, Fred It. Sansderso,n and Shyfat RIoy, Fnnd Trenis nn.l Fro'ip.t I n tndia,(1919). TABLE 7 Projections of Food Energv Deand B A S E S P R O J E C T I O N S 1971 1973/74 1975 1977 1979/80 1984/85 1989/90 1994/95 1999/00 … -------- - (calories per capita per day) - -------- - ---- ---- Prosent Study, No kedistribution Low (3.5%) 1/ 1,952 1,955 2,029 2,104 2,196 2.292 High (5.0%) 1/ 1,952 1,955 4/ 2,118 2,279 2,453 2,622 NCA Low 2/ 2,080 2,200 2,300 NCA High 2 / 2,080 2,480 2.600 Brookings Low 3/ 2,212 2,495 2,697 Brookings High 3/ ,212 2,733 3,403 FAO (1972-74) 1,967 WDR 2,201 Present Study, Redistribution 3.5% Growth; -0.5% Gini Change 1,952 1,963 2,045 2,128 2,229 2,335 Notes: 1/ Grovth rates refer to, total expenditure in the economy. Per capita expeniditure growth wotild be 1.62 in the low case and 3.1% in the high c3se, r,ver the entire projection period. There would be some variation in these figures for sub-periods. 2/ The NCA low case assumes 1% growth in per capita expenditure, and 2% growth in the high case. 3/ The Brookings low case (Alternative C) assumes 1.2% per capita income growth from 1975-80, 1.5% for 1980-90, and 2.0% for 1990-2003. The high easc assumesn 2.4X, 3.0%, and 4.0% growth for these periods. 4/ A 3.5% .xpenditure growth rate through 1979/80 was assumed for both the high and low cases, hence this figXtre is the same as above. Sources: NCA - Report of the National Commission on Agriculture, Part III, Demand and Supply (1976), Table 10.12. Brookings - The Brookings Institution, Fred H. Sanderson and Shymal Roy, Food Trends and Prospects for India (1979), Table 7.11 FAO - Food and Agricultural Organization of the U.N., Food Balance Sheets. 1972-74, Rome. WDR - World Development Report, 1980, The World Bank, Table 22. - 21 - is similar to the high case here and in the NCA study. However, their high projection appears to be-unrealistic. Per capita food energy demand would increase by about 15% by year 2000 at a 3.5% expenditure growth rate. 1/ A moderate rate of expenditure redistribution, i.e., a 0.5%/year drop in the Gini ratio, would leave the mean level of food energy demand virtually unaltered. However, a gradual increase in equality might have a more significant effect on the distribution of consumption. The next section presents evidence that the Gini ratio based on personal expenditures has actually been falling at this rate. The sensitivity of the projections to various assumptions is examined in Section IV. In general, the projections in Tables 4A and 4B appear consistent with actual consumption of some commodities in 1979/80. The projections into the 1980's and 1990's accord well with other research on consumer demand in India, expressed both in terms of quantities and in terms of calories per capita. III. INCOME REDISTRIBUTION Since demand is projected for each expenditure class using unique elasticities, income redistribution would alter the-aggregate. There are a number of measures of income distribution available to summarize the level of economic equality in a society: the Gini coefficient, Kuznets' Index, Theil's Index, the Pareto coefficient, the equally distributed equivalent, the coefficient of variation, and the standard deviation of the log-normal distribution, to name a few. Atkinson presents an interesting comparison of several of these.2/ Despite the ambiguities of crossing Lorenz curves inherent in the Gini coefficient, it remains the most widely recognized measure, and has been calculated frequently with historical data which facilitates trend comparisons. Since changes in income distribution affect the pattern of consumption, the projection 1/ It should be noted that the increase is moderated by the fact that the average energy requirement per capita is also rising slightly. As the population growth rate slows, the age structure of the popula- tion matures, and the mean requirement rises. 2/ Anthony B. Atkinson "On the Measurement of Inequality", Journal of Economic Theory, Vol. 2 (1970), pp. 244-257. Atkinson shows that the ranking of countries according to inequality obtained from the Gini ratio, the standard deviation of logarithms, and the coefficient of variation are similar to the rankings found with the equally dis- tributed equivalent measure using a different index of aversion to inequality. - 22 - programs were designed to model assumptions about an altered distribution and their impact on demand. Changes in the income distribution are summarized in terms of the Gini coefficient and shares of the total expenditures accounted for by given shares of the population. Redistribution is assumed to be proportionate to a population group's departure from economic equality. Crossing Lorenz curves (which may present problems when countries are compared with one another) are not, therefore, introduced. This section reviews an estimation procedure for the Gini ratio, makes historical and international comparisons, and reports the effects of redistribution on demand. The following section expands on the issue of redistribution in a sensitivity analysis framework, although the commodity focus is narrower. A. Gini Ratio Estimation and Comparisons. Gini ratios lie in a fairly tight band even when taken from a broad spectrum of countries. Ratios from a selection of countries which are based on similar years and population coverages are shown in Table 8. When countries are ordered by the ratios, a clear inverse pattern emerges between the Gini coefficient and the percent of income held by the least affluent 40% of the population. However, exceptions to this generalization can be found. Most of the estimates listed in the table rely on household income data. Changes in the distribution of individual consumption expenditure will obviously be related more closely to the demand pattern. Gini ratios derived from these data will be lower since savings behavior is disregarded. The Ranadive Gini ratio estimates for individual consumption in India are, therefore, the most pertinent series for this analysis (Table 9). They indicate an improvement in the distribution evidenced by a 8.3% drop in the Gini ratio over 15 years. Ahmed's figures on Gini coefficients derived from the personal income of individuals support this trend. Some of the series calculated by other investigators are more ambiguous in trend. - 23 - TABLE 8 INTERNATIONAL COMPARISONS OF INCOME DISTRIBUTION ORDERED BY THE GINI RATIO Income Share Held By Percent of Population Year Population a/ Gini Ratio Bottom 40% Top 20% Taiwan 1972 HH 0.284 22.3 37.2 Pakistan 1969/70 HH 0.336 20.2 41.8 United Kingdom 1968 HH 0.338 18.5 40.3 Yugoslavia 1968 HH 0.347 18.4 41.4 Korea (Rep. of) 1970 HH 0.372 17.7 44.5 Sri Lanka 1969/70 HH 0.377 17.8 44.9 Indonesia 1971 IR 0.463 17.3 52.0 El Salvador 1969 POP 0.465 12.4 50.8 India 1967/68 HH 0.478 13.1 53.1 Philippines 1971 HH 0.494 11.9 54.0 Chile 1968 HH 0.506 13.0 55.8 Ivory Coast 1970 IR 0.534 10.6 58.5 Mexico 1969 HH 0.583 10.2 63.2 Tanzania 1969 HH 0.597 7.8 63.3 Kenya 1969 IR 0.637 9.5 66.9 Brazil 1970 IR 0.646 8.1 67.3 a/ The coverage is national in all cases. IR: Income Recipient; HH: Household; POP: Population. Source: Adapted from Shail Jain, Size Distribution of Income: A Compilation of Data (Washington, D.C.: The World Bank, 1975) TABLE 9 SOME ESTIMATES OF OINI RATIOS FOR INDIA OJha and nhati Ahmed Ranadive Swamy Yer 1964 1971 1971 1971 1971 1971 1964 Pironal Income Pononal Incomo Peonal Income Personal Income Personal ltIcome Consumption Consuniptin households Individuas indIviduals Individuass houicholda especid ture *vpendJture misdivIdusla ht.useI,ulds 1951-52 . . . . . . 0.366 1952-53 . . . . . . 0.361 1953-54 043 0.359-0.374 0.437-0.511 0.336 0.369 1954-55 0 0 0.399-0.420 . 0.390 1955-56 - .30393-0.419 . 0.370 1956-57 0.341 . 0.4527 0.377-0.410 0.333 0.407 1957-58 . . . 0.371-0.391 0.432-0.540 . 0.398 1958-59 . . . . 0.383 1959-60 0.355-0.378 . 0.385 1960-61 . 0.4136 . . . 19(1-62 . 0.356-0.379 . 0.320 1962-63 0.385 . 1963-64 * 1964-65 . * 0.3873 * * 0.303 1965-66 . . . . . 1966-67 . . . . . * 1967-68 . . . . . . 1968-69 . . . . . 0.308 Sources P. D. Olba ad V. V. lBlutat Pattern of inCome ditttrbutIon In an uanderdeveloped economy: a cae study of India " In Anuricax &o_wik Xaview (Mcnasha (W lu.nsln)) Sep. 1964. p. 714: and Idem: * Pattern of Income distribution In India: 19513-5 to 1961.64 " (Pape presented at the Seminat on Income Distribution o0;nised by ttic Indian Statistical Institute. New Delhi. February 21-26. 1971). p. S. Mahfooz Ahmed: - SIw dbtribution of personal Income In India 1956-57. 1960-61 and 21'4-63 - (Pipcr presented at the Seminar .... 1971). table 6. K. R. Ranadive: - Pattern of Income distribution In India, 1953-54 to 1959-60 ;- InOuIlkiln of she Orford U.! ..rity Insl,iUfe of Economics and Statlstikc. Aug. 1968. p. 212 (distribution of Income of Individuals In 1953-54. 1954-55. 1995-56; 1957-58 and 193940: ranges are due tn rItinmites bjscd on different assumptions about savings and tax evasion); and Idem: " Distribution of income: Trends since planning" (Paper presented at the Seminar 1921). pp 16. 17 end 33 Subramanian Swamy: 'Structural changes and the distribution of Income by size: The ease of Indiea. In Review of Inro,ae arnd Wealth (New 2le'en (Cosinecitcut)), June 1967. p. 173. SOURCE: Felix Paukert, "Income Distribution at Different Levels of Development: A Survey of Evidence," International Labor Review, Aug.-Sept., 1973, pp. 97-124. - 25 - Gini ratios were estimated in the 28 thNSS Round for 1973/74, following Kakwani and Podder's approach.l/ A regression fit of the Lorenz curve was found using the following functional form: N =a V ( ff 4, where N = (P-E) IT = (P+E) V P = cumulative population shares E = cumulative expenditure shares The Gini ratio (G) is twice the area under the curve given by the integral from 0 to Nr2V G = 2a (1 )1++PB(1+c, 1+P) B(l+S, 1+P) is the Beta function. The program used to find the Gini ratio from the original data is listed in Appendix II.2/ The data from the urban and rural sectors were merged for the economy-wide estimate. 1/ N. C. Kakwani and N. Podder, "Efficient Estimation of the Lorenz Curve and Associated Inequality Measures from Grouped Observations," Econometrica, Vol. 44, No. 1 (January 1976), pp. 137-148. 2/ In RAL (Research Analysis Language), the Beta function is not avail- able, but it can be related to the Gamma function which is included in the syntax: B(M,N) = r(m). (n)/ (m+n) f(M) = J"exxm-ldx 0 - 26 - FIGURE 1 DISiRIBUTION OF INCOME AT DIFFEREN7 LEA ELS OF PER CAPITA GDP % Gini MEP .rjUo 100I Gini t-atio 050 ' *! fIncone reci.pviis 90 .. '. *0.43 40 0 7XX~~~~~~~~~~~~~~~~~~~~~~~~~~~~'04 i 38 60 . Top 20 % 't, *'0-t.4.3 , of Incomc reclptents ' : SO * T 3 < | *. ,r O.40 .- 32 40 \-*. ,0.X8: * 30 Middle 20 % _* of Income recipients 30 * r4 ..36 28 .5~ ~ ~ ~ ' Bottom 60 %. o1 Inecone recipients. 10 'rO 32 L 24 , '> O0 .4 .04_3_ C jlOa sto-200| $2.01-300 5301-5co S5S01D-n;a 5 w GDP .r (U.S. $ 1965) India GDP per capita (for 1973-74 in $U.S. 1965) SOURCE: Cross-sectional data on 56 countries presented by Paukert. (See source on Table 9.) NOTE: MEP is the maximum equilization percentage defined as the percentage of total income that would have to be shifted between quintiles to achieve equal income distribution. * 27 - TABLE 10 GINI COEFFICIENTS FOR INDIVIDUAL CONSUMPTION EXPENDITURE (1973/74) Rural Urban National 0.287 0.320 0.301 The national estimate is very close to the figure from Ranadive (0.308) and indicates a continuation of the trend of increasing equality. :Over .twenty years (1953/54 to 1973/74), the Gini ratio of individual *expeniditure has fallen at an annually compounded rate of 0.55%. The pattern is reinforced by observations of the ratio for intermediate years. 1/ An improving distribution is at variance with cross-sectional relationships between GDP per capita and the Gini ratio derived from total ,income (Figure 1). Countries at income levels similar to India's experience more inequality as income rises. The reconciliation of the longitudinal trend with the cross-sectional pattern probably lies in the difference-between consumption expenditure and incom"e. The multi-country study. was based on total income. The personal income distribution could be worsening (suggested by Ojha and Bhatt, Table 9)r while the consumption distribution improves. Some researchers, however, also show India's income distribution becoming more egalitarian which would make India somewhat unique when viewed against a background of cross-sectional data.2/ Changes in expenditure shares and Gini coefficients for three 1/ However, Gini ratios for the more recent period 1977/78 were estimated with the same technique using the decile distribution of private consumption from the 32nd NSS Round reported in the Sixth Five Year Plan, 1980-85. The rural, urban, and national ratios were 0.31, 0.34, and 0.32, which would imply a reversal of the trend. However, these estimates may not be comparable since it is not stated whether the deciles are based on the household or the individual population dis- tribution, and biases may have been introduced during interpolation to deciles. 2/ A recent Bank paper on income distribution in India concludes there has virtually been no trend. V. V. Bhanoji Rao, "EPD Income Distribu- tion Project Data on Income Distribution in India," Division Working Paper No. 1980-2, p. 19. However, Ahmend's figures show some improve- ment through the mid-1960's. - 28 - rates of change in the Gini ratio are presented in Table 11. The middle case, -0.5%/year, matches the historical movement. Somewhat over 2.0% of the total income is transferred from the upper 15% of the income distribution to the poorest 45% of the population by 2000 at this modest rate. A faster decline in the Gini ratio, -1.0%/year, would approximately double the size of the transfer. B. The Effects of Income Redistribution on Demand Projections. A more egalitarian income distribution would tend to reduce demand for some preferred and luxury items, and expand the consumption of basic. commodities, compared with a base case of an unaltered income distribution. 1/ A falling Gini ratio for personal expenditures would dampen the growth of demand for fruits and nuts, beverages, clothing, and milk. The demand for fuel and light, footwear, miscellaneous goods and services, rents and durable goods would also slacken freeing more revenue in the economy for the increased consumption of other goods. Projections for these items, however, are not reported since base-level quantities of consumption for most of these categories were not available or interpretable at a national level. 2/ The demand for clothing is affected more by redistribution than the other commodities studied. Most of the effects of moderate redistribution on consumption are relatively minor. Demand for foodgrains would rise 1.0% and 1.8% by 1984/85 given the present and an accelerated rate of improvement, respectively in the Gini ratio. The increase at the turn of the century would be 2.3 and 4.0%, respectively. Since the Gini ratio for personal consumption expenditures has actually been falling by 0.55% per year, the projection adjustments given in the left-hand columns of Table 12 may prove applicable. The projection tables for a falling Gini ratio, with 3.5% and 5.0% expenditure growth, appear in Appendix I. A more detailed discussion of the sensitivity of demand to income redistribution is contained in the following section. 1/ No distinction is made between income and expenditure for the purpose of this discussion. 2/ Although projections are not given in quantity or value terms, they are presented as multiplicative increases over the 1973-74 level. - 29 - TABLE 11 FUTURE EXPENDITURE SHARES AND GINI RATIOS GIVEN REDISTRIBUTION RATES R U R A L U R B A N Gini Ratio Bottom TOP Gini Ratio Bottom TOP Population Share 42.5 16.9 47.6 16.2 1973/74 Expenditure Share 24.8 33.7 26.3 36.0 Gini Ratio 0.287 0.320 1999-2000 Gini Change + 0.5%/Year Expenditure Share 22.3 36.1 23.4 38.7 Gini Ratio 0.327 0.364 Gini Change - 0.5%/Year Expenditure Share 27.0 31.7 28.7 33.6 Gini Ratio 0.252 0.281 Gini Change - 1.0%/Year Expenditure Share 28.9 29.8 31.0 31.4 Gini Ratio 0.221 0.246 Notes: Redistribution followed the rule: si = si-(AGIG)(P -S.) where S = starting expenditure share of ith group; i Si = expenditure share after redistribution a G/G = proportionate change in Gini ratio over the period; Pi = population share of the i th group. See Section II.C for further details. To avoid possible interpolation errors, the table is based on population share divisions aggregated from the reported data. (Interpolation would have resulted in exact decile or quintile shares.) It is demonstrated in the paper documenting the demand model that the redistribution rule satisfies the condition that the new Gini ratio G , based on expenditure shares S equals G+ANG. - 30 - TABLE 12 CHANGES IN QUANTITIES DEMANDED RESULTING FROM INCOME REDISTRIBUTION a/ Gini Ratio Declines Gini Ratio Declines 0.5%/Year b/ 1.0%/Year 1984/85 1999-2000 1984/85 1999/2000 (% Change) Pulses 0.8 1.8 1.4 3.1 Edible Oils 1.0 2.3 1.8 3.9 Meat, Fish, Eggs 0.3 0.4 0.3 0.7 Vegetables 0.8 1.8 1.5 3.2 Fruits, Nuts -1.3 -2.5 -2.5 -4.5 Sugar, Khandsari 0.0 0.5 0.2 0.9 Gur, Other Sugars 1.0 2.4 2.1 4.4 Spices 0.0 2.3 1.1 3.8 Beverages 0.0 -3.3 -2.3 -5.6 Tobacco, Pan, Intox 0.0 1.0 0.0 1.0 Clothing -2.9 -4.2 -4.7 -7.5 Milk -0.5 -0.7 -0.9 -1.3 Rice 1.2 2.8 2.3 5.0 Wheat 1.1. 2.8 2.2 5.0 Maize 0.4, 0.6 0.5 1.1 Sorghum & Millet 0.3 0.4 0.4 0.7 Sweeteners 0.7 1.6 1.3 2.8 Foodgrains 1.0 2.3 1.8 4.0 a/ Demand projections compared in this table assume a 3.5% total expenditure growth rate. b/ This rate corresponds to the historical trend over a 20-year period computed with personal consumption expenditure data. The effects of redistribution actually reflect changes in the marginal propensities to consume (MPC) at different income levels. Twenty percent of additional income would be spent on clothing in rural areas at 160% the mean income level in the survey (MPC = 0.20) whereas only 6% - 31 - would be allocated to clothing at 60% of the mean income level. 1/ Hence redistribution significantly reduces the demand for clothing. The corresponding high and low MPCs for gur and other sweetners are 0.01 and 0.03. This opposite ordering implies an elevated demand would accompany redistribution. Likewise, in a neutral case in which demand was virtually unaltered, the MPCs are similar across a range of incomes. The rural MPCs for meat, fish and eggs at 160% and 60% of the mean income, for example, are 0.027 and 0.034. IV. SENSITIVITY ANALYSIS Sensitivity analysis was performed on the projections for foodgrains, edible oils, sweeteners, and the share of wheat in foodgrains by varying the assumptions regarding population growth, urbanization, expenditure growth, and income distribution. The proportions of rice and pulses demanded out of total foodgrains were also considered. The population parameters were altered in a few discrete cases, but sensitivity analysis for expenditure growth and income distribution proceeded from a set of 27 projections with a blend of assumptions. Regressions were then run treating the set of demand levels as observations on a dependent variable. This approach yielded elasticities of future demand with respect to rates of expenditure growth and "income" 2/ redistribution. In most cases, sensitivity analysis was referenced to the projected demand in the year 2000. 3/ A. Population Growth and Urbanization. Population growth rates were varied by +0.1% and +0.05%, and urbanization by +1% of the total population per five years. 4/ A very high rate of urbanization was also tested in which the urban share was 2% higher than the base case per five years of projection. 5/ The span of the assumptions used for sensitivity 1/ The income levels selected for the comparison were chosen arbitrarily. The asymmetry of 60% versus 160% is intended to reflect some of the skewness of the income distribution. 2/ Technically, expenditure redistribution is being modeled since data savings behavior were not collected. 3/ The previous section discussed the effects of two rates of income redistribution on the complete set of commodities. 4/ In the "plus" case, for example, the urban share of total population would be 4.0% higher at the end of 20 years. 5/ See the right-hand column of Table 13. Table 13 POPULATION AND URBAN SHARE VALUES FOR SENSITIVITY ANALYSIS Population With Urban Share With Base Case Growth Rate Varied Urbanization Rate Varied 1/ Year Population Urban Share -0.1% -0.05% +0.05% +0.1% -1. O%5 yrs +1.0%/5 yrs +2.0%/5 yrs --million -------- .-(%)M _ ____-__ million…---------- ---------…M -(-%-)----- 1973/74 595.6 20.6 595.6 595.6 595.6 595.6 20.6 20.6 20.6 1979/80 672.2 22.3 668.2 670.2 674.2 676.2 21.3 23.3 24.3 1984/85 744.2 24.3 736.1 740.1 748.3 752.4 22.3 26.3 28.3 1989/90 820.5 26.9 807.5 813.9 827.1 833.7 23.9 29.9 32.9 1994/95 897.7 28.1 879.1 888.3 907.2 916.7 24.1 32.1 36.1 1999/00 973.6 30.0 948.7 961.0 986.4 999.2 25.0 35.0 40.0 …__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _(% )…__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Change in 2000 -2.6 -1.3 1.3 2.6 -5.0 5.0 10.0 1/ The adjustment for 1979/80 is spread over more than five years. - 33 - analysis is indicated in Table 13. At the extremes, the total population in year 2000 is varied by +2.6% and the urban share by -5% to +10%. The projections for total foodgrains demand vary in almost the same proportion as population. The composition of foodgrains demanded is hardly affected by altered population growth rates. Demands for sweeteners and edible oils are relatively insensitive to population assumptions compared with demand for foodgrains. These results are rather intuitive. The effects of urbanization rates are more interesting. Projections for total foodgrains, sweeteners, and edible oils are modified more by a 5% change in the urban share of population than by a 2.6% change in total population. Unfortunately, estimating the future urban share is more difficult and uncertain than forecasting the total population. Rapid urbanization lowers the growth in demand for foodgrains and sweeteners and raises the growth in demand for edible oils. The proportion of wheat demanded in total foodgrains would also increase somewhat, at the expense of demand for coarse cereals. 1/ Overall, the projections for these commodities are fairly insensitive to reasonable variations in population growth and urbanization rates, which supports the approach of using one set of population parameters for the primary presentation of results. 2/ Further testing confirmed that the effects of changing the population and urban share figures were independent and additive, and indicated that the compound result could therefore be inferred from Table 14. For instance, if rapid population growth coincided with faster than expected urbanization, the impact on foodgrain demand would be offsetting, resulting in a -0.3% change from the base case (+1.8% from population growth -2.1% from 1/ Caution must be used when interpreting the effects of urbanization since abstracting the differences in demand patterns between urban and rural consumers involves altered relative prices, subsistence consump- tion opportunities, income, and commodity availability, all of which affect the acquisition of new "tastes and preferences". In addition, urbanization will have unexplored second round effects on patterns of production and distribution, and on prices. To the extent that dif- ferences in consumption are explained by differences in income, income growth would have to be faster to allow for the acquisition of urban migration. Yet for the purposes of sensitivity analysis, urbanization and expenditure growth were varied independently. 2/ Therefore, complete sets of alternative projections were not reported in which population parameters were varied simultaneously with expen- diture growth rates. The number of cases under consideration was thereby reduced. Table 14 SENSITIVITY OF PROJECTIONS FOR YEAR 2000 TO ALTERED POPULATION AND URBAN SHARE ASSUMPTIONS 1/ Population Growth Rate Varied Urbanization Rate Varied Year 2000 (% per year) (% of Total Per 5 Years) Base ----------------------------- ------------------------ Case -0.1 -0.05 +0.05 +0.1 -1.0 +1.0 +2.0 Foodgrains (mmt) 190.66 187.31 188.98 192.34 194.00 194.71 186.71 182.87 % rice 43.8 43.8 43.8 43.8 43.8 43.9 43.7 43.6 % wheat 27.0 27.1 27.0 26.9 26.8 26.7 27.3 27.7 % pulses 10.6 10.7 10.7 10.6 10.5 10.6 10.7 10.8 Sweeteners (mmt) 25.93 25.85 25.89 25.97 26.00 26.48 25.41 24.91 Edible oils (mit) 6.56 6.52 6.54 6.59 6.61 6.39 6.73 6.89 ________(% )…(-- - - - - - -- - - - - M ---%) … --- Change in foodgrains -1.8 -0.9 0.9 1.8 2.1 -2.1 -4.1 1/ These projections all assume 3.5% total expenditure growth and no change in income distribution. - 35 - urbanization). The effects of population pressure on arable land might lead to the expectation that the rates of population growth and urbanization would be positively related. The foodgrain projections are satisfactorily robust with respect to popujlat-ion assumptions even if trends have a reinforcing effect on demand. Reinforcement would occur if population growth were accelerated and urbanization were retarded, or vice versa. Provisional census figures have recently been released reporting a population of 683.8 million as of February 1981. 1/ The census implies a higher-than-expected population growth rate of 2.2% over the seventies and about 0.1% higher for the last half of the decade than used in the demand projections. Revised population projections incorporating the census data have not appeared yet, but it is unlikely that they will reflect a 0.1% higher growth rate than earlier projections for the rest of the century. However, if this more rapid trend continues, demand for foodgrains may be 1-2% higher (2-4 million tons) by 2000. 2/ B. Expenditure Growth and Income Redistribution. The projections of foodgrain demand are sensitive to the expenditure growth assumptions. The demand for wheat and pulses as a proportion of total foodgrains in 2000 rises by about one-half of a percent (say 2 million tons) per one-half percent rise in the expenditure growth rate over the projection period. This generalization approximately holds for a wide range of growth rates extending from 2.5% to 6.0%. The proportion of rice in foodgrain demand is more stable and only begins to drop at the higher expenditure growth levels. These trends indicate that rapid economic expansion in real terms would stimulate a marked shift in consumption patterns in favor of wheat (Table 16) by the turn of the century. Future demand for sweeteners and edible oils is more critically influenced by expenditure assumptions than projected foodgrain consumption (Table 15). The higher expenditure elasticities for these two less essential foods are of course responsible for their sensitivity. 1/ Census of India 1981, Series 1, Provisional Population Totals. 2/ The assumptions regarding urbanization could not be checked against the provisional census figures since the urban population has not yet been reported. - 36 - TABLE 1 5 Sensitivity Comparisons Increase in Demand at 6% Expenditure Growth over Demand at 2.5% Growth 1984/85 1999/2000 …(%)…------ M ------------- Foodgrains 7.5 17.7 Sweeteners 18.2 76.2 Edible Oils 15.1 58.6 In order to study the interactions of income redistribution and expenditure growth in more detail, a set of 27 projections was produced using a grid of changes in the Gini ratio and expenditure growth rates. 1/ The projected levels of demand in 2000 were then regressed on these parametric variables. The Gini variable was converted to a multiplier over the time period to facilitate interpretation. For example, an original income redistribution variable was -0.5 signifying a 0.5% per year drop in the Gini ratio. The ratio is then 12.2% lower in the year 2000, so the value 0.878 was used in the regression. Expenditure growth was left in percent per year terms. The elasticities are shown in Table 17. According to these elasticities, if expenditure growth were 20% higher than the mean in the simulation (1.2 x 4.22% = 5%/year) foodgrain demand would increase by about 4% (20% x 0.20 = 4%), relative to the mean projection in the simulation set. This relationship is approximately confirmed by projection runs at 5% expenditure growth and a fixed Gini ratio. Interpreting the elasticities of demand with respect to Gini ratios is even more direct since the mean of the multiplicative variable in the parametric set is nearly one. 2/ Thus a 10% decrease from the mean in the regressions is in fact a 10% lower Gini ratio in year 2000. This improved distribution would raise foodgrain consumption by about 1.8%. The elasticities of projected demand with respect to modelling variables 1/ The previous section discussed the expenditure share implications of changes in the Gini coefficient (Table 11). 2/ This is the rationale behind the change in the variable definition. A multiplier of one of course corresponds to a zero percent change in the Gini ratio. Table 16 SENSITIVITY ANALYSIS OF EXPENDITURE GROWTH RATES 1/ Demand in million tons Given Expenditure Growth 2/ Expenditure Growth (%) 3/ 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 1984/85 Foodgrains 132.56 134.04 135.49 136.94 138.37 139.78 141.15 142.50 % rice 43.7 43.7 43.7 43.8 43.8 43.8 43.8 43.8 % wheat 24.7 24.9 25.0 25.2 25.3 25.5 25.6 25.8 % pulses 9.5 9.6 9.7 9.8 9.8 9.9 10.0 10.1 Sweeteners 15.26 15.65 16.04 16.43 16.82 17.22 17.62 18.03 w Edible Oils 3.83 3.91 3.99 4.08 4.16 4.24 4.33 4.41 1999/2000 Foodgrains 177.30 184.31 190.66 196.19 201.02 204.71 207.24 208.67 % rice 43.7 43.8 43.7 43.7 43.5 43.2 42.8 42.3 % wheat 25.8 26.4 26.9 27.5 28.0 28.6 29.1 29.6 % pulses 9.9 10.2 10.6 11.0 11.4 11.8 12.3 13.0 Sweeteners 21.65 23.75 25.93 28.20 30.55 32.99 35.52 38.15 Edible Oils 5.63 6.09 6.56 7.04 7.52 8.00 8.47 8.93 1/ The expenditure distribution is held constant. 2/ Except where demand is indicated as a percent of foodgrains. 3/ After 1979/80. Between 1973-74 and 1979-80, 3.5% growth was used in all cases. Table 17 ELASTICITIES OF PROJECTIONS Elasticity of Demand in Year 2000 Mean Demand With Respect to: 1/ In Simulation Set For: Gini Change Expenditure Growth -- million tons -- Foodgrains -0.18 0.20 197.84 Percent wheat -0.04 0.17 27.8 1/ Sweeteners -0.17 0.53 30.68 Edible Oils -0.12 0.64 7.58 Mean Parametric Values 2/ 0.948 4.22 1/ Percent 2/ Gini ratio change is measured as a multiplier over the period; expenditure growth is measured for the whole economy in percent per year. - 39 - compress a volume of sensitivity analysis into a readily interpretable form. Referring to Table 17, it can be concluded that: 1) A percentage change in the Gini ratio over the entire period would have an opposite and nearly equal effect on the demand for foodgrains as a similar proportionate change in the annual expenditure growth rate. (A 10% fall in the Gini ratio multiplier over the period would have an effect similar to a 10% rise in the mean expenditure growth rate from 4.2% to 4.6%: demand would increase by 2% in year 2000.) 2) The sensitivity of demand for foodgrains, wheat, I/ sweeteners, and edible oils is fairly similar with respect to changes in the income distribution. (An improved distribution would boost demand among the lower income groups for all items but would decrease demand among upper income groups especially for commodities with high marginal propensities to consume.) 3) The sensitivity of demand with respect to expenditure growth rates, unlike the response to redistribution, varies greatly between commodities, as expected. Appendix III contains the "data" used in the elasticity calculations, i.e., the set of parameter combinations and projection results. Graphing isoquants of projected demand against expenditure growth and altered Gini ratios is another approach to interpreting sensitivity analysis (Figure 2). The slopes of the isoquants describe the relative importance of the two modelling variables. The flatter the slope, the less important economic growth is relative to income redistribution since small changes in the Gini coefficient would offset larger changes in total expenditure and leave demand unaltered. The isoquant map conveys results from a large number of projection runs. The Gini ratio change scale used in Figure 2 is matched against annual change rates in Table 18. The discussion of historical rates of redistribution in the previous section highlights a scale value of 0.88 (-0.5%/yr) as a likely possibility. Although severe redistribution of income, improving or worsening, would have a decided effect on demand, the probable range of outcomes is narrow. The expenditure growth rate therefore remains the most critical and uncertain parameter in the projection model. l/ The demand for wheat would respond in a fashion similar to the demand for foodgrains in view of the low elasticity of the percent of wheat in the total. - 40 - FIGURE 2 PROJECTION ISOQUANTS FOR YEAR 2000 Gini 1-2 ~~~F O O D G R A I N S, __ _ 1.2 - a Gini Change 1.1 1.0 .-./ - --0.9 5/-- 0.7 - / i / , 07 I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I I - -Expenditure * ( 2.5 3.0 3.5 4.0 4.5 5.0 5.5 Growth Gini- 1.2 -- ' t ~S W EIE T E N E R S I f - P | i i e Expenditure j .-'A | ~2j5 3;0 315 4,0 4,.5 5t0 5.5 Growth ~!E D I B L E O I L S Gini -1.2 - W 4T N R Change . - , -'--1.0 ' - I -° -0.9 4. _ _ _ -~~~~-------4-- ~~~Expenditure 2,5 3.0 3.5 4.0 4.5 5.0 55 Gxowth NOrTES: Expenditure growth is measured as the annual percentage in- crease for the entire economy compounded from 1979-80 to 1999-2000. The change in the Gini ratio is measured as a multiplicative factor over the total period. The isoquants illustrate constant levels of demand. - 41 - TABLE 18 Correspondence of Multiplicative Factors and Annual Gini Ratio Change Rates Annual Percentage Change by 2000 Change 1.29 1.0 1.14 0.5 1.00 0.0 0.88 -0.5 0.77 -1.0 V. CONCLUSIONS The absence of prices is the most obvious and important limitation to projections based on cross-sectional expenditure data. Relative prices are necessarily assumed to be fixed, whereas they would respond to changing consumption patterns, production levels, factor prices, trade flows, and government policies in a number of areas. Not only are relative prices assumed fixed over the projection period, but prices between consumers in the survey should be the same for expenditures to carry a constant quantity interpretation. One step in overcoming this obstacle has been taken: separate urban and rural projections have been combined using quantity weights for some commodities. Key assumptions were that real per capita expenditure growth would be equal in rural and urban areas, and that any income distribution which did occur would follow a symmetric and proportional re-allocation rule in both sectors. Projections for 1979/80 for cereals, pulses, edible oils, and sweeteners are comparable with actual consumption if some allowances are made for the poor crop year. Demand in terms of calories per capita was closely aligned with other base period estimates and future projections, projections were fairly robust with respect to moderate variations in population growth and income redistribution. This study and projections by the National Commission of Agriculture indicate a 15% increase in per capita calorie demand to 2300 calories per day by year 2000 under the low economic growth alternative. - 42 - Foodgrain demand would grow by 2.6% through the mid-1980s, and then slow to an average rate of 2.3%. The total increase in the last two decades of the century would be 60-70% if real expenditure growth is 3.5-5.0% per year. The composition of foodgrains demanded would shift away from maize, sorghum, and millet in favor of wheat and pulses. If foodgrain demand grew by 60-70%, wheat demand would grow by 75-100%, and coarse cereals by 33-16%. There would be less increase in demand for the less: preferred grains if expenditure growth were more rapid. The proportion of rice demanded in foodgrains would remain almost fixed at 43-44% under a wide range of economic growth assumptions. At a 5% expenditure growth rate, the proportion of wheat demanded would increase 5% by 2000, and the share of coarse cereals would drop 7%. One percent of foodgrain demand would be 2 million tons in 2000, so the shift would be considerable. Faster urban migration would slightly dampen foodgrain demand, ceteris paribus. Demand for sugar/khandsari and edible oils is expected to be very buoyant even given slow expenditure growth. In the case of edible oils, this trend has strong implications for future prices and/or imports. Improvement in the expenditure distribution would affect demand according to differences in marginal propensities to consume at various, income levels. These differences cannot be inferred from expenditure elasticities at the mean. 1/ A continuation of the 0.5%/year drop in the Gini ratio would contract demand (relative to the constant ratio case) for fruits and nuts, beverages, clothing, milk and milk products by a few percent and slightly expand demand for edible oils, gur, rice and wheat. The predominant determinant of the projections and the primary source of uncertainty, is the expenditure growth rate. Elasticities of the projections with respect to the growth rate indicated that demand for foodgrains, sweeteners and edible oils would respectively be 4, 2, and 0.5 million tons higher at the turn of the century for each half of a percent increase in this critical parameter. 1/ The elasticities were one or above for edible oils, sugar/khandsari, and milk/milk products, but redistribution had positive, neutral and negative effects, respectively. - 43 - A P P E N D I C E S - 44 - API'ENDIX I Demand ProjecLions with Moderate Redistribution The following tables give the complete set of projections with low and high expenditure growth, and the historical rate of redistribution. FOR THE FOLLOWING PROJECTION' 7OTAL EXPEtoDrTURE GROWTH ASSUMPTION IS 3.5 PERCENT GINI RATIO CHANGE RATE IS -0.5 PERCENT OUTPUT IN MILtlION ME1RIC TONS 73/74 79/80 84/85 89/90 94/95 99/00 PULSES 3/ 8.71 11.40 13 29 15.42 17.89 20.61 EDIBLE OILS .2.67 3.37 4.03 4.81 5.69 6.71 MEAT. FISH, & EGGS 2.40 2.71 3.20 3.92 4.70 5.64 VEGETABLES 6.93 7.88 9.28 10.93 12.17 14.89 FRUIlS & NUTS 16.57 18 36 22 83 28.53 35.58 44 74 SUGAR & ''HANDSARI 4.52 5.25 6 44 7.91 9 61 11.69 CUR & OTHER SUJGARS 7 34 8 38 9.70 11.13 12.86 14.66 SPICES 0.69 0.77 0.89 1.03 1.17 1.33 BEVERAGES 0.32 0.34. 0 43 0.54 0.68 0.87 TOBACCO. PAN & INTOX. 0.46 0.51 0 60 0.71 0 84 0.99 CLOTHING 0.72 0.82 1.02 1.27 1.61 2.04 MILK 19.75 23.16' 2S.59 35.26 43 55 53.78 RICE 40.36 52.58 60.04 68.01 76.94 85.87 WHEAT 22.54 29.37 34.36 39.96 46.21 52 91 MAIZE 4.04 5.10 5.52 5.89 6.31 6.63 SORGHUM e MILLET 4/ 17.15 21.G2 23 58 25.43 27.38 23.95 OTHER 1/ - 1.00 1.05 1.29 1.59 1.95 2.42 SWEETENERS 11.85 13.63 16.15 19.03 22.47 26.35 FOODGRAINS 92.80 120 08 136.79 154.71 174.73 194.97 FOR THE FOLLOWING PROJECTION TOTAL EXPENDITURE GROWTH ASSUMPTION IS 5 PERCENT GINI RATIO CHANGE RAlE IS -O 5 PEOCENT OUTPUT IN MILLION METR!C TONS 73/74 79/80 1/84/85 89/90 94/95 99/00 PULSES 3/ 8.71 11.40 14.02 17.07 20.70 24.76 EDIBLE OrIS 2.67 3.37 4.28 5.40 6 68 8.19 MEAT, FISH. & EGGS 2.40 2.71 3.52 4.57 5.88 7 56 VEGETABLES 6.93 7.88 9.79 12 It 14.78 17.90 FRUITS & NUTS 16 57 18.3G 25 60 35 85 49 99 70.04 SUGAR & KHANDSARI 4.52 5.25 7.05 9.38 12.32 16.07 CUR & OTHER SUGARS 7.34 8.38 10.29 12.40 14.91 17.46 SPICES 0.69 0.77 0.93 1.10 1.29 1.48 BEVERAGES 0.32 0.34 0 .18 0.68 0.96 1.38 TOBACCO, PAN & INTOX. 0 40 0.51 O064 0 81 1.02 1 29 CLOTiHING 0.72 0 82 1.18 1 69 2.44 3.50 MILK 19.75 23.10 31.97 43 71 59.45 80.30 RICE 40.36 52 58 62 03 71 99 82 28 91.29 WIIEAT 22.54 29 37 36 12 43.67 51 90 60.26 MAIZE 4.04 5 10 5.49 5.80 6.11 6.24 SORGHIUM i MILLET 4/ 17.15 21 62 23.44 24.99 26 38 27.14 OTIIER / - 1.00 1.05 1.42 1.95 2 66 3.66 SWEEtENERS 11.85 13.6i 17.34 21.78 27.22 33.53 rOODGRAINS 92.80 120.08 141.09 163.53 187.37 209.69 1/ Irojcctions for "OLher" are niultiplicr.tivc factors on a 1973-74 base. 2/ Assunm:s 3,5% expenditure -rowth through 1979-30. 3/ Includes cereal substitutes and grams. 4/ Includes barley and other coarse cereals. NOTE: Grossing factors for non-household demand for foodgrains and edible oils were added. - 45 - APFENDIX II Gini Ratio Estimation Program The following is a listing of a program to estimate rural, urban, and combined Gini Ratios. The program is called "GINI.EST.PGM" and runs on RAL. 50 FETCH GINI EST 60 RPS:NUMBERS(14)-NUMBERS(14) 61 UPS-RPS 62 RES-RPS 63 UES-RPS 70 G=(0.0.O) 100 RP=WEIGHTRUR*PERSONPERHHR 200 UP=WEIGHTURB*PERSONPERHHU 300 RE=RP*SECTORTOTALR 400 UE=UP*SECTORTOTALU 500 FOR I = I TO 14 00 BEGIN 600 RPS(I)=SUM(RP(1 TO 1))/SUM(RP) 700 UPS(I)=SUM(UP(1 TO l))/SUM(UP) 800 RES(I)=SUM(RE(1 TO 1))/SUM(RE) 900 UES(I)=SUM(UE(1 TO 1))/SUM(UE) 0Oo END 1100 TPS=.794*RPS + 206*UPS 1200 TES=.743*RES + .257*UES 1300 VARP=("RPS`,"UPS"."TPS") 1400 VARE=("RES","UES"."TES") 1500 FOR 1= 1 TO 3 DO BEGIN 1600 P=VARP(I) 1700 E=VARE(I) 1800 ETA=(#P#-//E#)/SORT(2) 1900 FRUITPI=(#P#+#E#)/SQRT(2) 1950 FRUITPI=FRUITPI(1 TO 13) 1951 ETA=ETA(1 TO 13) 2000 RUN MULTREG ON LN(FRUITPI),LN(SORT(2)-FRUITPI),LN(ETA) & 2100 OPTIONS= SUPPRESS SAVE(COEFFICIENTS) 2200 G(I)=2=EXP(COEFFICIENTS(1)).(SORT(2))** & 2300 (l+COEFFICIENTS(2)+COEFFICIENTS(3))*GAMMA(1+COEFFICIENTS(2)) & 2400 *GAMMA(1+COEFFICIENTS(3))/GAMMA(2+COEFFICJENTS(2) & 2500 +COEFFICIENTS(3)) 2600 END 2700 DISPLAY G The data called in line 50 give urban and rural population and expend- iture shares, RP, UP, RE, and UE. Cumulative shares have variables ending in "S", and the total shares "TPS" and "TES" use weights given in lines 1100 and 1200. The rest of the syntax corresponds to the discussion in Section III, given the following: (Note that lines 60-70 simply load zeros into vectors) = ETA = FRUITPI a = ln(COEFFICIENT(l)) O = COEFFICIENT(2) = COEFFICIENT(3) The "GAMMA" in the program is the gamma function, not the gamma variable used in the text. - 46 - APPENDIX III Sensitivity Analysis Data The following is a listing of data from file PROJ.2000. Cl El G2 E2 F WPER E S -1.5 2.5 0.675 1.90 186.52 26.2 5.91 22.51 -1.5 3.5 0.675 2.45 200.99 27.3 6.90 26.93 -1.5 4.5 0.675 3.14 212.48 28.4 7 93 31.75 -1.5 5.0 0.675 3.56 216.52 28.9 8.45 34.32 -1.5 5.5 0.675 4.02 219.33 29.4 8.97 36.91 -1.0 3.5 0.770 2.45 198.29 27.2 6.82 26.67 -1.0 5.0 0.770 3.56 213.48 28.8 8.34 33.96 -0.5 2.5 0.878 1.90 181.08 26.0 5.75 22.00 -0.5 3.5 0.878 2.45 194.97 27.1 6.71 26.35 -0.5 4.5 0.878 3.14 205.71 28.2 7.69 31.06 -0.5 5.5 0.878 4.02 212.23 29.3 8.68 36.10 0.0 2.5 1.000 1.90 177.30 25.8 5.63 21.65 0.0 3.0 1.000 2.16 184.31 26.4 6.09 23.75 0.0 3.5 1.000 2.45 190.66 26.9 6.56 25.93 0.0 4.0 1.000 2.77 196.19 27.5 7.04 28.20 0.0 4.5 1.000 3.14 201.02 28.0 7.52 30.55 0.0 5.0 1.000 3.56 204.71 28.6 8.00 32.99 0.0 5.5 1.000 4.02 207.24 29.1 8.47 35.52 0.0 6.0 1.000 4.55 208.67 29.6 8.93 38.15 0.5 2.5 1.138 1.90 172.55 25.6 5.48 21.23 0.5 3.5 1.138 2.45 185.06 26.8 6.38 25.40 0.5 4.5 1.138 3.14 194.86 27.9 7 30 29.91 0.5 5.0 1.138 3.56 198.43 28.5 7.75 32.31 0.5 5.5 1.138 4.02 200.90 29.0 8.20 34.81 1.0 3.5 1.295 2.45 178.13 26.6 6.16 24.78 1.0 5.0 1.295 3.56 190 41 28.3 7.45 31.52 -0.5 5.0 0.878 3.56 209.69 28.7 8.19 33.53 Each line contains output from a projection run for the year 2000. A large number of projections created a simulation data set which was used to find elasticities of projections with respect to assumptions, and projection isoquants in Section IV, The variable definitions are: Gl Annual rate of change of Gini ratio for personal consumption expenditure, in percentage terms. El Annual percentage growth rate in total expenditure. G2 Multiplicative change in Gini ratio by 2000, E2 Multiplicative change in total expenditure by 2000. F Foodgrains projection for 2000, in mat. WPER Wheat percentage of foodgrains demanded in 2000. E Edible oils projection for 2000, in mmt. S Sweeteners projection for 2000, in mmt, - 47 - APPENDIX IV Foodgrain Equivalents Calorie Weights 1/ Pulses 0.97 Edible Oils 2.47 Meat, Fish, Eggs 0.42 Vegetables 0.08 Fruits, Nuts 0.17 Sugar, Khandsari 1.08 Gur, Other Sweeteners 0.98 Milk 0.17 Rice 1.00 Wheat 1.02 Maize 1.00 Sorghum, Millet 0.97 1/ Calorie weights are relative to rice which has about 357 kcals/100 g. Although there is some nutritional value in spices, beverages, and pan, they were assumed to make a negligible contribution to the diet. Source: FAO, Food Composition Tables for International Use, 1954. Foodgrain equiValents can be found by replacing the grossing factor vector in line 1380 of COMB.PROJ.PGM with these weights, with zeros inserted for other items. (Grossing factors are, thereby, implicitly assumed to be one since non-household demand is not included in estimates of calories per capita.) Line 1470 is replaced by a vector of ones, and the "Foodgrains" output can then be interpreted as rice equivalents in million tons. Multiplication by 9780.82 and division by the population in millions converts rice equivalents to calories per capita per day. - 48 - APPENDIX V List of Consumption Groups and Items - 3. I In NSS, d tta oni CQ5iUTIICI ccIpelditLire at c iist uII ly collcet cdI lot .I la t C numbe of ite m. But in plesentiig the I csni t- inil.i it itsim have been nicigeti toh III 21 hono,i4tn,,u' gi OlpS 'I Te gi o I,p, are ;1) ccrca Is, (2) gi A in, ,3) t c lea I su,tit utist., ( 1l Pluc a- Itl pI oduct-(s, (5j n 'k a id pt odutts, (G6 cdi, PC oil, (7) mcat, fith -I Id cgg (8) vc,cthlesIC%, (9) [liiit ai tt inlits, (10) S I.al , II) salt, (12) .phes, l3 beverages an iJf Icshnsiciuts, (14) pan, tobtcrcoarid Iti toxIc.itn Is, (15) fucl a idI Itlt, (16) clot Ih g, (17) footwear, 18) tilitccl laicous goods .niis SC ,ILCs, (,19) iciits, (20, taxcs, (21) ( tli.ib!c goods. Rcsults aI c also prcscnted for (') food-tottal, (ll) tiowi-food tottal alid (iII) tot,sl cotisutncexpetdititI C. 3.2 The groups of itcms of conlsumptioi as picsented in tlie table atc defiIed IIel il terms of their constitucits: 1. Cereats:iicc, vlscat,josvar,bbata, rnlaizc,batley,smallmillcts, tagi niul tllcii piodticts, 2. Gfan: bengal-giami anid its pioduicts, 3. Cerealsubstitules: Crcalsubstitute Iikc topioca,pea ctc., 4. Puleasandprod otsc :athl:r,tut,granr,tnooisg, inasoor, uid, khscs.irl, pea, soyabitiatsd othcr pulscs, and thici 1p oducts, 5. Milk andPlrosdurts liquid rilk(cow,buffalow, goatanlls othets),baby-foocl inik(conidensec po%vdei etc.) glice (cow and buffslo,, buttcer, tl sit, ghol, lassi and othel milk pioducts, 6. Edible oil: vanaspati, nintstatd oil, cocoanut oll, gingilIy oil, gtoutidnit oil, linsecd oil, scscfiit oil, other edible oil axid oil seeds, 7. Aleal, Jisit and eggs mcat (goat nieat, muttotn, becf, pork, buffalo inca t and other nmcat , Cgg poultry, fish (freih anid dried), biid and otlscis, 8. Vegetablcs potato, uniois, tomato, brinjal, cabbage, cauli-flo%ser, toot Negetables (arstr radish, etc.) Icafy vegetablCs atid other %cgetables, 9. Fruitsatd nuts :banana,oiange,lemon, mango, cocoanut,guava,pincappilc.grapes,othierfres fruits,cocoanutcop. Is,groundisut,c ashzvwtiut,d-teo,iaisin,otierdtyfruiltsatid nuts, 10. Su 9ar :sw!mi (fi Ctoils kh t-! :-:,C Zs!^-,..;' : -;8bu-I II. Sall sca salt, rock salt and other salt, 12. Spices cururinic, bl.tck peppet, pepper, dty chillies, giceti ceiillics, gsilc, tasmatiid,giniger, curiy powvder, otlter spices, 13. lBtosrages atl refresLnictils lea (,Io. ofCups), tet leCI, Cc,ffcc (t1o. of Clips5, coffcL potsdci . other d riskitig be-a,-i.,bIscitts, coafect noin,tcs et .. s.alted Icli ltiticILts, picpaicdb%ssets, cooked sticalst pickle, saucc, j tIltS andLI jeleI(s, OthCt pt ocCstd lod tteiiis, 14. Pan,tobacco antdints antls p.i iclaf, p fii linished,,supa i (betcl ilit,otlis ingredicnts fot pati) ls,i,cig icttet, Itftobicco, tookilt tob.tcco, (Ieitoot, slff, Zs d.tdls, kimarnin su,ti otlier tobi,c,o pio- ducts,opium, gatsji, rodlds, co.i,stt y liquoi,loicigi liqtior, othci drugs and intoxticasts, 15 Fuel atid lhgit coke, coal, hicssood, clectricit), gas, dlunig cake, charcoal, ketoscien, calllilc, nAtchcs, nsethylated spUlt, otlhcr fuel atd IgIt, IG. Clothing :cototl (mnilli sslie,rklloosrl,land-loom, klsadi) stool,al tlkrac oit om anyotlc, si,jtlletictextilc, parc tlk,ansdotlhcsttllclttdtiig itctnsol beddttga Lidulptholstcly, 17. Footwcar boot, shoe, slippei, sstidlc, chappal, swoodets sanidal etc., 18. 4Alscella,sjeous goods andsertices amusements (ciascissa, tlhcatre etc.) cducationl, niediciite, toilct at ticles, sunct latticlcs, coissunie scvicts, (oncyances, ctc., 19. Rt,it: rets on I ciidentnal house, re.id.sitial Lind anicl other consunisem goods (no coniputlatio ofremt fot rc,dent,lal houscs owned by thc ..ampic househioldl s%as m.sdc), 20. Taxes liecnce fees for keeping *,un, rAdio, cycle. tnotor car etc , tiltd other co,isuniser taxes tikc tihe JiLmaiCIpal taxes, toad taxes etc., but dueb not icitide incomc tax, 21. Durable gools furniture, itnuicil inistrunimits, ornaments, utensile anlc othes c.quipmcnt antd thcir I eairaiv, expctiscs including tic mnaintenaisce ofrCstdleiti.,l houses. l/ This is the original set of definitions for the expenditure survey, reproduced from the 28th Round of the National Sample Survey document. - 49 - REFERENCES Atkinson, Anthony B., "On the Measurement of Inequality," Journal of Economic Theory, Vol. 2 (1970), pp. 244-257. Blakeslee, Leroy, Heady, Earl, and Tramingham, Charles, World Food Production, Demand and Trade, Center for Agriculture and Rural Development, Iowa State University, Ames (1973). Census of India 1981, Series 1, Provisional Population Totals, GOI, New Delhi (1981). Desai, B. M., "Analysis of Consumption Patterns in India," Occasional Paper No. 54, Department of Agricultural Economics, Cornell University (August 1972). FAO, Food Balance Sheets, 1972-74, Rome. FAO, Food Composition Tables for International Use, Rome (1954). Hitchings, Jon, "Documentation of a Demand Projection Model Prepared for India," The World Bank ASADB (Divisional Paper, March 1981). Jain, Shail, Size Distribution of Income: A Compilation of Data, The World Bank (1975). Kakwani, N. C., and Podder, N., "Efficient Estimation of the Lorenz Curve and Associated Inequality Measures from Grouped Observations," Econometrica, Vol. 44, No. 1, (January 1976), pp. 137-148. Macgregor, John, "Agricultural Demand Projections for India," The World Bank, ASADB (Draft Divisional Paper 1979). Mellor, John W., "Agricultural Price Policy and Income Distribution in Low Income Countries," World Bank Staff Working Paper No. 214, (September 1974). National Commission on Agriculture, GOI, Demand and Supply, Part III, Delhi (1976). National Sample Survey Organization, Department of Statistics, GOI, Tables on Consumer Expenditure, 28th Round, No. 240, New Delhi (1977). - 50 - Paukert, Felix, '"Income Distribution at Different Levels of Development: A Survey of Evidence," International Labor Review, (August- September 1973), pp. 97-124. Ranadive, K. R., "Pattern of Income Distribution in India, 1953/54 to 1959/60, "Bulletin of the Oxford University Institute of Economics and Statistics, (August 1968). Rao, V. M., Food, Second India Studies, Delhi, Bombay: The MacMillan Co. of India, Ltd., (1975). Rao, V. V. B., "EPD Income Distribution Project, Data on Income Distribution in India," The World Bank,(Division Working Paper No 1980-2). Rojko, Anthony, et al; Alternative Futures for World Food, 1985, Vol. I, United States Department of Agriculture (1978). Sanderson, Fred H., and Roy, Shyamal, Food Trends and Prospects in India, Washington, D.C.: The Brookings Institution (1979). Scandizzo, Pasquale L., and Bruce, Colin, "Methodologies for Measuring Price Intervention Effects," World Bank Staff Working Paper No. 394 (March 1980). World Development Report 1980, The World Bank, (August 1980). PART II THE FOODGRAIN ECONOMY James Q. Harrison and John W. Wall South Asia Programs July 1981 51 - 52 - The Foodgrain Economy Table of Contents Page Introduction, Summary and Conclusions ............... 53 Supply-Demand Balances . . .57 Government Projections ... 58 Implications .. . 61 Table Annex 1 . .65 Table Annex 2 . ....66 - 53 - The Foodgrain Economy Introduction, Summary and Conclusions. Foodgrains dominate Indian agriculture in their share of the value of crop output, area cultivated and area irrigated (see Table 1), and in their overriding importance in Indian diets, particularly the diets of the poor. Developments in the foodgrain economy during the 1970s have raised hopes that a basic constraint limiting development in India -- domestic scarcity of staple food -- may have eased. In the last half of the 1970s, foodgrain production met domestic consumption, government stocks grew large, small quantities of grain were exported, and foodgrain prices fell relative to other prices in the economy. It is important to know whether this situation is a temporary result of unusually good weather or whether it is a lasting result of forces stimulating foodgrain production and controlling demand growth. Table 1 Pattern of Land Use and Output Mix by Major Crops (percent) Percent of Crop Share in: Gross Value of Total Area Total Area Percent of Crop Crop Output Cultivated Irrigated Area Irrigated (1977/78) (1977-79) (1978) (1978) Foodgrains 59.3 74.4 78.7 27.3 Oilseeds a/ 10.1 9.8 2.8 7.5 Sugarcane 6.3 1.8 5.6 79.9 Cotton 4.2 4.5 3.8 21.8 Other 20.1 9.5 9.1 24.7 Total 100.0 100.0 100.0 a/ Excluding cotton seed which appears separately in this table as cotton. Sources: Central Statistical Organization and Ministry of Agriculture. Graph 1 presents actual production and apparent consumption for the past twenty years along with alternative projections of supply and 54 Graph 1- FOODGRAIN PRODUCTION AND APPARENT CONSUMPTION o 0 N NQ PRODUCTION ; . .......... CONSUMPTON O - SUPPLY PROJ. HIGH 01 SUPPLY PROJ. LOW \ NQ ----------- DEMAND PROJ. HIGH ........... DEMAND PROJ. LOW // Z~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ;° _ X"" _ O tO _ 60 63 66 69 72 75 78 81 84 87 90 93 96 99 FISCAL YEAR - 55 - demand for foodgrains for the next twenty years.1/ The historical data reveal a convergence of the supply and demand lines in the last half of the 1970s. The projections suggest the possibility of continued self-sufficiency or even an emerging surplus in foodgrains in the future. Nevertheless, bad weather in 1979/80 resulted in a drop of 23 million tons or 17% from the record production of foodgrains achieved in 1978/79. Through the judicious use of foodgrain stocks built up over the past several years, the Government was able to contain the worst effects of the drought -- preventing serious famine, moderating foodgrain price increases, avoiding foodgrain imports, even managing to export a modest amount of grain. As creditable as this achievement is, the drop in production which made it necessary should serve as a reminder that the balance between the supply and demand for foodgrain is still precarious. While the performance of the recent past and the probable future trends point to India's self-sufficiency and emerging export potential in foodgrains, the balance remains delicate and a succession of poor monsoons could still result in a need for foodgrain imports to maintain consumer supplies or adequate buffer stocks. In this.context, the decision to resume imports in 1981/82 to rebuild stocks drawn down following the 1979 drought was a prudent measure to avoid speculative price pressures. The weaknesses of the 1981 monsoon made the need for imports even clearer. The resort to imports to maintain foodgrain supplies indicates the Government's continued commitment to use market mechanisms for this purpose rather than resorting to administrative controls on foodgrain movement and compulsory procurement. Changes in foodgrain prices will cause these projections to vary. The projections in Graph 1 have been made for illustrative purposes assuming prices remain essentially constant in real terms. In reality, prices will operate on both the supply and demand side, shifting both to offset persistent large imbalances. As can be seen in Graph 2, the real prices of foodgrain have varied in the past. They rose in real terms from the early 1960s (a period during which relative prices of agricultural commodities were among the lowest since Independence) through the unprecedented drought years early in 1967. They then fell in response to the Green Revolution until the spate of bad weather and constrained imports of the early 1970s, when they again began to rise. Good weather beginning in 1975, along with a well-designed pricing policy backed up with a large public stocks, initiated a fall in the real foodgrain price 1/ Apparent consumption is defined here as gross availability of foodgrains in the economy -- gross production plus net imports less changes in public stocks. - 56 - Graph 2 REAL PRICE OF FOODGRAINS RIAL PRICE OF FOODGRADS ..- ----- --- ---- TRED Ln 196 TO 90 o o 44~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~- 0- 60 62 64 66 68 70 72 74 76 78 so CALENDAR YEAR nffDI OF LMSM PRII oF P0OIMRAS DID Y STTU WHLEAL PRI:8 CM OF ALL NO9PODRI COM' - 57 - that has continued through 1980 despite bad weather and a short crop in 1979/80. The trend rate of decline of the real prices from 1968 through 1980 is 2.2% per annum. A continuation of a low, and even falling, real price of foodgrains is one option facing India. A lower real price has a beneficial impact on equity of consumption, as foodgrains occupy a dominant place in the consumption budget of the poorest families in India. By the same token, lower prices will stimulate demand and dampen the production increase so that the projected surplus would disappear. Supply-Demand Balances. Historical trend rates of growth of foodgrain production vary considerably from as low as 2% to as high as 3%, depending on the time periods chosen for calculating the rates. Finding rates as low as 2% is a result of including a very bad year at the end of the period and rates as high as 3% is a result of including a very good year at the end, as well as a bad year at the beginning. The supply lines in Graph 1 are based on two alternative trend rates, 2.3% per annum and 2.7% per annum. The lower rate is slightly below that calculated over the period 1967/68 through 1977/78, both fairly good years.l/ The high rate is that calculated over 1949/50 through 1979/80, a period which incorporates all of the information available. 2/ The demand projections presented in Graph I are taken from Jon Hitchings' "Demand Projections for India", the first paper in this volume. The method of projection and underlying assumptions are discussed in detail there. The main difference between the low and high cases is the assumed rate of growth of total consumption expenditure. The low case is based on a 3.5% per annum growth, which is just below the historical trend. The high case is based on a 5% per annum growth in total consumption expenditure, a rate roughly consistent with the Sixth Plan projections and achievable under stimulative economic policies. 1/ If production in 1978/79, a very good year, is added to the period of calculation, the trend rate becomes 2.7% per annum. If production in 1979/80, a very bad year, is added, the trend rate drops to 2.1%. If production in 1980/81 is taken to be 133 million tons and added, the trend rate is again 2.3%. 2/ This is also the rate calculated for the period 1960/61 through 1978/79. If production for the year 1979/80 is added, the rate falls to 2.6%. If production for 1980/81 (assumed to be 133 million tons) is added, the rate remains 2.6%. This indicates that the sensitivity in the rates to individual years falls as the period is lengthened. - 58 -- It should be noted that the high demand line probably is not economically consistent with the low supply line. Foodgrain production is such a large proportion of the value of agricultural output (c.60%) and agriculture constitutes such a significant portion of value added in the economy (c.40%) that to generate a 5% per annum growth in consumption expenditure in practice depends on a fairly rapid growth in foodgrain production. Slowly growing production of foodgrain dampens the rise in total income, and consumption expenditure, in proportion to its weight in value added. It also causes the price of India's main wage good (grain) to rise, dampening demand and raising costs of producing other goods so that overall growth is constrained. Moreover, foodgrain imports, when needed, take priority over other imports necessary for sustained rapid growth elsewhere in the economy. The Sixth Plan, which projects a 4.7% per annum growth in consumption expenditure through 1984/85 and 5% per annum from 1984/85 through 1994/95, projects foodgrain output to rise between 3.1% and 3.7% through 1984/85 and between 2.9% and 3.5% from 1984/85 to 1994/95. These are the rates of growth for foodgrain production internally consistent with the target rates of growth of expenditure. While it is logically conceivable for the economy to grow rapidly while foodgrain production lags, the history and structure of the Indian economy make this unlikely. Consequently, the relevant supply line to compare with the high demand line is. the high supply projection, as it is probably not possible to reach the high demand line if production is as low as on the low supply line. Government Projections. Government supply projections are higher than either of the trend lines discussed above (see Table 2). The National Commission on Agriculture (NCA) and the Planning Commission both project foodgrain production based on an analysis of projected land use, irrigation availability, technological potential and fertilizer consumption. In the cases of NCA supply projections, rough rules of thumb were used to convert these factors into expected foodgrain production. 1/ The production targets from the Sixth Five Year Plan 2/ are based on the production technology and inputs expected to be available during the Plan period, although the Plan document is not explicit about the methods used in the projections. 1/ In the NCA projections, one additional hectare of area under irriga- tion is assumed to add 0.5 tons of foodgrain production and one addi- tional ton of fertilizer applied to foodgrains is assumed to yield 10 additional tons of grain. (Report of the National Commission on Agriculture, 1976, Part III, pp. 96-98. 2/ Sixth Five Year Plan, 1980-85, p. 25 and p. 119. - 59 - The NCA supply projections assume little change in the total area under foodgrains from the 1970/71 base period to the year 2000, but the proportion of foodgrain area irrigated is assumed to rise from about 25% in 1970/71 to 42% by the year 2000. It is notable that while virtually all of the projected increase is from an increase in average yields rather than area, the yield increases are expected to be generated primarily by applying increased inputs of water and fertilizer, rather than by any substantial change in technology itself or any increased efficiency in input use. The assumed rates of growth of fertilizer (7% to 8% per annum) and irrigation (1.5 million hectares per year) are well below actual achievements in recent years. The two NCA projections differ in assumed rates of growth of fertilizer consumption between 1970/71 and 1984/85, with the low path assuming a rate of growth of 7.2% and the high path a rate of 10.2% per annum. For the entire 1970/71 to 2000 period, the rate of growth of fertilizer consumption is projected to be 6.3% to 7% per annum. Both of NCA's supply paths converge by the year 2000. The Sixth Plan targets, while lower than the NCA projections, are higher than past trends. The lower target of 149 million tons implies a rate of growth of 3.1%. While such growth rates have been achieved in past periods, as noted above such periods usually included a base year well below trend or an end year well above trend. The links between Government policies, programs and input supplies on the one hand and output on the other are not explicit in the Plan document. The Plan targets for agricultural input use are quite ambitious, with irrigated area expected to rise an average of 2.7 million hectares per year, fertilizer consumption to rise 12.9% per annum, gross cropped area to rise 10 million hectares and total area under high yielding varieties of foodgrains to rise more than 20 million hectares. Other agricultural support programs are expected to increase their coverage similarly. The NCA demand projections are considerably higher than those of Hitchings. The NCA projections are based on growth in consumption expenditure at lower rates than Hitchings' (NCA high: 4.7% per annum through 1980 and 3.5% per annum from 1980 through 2000; NCA low: 3.0% per annum through 1980 and 2.8% per annum from 1980 through 2000). Nevertheless, the projected quantities are higher due to a much higher starting base and a much greater provision for animal feed than in Hitchings' projections. The base year chosen for the NCA projections, 1971, had one of the highest per capita consumption levels in recent times. Starting from that base the NCA projects demand in 1979/80 at 133 (low) to 140 million tons (high). Hitchings' projections used 1973/74 as a base and project 1979/80 consumption at 120 million tons. Apparent availability (gross production plus net imports minus change in stock) was 114 million tons in 1979/80, well below trend due to the drought. Actual consumption no doubt was higher than 114 million tons, allowed by some drawdown of private stocks. Although the extent of drawdown of private stocks is not known, trend consumption cannot be much greater than the - 60 - Table 2 Alternative Projections of Foodgrain Supply and Demand through 1990/2000 (in million tons) 1979/80 1984/85 1989/90 1994/95 1999/2000 Supply Projections National Commission on Agriculture High 143 163 185 207 230 Low 135 150 176 203 230 Sixth Plan High 128 154 - 205 - Low 128 149 - 205 - Trend High (2.7% per annum) 125 143 163 186 213 Low (2.3% per annum) 125 140 157 176 197 Demand Projections National Commission on Agriculture High 140 163 183 204 225 Low 133 150 168 187 205 Hitchings High 120 140 161 184 205 Low 120 136 153 172 191 Sources: National Commission on Agriculture, Planning Commission and Jon Hitchings', "Demand Projections for India". - 61 - trend of availability. Trend availability in 1979/80 would be the same as trend production (125 million tons), given the lack of imports. Consumption in 1979/80 will be below trend due to the fall in income. Consequently, actual consumption in 1979/80 is likely to have been much closer to Hitchings' projected 1979/80 base of 120 million tons than the NCA's 133 to 140 million tons. The NCA projections incorporate provision of foodgrain for animal feeding, derived from an accelerated and modernized animal development program contained in the NCA report, that grows to reach 20 to 25 million tons by the year 2000. World Bank estimates, based on animal husbandry sector work, suggest animal feed requirements considerably lower than this, perhaps less than half. The Hitchings demand estimates presented here incorporate an allowance for seed, feed and waste of 12.5% of gross production. The provision for animal feed from this is generally placed at 5%, or 9.5 million tons and 10.2 million tons for the low and high projections, respectively, by the year 2000. The NCA also makes a slightly greater allowance for wastage, 4% (8 to 9 million tons by 2000) than Hitchings' allowance of 2.5% (4.8 to 5.1 million tons). In summary, it is unlikely that foodgrain consumption will reach the levels projected by the National Commission on Agriculture. The Sixth Plan's projection of demand for 1984/85 of 150 million tons is similarly unlikely in light of Hitchings' demand analysis. Consumption is likely to reach the levels projected under Hitchings' high demand alternative only if growth of foodgrain production is also fairly high. The high supply projection easily meets Hitchings' high demand projection, just as the low supply projection covers his low demand projection. Government projections of supply lie considerably above either of the supply projections. Although the Government's demand projections also lie considerably above Hitchings', they are probably even less plausible than those for supply. Implications. The projections suggest the real possibility of India moving into a period of sustainable foodgrain self-sufficiency or surplus. This outlook is substantially more optimistic than many past projections. As a new conclusion, it deserves to be treated with caution and interpreted with perspective. Foodgrain production will still depend on the moonsoon and a failure of the monsoon could lead India to import grain. As Graph 1 indicates, actual foodgrain production has fluctuated significantly around the trend. In some years, weather conditions are likely to make it necessary to draw down stocks and/or import, while in years of good monsoons stocks are likely to be replenished and/or some grain exported. The analysis contains new projections based on the prospects for population growth, household expenditure patterns and production which lead one to the optimistic conclusion as far as market supply/demand balances are concerned. The optimism of this conclusion should not translate into complacency about the prospects for eliminating - 62 - poverty and undernutrition in India by the year 2000. The projections imply rising per capita consumption of foodgrains, and this, together with the possibility of falling real prices for foodgrains discussed below, would have a beneficial impact on poverty and nutrition. But the projected rise in per capita consumption would take place from a base which is very low on average and pitifully low for the poorer segments of the population. Nor should there be complacency that the projected supply/demand balances will come about without continued efforts on the supply side to develop the technological base, input supply system, irrigation infrastructure and incentives needed for sustained production growth and, on the demand side, to moderate the rate of growth of population. Foodgrain self-sufficiency or surplus, if it comes about, will be an impressive achievement of a long standing goal of Indian economic development. It indicates that India, by continuing its major efforts in agriculture, may be in the desirable position of having a range of options which were not available during the period when domestic foodgrain supplies were inadequate. Among these options are the following: (a) Increase Consumption. A slowly falling real price of foodgrains would absorb a greater quantity of foodgrains in household consumption and slow production growth, so that the projected surpluses would not arise. The past several years have already seen a modest decline in real foodgrain prices and rise in per capita consumption, and the current level of household foodgrain consumption is approximately equal to a caloric-sufficient quantity, on average, for the Indian population. With unequal distribution of this consumption, this means undernutrition for lower income groups. It is the lowest income groups, who spend the greatest proportion of their income on foodgrains, for whom lower real prices are the most important. Effects of lower price on producer incentives are also important but so far at least technical change has offset the decline in producer prices and has kept profits sufficiently high to encourage greater production. (b) Exports of Foodgrains. The Bank projections are not reliable enough to predict exports, which are a small residual between two very large and fluctuating variables, demand and supply. Nevertheless, the prospects of foodgrain exports are real ones. The Sixth Plan projects the export of 3 million tons of rice by 1984/85 which, on the basis of the Bank projections, appears feasible from the point of view of domestic supply, keeping in mind the caveats concerning fluctuations in individual years. The prices at which the Government procures foodgrains relative - 63 - to world market prices become important in this regard. Currently, Indian procurement and market prices for rice are not higher than world market prices, and rice could be exported without losses to the Government. (c) Diversification in Agriculture. The demand projections contained in the first section of the paper indicate the demand for several other important agricultural commodities will grow faster than for foodgrains (see Table 3). Most farmers grow a variety of crops in which foodgrains are dominant (Table 1). Changes in foodgrain production technology or prices can increase the relative profitability of foodgrains, causing farmers to plant a greater proportion of their area to them. Alternatively, farmers who see greater profit opportunities in other crops can apply yield-increasing technology to meet their own foodgrain needs on less area, thereby freeing land for other crops. The introduction of shorter duration varieties of foodgrains (especially paddy) gives farmers greater opportunities for planting a second crop. Relatively small shifts in foodgrain area can mean substantial shifts in area and production of other crops. For example, 1% of the foodgrain area is equivalent to 8% of the oilseed area, 17% of the cotton area or 40% of the sugarcane area. Thus management of foodgrain supply and demand is of central importance for both foodgrains and for most other agricultural commodities as well. Moreover, if the foodgrain situation is under control, India's substantial agricultural infrastructure, with its technical (research and extension), physical (irrigation and other investments), logistical (input supply and distribution) and economic (pricing policy, procurement and distribution) dimensions, has more flexibility to meet the needs of other crops. - 64 - Table 3 Projected Annual Growth Rates of Demand for Selected Commodities (in percent) 1979/80 to 1984/85 1984/85 to 1999/2000 Low High Low High Total Sweetners a/ 3.4 4.9 3.3 4.4 Edible Oils 3.5 4.8 3.4 4.3 Clothing 4.8 7.8 4.8 7.6 Foodgrains 2.6 3.2 2.3 2.6 a/ Sugar, khandsari, gur and other sweetners. Source: Jon Hitchings "Demand Projections for India". None of these options will be available to India without sustained efforts to raise foodgrain yields. Any significant slackening of the rate of growth of productivity would return the foodgrain.economy to its previous, unhappy position of scarcity, fluctuating prices and large cereal imports. Continued work on basic and adaptive research, seed development, irrigation and extension will be required to fulfil the promise of an emerging foodgrain self-sufficiency or surplus. Annex Table I AREA, PRODUCTION AND YIELD OP FOODGRAINS, 1949/50 TO 1979/80 (Area - '000 Hectares Production - '000 Tons Yield - Kg per Hectare) RICE WHEAT COARSE CERALS PULSES TOTAL FOODGRAINS A P Y A P Y A P Y A p Y A P Y 1949/50 30519 23542 771 9758 6391 655 38836 16824 433 20167 8159 405 992801 54916 553 30810 20576 668 9746 6462 663 37674 15376 408 19091 8411 1441 97321 50625 1522 52 29830 21300 714 9471 6183 653 3988 16093 414 18775 6420 449 96961 51996 1536 53 29969 22899 764 9828 7501 763 42446 19612 462 19845 9189 463 102088 59201 580 54 31289 28214 902 10691 8017 750 45366 22972 506 21729 10619 489 109065 69921 640 t65 30764 25219 820 11259 9043 903 43921 22823 520 21914 10950 500 107858 68035 631 56 31521 27557 874 12367 8760 708 43456 19498 448 23216 11045 476 110560 66850 605 i; 32277 29037 800 13524 9403 695 42019 19864 473 23316 11551 495 111136 69955 629 B.E1 32298 25525 790 11730 7998 682 42914 21226 495 22538 9562 424 109480 64311 587 33172 30847 930 12617 9958 7 44664 23187 519 2431t 13149 541 114764 77141 672 no 33820 31676 937 13360 10324 772 43790 22973 522 24833 11799 475 115823 76672 662 61 34128 34574 1013 12927 10997 851 44963 23743 1529 23563 12704 539 115581 82018 i710 62 34694 35663 1028 13570 12072 990 44725 23216 519 24243 111755 485 117232 82706 '705 63 35695 33217 931 13590 10776 793 44294 24630 556 24265 111528 475 117844 80151 '680 64 35809 36999 1033 13499 9853 730 43927"23718 540 24186 10073 !416 117421 90642 16B7 65 36462 39308 1078 13422 12257 913 44353 25374- 572 23875 12417 520 118112 89356 757 66 35470 30589 862 12572 10394 827 44343 21420 483 22718 9944 438 115103 72347 629 67 35251 30438 863 12838 11393 887 45092 24053 533 22121 8347 377 115302 74231 644 . 68 36437 37612 1032 14998 16540 1103 47337 28798 608 22649 .12102 534 121421 95052 783 69 36967 39761 1076 15958 18651 1169 46241 25183 545 21264 '10418 490 120430 94013 791 70 37680 40430 1073 16626 20093 1209 47241 27287 578 22023 11691 531 123570 99501 805 -1 37592 42225 1123 18241 23832 1307 45949 30547 665 22534 11819 524 124316 108422 872 72 37758 43068 1141 19139 26410 1380 43575 24596 564 22151 11094 501 122623 105168 858 73 36688 39245 1070 194631 24735 1271 42211 123139 548 20915 9907 474 119277 97026 813 74 38286 44051 1151 19853 21778 1172 46242 '28828 623 23427 10008 427 126538 104665 827 75 37999 39579 1045 18010. 24104 1338 43152 26129 606 22024 10014 *455 121075 998261824 76 39475 43740 1235 20454 28846 1410 43798 35409 808 24454 13039 533 128181 121034 944 77 33511 41917 1088 20922 29010 1387 46940 28979 615 22963 11361 494 124356 111167 894 78 40292 52671 1308 21456 31749 1480 42280 30014 710 23497 11973 510 127515 126407 991 79 40480 53770 1328 22640 35510 1568 42230 30440 721 23660 12180 515 129010 131900-1022 so 38980 42190 1082 21960 31560 1437 41220 26730 1648 21750 8370 385 123910 108950 878 Annual Z Rates of Growth 1949/50 - 1979/80 0.9 2.6 1.7 2.8 5.9 3.0 0.2 1.8 1.6 0.3 0.4 0.1 0.8 2.7 1.9 1949/50 - 1964/65 1.3 4.0 2.6 2.6 4.2 1.6 0.9 2.8 1.9 1.7 2.4 0.7 1.4 3.4 2.0 1967/68 - 1979/80 0.5 1.8 1.2 3.0 5.3 2.3 -0.9 0.9 1.8 0.5 -0.6 -1.1 0.4 2.1 1.7 1967/68 - 1978/79 Q.5 2.4 1.7 3.2 5.7 2.5 -0.8 1.3 2.1 0.8 0.5 -0.3 0.5 2.6 2.1 Sources: Estimates of Area and Production of Principal Crops in India, 1978-79 published by Directorate of Economics and Statistics, Ministry of Agriculture; and, Economic Survey, 1980-81, published by Government of India. Annex Table 2 FOODORAIN AVAILABILITY Calendar Foodgrain Net Changes in Gross Net Per Capita Year Production d/ Imports Public Stocks Availability a/ Availability b/ Net Availability c/ - ------------------------ Thousand Tons ------------------------- Grams per Day 1950 54916 2160 -870 57946 51082 390 51 50825 4800 590 55035 48682 365 *52 51996 3930 620 55306 48807 360 '53 59201 2040 -480 61721 54321 394 54 69821 830 200 70451 61723 440 55 68035 600 -740 69375 60871 426 56 66850 1390 -600 68840 60484 415 '57 69855 3630 860 72625 63893 429 58 64311 3210 -270 67791 59752 393 59 77141 3860 490 80511 70868 456 60 76672 5130 1400 80402 70818 446 61 82018 3490 -170 85678 75426 465 62 82706 3640 -360 86706 76368 460 63 80151 4550 -20 84721 74702 440 64 80642 6260 -1240 88142 78062 449 1 65 89356 7450 1060 95746 84577 475 0' 66 72347 10340 140 82547 73504 404 1 67 74231 8660 -260 83151 73872 396 68 95052 5690 2040 98702 86821 455 69 94013 3850 460 97403 85651 439 70 99501 3580 1120 101961 89523 448 71 108422 2030 2570 107882 94329 461 72 105168 -490 -4690 109368 96222 461 73 97026 3590 -310 100926 88798 417 74 104665 4830 -400 109895 96812 445 75 99826 7390 5560 101656 89178 402 76 121034 6440 10270 117204 102075 451 77 111167 410 -1250 112827 98931 429 78 126407 -1000 -270 125677 109876 468 '79 131900 -940 360 130600 114113 475 80 108850 -420 -5670 114100 100494 410 a/ Gross production plus net imports minus stock changes. b/ Gross production minus 12.5% allowance for seed, feed and waste plus net imports minus stock changes. c/ Net availability divided by mid-year population divided by 365 days. d/ Fall-harvested crops of the previous year are added to the Spring-harvested crops of the calendar year shown. Part III THE VEGETABLE OIL ECONOMY John W. Wall June 1981 67 - 68 - The Vegetable Oil Economy Table of Contents Page No. Introduction .. ............. 69 Supply Trends. . . . 70 Demand Conditions . ........ . . ..... . . .......... 71 Supply/Demand Balance . . . .71 Supply Prospects ....74 Oilseeds Production . . . .74 Oilseeds Processing . . . .76 Pricing Policy... 78 Medium-Term Management of the Vegetable Oil Economy and the Plan............. . 82 - 69 - Introduction. Oilseeds constitute one of the most important groups of agricultural commodities in India. Oilseeds (including cottonseed and coconut) occupy about 15% of gross cropped area and contribute about 10% of the value of output from agriculture. The oilseeds group contains a large number of field crops: groundnut, rape and mustard, castor, sesame, linseed, niger, safflower, sunflower, cottonseed and soyabean. In addition vegetable oil is extracted from rice bran and a variety of tree crops: coconut, oil palm and various forest seeds such as sal and neem. Vegetable oil is an item of mass consumption in India and is the main source of the basic human require- ment for essential fatty acids. Most vegetable oil is consumed directly as food, although the demand for vegetable oil for soaps and paints is rising more rapidly than for food purposes. A growing proportion of edible oil, reaching about 17% in 1979, is consumed in the form of hydrogenated oil, mainly in urban areas and northern parts of India. The remainder is consumed as either unrefined or refined liquid oil. Total consumption of vegetable oils of around 6 kg per capita is low compared with over 25 kg per capita in developed countries and over 10 kg per capita for the world average. Oilseeds and vegetable oils loom large in India's foreign trade. India is a traditional exporter of oilseed cakes and meals, particularly deoiled groundnut and cottonseed meals; of speciality oils such as castor and linseed; and of handpicked, selected (1iPS) groundnuts. Recently, deoiled rice bran has become a major export. During the period 1975/76 to 1978/79, exports of oilseeds and their derivatives averaged over US$200 million per year, about 14% of all agricultural exports and 3% of all exports. Since 1976/77, imports of vegetable oils have grown large, to 1 million tons or more per year with import values of US$600-800 million through 1979/80. Next to petroleum vege- table oils have been the largest item in the import bill, replacing cereals, which had been the largest item of food imports up to 1975/76. Almost half of all vegetable oil produced in India is derived from groundnut (see Table 1). Groundnut is grown mainly as a rainfed crop in semi-arid areas of western and southern India--particularly Gujarat, Tamil Nadu, Andhra Pradesh and Karnataka--in the kharif (June to October) season. About 11% of groundnut area and about 19% of groundnut produced is planted in January and February and harvested in May (hot weather or summer groundnut). Rape and mustard seed is the next largest source of oil. Rape/mustard is grown in the rabi season (October to April) in the northern and eastern parts of India, particularly in Uttar Pradesh, which alone accounts for more than half of total area. Cotton is grown mainly in Gujarat, Punjab, Maharashtra and Karnataka in the kharif season, mainly for cotton fiber and only secon- darily for the seed. A lower proportion of cottonseed is crushed for oil than other oilseeds as a higher proportion of the seed is fed directly to cattle. Nevertheless an increasing proportion of cottonseed is being processed to extract its oil. Coconut is a perennial crop grown in tropical coastal areas of India, mainly in Kerala, which accounts for more than half of total coconut production, followed by Tamil Nadu and Karnataka. Linseed is grown mainly in Uttar Pradesh, Madhya Pradesh and Maharashtra. Castor is grown mainly in Gujarat and Andhra Pradesh. - 70 - Table 1: OILSEEDS AND VEGETABLE OIL PRODUCTION BY SOURCE, 1978/79 (in %) % of Oilseed % of Gross % of Domestic Oilseed Oil Cropped Area /a Cropped Area Oil Production Yields Yields kg/ha kg/ha Groundnut 28 4.4 47 846 201 Rape/Mustard 13 2.1 18 528 161 Sesamum 9 1.4 5 221 82 Safflower 3 0.4 1 301 104 Niger 2 0.4 1 243 n.a. Soyabean 2 0.3 1 800 /c 120 /c Sunflower 1 0.1 1 800 /c 275 /c Cottonseed 30 4.7 8 336 44 Coconut 4 0.6 5 5,126 /b 615 Linseed 7 1.2 5 254 88 Castor 2 0.3 3 528 183 Rice Bran n.a. 23.2 2 n.a. n.a. Tree Crops n.a. n.a. 3 n.a. n.a. /a Excludes area planted to rice and tree crops other than coconut. /b Nuts per hectare. /c Yield per hectare assumed to be 800 kg of seed. Source: Ministry of Agriculture, World Bank Estimates. Supply Trends. The growth in production of oilseeds has tended to slow over the last decade (see Table 2). Through the mid-1960s oilseeds yields grew very little and almost all growth in production was due to growth in area. Since then, the rate of growth of yields has improved, but not enough to offset the much slower growth in area. Table 2: ANNUAL GROWTH RATES OF AREA, PRODUCTION AND YIELD OF OILSEEDS 1949/50-1978/79 (in %) 1949/50-1978/79 1949/50-1964/65 1967/68-1978/79 Oilseeds Area (A) 1.28 2.65 0.24 Yield (Y) 0.49 0.19 1.25 Production (P) 2.13 3.14 1.61 Source: Based on index numbers of area, production and yield of the oilseed group as given in "Estimates of Area and Pro- duction of Principal Crops in India 1978-79", Directorate of Economics & Statistics, Ministry of Agriculture, Government of India, 1980. The oilseed group contains groundnut, sesamum, rapeseed and mustard seed, linseed, castorseed, safflower, nigerseed, coconut and cottonseed. - 71 - During the period 1967/68 to 1978/79, domestic production of vege- table oil grew on average by 2.5% per annum, substantially faster than total supply of oilseeds. This is due to several factors. An increasing proportion of some oilseeds, mainly cottonseed, has been processed for oil. The intro- duction of solvent extraction methods has meant an increasing proportion of the oil in the seed has become available for use. Finally, there has been a growing extraction of oil from sources other than those included in the above definition of oilseeds. These other sources are minor oilseeds of tree origin such as sal, neem and kusum, and rice bran. Demand Conditions. Household,consumption data indicate that demand for vegetable oils for food increases at roughly the same rate as income, pro- viding relative prices are constant. The growth in consumption would be dam- pened by a relative price rise since households would substitute other goods. With per capita income growing at 1.5% per annum and population growing at 2% per annum or more, growth in total demand would exceed 3.5% per annum, assum- ing no relative price movements. However, from the early 1960s through the mid-1970s, growth in demand of vegetable oils was, constrained by growth in supply, which was lower than 3.5% per annum. Vegetable oil imports were insig- nificant between 1960/61 and 1975/76, so fluctuations in domestic production directly caused fluctuations in domestic supply and domestic prices. In years of short harvests, prices rose rapidly to choke demand back to available supply. Through the mid-1970s, there was a tendency for relative prices of vegetable oils to rise, although the fluctuations were so wide as to obscure any strong trend. These relationships can be seen in Graphs 1, 2 and 3. Starting in 1976/77, large scale imports of edible oil allowed demand to rise by more than domestic supply. Imports in the oil year (November to October) of 1976/77 through 1979/80 were one million tons or more per year, comprising one-fourth or more of domestic supply. Per capita availability rose from an average of five kg for the period 1974 to 1976 to an average of over six kg for period 1977 to 1980. These imports of vegetable oil kept the vegetable oil prices from rising significantly relative to other commodities despite low oilseed production during this period. Indeed some have argued that the large scale import of vegetable oils had adversely affected the production of oilseeds by keeping their prices low. The imports now also contribute to India's very large trade deficit which must be reduced. Supply-Demand Balance. Assuming stable relative vegetable oil prices and based on past trends, World Bank projections of demand lie above projections of supply for the foreseeable future (see Graph 3). The supply projection is based on a simple extrapolation of past trends. DL, the low demand projection, is based on a 3.5% per annum rate of growth in total expen- diture. DH, the high demand projection, is based on a 5% per annum growth in total expenditure. With either demand line the gap between supply and demand becomes quite large. By 1990, filling the gap through imports would mean spending between US$3-4 billion (6-8% of total imports) and absorbing between 8-10% of total projected world exports of vegetable oils. The alternative to importing to fill the gap is to increase the rate of growth of domestic supply and decrease the rate of growth of demand. Decreasing the rate of growth of demand in a country with such low per capita consumption as in India would not be desirable in itself but a consequence of pursuing other policies. - 72 - Graph 1 110 - 100 A Vegetable oils 90 Retative 80 s / V \/ Price Index" 70 _ V 1970/ 71 = 100 19 61 62 64 66 68 70 72 74 76 78 80 (J INDEX OF WHOLESALE PRICES OF VEGETABLE OILS DIVIDED BY INDEX OF WHOLESALE PRICES OF ALL COMMODITIES. Graph 2 Consumption of '7 - Vegetabte oits 6 per capita 4. w- - N 0V1 g l ........ ^<- ~o r3 w4 S V < i * N - C C N C C - C N C C C C C - C _ C C N N C C N C C - C C N C N - C _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ C_C C_ _ _ _ _ _ __C N_C C_ _C_ _N_ _C_ C N N N C N Cs * Ni * N 4i N N C C N C CsN . 4 4 N N C 4 . C 4 N C C s w eg > 3 b S u o > t TABLE IC IMPORTS OF VEGETABLE OILS 1960/61 - 1978/79 Unit, THOUSAND TONNES ON OIL YEAR BASIS (i.e Julv - June basis) a/ OIL YEAR GROUND RAPE & SOYBEAN PALM OIL COCONUT COTTONSEED SUNFLOWER SESAME SAFFLOWER NICERSEED SUBTOTAL LINSEED CASTOR RICEBRAN TREECROP GRAND (July-June) NUT OIL MUSTARD OIL 4/ OIL OIL OIL OIL OIL OIL EDIBLE OILS OIL OIL OIL OIL TOTAL (1) (2) OIL (3) (4) (5) (6) (7) (8) (9) (10) (11) (2) to (11) = (12) -TIT 14) (TT5) (16) (T2co (16) 1960/61 N.S. - 0 5 33.7 - 0.3 - - - - 34.5 - - 34.5 1961 /62 0.1 _ 1.7 37 B 1.1 1.0 - _ _ 41 7 - - _ _ 41.7 1962 /63 N.S. - 0.7 29.2 3.6 0.6 - - - - .34.1 - - - 34.1 1963/64 N.S. 0.1 2.1 35.5 1.1 - _ _ _ 38 B N.S. - _ _ 38.8 1964/65 - 0.4 17.6 24 6 0.1 2.0 _ _ 44 7 N.S. - - - 44.7 1965 /66 - 0.4 42.9 6.9 - 6.1 N.S. _ _ _ 56.3 N,S. - _ _ 56.3 1966/67 - 0.3 31.9 10.6 - - N,.S _ _ _ 42 8 N.S. - _ 42.8 1967 /68 - 0.3 49.8 5.0 - - - _ _ 55.1 - - _ 55.1 1968/69 - 01 50.7 0.6 - - N,5. - - - 51 4 NS, - - - 51.4 1969/70 - 0.1 74.1 0.3 0.4 - 09 - 7 _ 15.8 N.S. - _ _ 75.8 1970 /71 - N,s, 84.6 0 5 1.1 _ 2.8 _ _ 89 0 0 1 - - _ 89 1 1971 /72 - N.S. 88.7 0.7 - - - - - - 89 4 0.3 - - - 89.7 1972 /73 - 7.1 52.8 24.2 - - - - - - 84.1 0 2 - _ 84.3 1973 /74 - 21 0 49.0 72.3 - - - - - _ 142 3 0.2 - - - 142 5 1974 /75 0.5 1 5 12.0 14 9 - - - _ 28.9 0 3 - _ 29.2 1975 /76 6.2 10 0 24 6 21.3 N.S. - - _ _ 62.1 0.5 - - - 62 6 1976 /77 35 2 113.2 152.4 145 0 5.3 - 7.1 - _ _ 458.2 0.8 - - - 459.0 1977 /78 66 9 352.6 347.0 444.4 16 5 0.1 29 1 - - 1256.6 1.0 - _ _ 1257 6 1978 /79 bj 8.1 213.2 331.2 375.0 2.8 - 23.0 - - - 953.3 N.A N A. N.A. N A. N.A. 1979/80 c/ 3.0 88 7 250.0 282.3 - - - 624.0 N.A. N.A. N.A. N.A. N.A. a/ See Table ID. b/ Calculated by aultiplying the imports during the periods April 1978 - March 1979 and April 1979 - February 1980 by 0.75 and 0.273 respectively and then adding them. In other vords, the figures for the period April 1979 to February 1980 were blown up by a factor (12 e 11 - 1.0909) to get the figures for the financial year 1980 (i.e. April 1979 - March 1980). c/ Relates to the period April 1979 - February 1980. df Includes palm kernel oil. TABLE ID EXPORTS OF VEGETABLE OIL (ON OIL YEAR BASIS, i e, JULY - JUNE) a/ Unit: Thousand Tonnes OIL YEAR GROUND RAPE/11USTARD SESAME SAFFLOWER NIGERSEED SOYBEAN SUNFLOWER COTTONSEED COCONUT SUBTOTAL LINSEED CASTOR RICEBRAN TREECROP GRAND TOTAL NUT OIL OIL OIL OIL OIL OIL OIL OIL OIL EDIBLE OILS OIL OIL OIL OIL (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (2) to (10) (11) (12) (13) (14) (15) (11) to (15). (16) 1960/ 61 8.15 0.20 N.S. - - - - N.S. 8.35 5.65 41.25 - 0.03 55.28 1961/ 62 22.75 0.20 N.S. - - - - - N.S. 22.95 0.80 26.20 - 0 02 49.97 1962/ 63 70.83 0 20 N.S. - - - N.S. 71.03 1.03 31.43 - 0.02 103 51 1963/ 64 78.60 0.23 N.S. - - - - - 0.03 78.86 0.88 31.38 - 0.04 111.16 1964/ 65 8.53 0 38 N.S. - - - - - 0.13 9 04 0.93 17.23 - 0.02 27.22 1965/ 66 1.50 0.53 - - - - _ 0.20 2.23 0.30 8.30 - 0.02 10 85 1966/ 67 0.75 0.30 0.10 - - - - _ 0.18 1 33 N.S. 1.85 - N.S. 3.18 1967/ 68 0.25 0.38 0.10 - - _ _ 0.10 0.83 N.S. 14.23 - N.S. 15.06 1968/ 69 0.10 0.58 N.S. - - - - - 0.10 0.78 0.05 30.78 - - 31.61 1969/ 70 0.13 0.43 N.S. - - - - - 0.10 0 66 0.15 14.10 - - 14 91 1970/71 0.18 0.28 N.S. - - - - - 0.08 0.54 N.S. 16.18 - - 16 72 1971/ 72 0.10 0.80 N.S. - - - - - 0.02 0.92 N.S. 24 85 - - 25 77 1972/ 73 0.10 1.33 N.S. - - - - - 0.48 1.91 N.S. 43.05 - 0.01 44.97 1973/ 74 0.10 0.20 N.S. - - - - - 010 0.40 4.91 31.43 - 0.15 36.85 1974/ 75 0 10 0.20 0.15 - - - - - 0.10 0.55 22.06 21.00 - 0.58 44 19 1975/76 0.93 0.18 0.45 - - - - - 0.08 1.64 32.50 27.35 - 1.12 62.61 1976/ 77 2.63 0.13 0.10 - - _- - N.S. 2.86 31.58 34 98 - 2.30 71.66 1977/ 78 0.33 0 15 0 15 - - - - - - 1.03 c/ N.S, 14.00 - 2.69 17.72 1978/ 79 b/ 0.43 N.S. 0.01 - - - - - - 2 42 S/ N.S. 5.36 - 1.85 9.63 a/ Obtained by multiplying the export figures for the financial years t and t + I by 0.75 and 0.25, respectively, and adding them For example, to get the figures for oil year 1960/61, we multiply the export figures for the financial years 1960/61 and 1961/62 by 0.75 and 0.25, respectively, and then adding them, b/ relates to the financial year 1978-79, C/ indudes exports of pal. oil. - 90 - Table 2A INDIA - Oilseeds Production - Groundnut Area Production Yield Thousand Thousand tons hectares (in the shell) kg/hectare 1949/50 3979 3433 863 1950/51 4494 3481 775 1951/52 4917 3192 649 1952/53 4795 2929 611 1953/54 4247 3445 811 1954/55 5541 4245 766 1955/56 5133 3862 752 1956/57 5532 4369 783 1957/58 6420 4710 734 1958/'59 6251 5178 828 1559/60 6442 4562 708 1960/61 6463 4812 745 1961/62 6889 4994 725 1962/63 7283 5064 695 1963/64 6886 5298 769 1964/65 7376 6004 814 1965/66 7698 4263 554 1966/67 7299 4411 604 1967/68 7553 5731 759 1968/69 7088 4631 653 1969/70 7125 5130 720 1970/71 7326 6111 834 1971/72 7510 6181 823 1872/73 6990 4092 585 1973/74 7024 5932 845 1974/75 7063 5111 724 1975/76 7222 6754 935 1976/77 7043 5264 747 1977/78 7029 6087 866 1978/79 (Revised) 7433 6208 835 1979/80 (Final) 7238 5772 797 Compound Growth Rates % % % 1949/50-1978/79 1.8 2.0 0.2 1949/50-1964/65 3.9 4.0 0.1 1967/68-1978/79 -0.2 1.3 1.5 1949/50-1979/80 1.7 1.9 0.2 1967/68-1979/80 -0.1 1.2 1.3 Source: Estimates of Area and Production of Principal Crops in India .1978-79, Ministry of Agriculture and World Bank estimates. - 91 - Table 2B INDIA - Oilseeds Production - Rape & Mustard Seed Area Production Yield Thousand Thousand tons hectares kg/hectare 1949/50 1935 806 417 1950/51 2071 762 368 1951/52 2401 943 393 1952/53 2105 858 408 1953/54 2244 872 389 1954/55 2439 1037 425 1955/56 2556 860 336 1956/57 2539 1043 411 1957/58 2412 933 387 1958/59 2447 1042 426 1559/60 2910 1063 365 1960/61 2883 1347 467 1961/62 3168 1346 425 1962/63 3127 1303 417 1963/64 3046 915 300 1964/65 2910 1474 507 1965/66 2913 1298 446 1966/67 3006 1228 408 1967/68 3244 1568 483 1968/69 2870 1347 469 1969/70 3172 1564 493 1970/71 3323 1976 594 1971/72 3614 1433 396 1872/73 3319 1808 545 1973/74 3457 1704 493 1974/75 3680 2252 612 1975/76 3339 1936 580 1976/77 3129 1551 496 1977/78 3584 1650 460 1978/79 (Revised) 3544 1860 525 1979/80 (Final) 3475 1433 412 Compound Growth Rates % % % 1949/50-1978/79 1.9 3.1 1.3 1949/50-1964/65 3.0 3.4 0.4 1965/66-1978/79 1.3 2.6 1.3 1967/68-1978/79 1.0 1.7 0.7 1949/50-1979/80 1.8 2.9 1.1 1965/66-1979/80 1.2 1.8 0.6 1967/68-1979/80 0.9 0.8 -0.2 Source: Estimates of Area and Production of Principal Crops in India 1978-79, Ministry of Agriculture and World Bank estimates. - 92 - Table 2C INDIA - Oilseeds Production - Sesamum Area Production Yield Thousand Thousand tons hectares kg/hectare 1949/50 2046 438 214 1950/51 2204 445 202 1951/52 2405 452 188 1952/53 2377 471 198 1953/54 2570 563 219 1954/55 2626 603 230 1955/56 2293 467 204 1956/57 2172 438 202 1957/58 2094 359 171 1958/59 2250 514 228 1559/60 2136 370 173 1960/61 2169 318 147 1961/62 2252 372 165 1962/63 2552 492 193 1963/64 2412 439 182 1964/65 2486 484 195 1965/66 2509 424 169 1966/67 2794 416 149 1967/68 2654 445 168 1968/69 2423 422 174 1969/70 2309 448 194 1970/71 2433 562 231 1971/72 2392 449 188 1872/73 2288 385 168 1973/74 2386 485 203 1974/75 2234 392 176 1975/76 2170 479 221 1976/77 2279 422 185 1977/78 2384. 520 218 1978/79 (Revised) 2389 514 215 1979/80 (Final) 2384 371 156 Compound Growth Rates % % % 1949/50-1978/79 0.1 0.0 -0.1 1949/50-1964/65 0.3 -0.9 -1.2 1965/66-1978/79 -1.1 0.9 1.9 1967/68-1978/79 -0.7 0.7 1.6 1949/50-1979/80 -0.1 -0.1 -0.2 1965/66-1979/80 -0.9 0.2 1.1 1967/68-1979/80 -0.6 0.1 0.4 Source: Estimates of Area and Production of Principal Crops in India 1978-79, Ministry of Agriculture and World Bank estimates. - 93 - Table 2D INDIA - Oilseeds Production - Linseed Area Production Yield Thousand Thousand tons hectares kg/hectare 1949/50 1521 418 275 1950/51 1403 367 262 1951/52 1380 333 241 1952/53 1362 372 273 1953/54 1387 385 278 1954/55 1361 390 287 1955/56 1529 420 275 1956/57 1682 390 232 1957/58 1283 259 202 1958/59 1595 452 283 1559/60 1974 446 226 1960/61 1789 398 222 1961/62 1977 463 234 1962/63 1904 430 226 1963/64 1995 379 190 1964/65 2042 494 242 1965/66 1723 331 192 1966/67 1495 260 174 1967/68 1777 438 247 1968/69 1697 329 194 1969/70 1803 469 260 1970/71 1897 474 250 1971/72 2064 529 256 1872/73 1726 428 248 1973/74 2038 504 247 1974/75 2071 564 272 1975/76 2119 598 282 1976/77 1888 419 222 1977/78 2010 527 262 1978/79 (Revised) 2092 535 256 1979/80 (Final) 1641 270 165 Compound Growth Rates _ % % 1949/50-1978/79 1.4 1.2 -0.2 1949/50-1964/65 2.9 1.2 -1.6 1967/68-1978/79 1.5 2.5 1.0 1949/50-1979/80 1.3 0.8 -0.4 1967/68-1979/80 0.7 0.1 -0.6 Source: Estimates of Area and Production of Principal Crops in India 1978-79, Ministry of Agriculture and World Bank estimates. - 94 - Table 2E INDIA - Oilseeds Production - Castorseed Area Production Yield Thousand Thousand tons hectares kg/hectare 1949/50 590 130 220 1950/51 555 103 186 1951/52 582 108 186 1952/53 536 104 194 1953/54 545 105 193 1954/55 555 124 223 1955/56 574 125 218 1956/57 569 124 218 1957/58 447 89 199 1958/59 458 112 245 1559/60 492 115 234 1960/61 466 107 230 1961/62 486 109 224 1962/63 469 99 212 1963/64 484 102 211 1964/65 441 107 242 1965/66 405 80 199 1966/67 401 110 274 1967/68 439 121 276 1968/69 394 116 294 1969/70 402 123 306 1970/71 439 136 310 1971/72 453 154 340 1872/73 426 145 341 1973/74 546 229 419 1974/75 590 210 356 1975/76 375 143 381 1976/77 496 179 361 1977/78 380 217 572 1978/79 (Revised) 447 229 512 1979/80 (Final) 438 233 532 Compound Growth Rates % % % 1949/50-1978/79 -1.0 2.1 3.1 1949/50-1964/65 -1.8 0.7 1.1 1967/68-1978/79 0.6 5.9 5.4 1949/50-1979/80 -0.9 2.3 3.2 1967/68-1979/80 0.4 5.9 5.5 Source: Estimates of Area and Production of Principal Crops in India 1978-79, Ministry of Agriculture and World Bank estimates. - 95 - Table 2F INDIA - Oilseeds Production - Five Major Oilseeds (Groundnut, Rape/Mustard Seed, Sesamum, Linseed and Castorseed) Area Production Yield Thousand Thousand tons hectares kg/hectare 1949/50 10071 5225 519 1950/51 10727 5158 481 1951/52 11685 5028 430 1952/53 11175 4734 424 1953/54 10993 5370 488 1954/55 12522 639,9 511 1955/56 12085 573,4 474 1956/57 12494 6364 509 1957/58 12656 6350 502 1958/59 13001 7298 561 1559/60 13954 6556 470 1960/61 13770 6982 507 1961/62 14772 7284 493 1962/63 15335 7388 482 1963/64 14823 7133 481 1964/65 15255 856,3 561 1965/66 15248 6396 419 1966/67 14995 6425 428 1967/68 15667 8303 530 1968/69 14472 6845 473 1969/70 14811 7734 522 1970/71 15418 9259 601 1971/72 16033 8746 546 1872/73 14749 6858 465 1973/74 15451 8854 573 1974/75 15638 8529 545 1975/76 15225 9910 651 1976/77 14834 7834 528 1977/78 15386 9001 585 1978/79 (Revised) 15904 9347 588 1979/80 (Final) 15175 8078 532 Compound Growth Rates % % % 1949/50 - 1978/79 1.3 2.0 0.7 1949/50 - 1964/65 2.7 3.3 0.6 1967/68 - 1978/79 0.2 1.5 1.3 1949/50 - 1979/80 1.3 1.9 0.6 1967/68 - 1979/80 0.1 1.1 0.9 Source: Estimates of Area and Production of Principal Crops in India 1978-79, Ministry of Agriculture and World Bank estimates. - 96 - Table 2G INDIA - Oilseeds Production - Cottonseed Area Production Yield Thousand Thousand tons hectares kg/hectare 1949/50 4926 914 186 1950/51 5882 1014 172 1951/52 6556 1090 166 1952/53 6359 1111 175 1953/54 6987 1372 196 1954/55 7546 1479 196 1955/56 8086 1391 172 1956/57 8019 1638 204 1957/58 8014 1651 206 1958/59 7964 1623 204 1559/60 7295 1223 168 1960/61 7610 1865 245 1961/62 7978 1614 202 1962/63 7730 1842 238 1963/64 8221 1912 233 1964/65 8365 2000 239 1965/66 7962 1614 203 1966/67 7836 1803 230 1967/68 7995 1967 246 1968/69 7596 1848 243 1969/70 7731 1890 244 1970/71 7605 1621 213 1971/72 7800 2363 303 1872/73 7679 1950 254 1973/74 7574 2146 283 1974/75 7562 2439 323 1975/76 7350 2027 276 1976/77 6885 1987 289 1977/78 7866 2483 316 1978/79 (Revised) 8119 2734 237 1979/80 (Final) 8078 2643 327 Compound Growth Rates _ % % 1949/50-1978/79 0.6 2.7 2.1 1949/50-1964/65 2.4 4.5 2.7 1967/68-1978/79 0.2 2.8 3.0 1949/50-1979/80 0.6 2.7 2.1 1967/68-i979/80 0.2 2.9 2.9 Source: Estimates of Area and Production of Principal Crops in India 1978-79, Ministry of Agriculture and World Bank estimates. - 97 - Table 2H INDIA - Oilseeds Production - Coconuts Area Production Yield Thousand Thousand tons hectares kg/hectare 1949/50 569 3448 5785 1950/51 622 3582 5759 1951/52 630 3606 5724 1952/53 651 4498 6909 1953/54 663 4649 7012 1954/55 641 4409 6878 1955/56 647 4226 6532 1956/57 657 4383 6671 1957/58 666 4455 6689 1958/59 690 4589 6651 1559/60 715 4734 6621 1960/61 717 4639 6470 1961/62 723 4478 6194 1962/63 798 5017 6288 1963/64 798 4725 5920 1964/65 848 5043 5950 1965/66 884 5035 5698 1966/67 893 5192 5814 1967/68 924 5321 5760 1968/69 988 5546 5613 1969/70 1033 5859 5670 1970/71 1046 6075 5811 1971/72 1088 6124 5626 1872/73 1099 5997 5456 1973/74 1102 5851 5309 1974/75 1116 6030 5401 1975/76 1070 5829 5449 1976/77 1075 5765 5366 1977/78 1057 5413 5121 1978/79 (Revised) 1067 5471 5127 1979/80 (Final) n.a. n.a. n.a. Compound Growth Rates % % 1949/50-1978/79 2.5 1.6 -0.8 1949/50-1964/65 2.0 2.0 0.0 1965/66-1978/79 1.5 0.6 -0.9 Source: Estimates of Area and Production of Principal Crops in India 1978-79, Ministry of Agriculture and World Bank estimates. - 98 - Table 2I INDIA - Oilseeds Production - Nigerseed Area Production Yield Thousand Thousand tons hectares kg/hectare 1964/65 487 97 199 1965/66 522 91 173 1966/67 473 79 166 1967/68 468 98 209 1968/69 441 89 203 1969/70 482 98 203 1970/71 489 128 261 1971/72 494 116 235 1972/73 467 93 198 1973/74 552 136 246 1974/75 621 132 213 1975/76 615 151 245 1976/77 568 113 199 1977/78 (Revised) 609 148 243 1978/79 (Revised) 612 147 240 1979/80 (Final) 576 96 167 Compound Growth Rates % % % 1964/65-1978/79 2.1 3.9 1.8 1967/68-1978/79 3.1 4.1 0.9 1964/65-1979/80 1.9 2.9 1.0 1967/68-1979/80 2.7 2.5 -0.2 Source: Estimates of Area and Production of Principal Crops in India 1978-79, Ministry of Agriculture and World Bank estimates. - 99 - Table 2J INDIA - Oilseeds Production - Safflower Area Production Yield Thousand Thousand tons hectares kg/hectare 1965/66 462 69 149 1966/67 478 72 151 1967/68 513 78 152 1968/69 578 94 162 1969/70 580 142 245 1970/71 588 154 262 1971/72 598 131 218 1972/73 423 82 193 1973/74 614 191 311 1974/75 648 212 327 1975/76 674 238 354 1976/77 683 220 322 1977/78 (Revised) 707 188 266 1978/79 (Revised) 703 209 297 1979/80 (Final) 708 231 326 Compound Growth Rates % % % 1965/66-1978/79 3.0 9.4 6.4 1967/68-1978/79 2.7 8.6 5.9 1965/66-1979/80 2.9 9.0 6.1 1967/68-1979/80 2.6 8.2 5.5 Source: Estimates of Area and Production of Principal Crops in India 1978-79, Ministry of Agriculture and World Bank estimates. - 100 - Table 2K INDIA - Oilseeds Production - Sunflower Area Production Yield Thousand Thousand tons hectares kg/hectare 1977/78 385 160 416 1978/79 390 165 423 1979/80 350 150 429 Source: World Bank estimates of Area and Production of Principal Crops in India. - 101 - Table 2L INDIA - Oilseeds Production - Soyabean Area Production Yield Thousand Thousand tons hectares kg/hectare 1969/70 24 11 458 1970/71 30 18 600 1971/72 32 20 625 1972/73 35 25 714 1973/74 90 30 333 1974/75 90 35 389 1975/76 100 70 700 1976/77 200 150 750 1977/78 225 180 800 1978/79 275 220 800 1979/80 400 300 750 Compound Growth Rates % % 1969/70-1978/79 2.9 3.4 0.5 1969/70-1979/80 2.9 3.4 0.5 Source: World Bank estimates. - 102 - TABLE 211 Summary Table Giving the Annual Average Growth Ratea - By Type of Oilseeds (In Percentage) Type of (…------------------------PERIOD…--------_-______________________ Oil Seed 1949/50 to 1967/68 to 1949/50 to 1967/68 to 1949/50 to 1964/65 1978/79 1978/79 1979/80 1979/80 Ground Area 3.9 -0.2 1.8 -0.1 1.7 Nut Production 4.0 1.3 2.0 1.2 1.9 Yield 0.1 1.5 0.2 1.3 0.2 Rape & Area 3.0 1.0 1.9 0.9 1.8 Mustard Production. 3.4 1.7 3.1 0.8 2.9 Yield 0.4 0.7 1.3 -0.2 1.1 Sesamum Area 0.3 -0.7 0.1 -0.6 0.1 Production -0.9 0.7 0.02 -0.1 -0.1 Yield -1.2 1.6 -0.1 0.4 -0.2 Linseed Area -2.8 1.5 1.4 0.7 1.2 Production 1.2 2.5 1.2 0.1 0.8 Yield 1.6 1.0 -0.2 -0.6 -0.4 Castor Area -1.8 0.6 -1.0 0.4 -0.9 Seed Production -0.7 5.9 2.1 5.9 2.3 Yield -1.1 5.4 3.1 5.5 3.2 Five Area 2.7 0.2 1.3 0.1 1.3 Major Production 3.3 1.5 2.0 1.1 1.9 Oilseeds Yield 0.6 1.3 0.7 0.9 0.6 Cocoa Area 2.0 0.9 2.5 NA NA Nut Production 2.0 -0.1 1.6 -DO- Yield n.s. -1.0 -0.8 -DO- Cotton Area 2.4 -0.2 0.6 0.02 0.6 Seed Production 4.5 2.8 2.7 2.9 2.7 Yield 2.1 3.0 2.1 2.9 2.1 Saf- Area 2.7 2.6 flower Production 8.6 8.2 Yield 5.9 5.5 Niger- Area 3.1 2.7 Seed Production 4.1 2.5 Yield 0.9 -0.2 Soya- Area 29.0 29.0 Beanc Production 34.0 34.0 Yield 5.0 5.0 a/ Annual growth rates obtained by fitting a least square exponential curve to data on area, production, and yield of oil seeds. - 103 - Annex Table 3A Wholesale Prices of Groundnut Kernels and Oil 1960-1980 Indian -/ and International Markets (Rs/ton and US$/ton) CIF CIF Bombay Bombay Rotterdam Bombay Bombay Rotterdam Calendar Ground- Ground- Ground- Year Groundnut Groundnut Groundnut nut Oil nut Oil nut Oil Rs/ton US$/ton US$/ton Rs/ton US$/ton US$/ton 1957 660.4 139 - 1491.5 313 - 1958 698.7 147 - 1479.7 311 - 1959 758.5 159 - 1567.1 329 - 1960 878.2 184 197 1858.8 390 326 1961 981.7 206 196 2078.8 437 331 1962 908.8 191 171 1875.4 390 275 1963 562.5 118 172 1787.1 375 268 1964 1161.3 244 187 2288.8 481 315 1965 1312.5 276 206 2707.1 568 324 1966 1906.7 311 190 4175.0 681 296 1967 1770.8 236 180 3851.0 513 283 1968 1420.0 189 167 3023.3 403 271 1969 1973.3 263 207 4246.3 566 332 1970 2240.8 299 230 4831.3 644 379 1971 1928.0 257 249 4148.3 553 441 1972 2069.2 272 261 4353.3 573 426 1973 3583.3 463 393 7271.3 939 546 1974 3741.7 462 607 8312.7 1026 1077 1975 3049.2 364 452 6978.8 833 857 1976 2722.5 304 424 5431.3 606 741 1977 3975.0 455 551 8295.8 949 852 1978 3257.5 398 621 7085.8 865 1079 1979 3823.3 471 565 8625.0 1061 888 1980 4544.1 576 493 (E) 10058.4 1274 863 a/ The average price for the year is based on average of month-end prices in Indian market. (E) Estimate. Sources: Ministry of Agriculture and World Bank. - 104 - Annex Table 3B Wholesale Prices of Groundnut (Deoiled) Oilcake Indian -/ and International Markets (Rs/ton and US$/ton) CIF Calendar Bombay Bombay Rotterdam Year Groundnut Groundnut Ground- Oilcake Oilcake nut Meal (Decorticated) (Decorticated) Rs/ton US$/ton US$/ton 1957 217.8 46 1958 269.5 57 1959 318.6 67 - 1960 336.8 71 98 1961 358.3 35 93 1962 388.5 82 102 1963 398.6 84 196 1964 460.7 97 108 1965 513.3 108 119 1966 626.4 102 111 1967 653.7 87 127 1968 546.1 73 112 1969 676.6 90 111 1970 714.3 95 123 1971 573.9 77 98 1972 803.0 106 122 1973 1420.4 183 266 1974 1273.3 157 174 1975 1032.5 123 140 1976 1142.9 136 178 1977 1645.0 188 218 1978 1307.3 160 204 1979 1416.7 174 210 1980 1574.6 199 220 a/ The average price for the year is based on average of month-end prices in Indian market. Sources: Ministry of Agriculture and World Bank. - 105 - Annex Table 3C Wholesale Prices of Rape/Mustard Seed Cake Indian and International Markets (Rs/ton and US$/ton) FOB Ex-Mill Kanpur Kanpur Hamburg Calendar Mustard Mustard Rape Seed Year Seed Cake Seed Cake Meal Rs/ton US$/ton US$/ton 1957 293.1 62 1958 297.6 62 1959 286.9 60 1960 265.8 56 1961 279.5 59 1962 295.5 62 1963 313.7 66 72 1964 430.6 90 71 1965 470.0 99 74 1966 464.5 76 68 1967 552.1 74 70 1968 576.6 77 73 1969 562.1 75 75 1970 515.8 69 84 1971 533.4 71 71 1972 616.5 81 90 1973 755.8 98 178 1974 1029.2 127 143 1975 705.8 84 128 1976 801.6 89 154 1977 1206.7 138 169 1978 916.3 112 168 1979 1143.3 141 177 1980 1400.0 b/ c/ 177 204 a/ The average price for the year is based on average of month-end prices in Indian Market. b/ Based on data for 11 months of the year, the date for the month of June 1980 being "not recorded". c/ Data for May 1980 relates to last but one week. Sources: Ministry of Agriculture and World Bank. - 106 - Annex Table 3D Wholesale Prices of Rape and Mustard Seed and Oil Indian -/ and International Markets (Rs/ton and US$/ton) CIF Euro- pean Ports FOB Ex-Mill Calendar Kanpur Canadian Kanpur Dutch Year Rape/ Rape/ Mustard Mustard Rapeseed Mustard Seed Mustard Seed Rapeseed Oil Oil Oil (Yellow) (Yellow) Rs/ton US$/ton US$/ton Rs/ton US$/ton US$/ton 1957 942.3 198 - 2210.0 464 1958 850.1 179 - 1878.9 395 1959 833.5 175 - 1813.0 381 - 1960 862.7 181 160 2029.5 426 219 1961 960.8 202 165 2144.4 450 280 1962 1013.7 213 127 2385.2 501 221 1963 970.9 204 146 2186.7 459 215 1964 1322.2 278 134 3157.5 663 252 1965 1489.9 313 123 3420.5 718 263 1966 1637.4 267 130 3761.9 614 244 1967 2003.8 267 122 4671.4 623 206 1968 1588.9 212 106 3749.1 500 161 1969 1756.9 234 111 4051.4 540 200 1970 1960.0 261 142 5082.0 678 293 1971 2023.0 270 142 4802.4 640 295 1972 2136.7 281 132 5167.6 680 232 1973 2746.7 355 254 6631.4 857 395 1974 3820.0 471 261 9420.8 1163 745 1975 2466.7 294 293 5841.7 697 551 1976 2564.2 286 246 5918.3 661 415 1977 4316.6 494 312 10310.0 1180 584 1978 4101.7 501 297 9420.0 1150 597 1979 3818.3 470 313 9508.3 1170 635 1980 4959.1 628 311 12422.7 1574 570 a/ The average price for the year is based on average of month-end prices in Indian market. Sources: Ministry of Agriculture and World Bank. Annex Table 4: FARM HARVEST PRICES OF FIVE MAJOR OILSEEDS (Rupees Per Tonne) Groundnut Rape and Mustard Sesamum Linseed Castor Seed …_______…__ …-… - ---- - ---- -- --- - -_---_-----__ ___-__ __ __ __-_ __ __-----_-__----_-_ _-- -…- Andhra Tamil Uttar Uttar Tamil Uttar Madhya Gujarat Pradesh Nadu Pradesh Rajasthan Pradesh Rajasthan Nadu Pradesh Pradesh Cujarat -- ____ - _____ _ __ __ - - ------ - ------ -_ … - ---- 1962/63 586.00 542.20 501.70 741.30 878.20 851.50 927.00 911.60 605.50 538.40 564.00 1963/64 569.30 578.90 584.40 905.40 1,114.60 893.00 1,109.50 1,037.70 702.40 651.10 634.50 1964/65 727.50 825.20 744.60 1,007.30 1,429.40 1,249.50 1,431.50 1,138.60 819.10 779.70 761.60 1965/66 1,022.60 1,180.90 984.40 1,199.10 1,599.40 1,487.80 1,704.40 1,580.60 1,190.40 1,265.10 1,114.70 1966/67 1,247.70 1,342.60 1,031.50 1,663.00 1,841.30 1,804.00 1,858.90 1,882.00 1,609.60 1,726.90 1,431.10 1967/68 954.20 955.10 964.50 1,244.60 1,479.00 1,670.80 1,603.80 1,673.70 1,027.60 1,020.30 1,107.50 1968/69 1,074.70 1,089.90 976.80 1,344.60 1,449.00 1,561.00 1,730.60 1,722.40 1,197.20 1,186.70 1,034.90 1969/70 1,378.80 1,280.20 1,307.10 1,469.00 1,640.20 1,654.60 2,027.90 2,024.90 1,418.10 1,533.90 1,402.70 1970/71 1,408.00 1,314.20 1,184.30 1,588.00 1,715.20 2,065.50 2,092.40 2,030.20 1,473.90 1,416.60 1,337.80 1971/72 1,324.40 1,220.70 1,187.10 1,691.90 1,816.80 2,059.30 2,228.90 2,134.60 1,489.10 1,395.20 1,251.90 1972/73 1,663.00 1,849.80 1,311.40 1,844.30 2,202.10 2,562.60 2,898.30 2,719.90 1,769.90 1,786.30 2,209.40 1973/74 2,318.80 2,233.70 2,172.70 3,122.70 2,674.30 3,200.70 3,578.70 3,373.60 3,034.50 2,964.10 2,374.40 1974/75 2,388.20 2,590.60 3,128.50 2,310.30 2,555.10 3,456.50 4,172.00 3,541.20 2,128.00 2,195.50 1,977.30 1975/76 1,660.10 1,454.90 2,386.70 1,632.10 1,760.60 2,681.10 2,878.90 2,888.20 1,596.50 1,580.20 1,531.10 1976/77 1,817.20 2,347.10 2,830.30 3,344.60 3,324.30 3,559.80 3,400.70 3,368.10 3,077.80 3,056.70 2,329.30 1977/78 2,215.10 2,085.50 2,905.30 3,102.60 3,144.60 4,114.00 4,082.10 3,291.10 2,734.50 2,546.10 2,097.00 1978/79 n/a 1,998.70 1,242.20 n/a 3,231.00 n/a 3,810.00 n/a n/a 2,553.40 n/a Source: Agricultural Situation in India, Various Issues PART IV THE SUGAR ECONOMY James Q. Harrison June 1981 109 - 110 - The Sugar Economy Table of Contents Page No. The Role and Scope of the Sugar Economy .... ................ 111 Market Structure and Public Policies ....................... 111 Short-Term Market Instability .............................. 113 Long-Term Supply and Demand Trends ......................... 116 Sugar Production and Export ................................ 122 Policy Implications ........................................ 124 Annex . ..................................................... 126 Annex Tables ........... .................................... 130 - ill - THE SUGAR ECONOMY 1. The Role and Scope of the Sugar Economy. India is the world's lar- gest producer of sugarcane. Although sugarcane covers only 1.8% of India's gross cultivated area, it is one of the most important non-foodgrain crops, for it contributes about 6.3% of the gross value of agricultural output, about 3.5 times as much per hectare as the average of all other crops. About 80% of the sugarcane area is irrigated which is equivalent to about 5.6% of India's total irrigated area. Fertilizer used on cane was about 84 kg of nutrients per hectare in 1975/76, about four times the average for all crops. While sugarcane is grown in nearly every state, its greatest concentrations are in U.P. and Maharashtra where 45% and 14%,, respectively, of India's cane is pro- duced. Sugarcane is cultivated by about 25% of India's 80 million farmers, on farms of all sizes, though the proportion of area devoted to cane is somewhat higher on the smaller farms. 2. Sugarcane processing, including the production of mill sugar, khand- sari and gur, is a major industry. It employs nearly 8% of all manufacturing workers and about 3.6% of the capital invested in manufacturing to produce about 2% of the total value added in manufacturing. Throughout the two decades before 1980, India has been an exporter of sugar, but exports were generally low before 1974/75, when they rose rapidly, peaking at levels of over 1 million tons in 1975/76, providing 12% of India's foreign exchange earnings. Since then exports have been much lower and highly volatile. In 1980, India is likely to be a net importer of sugar. Sugar is a large item of personal con- sumption accounting for about 5% of private consumption expenditures. The 7.7% weight of sugar products in the wholesale price index suggests its sensi- tive role. In 1979/80, the rise in the prices of sugar products alone accounted for about 30% of the entire 20% rise in the wholesale price index of all com- modities. Rising prices of sugar thus provoke sharp consumer reactions and give rise to considerable political controversy. The high volatility of output and prices which often characterizes the sugar economy in India (and other coun- tries as well) is thus a matter of serious concern. The underlying importance and visibility of the sugar economy have led to extensive and varied Government interventions in its operations. 3. Market Structure and Public Policies. Virtually all sugar products in India are based on sugarcane. The structure of the market for sugarcane and its products and the network of public policies governing sugar production strongly influence the performance of the sugar economy. In contrast with many countries, sugarcane is not grown in large blocks on plantations as a monocrop, but rather in relatively small patches forming a part of the crop- ping pattern of millions of individual farmers. Thus cane production is scattered rather widely among vast numbers of independent farmers who grow a range of other crops. Sugarcane production is highly seasonal with the bulk of the crop in most areas harvested in the December-March period. The sugar content of cane in most areas drops sharply with the onset of warmer weather. The sugar content and weight of cane drop rapidly after harvesting, making it necessary to process the cane as soon as possible, usually within two days of cutting. The production of sugar products is thus also highly seasonal with 70% of sugar being produced between December and March. Cane is a relatively long-duration crop, standing in the field 9 to 18 months between planting and - 112 - harvesting. It is a common practice to allow a second or ratoon crop of cane to grow from the stubble of a harvested field. Thus there is a considerable time lag between the decision to plant an area under cane and the full effect of that decision on the supply of cane. 4. An estimated 11-12% of the cane crop is used for seed, animal feed and direct human consumption, or disappears as post-harvest loss. On average, about 46% of the crop is used to produce gur, which consists of condensed and solidified cane juice produced with very simple equipment in rural areas, often by farmers themselves. Gur typically weighs about 10% of the weight of the cane used to produce it and contains about 60-65% sucrose. It is by far the most important sweetener, particularly for the rural population and for low income consumers. Gur's weight in the wholesale price index is more than twice that of sugar and nearly ten times that of khandsari, yet it has attracted the least direct public intervention, a pragmatic recognition that efforts to control such large numbers of tiny scattered units would be imprac- tical. About 8-9% of the cane crop is used for khandsari, a white crystal sweetener similar to mill sugar but produced through a smaller-scale, less technically efficient process. The quality of khandsari is generally consi- dered inferior to mill sugar and this is reflected in the 15-20% price dif- ferential between the two products. An average of about 34% of cane crop is used by large-scale sugar mills to produce sugar, often referred to as "mill sugar" to distinguish it from other sweeteners. Sugar mills are much larger in scale than khandsari units (the average sugar mill produces about 20,000 tons of sugar a year while a large khandsari unit produces about 750 tons), and more efficient physically, extracting 9.5-10.5% of the weight of cane as sugar compared with 6-7% for khandsari units. 5. Gur and khandsari units are virtually free from controls on the prices they pay for their cane or the prices they charge for their products. Gur production is not taxed and while khandsari is taxed, it is at much lower effective rates than mill sugar (in 1977/78, excise tax revenues on khandsari were Rs 93 per ton compared with Rs 466 per ton of sugar). Although the sugar mills are neither the largest consumers of cane nor the major producers of sweeteners, the importance of their product on the urban market, their size and their manageably small numbers, have made them the main focus of govern- ment intervention in the sugar economy. The primary instruments of government control have been: (a) setting a minimum price for cane purchased by the mills; (b) varying controls over the monthly releases; (c) requiring mills to sell to the Government a fixed proportion of their production at a fixed price, usually below the market price; and (d) imposition of excise taxes of varying levels and partial excise tax rebates. The blend of these various policies has changed significantly at least ten times since 1960/61 and many more times than that if one counts changes in the formulae for calculating the fixed sugar prices or excise taxes. The present policy, and the one which was dominant through most of the 1970s, is partial control: minimum prices which mills must pay for their cane are set by the Central Government and - 113 - often augmented by State Governments; mills must sell the bulk of their pro- duction (usually and at present 65%) to the Government at fixed prices, 1/ while the remainder is sold in the free market, with the Government control- ing the monthly releases of both levy and free market sugar. The levy sugar is sold in government fair price shops at the same fixed price all over India, with this retail price based on the average ex-factory price plus some allowances for transportation and distribution. 6. Short-Term Market Instability. The aspects of market structure dis- cussed above help explain the large fluctuations in the production of sugar- cane and sugar products over the past several years. Variations in weather combine with the lagged response of cane production to prices (see para 3 above) to produce oscillations in the production of cane. A short crop sends prices up, but the full impact on production is felt only one to two years later due to the long time between planting and harvesting, and the practice of ratooning. These oscillations, however, are amplified by the controls on sugar mills so that the fluctuations in sugar production are even more violent than those in cane production. The reason for this is clear: in years of a short crop, gur and khandsari producers are able to offer higher prices to far- mers, and since the prices of gur and khandsari are not controlled, pass the cost on to the consumers. Sugar manufacturers lack this flexibility since the price of the bulk of their output is fixed and since their profit margins are already under pressure due to high excise taxes. Consequently, in such years farmers sell as much cane as they can to the gur and khandsari producers, causing mill sugar production to fall proportionally more than cane production. When the cane crop is large the reverse takes place. The cane price drops, and the gur and khandsari units, unbound by the statutory minimum cane price, pay as little as the market will bear. Since they are competitive, much of the cane cost reduction is passed on to consumers. The mills, in contrast, are bound to pay the statutory minimum price, so farmers clamor to sell as much of their cane to the mills as possible. In such years the mills are often obligated to extend the crushing season into unprofitable months and to build up stocks which are a drain on profits. The relationship between the size of the cane crop and the proportion of it that is crushed by sugar mills emerges clearly in Graph 1. 2/ A measure of the instability in sugar production generated by this set of policy instrumunts is that while the standard deviation of cane production is only about 18% of its mean, for sugar production it is 31% of its mean. 1/ These ex-factory selling prices for levy sugar are calculated separately for each of 16 sugar zones, adding to the zone's average cane cost (including cane purchase taxes if any) and the zone's average cost of converting cane to sugar, a fixed return equal for all zones. In 1979/80 these prices ranged from Rs 2.44/kg in Gujarat to Rs 3.06 in South Bihar. Rs 0.23/kg was calculated "return" for all zones. 2/ The asterisks in the graph are data points and the diagonal solid line is an equation fitted to those points relating the total cane (TC) crop (in million tons) to the percentage of cane sold to the mills, PMC. The equation (PMC = 7.3 + 0.18TC) explains 84% of the variation in PMC. See Annex. - 114 - 7. Cane production rose very rapidly after 1974/75. As expected, prices of khandsari and gur began falling from 1975 and 1976, respectively, reflecting the emergence of excess supplies of cane. This should have reduced the average prices cane farmers received, but this signal was obscured by two factors: farmers increased the proportion of cane sold to mills (since the fixed statutory minimum price was much higher than the open market price offered by gur and khandsari units), and reinforcing this, a number of state governments required the mills to pay "state advised prices" which were sub- stantially above the minimum. The results can be seen in Graph 2. Even though the index of gur prices was falling after 1976, the index of average realized cane prices 1/ fell only slightly in 1975/76 and rose to a then record level in 1976/77. Thus, even though gur prices fell sharply in 1977, farmers were encouraged by the cane price signals (and excellent weather in the 1977 planting season) to plant a record area and achieve a record produc- tion of 177 million tons in 1977/78. This was 16% higher than the 1976/77 crop of 153 million tons which itself had been substantially higher than any previous crop. Naturally, with gur and khandsari prices already falling, farmers made every effort to sell the greatest possible quantity of their cane to the mills. 8. Urged by State and Central Governments, the mills extended the crush- ing season and produced 6.46 million tons of sugar in 1977/78--33% more than the previous year, and about 16% above rated capacity. Despite this, farmers average cane realization fell sharply because of the very low prices offered by gur and khandsari units. In some areas farmers reported prices of about Rs 40 per ton, while in others they were not able to sell their cane at all either burning it or leaving it in the ground for next year. Even farmers selling to mills did not get full payment because as the mills' stocks of unsold cane peaked over the year to 4.6 million tons, many mills simply did not have the funds to pay for the cane, despite obtaining large and costly bank advances against these stocks. As this large sugar glut emerged, in August 1978, the Government announced the complete decontrol of sugar. Sugar prices dropped 18% over the next six months as many mills sought to liquidate their stocks before the price dropped further. This was possible since the Government had also lifted all restrictions on monthly sales. For a while, sugar was sold at less than the levy price, and even substantially below cost. In March 1979, the sugar industry introduced voluntary distribution controls and by June 1979, the Government had reintroduced controls on releases followed by full price control in September 1979. Due to low prices, internal consump- tion of sugar in 1978/79 was a record 6.2 million tons, necessitating a large drawdown in stocks, despite production of a near record 5.8 million tons of sugar in that year. By the end of the 1979 season, stocks of sugar were down to 2.05 million tons. 1/ Since the official wholesale price index reflects only the statutory minimum price, an index of "average realized cane prices" was calculated as an average of the gur price index at harvest time and the index based on the statutory minimum price, weighted by the proportion of cane sold to the mills. This index does not reflect the state advised prices or the higher prices offered by cooperatives. *Some states require mills to pay 20%-30% higher than the statutory minimum. Where cooperatives 6perate, mill profits are normally passed on to grower members as higher cane prices since agricultural income is not taxed. Both these factors further distort the prices perceived by farmers. - 115 - GRAPH MILL'S SHRRE IN TOTRL CANE PRODUCTTON 90.0 110.0 130.0 150.0 170.0 z 38.0 38.0 36.0 36.0 34.0 34.0 o 32.0 32.0 z 30.0 30.0 CC n 28.0 28.0 o 26.0 26.0 0~ 24.0 _24.0 '22.0 j , , iz,, , ,,l,,,l,,,l,z _ 22 .0 90.0 110.0 130.0 150.0 170.0 TOTRL CRNE CROP (MN. TONS) PRICE INDEXES FOR GUR RND CRNE GRPH2 GUR --- CANE 1960 1964 1968 1972 1976 1980 x 375.0 _ I I I I I I I - 375.0 LU 350.0 350.0 z 325.0 325.0 ` 300.0 300.0 Lu 275.0 275.0 - 250.0 250.0 w 225.0 A 225.0 200.0 200.0 L*1 0 175.0 cr 150.0 150.0 w 125.0 _ / 125.0 IJ100.0 - 100.0 r 75.0 _ 75.0 :K 50.0 50.0 1960 1964 1968 1972 1976 1980 CROP YERR (OCTOBER - SEPTEMBER) - 116 - 9. Meanwhile, the farmers reacted to the low 1978 cane prices by plant- ing little new area to cane and relying mainly on a ratoon crop, bringing cane production down to 156 million tons in 1978/79. This was still a large crop, and with gur and khandsari prices still low, farmers sold a record 38.2% of their crop to the mills. Their average realizations, however, were lower than in 1978. There were signs of gur prices beginning to rise but these were not as clear as they might have been due to the unusually high availability of unusually cheap mill sugar. The full impact of the effective cane price reductions was felt at the time of planting the 1979/80 crop. With little new planting the previous year, there was little scope for a ratoon crop and little incentive for new planting. Weather conditions were unfavorable. Area and yield both dropped sharply, resulting in a crop of 128 million tons, the lowest in seven years, and 38% below the 1977/78 record. By harvest time, gur and khandsari prices had risen sharply, resulting in only 30% of the crop being sold to the mills. Mill sugar production fell to 3.9 million tons, 40% below the 1977/78 record. During the year the full price controls on sugar were lifted and partial controls (65% levy, 35% open market) went into effect December 1979 and continue to date. From November 1979 through most of 1980, sugar prices rose at unprecedented rates and on occasion sugar disappeared entirely from the open market. While sugar prices rose by 52% between September 1979 and September 1980, gur prices rose 110%. Although there was press speculation of unfair trade practices in the sugar industry, the far more rapid and significant rise in gur and khandsari prices --where trade is decentralized and competitive--suggests the real reason for the price increase was a severe shortage of cane, a shortage brought on, para- doxically, by the glut two years earlier. The rise in prices of sweeteners alone accounts for 35.2% of the rise in the general wholesale price index during that period. 10. With cane prices high in the planting season for the 1980/81 crop, cane production is expected to recover to about 155 million tons. Poor weather conditions at planting time (February-March 1980) probably prevented a more rapid recovery. With even higher statutory prices and high gur and khandsari prices in the early months of 1981, 1/ one can expect the 1981/82 crop to be even larger--possibly on the order of the 1977/78 record. The following sections assess the longer-term supply and demand prospects. It is worth noting, however, that the structural causes of the volatility in the cane and sugar products markets remain. Without an appropriate balance between the price of cane and its products and between the price of cane and alternative crops, it is difficult to see how these wide oscillations can be avoided. Ways of obtaining such balance are discussed below in the section on policy implications. 11. Long-Term Supply and Demand Trends. The production of sugarcane has grown significantly over the past two decades. As Table 1 indicates, however, the aggregate growth trends have varied among regions and between decades. Graph 3 shows very sharp year-to-year fluctuations around these longer-term trends. Over the past twenty years as a whole, the rate of growth of cane production has been 2.70% per annum. Over the last decade, it has 1/ The Government raised the statutory minimum price of cane by 25% in 1979/80 and by another 4% in 1980/81. The price index for gur in December 1980 was about 50% higher than it was a year earlier. - 117 - accelerated to 3.78% per annum. The increase in the rate of growth, however, has been due to an increase in the growth of area from 1.44% per annum to 2.75% per annum, while growth in yield has been somewhat lower in the past 10 years than for the 1960/61-1978/79 period as a whole. It is particularly disturbing that yields have virtually stagnated in U.P. and the tropical states other than Maharashtra, which together account for 70% of India's cane production. Table 1: GROWTH TRENDS IN SUGARCANE AREA, YIELD AND OUTPUT Growth Rates (% p.a.) % of All-India 1960/61-1978/79 1967/68-1978/79 Region Production (1978/79) Area Yield Output Area Yield Output All-India 100% 1.44 1.27 2.70 2.75 1.05 3.78 Uttar Pradesh 40% 1.14 0.74 1.94 3.33 0.35 3.79 Other Sub-Tropical 16% 0.13 1.48 1.61 1.24 1.41 2.64 Maharashtra 14% 2.66 1.42 4.08 2.54 4.19 6.72 Other Tropical 30% 3.86 0.41 4.26 3.12 0.17 3.29 Source: Ministry of Agriculture. GRAPH 3 RLL INDIR - SUGRRCRNE PRODUCTION 1960 1964 1968 1972 1976 1980 18C.C I 180.0 170.0C 170.0 () 160.0 160.0 z CD 150.0- 150.0 140.0 140.0 130.0 130.0 120.0 120.0 E 110.0 110.0 100.0 100.0 90.0 90.0 1960 1964 1968 1972 1976 1980 CROP YERR (OCTOBER - SEPTEMBER) - 118 - 12. The profitability of sugarcane cultivation has been a major force behind its rapid growth. Farm management studies in the late 1960s and early 1970s show sugarcane earning two to five times more than most competing crop combinations, more than enough to compensate for sugarcane's longer duration (9-18 months). Notable technological advances in some alternative crops-- mainly foodgrains--and a gradual, although not statistically significant, decline in cane prices relative to some competing crops, have tended to narrow sugarcane's competitive edge somewhat. Still, partial cost of cultivation and crop budget data, combined with the continued strong growth in cane production even during the green revolution period, indicate that sugarcane is substan- tially more profitable than most other crop combinations in most years. 13. There are striking regional differences in cane production trends. Cane is grown in two broadly different zones--sub-tropical (predominately UP) and tropical. Over the past two decades, production in the tropical zone has grown more than twice as fast as in the sub-tropical zone. Growing conditions in the tropical zone are technically superior for cane cultivation, as Table 2 suggests. Cane growers in the south enjoy a longer productive growing season giving more cane per unit of land and more sugar per unit of cane. The apparent technical superiority of the tropical belt, however, does not tell the whole story. Although available cost of cultivation data are fragmentary, they suggest that the much higher costs of growing cane in the tropical zone substantially offset the region's greater technical efficiency. The duration of a significant portion of tropical zone cane crop is 12-18 months, versus 9-10 months in the sub-tropical zone. Much of the cane grown in the sub-tropical zone is grown in relatively fertile soil requiring little or no irrigation. Table 2: CANE YIELDS, SUGAR CONTENT AND SUGAR YIELDS (Average for 1976/77-1978/79) World All-India U.P. Maharashtra Cane Yield (tons/ha) 56.1 53.20 43.30 92.10 Sugar Content (%) 10.0 9.76 9.38 10.89 Sugar Yield (tons/ha) 5.6 5.20 4.10 10.00 Source: Food and Agriculture Organization and Ministry of Agriculture. 14. This point is brought out in Table 3 which compares cost of pro- duction and net return estimates for Maharashtra and Uttar Pradesh. In recent years the cost of cultivation in Maharashtra has been about twice the U.P. level so that net returns are not significantly different, given the inexact nature of the data. To some extent the cost of production data may overstate the eco- nomic production costs both because of the methodology used (e.g., labor valued at market wages) and because those providing the data (the farmers) may know they are used to justify increases in the minimum sugarcane price. On the other hand, other costs, particularly water, may not be valued at their full opportunity cost. This would further reduce Maharashtra's apparent comparative advantage since the Maharashtra crop requires substantially more water because of its longer duration and the soils and climate in which it is grown. - 119 - Table 3: SUGARCANE PRODUCTION COSTS AND RETURNS 1976/77 Maharashtra Uttar Pradesh Gross Returns (Rs/ha) 8,939 5,576 Costs (Rs/ha) 4,164 1,849 Profit (Rs/ha) 4,775 3,727 Yield (tons/ha) 65.9 41.9 Costs (Rs/ton) 63.2 44.1 Profit (Rs/ton) 72.5 89.0 Source: Ministry of Agriculture. 15. As noted earlier, the longer-term growth in India's sugar production is marked by sharp year-to-year fluctuations. Measures to reduce such insta- bility are discussed below. Here an effort is made to discuss the level and adequacy of the long-term growth rates. Graph 4 shows actual production and domestic consumption of all sweeteners from 1960 to 1980 and alternative projections of. production and domestic demand from 1980 to 2000. The three production projections are based on extrapolations of: (a) the higher his- torical growthrate (3.8%) in cane production experienced since 1967/68 (the high production projection); (b) the lower growth rate (2.7%) experienced over the period since 1960/61 (the low projection); and (c) an equation relating cane production to a time trend and lagged cane prices relative to prices of competing crops (the medium projection). The consumption projections are based on: (a) an extrapolation of the historical growth rate in consumption (2.44%) (the lowest projection); (b) estimates of per capita expenditure elasticity of demand for sweeteners based on NSS cross-sectional survey data and World Bank population projections assuming, alternatively, 5% and 3.5% rates of growth in total expenditures; and (c) estimates of expenditure and price elasticities for sweeteners derived from time series data, again assum- ing 5% and 3.5% as alternative expenditure growth rates. 1/ 1/ Data and methodology underlying the projections are given in the Annex. - 120 - GRAPH 4 PRODUCTION RND CONSUMPTION OF SWEETENERS PRODUCTION CONSUMPTION 1960 1965 1970 1975 1980 1985 1990 1995 2000 35.0 11111111111111 jI 11111111111111 II 1111_ 35.0 ° 32.5 32.5 2 30.0 30.0 27 .527. 25 .0250 , 22.5 22.5 I- ' 20 .0 20 0 U17 .5 17 .5 c 15 0 1 5 °0 12 .5 12 .5 o10.0 _ 10 .0 Mc 7 .5 ~~~~~~~~~~~7 .5 1960 1965 1970 1975 1980 1985 1990 1995 2000 CROP YERR (OCTOBER - SEPTEMBER) 16. The projections suggest that if past growth rates in cane production can be sustained, serious chronic shortages of sweeteners are not likely to emerge. 1/ The projection of recent growth in cane production more than covers even the highest demand scenario. The two highest demand projections assume an unprecedented growth in expenditure of 5% per annum sustained over 20 years and may be assumed to represent the upper limit to likely demand. Even if the highest demand projection is taken as given, cane prices would need to rise relative to competing crops by only about 0.4% per annum above present levels to bring the medium production projection into balance with 1/ It should be emphasized that this statement refers to average trends. If the wide fluctuations in production described above in paras 6 to 10 remain, there will be years of acute shortage (as well as years of excess supply). - 121 - this very high demand projection. 1/ If the lower expenditure growth assump- tions are taken as more probable--and they certainly are more consistent with historical behavior--then the situation would be even more comfortable. 17. The tapering off of yield growth, particularly in the main cane pro- ducing areas (para 13 above), is a cause for concern and raises the question whether the future scope for expanding production is great enough to allow past trends to continue. Nevertheless, even if sugarcane only maintains its share of irrigated area (5.6%) and yields do not increase at all, the 13.6 million ha additional irrigation projected for the Sixth Plan would imply cane production increases by 1984/85 greater than the highest production projection on the graph. Moreover, an analysis of present cane yields and cultivation practices suggests there is ample scope for further yield increases and/or cost reducing innovations which would maintain the relative profitability of cane production, stimulating both higher production per hectare and further area growth as well. A range of fairly well-tested simple techniques are known which would substantially increase yields of both newly planted and ratooned cane. These include careful selection and treatment of planting material, appropriate spacing, weeding, combined with moderate doses of fertilizer and, for the ratoon crop, close shaving, trash burning and gap filling. When such techniques were systematically adopted by farmers in Chambal, Rajasthan, an area not specially suited for cane cultivation, yields rose over 45%. 2/ Research findings suggest substantial water savings can be obtained through skip-furrow irrigation. While farmers who do not pay for water on a volumetric basis may have little incentive to adopt this practice, many farmers who irri- gate with private pumps can cut costs substantially. Research also indicates that inter-cropping with such crops as oilseeds, potatoes and wheat gives virtually a full yield of each crop and helps control weeds during the early stages of cane growth. The joint profitability of the crops is substantially enhanced. 3/ These and a number of other promising innovations have yet to be widely adopted and adapted to farmers' field conditions. 18. This suggests strongly that where a strong link can be established between farmers, extension workers and research, cane yields and profits can rise. Unfortunately, as the yield growth data suggest, in some areas this is not happening. Research stations often have useful results on hand which have not reached farmers. Yet some of the farmers' constraints are not fully reflected in research stations, and a disproportionate research effort is often devoted to topics (e.g., sugarbeet) which have little relevance to most farm conditions. In some of the most important cane growing areas, the link between extension agents, research and the farmers on technical matters is very weak. There are good reasons for this as virtually all the time of 1/ Such a rise in cane prices would also tend to reduce demand implying that an equilibrium would be reached without cane prices rising even that much. 2/ Government of Rajasthan, Increased Sugarcane Production in Chambal Command Area, 1977-78, p.2. 3/ Sugarcane Research Institute, Shahjahanpur; Indian Institute of Sugar Research, Lucknow. - 122 - special cane development staff is taken up in necessary administrative and logistical matters (such as organizing cane deliveries). In one major cane growing area none of the farmers interviewed had received technical advice on improved techniques and varieties. Moreover, since cane is only one of many crops grown by the farmers, an extension service focussing on cane alone is not ideal for the job. Fortunately, most states in India are in the process of reorganizing extension services to provide a strong channel for effective technical advice to farmers and feedback to research. Such a reform should strengthen the technical basis for growth not only of sugarcane but also for all other crops. The Plan program for improving the technological base of cane cultivation should be designed to reinforce these extension reforms rather than strengthening special schemes dealing with cane alone. One way to do this would be to use funds earmarked for centrally designed demonstrations for field trials adapted to local conditions. Where the extension reforms are operative in major cane areas, consideration could be given to adding a tech- nical subject matter specialist in sugarcane to the subject matter specialists already in place. 19. Sugar Production and Export. The present profitability and future technical possibilities make it probable that sugarcane production can con- tinue to grow at rates similar to those in the recent past. If so, cane pro- duction can exceed domestic demand, offering export possibilities. To give an approximate idea of the orders of magnitude of potential exports, the gaps between two alternative production and demand scenarios are given in Table 4 These figures are not meant to indicate projections of actual exports---they are simply not that precise. Moreover, the gap between supply and demand on which they are based is not likely to be closed only by exports--some fall in relative sugar prices and consequent increase in demand and slower supply growth would also narrow the gap. As the projections are pushed further into the future, the mix of developments which will bring about equilibrium becomes increasingly speculative. Table 4: SUGAR EXPORT POTENTIAL (million tons) Production Projection: High Growth Medium Growth Demand Elasticity: NSS Time Series Expenditure Growth: 5% p.a. 3.5% p.a. Year 1985 1.8 1.9 1990 1.7 2.5 1995 1.5 3.0 2000 1.7 4.3 Source: World Bank estimates. (See Annex, para 6.) 20. The table does suggest that India's objective of bringing sugar exports to a level of one million tons per year by 1985 is feasible in terms of total balances of supply and demand for sweeteners. The high NSS demand estimate, combined with the increasing proportion of sugar in total sweetener consumption, suggest an upper limit to domestic demand for sugar of about - 123 - 6.8-6.9 million tons in 1985. This is consistent with the targeted increase in installed sugar mill capacity to about 8.04 million tons by 1984/85. The targeted investment of about Rs 8,050 million would increase total crushing capacity by about 171,250 tons of cane per day, with about 60% of new capacity to be from new plants and 40% from expansions of existing mills. While this is higher than the 150,000 tpd capacity strictly needed to reach 8.04 million ton production level, it is probably wise to allow for some slippage and cost increase. Most of the investment is likely to be in the cooperative sector since government policy offers cooperatives highly favorable financing terms. If Plan resources are constrained, incentives to encourage a higher proportion of private investment could be considered. Careful consideration should also be given to increasing the proportion of expansions of existing plants in the total increase in capacity. Expansions cost 40% less than new plants per ton of capacity, and probably involve shorter gestation periods and fewer logis- tical problems. The locations of new plants will have to be selected care- fully to ensure adequate cane supplies. The reintroduction of incentives for new mills and mill expansion 1/ combined with the recovery of sugar prices has increased the attractiveness of investments in this sector substantially, but the violent fluctuations in cane supplied to the mills will remain a major disincentive unless remedial measures are taken. 21. The volatility of sugar both in India and in the world market makes it difficult to project the competitiveness and profitability of India's sugar exports. Both Indian and world sugar markets appear cyclical and the profitability of India's sugar exports depends on India's cycle being in or out of phase with the world market. Prior to 1973/74 the unit value of India's exports was generally below both domestic market prices and the ex-factory levy prices. In 1974 and 1975 world market prices were high and resulted in sharp increases in the volume and value of India's sugar exports. Between 1976 and 1979 world prices were generally lower than Indian domestic price but higher than ex-factory levy prices (which exclude excise taxes). In 1980 world sugar prices were high and are expected to remain well above Indian production costs through 1981. Indian cane yields and sugar production costs appear fairly competitive with other suppliers. From the Annual Survey of Industries 1977/78 and from the calculations used to determine the ex-factory levy price, Indian sugar production costs appeared to be about US$171-208 per ton of D-29 sugar in 1977/78, a year when developing country sugar production costs were estimated at US$220-310 per ton for raw sugar of lower quality. 2/ World market prices for raw sugar in 1985 are estimated at US$412 per ton in 1980 dollars compared with 1980 Maharashtra production costs of about US$320 for a more refined grade of sugar. 3/ Continued improvements in cane produc- tivity, such as those discussed earlier, would help ensure the continued 1/ New plants and expansions of existing plants intially are allowed to sell a greater proportion of their sugar in the free market, with the propor- tion falling gradually over several years to the same level as that of existing plants. Moreover, they face lower excise taxes intially. 2/ IBRD, Price Prospects for Major Primary Commodities, January 1980. 3/ IBRD, Commodity Price Forecasts - Updating, November 12, 1980, p. 4. - 124 - competitiveness of India's sugar exports by keeping costs down. The availabil- ity of a buffer stock would give India greater control over the timing of sugar exports to take advantage of fluctuations in world market prices. 22. Policy Implications. Sugarcane and sugar products policy in India has been the subject of many detailed studies, each of which called for a long-term, stable policy. The very volatility of the sugar economy, however, has prevented this, and as noted earlier, sugar policies in India have been changed frequently and substantially. As the sugar economy emerges from the acute shortage and rapid price rises of the past year it is difficult to look ahead to implementing longer-run policies. Yet successful implementation of longer-term policies could also contribute to resolving the short and medium- term problems of supply management. 23. Sugarcane price policy is clearly central to the sugar economy. The recent past has suggested that the prices set by the Agricultural Prices Commission (APC) provided a reasonable stimulus to cane production. The price, which has been uniform across states but linked to the sugar content of cane, helps promote locational efficiency. The state advised prices and cooperative profits distribution to cane growers, however, counteracted the APC efforts and encouraged excess production in 1977/78 which led directly to the short- ages and price rises in 1980. Hopefully, this experience will lead to greater reliance on the APC's technical judgment and a realization that unrealisti- cally high cane prices do not really benefit most farmers in the long run. While it will be difficult to resist political pressures from farm groups to raise prices, avoiding excessively high minimum support prices will be essen- tial if the 1977/78 experience is to be avoided in the near future. 24. A well executed cane price policy is an essential element but only one element in a longer-term policy. Weather and price variations in other sugarcane uses will continue to cause year-to-year variations around longer- term trends in cane production. Due to the duration of the cane crop and ratooning, these variations will tend to take time to even themselves out by which time new shocks will have destabilized the system. A buffer stocking policy implemented at a suitable time would offer the Government a useful tool for controlling supply variations and dampening the price swings, thereby reducing one of the sources of instability in cane production. Moreover, a buffer stock could give the Government flexibility in the timing of its export sales to take advantage of world market conditions. The suggestion of a buffer stock dates back at least as far as the 1965 Sugar Enquiry Commission Report and is included in the Sixth Plan. Concern over the cost of such a policy has been one major reason why it has not been tried. Also, without an appropriate cane price one could envision ever-increasing stocks of sugar accumulating with the Government. With an appropriate cane price policy, however, the costs of such a buffer stock could be kept reasonably low. By keeping the stocks with the mills themselves (which managed stocks over 4 million tons in 1978) the Government could restrict its financial outlay to the interest charges. Since India is just emerging from a sugar shortage, now is obviously not the time to start a buffer stock. Next year, however, the cane crop is likely to be much larger, perhaps approaching the 1977/78 level. World market prices may be such that excess sugar production could be profit- ably exported, but ISA quotas may limit such possibilities. In any case it - 125 - would seem to make sense for the Government to be ready from next year to start establishing a buffer stock, should the domestic crop be large enough. 25. In years of a short cane crop, mills will continue to have difficul- ties in competing with gur and khandsari units since controls on sugar prices put effective limits on prices mills can offer for cane. In this context and after building up a sugar buffer stock, the possibility of gradually relaxing and eventually removing the sugar levy system should be considered carefully. While the rather disastrous experience with the recent decontrol may make it difficult to advocate decontrol now, it is clear in retrospect that the timing (when sugar stocks were over 4 million tons) and abruptness (including decon- trol of monthly releases) of the 1978 decontrol had more to do with the results than decontrol itself. If the Government maintained buffer stocks plus controls on monthly releases, such wild market fluctuations would not be possible. In addition to allowing mills to compete more effectively for cane, such a decontrol would remove some of the locational inefficiencies built into the present levy price system which pays higher prices for levy sugar to mills located in areas with higher cane costs and which have higher conversion costs. With decontrolled prices, mills would face the same prices, which would encour- age new investment in low cost areas. In striving for better locational efficiency the Government should also consider whether the present system of greater investment incentives (in the form of lower proportions of levy sugar) in areas with lower sugar recovery is in the long-term interest of the sugar industry. 1/ 26. The mix of policies--an appropriate minimum price for cane, buffer stocking, and gradual relaxation of the levy system--would go far in reducing the year-to-year fluctuations in cane production and cane available to the sugar mills. This would reduce' the market risks faced by farmers and mills alike and would help to ensure that the Plan's investment goals in this indus- try are met. Under this system the mills would be more certain of a steady supply of cane and would not bear the entire burden of excessive stock buildup should it develop in years of a large crop. The reciprocal instability in the gur and khandsari sectors, which mirrors the instability in sugar would also be reduced. By reducing the proportion of levy sugar gradually, incentives for shifts to more efficient locations would build up gradually so that such shifts would also be gradual, avoiding the hardship of a rapid dislocation. 27. The long-term competitiveness of India's sugar economy depends on its technical base. Continued emphasis should be placed on improving research and extension efforts in both the cane production and the cane processing fields. The general programs now underway to improve agricultural research institutions and extension services in most states in India should continue to receive high priority. 1/ It is by no means clear where locational advantages really lie. We have seen that Maharashtra's high yields and sugar recovery may be offset by higher costs. Moreover, changes in technology and relative prices may cause locational advantage to shift over time. - 126 - THE SUGAR ECONOMY ANNEX Supply and Demand Projections for Sugarcane 1. Sugarcane Production: Three equations were estimated to project sugarcane production: (i) an estimation of the rate of growth achieved over the 1967/68-1978/79 period: ln (SCP) = 4.65520 + .03785(T); R2 = .69 (84.67) (5.066) where SCP equals sugarcane production in millions of tons and T = time (1967/68 =1...). Fiaures in parentheses are t statistics. ii) an estimation of the rate of growth over the 1960/61-1978/79 period: ln (SCP) = 4.5447 + .02700(T); R2 = .66 (88.49) (5.995) where T = time (1960/61=1,...) iii) an equation including the response of production to relative prices as well as its trend rate of growth: ln (SCP) = 4.455 + .02995(T) + .4487(ln(Pc)); R2 = .72 (77.7) (7.10) (3.086) where T = time (1960/61=1,...) and Pc = an index of the average realized cane price for the preceding two years deflated by the price index of major competing crops (foodgrains, oilseeds, cotton). The average real- ized cane price index was constructed by calculating a weighted average of the index of statutory minimum cane prices (the official cane price index) and the price index for gur during the main harvest months (Decem- ber-March). The weights were the proportion of cane crushed by mills and the proportion crushed by gur and khandsari units. For projection pur- poses, the 20 year average value for Pc was used (1.122). The value for Pc as of January 1981 was about 1.3183. This would imply a 1981/82 crop of over 180 million tons. Using the 20 year average Pc(l.122) shifts the production projections down from what they would have been had present price levels been used. Thus this projection has an implicit assumption of a 15% fall in relative cane prices. 2. Sweetener Production To relate a given level of cane production to production of sweeteners, the following equation was estimated: TSP = .08774 SCP; R2 = .9996 (218.14) although this equation gives an excellent fit to the historical data it is somewhat conservative to use it in the projections because the proportion - 127 - of sugarcane being used for mill sugar is rising and mill sugar production has higher extraction rates in terms of sugar (sucrose) than gur and khandsari. Thus the sucrose content in the total volume of sweeteners would tend to rise. Moreover, technical improvements in the sugar content of cane and in extraction rates are probable during the projection period. Consequently, the use of this fixed coefficient gives a somewhat downward bias to the sweetener supply projections. 3. Sweetener Consumption Total consumption for all sweeteners was estimated in several different ways: i) The historical trend was estimated using the equation: ln(TSC) = 9.006 + .02444(T); R2 = .83 (293.6) (9.54) where TSC is total sweetener consumption and T is time (1960/61=1). ii) Estimates of the relationships between total consumption of sweeteners, total expenditures on all commodities and real prices of sweeteners. These estimates were based on time series data for the 1960/61-1978/79 period: ln(TSC) = 5.01182 + .7631 ln(E) - .2178 ln(Ps); R2 = .98 (26.74) (23.10) (9.07) where E is total private consumption expenditure in billions of 1970/71 Rupees (from CSO, National Accounts Statistics) and Ps is the weighted average price index for sweeteners deflated by the price index for all other commodities. For projection purposes, E was assumed to grow alternattively at 5% and 3.5% from its 1976/77-1979/80 average base of Rs. 356 billion. Ps was assumed to be at its 20 year average of 1.294 whereas its actual 1979/80 value was 1.395. Using the 20 year average Ps shifts the demand projections up somewhat. iii) The projections given in Tables 4A and 4B of Jon Hitchings paper, Demand Projections for India, were used to provide another set of demand projections after adjustments. The figures in Tables 4A and 4B of that paper were used to calcu- late separate growth factors for sugar and khandsari and gur. Khandsari consumption (about 1 million tons in 1973/74) was substracted from sugar and added to gur. The growth factors were then applied to the 1973/74 base figures of 3.52 million tons sugar and 8.35 million tons gur and khandsari. The total sweetener demand projections thus calculated were then adjusted - 128 - downward slightly to be consistent with the 13.1 million ton 1979/80 demand estimated from equation 3(ii) above. The adjusted projections are given below in paragraph 5. 4. Sugar Consumption The share of sugar in total sweetener consumption is growing, reflecting changing consumer preferences. To project the quantity of sugar in total consumption the following equation was estimated from 1960/61-1980/81 historical data: ln(SPTSC) = -1.4034 + .02076 T; R2 = .54 (27.06) (4.80) where SPTSC is sugar as a percent of total sweetener consumption. This yields the following proportions of sugar for the future: 1985 41.3% 1990 45.8% 1995 50.8% 2000 56.4% 5. Production and Consumption Projections The above equations and assumptions yield the following projections for production and consumption. PRODUCTION OF SWEETENERS (million tons) RECENT TREND LONGER TERM TREND LONG TERM TREND WITH PRICE EFFECT (equation l(i)) (equation 1(ii)) (equation l(iii)) 79/80 15.1 14.2 14.5 84/85 18.2 16.2 16.8 89/90 22.0 18.6 19.5 94/95 26.6 21.2 22.7 99/00 32.2 24.3 26.4 - 129 - CONSUMPTION OF SWEETENERS (million tons) Historical Time Series NSS Trend (equation 3(ii)) (para 3(iii)) (equation 3(i)) Expenditure Growth Expenditure Growth 3.5% 5.0% 3.5% 5.0% 1979/80 13.3 13.1 13.1 13.1 13.1 1984/85 15.0 14.9 16.0 15.4 16.4 1989/90 17.0 17.0 19.3 18.0 20.3 1994/95 19.2 19.4 23.2 21.0 25.1 1999/00 21.7 22.1 28.0 24.4 30.5 6. Export Projections Alternative projections of sweeteners available for export were estimated by a) subtracting the NSS high consumption projection from the "recent trend" production projection (equation 1(i) less 3(iii)) and b) subtracting the time series, 3.5% expenditure growth demand projection from the "long term trend with price effect" supply projection (equation l(iii) less 3(ii)). Total domestic sugar consumption was estimated by applying the proportions of sugar to total sweeteners estimated in paragraph 4 above to the total domestic sweetener consumption projection. This amount was then added to the export projection to given an estimate of total sugar production requirements, for comparison with planned sugar production capacity. Annex Table1 (Area '000 hectares) Sugarcane Production Yield - tons per hectare Production - million tons) b/ R Crop All-India Uttar Pradesh Other Sub-Tropical Maharashtra Other Tropical - Year Area Yield Production Area Yield Production Area Yield Production Area Yield Production Area Yield Production 1960161 2415 45.5 110.0 1330 41.0 54-5 624 36.5 22.8 155 78.1 12.1 306 67.3 20.6 1961/62 2455 42.3 104.0 1362 37.7 51.3 655 31.5 20.6 154 71.4 11.0 284 74.3 21.1 1962/63 2242 41.0 91.9 1262 34.2 43.2 564 31.2 17.6 135 79.3 10,7 281 72.6 20.4 1963/64 2249 46.3 104 2 1229 38.7 47.5 529 35.5 18.8 137 83.9 11.5 354 74.6 26.4 1964/65 2603 46,8 121 .9 1390 40.4 56.2 629 37.8 23.8 159 68-6 10.9 425 72.9 31.0 1965/66 2836 43.7 124.0 1490 38.0 56.6 752 33.5 25.2 171 64.3 11.0 423 73.8 -31.2 1966/67 2301 40.3 92.8 1190 33.1 39.4 603 29.5 17.8 161 62.1 10.0 347 73.8 25.6 1967/68 2047 46.7 95.5 998 38.0 37.9 490 34.9 17.1 165 63.6 10,5 394 76.1 30.0 1968/69 2532 49.2 124.7 1203 42.0 50.5 631 35.3 22.3 190 66.3 12.6 508 77.4 39.3 1969/70 2749 49.1 135.0 1376 44.1 60.7 680 38.7 26,3 222 65.0 14.4 471 71.3 33.6 1970/71 2615 48.3 126.4 1345 40.6 54.7 624 39.7 24.8 217 68.0 14.8 429 74.8 32.1 1971/72 2390 47.5 113.6 1274 38.7 49.4 516 37.6 19.4 182 63.2 11.5 418 79.7 33.3 1972/73 2452 50.9 124.9 1308 43.4 56.7 536 39.9 21.4 146 81.6 11.9 462 75.5 34,9 1973/74 2752 51.1 140.8 1473 41.3 60.8 576 41.5 23.9 165 78.6 12.9 538 80.3 43.2 1974/75 2894 49.9 144.3 1492 41.2 61.5 641 39.6 25.4 185 92.8 17.2 576 69.8 40.2 1975/76 2762 50.9 140.6 1441 40.5 58.4 611 42.1 25.7 212 89.1 18.9 498 75.5 37.6 1976/77 2866 53.4 153.0 1456 42.9 65.2 621 41.4 25.7 241 89.2 21.5 548 74.1 40.6 1977/78 3151 56.2 177.0 1637 46.9 76.8 658 43.8 28,8 *246 94.8 23.3 610 78.9 48.1 1978/79 3119 50.2 156.4 1635 38.3 62.6 654 38.7 25.3 244 92.1 22.5 586 78.5 46.0 1979/80a/ 2700 47.4 128.0 1442 - - 546 - - 236 - _ 476 - - a/ Preliminary estimate. b/ Includes Andaman and Nicobar Islands, Assam, Bihar, Delhi, Haryana, Himachal Pradesh, Jammu and Kashmir, Madhya Pradesh, Manipur, Meghalaya, Mizoram, Nagaland, Punjab, Rajasthan, Tripura and West Bengal. c/ Includes Andhra Pradesh, Gujarat, Kerala, Karnataka, Orissa, Pondicherry, Goa Daman and Diu and Tamil Nadu. Source: Estimates of Area and Production of Principal Crops in India, Directorate of Economics and Statistics, Ministry of Agriculture, Government of India, various years from 1960-1980. Annex Table 2 Production and Consumption of Sweeteners Percentage of Mill Sugar Gur and Khandsari Cane used by Opening Total Per Capita Production/ Per Capita Crop Mills Stocks Production Exports Imports Consumption Consmption Cons ption Cons tion Year ------------------------- rOOtons -------------------- (kgs) u tons (kgs) 1960/61 28.2 742 3021 194 - 2087 4.67 6687 14.95 1961/62 26.9 1482 2729 354 - 2601 5.68 6432 14.05 1962/63 22.6 1256 2139 566 - 2502 5.34 6074 12.96 1963/64 24.7 327 2573 243 - 2326 4 .85 6667 13.90 1964/65 27.4 331 3232 267 - 2437 4 .96 7205 14.68 1965/66 29.4 859 3541 392 - 2792 5.56 6911 13.75 1966/67 23.3 1216 2151 235 - 2595 5.04 6125 11.91 1967/68 23.7 537 2248 139 - 2211 4.20 6305 11.98 1968/69 30.2 435 3559 79 - 2609 4.84 7346 13.63 F 1969/70 33.9 1306 4262 217 - 3261 5.91 7401 13.41 1970/71 30,2 2090 3740 395 - 4025 7.15 7437 13.22 1971/72 27,3 1410 3113 144 - 3780 6.59 6998 12.19 1972/73 32.4 599 3873 97 3511 6.00 7030 12.01 1978/74 30.0 864 3948 415 - 3519 6.08 8336 14,39 1974/75 33.6 878 4797 924 - 3457 5,68 8063 13.24 1975/76 29.8 1294 4262 1021 - 3691 5.94 8367 13.47 1976/77- 31.9 844 4840 312 - 3750 5.92 8841 13.95 1977/78 38.0 1622 6461 202 - 4591 7.10 9088 14.05 1978/79 38.2 3290 5841 863 - 6222 9.44 7936 12.04 1979/80a/ 30.5 2046 3900 290 192 5300 7.88 6700 9.97- a/ Preliminary estimate. Source: Cooperative Sugar Directory and Yearbook, 1979, National Federation of Cooperative Sugar Factories, Ltd., 1980. Annex Table 3 Population. Expenditure and Price Data Private Final a/ Wholesale Price Index (FY 1970/71 - 100) Consumption Competing c/ Crop Population Expenditure Adjusted Crops d/ e/ Year (June, millions) (Rs billion, 7Q/1 prices) All Commodities Sugarcane Sugarcane b/ (Prior Year) Sugar Gur Khandsari Sweeteners Non-Sweeteners 1960/61 447.2 212.0 55.8 58.6 54.0 48,.3 64.4 54.2 - 57.5 55.7 1961/62 457;8 216.0 56.0 58.6 S1.S 47.6 64.6 58.9 63.8 61.0 55.6 1962/63 468.5 220.4 58.8 58.6 66.6 49.1 67.6 88.1 76.7 81.1 57.1 1963/64 479.6 226.1 63.8 67.3 85.3 51.0 74.2 99.2 96.1 91.4 61.6 1964/65 490.9 245.7 70.0 67.8 78.7 60.0 76.8 83.9 87.1 82.0 69.1 1965/66 502.6 238.4 77.4 67.8 68.5 70.9 80.0 77.8 92.7 79.5 77.2 1966/67 514.4 243.8 89.0 75.6 109.2 78.1 86.8 162.8 154.2 139.2 85.1 1967/68 526.5 262.7 91.9 100.1 176.8 99.1 98.9 234.4 231.4 193.2 84.0 1968/69 538.9 270.6 92.8 100.1 122.8 99.7 100.4 132.4 157.8 124.4 90.3 a 1969/70 551.7 280.8 97.4 100.1 89.9 94.6 99.7 88.9 101.0 93.0 97.7 1970/71 562.7 298.4 102.6 100.0 105.3 100.2 106.2 129.8 107.1 121.1 101.2 S 1971/72 574.0 307.0 109.8 101.1 140.1 99.4 137.3 180.7 161.8 166.3 105.4 a 1972/73 585.4 300.7 125.7 118.9 168.9 106.6 160.6 208.5 213.7 194.6 120.3 1973/74 579.1 308.8 160.1 115.2 165.0 128.6 160.0 200.6 233.7 190.5 157.7 1974/75 609.1 310.9 177.0 121.8 171.7 169.4 173.3 231.5 250.8 214.9 174.0 1975/76 621.3 334.5 171.4 126.3 161.6 188.2 169.8 235.4 231.1 215.6 168.0 1976/77 633.7 331.9 183.5 125.2 181.2 149.8 167.6 217.8 212.0 202.2 182.0 1977/78 646.7 366.5 184.5 124.9 151.1 166.9 154.6 163.6 16Q.9 160.7 186.4 1978/79 659.3 384.3 197.3 147.5 143.9 172.3 153.0 188.4 148.0 174.6 199.1 1979/80 672.2 366.4 237.1 184.5 265.6 173.6 211.6 375.9 3 321.6 230.5 a/ riscal year data. b/ Represents the sum of the result of multiplying the WPI of sugarcane by the proportion of cane used by mills plus the result of multiplying. the WPI of , u (December-March) by the remainder of one minus the proportion of cane used by mills. p rceant6 a weighted average of the WPIs of foodgrains (82Z), oilseeda (137.), and cotton (5%), legged by one year. 4 Aclud.a supr, VW Apd Waudpari, 3/ lipresj~ate ~oia11 eitive except sugar. gur, and khandsari. Sotrcaes: Index ! a.ro 'of Wholesale Prices, Weekly Data, Economic Intelligence Service; Wholesale Price Statistics, 1947-1978, Volume II, by H. L. Chandhjk, Economic snd Scientific Researdh Foutrdation; National Az6,conts SWatidti6cs, Central Statistical Orginfzation; "Quick Estimates", CSO, January 27, 1981; World Blank estimates. Annex Table 4 Sugar Trade Data World (ISA daily) Weighted Average Levy Sugar Sugar Sugar Sugar Exports Price, f.o.b. major Average Ex-Factory Crop Exports Exports Exchange Exports Unit Value Calendar Caribbean parts Price Price Year (000 tons) (Rs million) Rate (US$ million) ($/ton) Year ( ($/ton)b/ ($/ton)R/ 1960/61 56.0 24.8 4.762 5.2 92.9 1961 69 214.7 1961/62 284.8 145.7 4.76' 30.6 107.4 1962 60 213.7 1962/63 418 . 0 1 70. 1 4. oD_ 35.7 85 .4 1963 61 216. , 1963/64 432.3 257.3 4.76_ 54.0 124.9 1964 184 23 0. 3 1964/65 282.1 189.2 4.762 39.7 140.7 1965 127 2S2.3 1965/66 314.1 15 .1 4. 762 24.2 77.0 1966 44 25.,3 _ 3 1966/67 356.7 167.0 7.000 23.5 67.0 1967 40 183.7 _ 1967/68 228.0 159.S 7.50i' 21.3 93.4 1968 42 189.4 - w 1968/69 100.0 102.1 7.500 13.6 136.0 1969 42 213.8 - 1969/70 82.0 85.5 7,500 11.4 139.0 1920 71 212.1 1970/71 348.0 276.0 7.50O0 36.8 105.7 £971 81 210.8 - 1971/72 317.0 302.2 7.444 40.6 128.1 1972 99 240. J - 1972/73 102.0 133.0 7.706 17.3 169.6 1973 160 301.6 186.8 1973/74 252.5 426.9 7,791 54.8 217.0 1974 208 321.5, 185. j 1974/75 69S.0 3390.0 7.976 425.0 6111.5 1975 b54 330.2 196.8 1975/76 1201.0 4723.0 8. c- .3 545. 8 454.5 1976 449 3i3.3 172. 3 1976/77 5B0.0 1481.0 8.73; 165.7 285.7 1977 255 306.8 165.4 1977/78 70.0 95.0 B. 65I3 22.8 325.7 1978 175 302.7 I81.I 1978/79 73a.0 1319.0 8.206 160.7 217.8 1979 172 29 0.0 207.9 1979/80a/ 564.6 1199.0 8.100 148.0 262.1 1980 213 313.9 2 31 . 'i a/ Preliminary estimate. b/ Calculated from calendar year Wholesale Price Index for sugar by dividing index number by a factor of 0.063 ,which represents the relationship between the WPI and the weighted average of the market prices used to estimate the WPI. c/ For D-29 grade sugar, ex-Maharashtra factories. Source: Cooperative Sugar Directory and Yearbook. 1979, National Federation of Cooperative Sugar Factories, Ltd., 1980. RECENT PAPERS IN THIS SERIES No. TITLE OF PAPER AUTHOR 470 Industrial Country Policy and Adjustment to Imports J.M. Finger from Developing Countries 471 The Political Structure of the New Protectionism D. Nelson (consultant) 472 Adjustment to External Shocks in Developing Countries B. Balassa 473 Food Policy Issues in Low-Income Countries E. Clay (consultant) 474 Energy, International Trade, and Economic Growth A. Manne (consultant) 475 Capital-Importing Oil Exporters: Adjustment Issues A.H. Gelb and Policy Choices 476 Notes on the Analysis of Capital Flows to Developing R.C. Bryant (consultant) Nations and the 'Recycling' Problem 477 Adjustment Experience and Growth Prospects of the F. Jaspersen Semi-Industrial Countries 478 Trade Policy Issues for the Developing Countries I. Frank (consultant) in the 1980s 479 Trade among Developing Countries: Theory, Policy 0. Havrylyshyn Issues, and Principal Trends (consultant) M. Wolf 480 Trade in Services: Economic Determinants and A. Sapir (consultant) Development-Related Issues E. Lutz 481 International Migrant Workers' Remittances: Issues G. Swamy and Prospects 482 Private Bank Lending to Developing Countries R. O'Brien (consultant) 483 Development Prospects of the Capital Surplus R. Hablutzel Oil-Exporting Countries 484 Private Capital Flows to Developing Countries and A. Fleming Their Determination 485 International Adjustment in the 1980s V. Joshi (consultant) - 2 - No. TITLE OF PAPER AUTHOR 486 Adjustment in Low-Income Africa R. Liebenthal 487 A Comparative Analysis of Developing Country C. Wallich Adjustment Experiences in the 1970s: Low-Income South Asia 488 Developments in and Prospects for the External N. Hope Debt of the Developing Countries: 1970-80 and Beyond 489 Global Energy Prospects B.J. Choe A. Lambertini P. Pollak 490 Sociocultural Aspects of Developing Small-Scale R. B. Pollnac (consultant) Fisheries: Delivering Services to the Poor 491 The Educational Use of Mass Media S. Futagami et al 492 The Political Market for Protection in T. Anderson Industrial Countries R. Baldwin 493 On Protectionism in the Netherlands K. Koekkoek 494 Patterns of Barriers to Trade in Sweden L. Lundberg 495 Social Infrastructure and Services in Zimbabwe R. Faruqee 496 Economic Aspects and Policy Issues in Groundwater I. Carruthers R. Stoner (consultants) 497 Labor Productivity: Un Tour D'Horizon S. Horton (consultant) T. King 498 The Determinants of Labour Earnings in Developing R. Mohan Metropoli: Estimates from Bogota and Cali, Colombia PUB HG3881.5.W57 W67 no.500 Harrison, James Q. Irxiia, denand ard supply prospects for agriculture /