|
II.4. Future CEREALS Demand
4.1. The Committee has obtained details of and projections from a model developed at the Centre for Economic and Social Studies (CESS), Hyderabad, using a piece-wise Linear Expenditure System[3], and has analysed this:
Based on NSS data, this model allows tastes to change over time, from food to non-food, from cereals to other food and from coarse cereals to rice and wheat. Moreover, the model is able to incorporate income distribution shifts.
The results of this model supports the view that per capita cereals consumption is likely to increase as per capita incomes rise, rather than the opposite view that rising incomes will cause cereals consumption to fall.
However, not only is the income elasticity of demand (or response of cereals consumption to higher income) declining and is now quite low, there is clear evidence that changes in tastes have caused a shift from coarse cereals to rice and wheat, from cereals to other food and from food to non-food.
Moreover, the model indicates that cereals consumption has fairly strong negative response to cereals prices.
Cereals prices rose 30 per cent more than the general price level between 1990-91 and 1999-00 and, according to the model, would have caused per capita demand to decline by at least 9 per cent. This is exactly the magnitude of actual decline shown by the NSS for total cereals during the period.
Also, since per capita consumption of coarse cereals fell 40 per cent and of rice and wheat only 3 per cent, there is confirmation not only of taste shifts from coarse cereals but also of the model prediction that per capita rice/wheat consumption would have risen about 7 per cent during the 1990s had relative cereals prices not increased.
A major implication from the model is that the high cereals price rise during the 1990s not only restrained cereals demand but, since cereals are the cheapest source of nutrients, may also have affected the nutrition condition of the poor adversely. This puts in perspective the observed dietary diversification, although the model itself shows that rising incomes will cause large increases in the demand for non-cereals food, particularly livestock and horticulture products.
4.2. Projections to 2020 have been made using this model taking as base 1998, when NSS cereals consumption was the lowest. Consumer expenditure is assumed to grow at 5 per cent per annum in real terms and 2020 population is assumed at 1343 million, of which 33 per cent is urban. On this basis:
The food use of rice and wheat is projected at 132 million tonnes in 2000 (roughly actual NSS consumption in 1999-2000 and official availability in TE 2000), 170 million tonnes in 2010 and 211 million tonnes in 2020.
The food use of coarse cereals is projected at 13 million tonnes in 2000 (the NSS actual but much less than official availability), rising to only 14 and 16 million tonnes in 2010 and 2020.
Total food use of cereals is thus projected to 184 million tonnes in 2010 and 227 million tonnes in 2020.
This is lower than behaviouristic projections of human cereals consumption (188 million tonnes for 2006-07) made by the Tenth Plan Working Group on Crop Husbandry, Demand & Supply Projections and Agricultural Inputs.
The implied per capita consumption is 169 kpca in 2020, which is 14 per cent higher than the NSS estimate for 1999-2000 but around levels actually reached in the late 1980s and early 1990s by both NSS and official availability.
However, the estimate for 2020 has a margin of error, of almost 10 million tonnes either way. For example, this could be 4 per cent higher since real cereals prices have declined 13 per cent since 1999. Conversely, if the tastes shift from coarse cereals continues, demand in 2020 could be about 10 million tonnes less than projected above.
As against the projected growth rate of food demand of 2.2 per cent per annum for cereals, the model predicts 5.2 per cent growth for milk, 5.0 per cent for meat, 4.6 per cent for fruits and vegetables, 4.0 per cent for sugar and 3.8 per cent for edible oils. The implied growth rate of per capita demand for non-cereals food is higher than in the past.
4.3. The results from this model have been compared to other available projections. Keeping data caveats in perspective, the Committee has reviewed all recent estimates of future cereals demand. It has found that although there is general agreement that the rate of growth of cereals demand in the future will be much less than in the past, there are considerable differences regarding magnitudes. Currently available projections of cereals demand in 2020 vary on each of the essential elements - population, per capita direct human consumption, and the requirement for animal and poultry feed:
The highest estimates are from Bhalla, Hazell and Kerr (1999)[4] who place food and feed demand for cereals in 2020 at 246 and 50 million tonnes respectively, assuming population at 1329 million. The implied per capita cereals demand for food is 185 kpca which, although only about 10 per cent higher than per capita consumption in the late 1980s, is 20 per cent higher than actual average consumption in TE 2000. This is, however, in line with the Food and Agriculture Organisation (FAO)’s latest forecast, placing per capita food demand for cereals in South Asia at 186 kpca in 2015 and 192 kpca in 2030[5]. The feed demand projected by these authors is also considerably larger than in the other estimates.
Using the same population assumption as Bhalla et al., the International Food Policy Research Institute (IFPRI)’s IMPACT model forecasts 2020 food and feed demand at 224 and 13 million tonnes[6]. The feed projection from this model is the lowest of all available estimates. This assumes that present feed use is only 2.5 per cent of production and, as discussed above, may be an underestimate if coarse cereals are being diverted from food to feed. The food demand from this model works out to 168 kpca. This was actual consumption in India in the late 1980s, but is about 10 per cent higher than actual consumption in TE 2000.
Bansil (1999) projects food and feed demand for 2020 at 228 and 31 million tonnes respectively, assuming population at 1360 million tonnes. In per capita terms, the food demand is almost exactly the same as in the IMPACT model above.However, feed demand is put considerably higher than in IMPACT, though much less than in Bhalla et al.
Kumar (1998) projected food and feed demand at 238 and 15 million, using a 2020 population of 1421 million[7]. In per capita terms, this is also similar to the IMPACT projection for both food and feed. However, a later projection from the same source (Paroda and Kumar, 2000) has revised the food demand sharply downward to only 152 kpca. This lower level is intermediate between the NSS and availability estimates for 1999-2000.
The lowest food estimate is from Dyson and Hanchate (2000)[8] who put food and feed demand at 194 and 30 million tonnes, projecting 2020 population at 1315 million. This basically accepts Bansil’s feed estimate, but for food, state-wise NSS 1993-94 estimates were adjusted to the 7 day reference period and projected using past NSS growth rates of per capita consumption. Since NSS consumption has declined over time, projected per capita food demand for cereals in 2020 is 15 per cent less than the assumed base. At 147 kpca, this is equal to the NSS estimate for 1999-00.
Thus, none of the projections expect per capita food demand for cereals to fall below its current low level but none except Bhalla et al and the FAO expect it to rise much above the 166 kpca level to which both NSS consumption and official availability had converged at the end of the 1980s. Also, despite large differences, all the estimates predict higher feed demand than in the past.
4.4. The CESS projection of 2020 food demand for cereals is thus in agreement with IMPACT, Kumar and Bansil, and half-way between Bhalla et.al and FAO on the one hand and Dyson-Hanchate and Paroda-Kumar on the other. It should be noted, however, that CESS begins with a lower base than the other projections, most of which start with a base food consumption of cereals at around 167 kpca which remains constant over time. Like CESS, Bhalla et al and FAO have per capita food demand increasing over time but their higher base leads to higher projections for the future. Dyson-Hanchate, on the other hand, start with even higher base consumption and project very large decline in per capita consumption over time. Hence, despite the similarities in projected values, there are differences, since unlike CESS most of the other models do not incorporate price effects explicitly.
4.5. However, the CESS model does not make any projections of the feed demand for cereals, on which there is a very large gap between the IMPACT estimate of 13 million tonnes in 2020 and Bhalla et al’s 50 million tonnes. Differences here not easy to resolve since there is no reliable data available. Although the IMPACT (and Kumar 1998) feed projections are relatively low, these actually involved a tripling between 1993 and 2000 of the feed use assumed as base. This base was, however, nothing but a convention followed since the 1950s of treating feed use as 2.5 per cent of gross output, along with similar norms of 5 per cent each for seed and waste. On the other hand, Bhalla et al had originally arrived at a much larger feed demand than in their final estimate, assuming structural changes in the livestock sector and based on data from neighbouring countries in Asia. Bansil had questioned this, and arrived at his estimate of 31 million tonnes based on norms of feed requirement to final demand for livestock products, derived from India’s organised sector. He had, however, indicated that this estimate was likely to be an upper limit since cereals use for feed in the unorganised sector is lower. Bhalla et al’s final estimates are derived by applying Bansil’s norm to their higher estimates of demand for livestock products. One consequence of this recent debate is that it is now generally agreed that the existing norm is an underestimate and that at least a part of what is treated as waste is actually feed.
4.6. The Tenth Plan Working Group on Crop Husbandry, Demand & Supply Projections and Agricultural Inputs has in fact reversed the norm, taking 5 per cent as feed and 2.5 per cent as waste. This increases the estimate of present feed use from about 5 million tonnes to about 10 million tonnes. But even this may be an underestimate since there is currently a gap of about of about 10 million tonnes between coarse cereals availability and their consumption according to the NSS, which has appeared entirely during the 1990s. If this is treated as feed and added to whatever is considered the correct feed component in the current norm for feed plus waste, feed use has grown from 4-8 million tonnes in the end 1980s to 15-20 million tonnes at the end of the 1990s, i.e. at a minimum growth rate of 6 per cent per annum which, although high, is in line with actual growth of milk and meat production.
4.7. Applying the norms used by Bhalla et al (based on Bansil) to actual and projected milk/meat production in 1990, 2000 and 2020 from the CESS model gives estimates of 7,12 and 33 million tonnes for feed requirement to produce milk/meat. Total food and feed demand for cereals in 2020 from the CESS model then works out to 260 million tonnes, halfway between the Bhalla et al and Dyson-Hanchate estimates of 296 and 224 million tonnes respectively.
II.5. IMPLICATIONS FOR GRAIN POLICY
5.1. Taken in conjunction with the much higher growth of non-cereals food consumption predicted not only by CESS but also the other models, the estimates above imply considerable increase in the demand on existing land and water resources for meeting the required growth of food demand from domestic supply. Estimates of cereals supply in 2020 do however suggest that there is no reason for undue pessimism. For example Kumar (2001) projects the supply of rice and wheat at 115 and 105 million tonnes in 2020 as against respective food demand of 119 and 92.million tonnes projected by CESS. This implies shortage for rice and excess for wheat. Like demand estimates, supply estimates for 2020 also vary from a low of around 234 million tonnes to over 300 million tonnes. The consensus appears to be that India will be able to produce enough cereals to meet of food and feed demand around 260 million tonnes by 2020 provided that there is adequate increase in factor productivity through continuous genetic improvements and greater efficiency input use. Nonetheless since changing demand patterns will require much higher diversion of land and water resources to other crops, the prospect will require considerable production effort and should not be cause for complacency. Nonetheless there is no basis at present to assume any large shortages to emerge over the next 20 years, although available estimates certainly do not indicate that surpluses as reflected in the current magnitude of stocks will be permanent.
5.2. This assessment implies that the present situation of large stocks cereals is likely, if properly managed, to give way to much greater balance between supply and demand over the coming decade. This should be accompanied by a rapid pace of diversification of both agricultural production and the pattern of food consumption. Also, within cereals, feed demand is likely to increase much more rapidly then food demand. In terms of agricultural policy the priorities are therefore to encourage diversification of area from cereals to higher value crops while intensifying yield improvement in the case of cereals crops. From the point of view of the policy framework the most important implication of this is that, unlike in earlier years, price policy should not encourage greater area under rice and wheat. However, investment and technology support needs to be intensified in order to get the required yield increase, the potential for which is greatest in Eastern and Central India. The price environment would need to be more conducive for the growth of cereals yield in the lower Gangetic plane and of non-cereals agriculture elsewhere.
5.3. Certain other implications of available supply and demand projections need also to be noted:
Available price elasticities of demand for rice and wheat place these around 0.3 - 0.5 which imply that a 10% price increase is likely to reduce the demand for rice and wheat by about 5 million tonnes currently and about 9 million tonnes in 2020. Thus, the price scenario can be quite critical to actual demand outcomes.
On the basis of the available NSS data it appears that although the income elasticity for cereals has declined considerably over the years it is still fairly high for the poorer segments of the population with the bottom 40% consuming about 1/3rd less then the top 20%. However, this is very low for richer groups and is negative among the rich in urban areas.
It is significant to note in this context that although the demand for rice and wheat has tended to decline over the 90’s for the population as a whole, it has increased among the bottom 40% of the population.
Consequently a thrust on poverty removal can significantly increase cereals demand and conversely an increase in poverty could decrease cereals demand. Even at the high prices of 1999-2000 cereals demand would have increased by about seven million tonnes, or by more than 5%, if the population below poverty were raised to the poverty line.
There are very large inter-regional variations in cereals demand. Per capita cereals consumption appears to be low in the relatively richer states such as Punjab, Maharashtra and Gujarat and high in poorer states like Orissa and Bihar. As a result, increased regional equality may also lead to higher cereals demand while greater regional inequality would tend to reduce such demand.
The most important implication of these variation in cereals demand across regions and income groups is that any increase in cereals price either due to shortage or otherwise hurts the poorer segments of society most; and simultaneously that food demand of cereals is now even more dependent on whether incomes accrue to the poor.
Conversely, food demand for cereals may fall if future income growth bypasses rural areas and is concentrated among relatively better-off urban groups. However, this will increase feed demand more.
In fact, if the pattern of growth observed since 1990 continues, the real increase in cereals demand will be for feed. Thus, quality issues will become more important, and it will be necessary to ensure better segregation of stock and easier release for feed rather than food use.
Nonetheless, with most of the poor living in rural areas and over 75% of cereals demand originating in rural areas the trends in rural incomes will be critical in determining cereals demand. Since India’s nutritional outcomes remain appalling, what happens to rural incomes will also determine the extent of hunger and the quality of future generations.
5.4. This aspect, that cereals demand will depend essentially on what happens to rural incomes, highlights the central importance of an integrated approach to diversifying agriculture will require a larger growth in cereals output then if agriculture was stagnant. Also, the process of diversification agriculture would require that those who move out of cereals production and rely on other crops can be assured of their food supply. This involves more than just adequate cereals production but also a stable relationship between cereals prices and the prices of the other crops. Thus, the need for a stable price environment, particularly in the matter of inter-regional variations in cereals supply and of inter-crop parties is a necessary requirement for achieving the desired direction of agriculture growth.
5.5. Especially in the present situation of large food stocks, it is also necessary to stress that a very significant dent on hunger and malnutrition can be made immediately with the stocks available. The annual stock accumulation of 7-10 million tonnes since 1998 represents roughly the requirement to bring cereals consumption of the poor and the destitute to nutrition norms. This can, however, be accomplished only through a suitable delivery system which targets the food to those in need and without adequate income. Leakages into the open market without the requisite demand will only bring down prices. Thus, the ability of cereals producers to obtain adequate returns will depend on reforms in the delivery system and on income creation schemes targeted at the poor. This is clearly where public policy will have to play a critical role, but it should also be understood that if this is done, there will be no surpluses and the importance of the production effort will become more significant.
5.6. In this context, it should be observed that world cereals trends have shown a marked slow-down in production recently (chart 2.10). Also, the FAO projects developing country cereals imports increasing from 107 million tonnes in 1995-97 to 198 and 270 million tonnes in 2015 and 2030, with very large shortages in Near East/North Africa and East Asia. In this scenario, North America and Europe will monopolise supplies in world trade and may cut the huge farm subsidies that they give currently. World cereals prices would then average at least at mid-1990s levels, i.e. twice present levels. Much will depend on biotechnology, and whether this will reverse the present trend for per hectare world yields of both rice and wheat to grow slower than world population. In this situation, India cannot afford to neglect the priority of cereals self-sufficiency which, despite present high stocks, cannot be taken for granted.
CHART: 2.10
[1] Svedberg, Peter (2001): Hunger in India – Facts and Challenges
[2] Chakravarti, Lalita (2001): Biological Stress signals in the Millet zone of India
[3] The Committee is thankful to Prof. R. Radhakrishna and Dr. C. Ravi for making the model and its results available.
[4] Bhalla, G.S., P. Hazell and J. Kerr (1999): Prospects for India’s Cereal Supply and Demand to 2020. Food, Agriculture and Environment Discussion Paper No. 29, International Food Policy Research Institute, Washington.
[5] FAO (2000): Agriculture towards 2015/30, the Interim Report, Rome.
[6] The IMPACT results are from Rosegrant, M.R., M. Agcaoili and N. Perez (1995): Global food projections to 2020: Implications for Investment. Food, Agriculture and Environment Discussion Paper No. 5, International Food Policy Research Institute, Washington.
[7] Kumar, P. (1998): Food Demand and Supply Projections for India. Paper No. 98-01, Indian Agricultural Research Institute, New Delhi
[8] Dyson, T. & A. Hanchate (2000): India’s Demographic and Food Prospects, EPW, November
| 
