Economic implications of farm size distribution, tenancy, cropping patterns and income distribution in Bihar's agriculture

ECONOMIC IMPLICATIONS OF FARM SIZE DISTRIBUTION, TENANCY, CROPPING PATTERNS AND INCOME DISTRIBUTION IN BIRAR'S AGRICULTURE by Hiromitsu Kaneda Series Studies in Employment and Rural Development No. 36 Division Employment and Rural Development Department Development Economics Development Policy Staff International Bank for Reconstruction and Development The author prepared this paper in the context of the foodgrain production review mission to Bihar State, India. The author is solely responsible for the views, analyses and conclusions expressed in the paper. This paper may not be published nor may it be quoted as representing the views of the Bank and its affiliated organizations. Washington, D.C., January 1977 ECONOMIC IMPLICATIONS OF FARM SIZE DISTRIBUTION, TENANCY, CROPPING PATTERNS AND INCOME DISTRIBUTION IN BIHAR'S AGRICULTURE Table of Contents Page No. 1. The Poverty of Rural Bihar . . . . . . . . . . . . . . . . . 1 2. The Performance of the Foodgrain Production Subsector . . 2 3. Distribution of Land Holdings and Tenure . . . . . . . . . . 3 3.1 Changes in Distribution of Holdings, 1961/62 to 1971/72 . . . . . . . . . . . . . . . . . 3 3.2 Changes in Tenurial Relations, 1961/62 to 1971/72 . . . . . . . . . . . . . . . . . 5 3.3 Special Characteristics of Tenancy in Bihar . . . . . . 7 4. Distribution of Estimated Cropping Incomes in Bihar . . . . 7 4.1 Estimation Procedures . . . . . . . . . . . . . . . . . 8 4.2 The Incidence of Rural Poverty 9 5. Economic Implications . . . . . . . . . . . . . . . . . . . 11 List of Tables Table No. Title Page No. 1 Distribution of Household Ownership Holdings and Operational Holdings by Size . . . . . . . . . 4 2 Changes in Tenurial Relations by Size of Ownership Holding, 1961/62 and 1971/72 . . . . . . 6 3 Estimated Net Income from Cropping Activities by Size of Holdings, Average for Bihar State. . . . 10 ECONOMIC IMPLICATIONS OF FARM SIZE DISTRIBUTION, TENANCY, CROPPING PATTERNS AND INCOME DISTRIBUTION IN BIRAR'S AGRICULTURE 1. The Poverty of Rural Bihar By almost any index of economic development, Bihar ranks one of the lowest among the states in India. Per capita income of Bihar in current prices was Rs. 489 in 1972/73 when the national figure stood at Rs. 698. In 1960/61 prices Bihar's per capita income was Rs. 230, while that of the nation was Rs. 337. It is worth recalling that a somewhat arbitrary poverty line of Rs. 20 per capita per month in 1960/61 prices was recommended by a group of experts appointed by the government. During the period between 1960/61 and 1972/73 (the latest year for which the income data are available) there was no clearly discernable trend in real income per capita in the state, although the incomes measured in current prices rose at the rate of 8 percent per annum (i.e., the rate of inflation). At this rate of increase in prices the poverty line of Rs. 240 per capita per year in 1960/61 should correspond to Rs. 690 in 1972/73. Some Indian scholars have used Rs. 15 per capita per month in 1960/61 prices for rural areas acknowledging that the prices of major components of the consumption basket are lower in rural areas than in urban. If this figure is accepted, the rural poverty line in India in current prices would be (assuming the 7 percent rate of inflation) Rs. 434in 1972/73 and Rs. 674 in 1975/76. Reflecting the low income levels the latest consumer expenditure survey (1969/70) indicates that the Engel ratio (proportion,of expenditure devoted to food) of rural households in Bihar averages 78 percent (ranging from the high of 84 to the low of 61 percent) and that for the urban house- holds averages 70 percent. It is significant to note that 51 percent of consumption expenditure of rural households and 35 percent of that of the urban counterparts in Bihar are devoted to foodgrains. It is not surprising, therefore, to find that the income elasticities of demand for food are .90 for rural and .86 for urban households in the state. More pertinent is the finding that the income elasticities of demand for foodgrains amount to .83 in rural Bihar and .49 in urban areas, respectively. In the year of the consumption survey the "poverty-line income" in rural areas (computed on the assumptions used earlier) was Rs. 397 per capita in 1969/70 prices. If we divide the income classes (by monthly per capita expenditure) into two groups, one above the poverty line income and one below that, and compute income elasticities of demand for foodgrains, we find that those below the annual per capita income of Rs. 397 would have raised their con- sumption expenditure for foodgrains by somewhat more than one percent for -2- a one percent rise in income and those above the poverty line by approxi- mately a half that percentage for a one percent rise in income. Simply put, this means that for those below the poverty line in the rural areas of Bihar,foodgrains (defined to include: rice, wheat, jowar, bajra, maize, barley, small millets, ragi and their products) are luxury goods whose consumption would rise proportionally faster than a rise in expendable resources. It is immediately apparent that the depth of the poverty of rural Bihar compels the state's foodgrain production increases to be made accessible to the majority of the 90 percent of the state's population residing in rural areas. 2. The Performance of the Foodgrain Production Subsector Since the early 1960's the state's population has grown at the annual rate of about 2.0 percent. With approximately 62 million people in the state in 1976 it is expected that the population is to increase at the rate of some 2 percent per annum in-the coming decade or so, adding roughly 5 million persons every five years. .However, since 1960 Bihar's agricultural output has grown at a lower pace than has population, 1.9 percent as compared with 2.2 percent per annum. Thus, Bihar has had a deteriorating per capita food availability. This has been especially pronounced since the 1970's. According to GOB statistics the output of oodgrains averaged 8.4 million tons for 1970/71 and 1971/72 while the foodgrain output in the most recent crop years 1973/74 and 1974/75 averaged only 8 million tons. In the mean- time, the state's population has increased by some 3 million. Even in 1970/71, according to a calculation by GOB, the "nutritional requirements" for food- grains in Bihar amounted to 10.9 million. There is no denying the fact that under-nutrition as well as mal-nutrition exist widely in the state. For the period since 1960 the performances in cropped acreage, yield and output have differed among different foodcrops. Rice production has been stagnant at about 4.5 million tons with an area of 5.2 million hectares in 1960/61 and about the same in 1974/75. Yields and area for wheat have increased steadily at 6 and 8 percent per annum, respectively. Since the mid-1960's the rates of increase have accelerated to 9 and 10 per- cent per year. In 1974/75 wheat production was 1.7 million tons and the area planted to the crop was about 1.5 million hectares up from 660,000 hectares in 1960/61. Potatoes were the only other major food crop whose area reached 114,000 hectares in 1974/75, having grown by about 2.8 percent per annum during the period under discussion. However, potato yields have shown no significant trend. Among foodgrains ragi and barley, Khariff and Rabi crops, respectively, have steadily lost acreage and production. Among pulses the most important grain crop has been losing acreage at the annual rate of some 6 percent. It should be noted that gram and other pulses have income elasticities greater than unity in both rural and urban areas. It is important to note that there are substantial year-to-year fluctuations in food production due mainly to the unreliable rainfall -3- patterns that often result in droughts or floods. The stagnation in the yield performance of major foodgrains, such as rice and maize, ,should be noted in conjunction with considerable annual variations in their yields. Taking the average yield over the five year period between 1969/70 and 1974/75 as the base and constructing indices of annual yields, one finds that rice yields fluctuate for more than 10 percent (up or down on the previous year's level) eight out of fifteen years, while maize yields do the same ten out of the fifteen years. Perhaps it is equally important to note that the average yields by Bihar's agricultural districts (the old seventeen districts) of major crops, including wheat and potatoes as well as rice and maize, diverge substantially for any given year. The annually computed coefficients of variation (standard deviations divided by the annual means) range between .16 to .46 for rice, between .11 to .46 for wheat, .17 to .50 for maize, and .24 to .67 for potatoes. The variability among district-level yields, moreover, is not at all lessening over the years under disucssion for any of the crops. 3. Distribution of Land Holdings and Tenure There is little'additional cultivable land left in Eihar. At the time of the Agricultural Census of 1970/71 the average size of an operational holding varied from the low of one hectare or less in the mid-state districts on either side of the Ganges (such as Saran, Vaishall, Nalanda, Muzaffapur, and Samastipur) to the high of 3 hectares in Ranchi. The average for the state as a whole was 1.5 hectares per holding with about 5.5 persons in a household. The unfavorable (and deteriorating) land/man ratio in Bihar is further accentuated by the extremely skewed distribution of land. According to the Agricultural Census, although farmers with less than 1 hectare account for 64 percent of all the holdings in the state, they farm only 16 percent of the land. Farmers with more than 2 hectares account for about 21 percent of the holdings an,' control 70 percent of the cultivated land. Thus, approximately 80 percent of operational holdings are less than 2 hectares in size. In addition, more than 70 percent of the state's working popu- lation are "marginal farmers" (with less than 1 hectare) or landless agri- cultural workers. According to the Census of 1971, agricultural laborers numbered 6.8 million in the state comprising approximately 39 percent of the state's total work force. 3.1 Changes in Distribution of Holdings, 1961/62 to 1971/72 According to the National Sample Surveys of 1961/62 and 1971/72, the distribution of both ownership holdings and operational holdings changed towards lower size groups. Allowances for sampling errors notwithstanding, the change over the ten-year period is clearly discernable in Table 1, which contrasts the patterns of distribution by the size of holdings. The lowest size group, those with less than 1 hectare, increased by 4 percent in number of holdings as well as in the cultivated area they control. On the other hand, it is clearly seen that the groups with more than 4 hectares decreased their relative importance rather significantly during the decade. -4 Table 1: DISTRIBUTION OF HOUSEHOLD OWNERSHIP HOLDINGS -.-AND -OPERATIONAL HOLDINGS BY SIZE 1961/62 and 1971/72 Bihar State, Rural Household Own- 1961/62 1971/72 ership Holding Percent of Percent of Percent of Percent of Size :in hectares Number of h1dgs Area Number of h1dgs Area Less than 1.0 68.0 14.2 72.7 18.2 1.0 - 2.0 15.8 20.2 15.1 23.4 2.0 - 3.0 6.7 15.7 6.3 17.3 3.0 ~-~.0 3.3 10.2 2.9 10.8 L.0 - 10.0 5.2 26.3 2.7 23.6 10.0 and more .9 13.5 .L 6.7 Total 100.0 100.0 100.0 100.0 O'peira- .. tional Holding Size in hectares Less than 1.0 54. 14.9 58.9 18.1 1.0 - 2.0 23.0 21.5 23.3 26.2 2.0 - 3.0 9.9 15.7 8.5 17.0 3.0 - 4.0 4.7 10.6 4.3 12.0 4.o - 10.0 7.1 26.0 6.5 21.0 10.0 and more 1.0 11.3 .4 5.7 Source: Government of India, Ministry of Planning, Department of Statistics, National Sample Survey Organisation, National Samole Survey, No. 144, Seventeenth Round, 1961-62, and No. 215, Twentysixth Round, July 1971- September 1972. -5- Overall, it appears that the distribution has become a little less inequitable. Nonetheless, the stark reality of the change is that the overwhelming majority of small and marginal farmers just increased in number over the decade. Significant also in this regard is the change in the magnitude of the landless workers in agriculture. Due to defini- tional changes, figures from the Census of 1961 and that of 1971 are not strictly comparable. It is worth noting, however, that the number of agricultural laborers was reported to have been 4.4 million in 1961 and 6.8 million in 1971. 3.2 Changes in Tenurial Relations, 1961/62 to 1971/72 The 1970/71 Agricultural Census of Bihar reports only 0.4 percent of the holdings on rent. It is well known, however, that the tenurial data in the Agricultural Census do not reflect the reality. As the census was in fact a compilation of the existing data from the land and revenue records rather than a direct enumeration from the field enquiries, the census figures under-record the actual incidence of tenancy. The reasons for under-recording are well known also. The sharecroppers are reluctant to come forward and record their claims for fear of eviction. In addition, the common practice of oral agreements with respect to tenancy tends to make tenanted land appear as owner-cultivated. Given the increasingly large number (and proportion) of marginal farmers and agricultural laborers, it is not surprising that the system of sharecropping is widespread in Bihar. It is reported that sharecropping is preponderant particularly in the north of the Ganges in such districts as (old) Darbhanga, Madjubani, Purnea and Champaran, where large landowners are said to maintain strongholds. It is extremely important to know how the incidence of share- cropping and tenancy in general has changed over the recent decade or so. Unfortunately, desirable data are hard to come by. Comparing the National Sample Survey data for 1961/62 and for 1971/72 (Table 2), one observes that the number of holdings leasing out owned land increased from about 7 percent to some 16 percent and that the area they lease out did also increase from about 4 percent to about 7 percent. Looking at the picture from the other side, the sample survey of 1971/72 indicates that almost 40 percent of operational holdings in Bihar leased in land under some leasin- terms, 28 percent of which were "for share of produce" (i.e., sharecropping). These operational holdings on average had 14 percent of their acreage under some form of lease, of which some 11 percent were for produce share. It is important to note that a substantial proportion of the smallholders are tenants. In 1971/72 about 1/3 of the operational holdings with less than 2 hectares (more strictly, 0.5 acres to 4.99 acres) were sharecroppers. If we include those who lease in land under some other terms (in addition to share- croppers), the proportion of tenants in this size group rises to almost 45 percent. -6- Table 2: CHANGES IN TENURIAL RELATIONS BY SIZE OF OWNERSHIP HOLDING 1961/62 and 1971/72 (In percentages) Bihar, Rural Household Owner- 1961/62 1971/72 ship Holding Households leas- Uwned Area Households Ow'cied Area Size in acres ing out land leased out leasing cut leased out 0.01 - 0.49 5.11 4.57 7.88 12.22 0.50 - 0.99 7.96 4.31 21.17 15.11 1.00 - 1.24 17.28 9.31 ) 6.30 ) 1.90 1.25 - 2.49 19.06 6.85 2.50 - 4.99 7.65 2.L9 20.13 5.82 5.00 - 7.49 10.23 3.43 26.24 8.39 7.50 - 9.99 11.15 2221 22.22 3.55 10.00 - 12.49 8.33 1.92 27.68 4.42 12.50 - 14.99 14.29 5.49 o0.32 12.34 15,00 - 19.99 16.67 3,52 29.75 6.20 20.00 - 24.99 21.15 3.56 40.66 6.37 25.00 - 29.99 33.33 11.61 24.11 2.69 30.00 - 49.99 21.21 4.90 15.67 1,10 50M00 and above 25.00 7.95 40.74 11.52 All Sizes 7.30 3.56 15.80 6,78 Source: Same as Table 1. Note: Figures are in percentages of holdings or owned area belonging to the same size group. -7- 3.3 Special Characteristics of Tenancy in Bihar Table 2 offers data for a highly important observation with regards to the nature of tenancy in Bihar as well as the nature of its change in recent years. First, it is significant that in 1971/72 almost 20 percent of the size group owning less than 2 hectares leased out-land. amounting to some 6 to 10 percent of their owned land. This means that in addition to the problems of insecurity of tenancy that depends mostly on oral contracts, whereby sharecroppers' rights and claims can be resisted by the landowners' implied threat of eviction, a sizeable part of tenancy in Bihar is subject to a system of petty landlords, who themselves are too small to have any positive impact on the agricultural practices of their tenants. Secondly, looking at the decade's change in the pattern of owner- ship holdings that lease out, one must conclude that more small landowners have come to lease out in recent years than before. Lacking the data that indicate the change over the decade in the distribution pattern of holdings, that lease in, we do not know for sure to whom the incremental amounts of the small landowners' leased-out land went. In view of some observations made in Punjab and Haryana, a presumption is strong that some (many, or a majority?) of the small landowners are leasing out to the larger ones. This may be a result of the situation where the small landowners find it more difficult to have access to economic and public service inputs (e.g., water, power, extension services and credits, etc.) than their larger counterparts. Or, this may indicate that the small can less afford to bear the risks attendant to modern agricultural practices than the large. The observation is consistent nonetheless with the greatly increased number of landless workers in agriculture in the state. 4. Distribution of Estimated Cropping Incomes in Bihar According to the official estimates, the income originating in Bihar's agriculture amounted to Rs. 12,740.4 million in 1972/73, while it was worth Rs. 9,728.0 million in 1970/71 in current prices. These figures exclude such allied activities as animal husbandry, forestry, and fisheries. The figure for 1970/71 implies nonetheless that in the year of the Agri- cultural Census the average current income per holding was Rs.1,285 from cropping activities. Assuming that the number of holdings in Bihar has grown at the rate of one hundred thousand a year, the cropping income per holding in 1972/73 can be computed similarly to be Rs. 1,640. If we include the incomes from the allied activities, the amount of this broader concept of agricultural income reaches Rs. 1,990 per holding and Rs. 360 in per capita terms. It is to be noted that this figure is some 10 percent lower than the rural poverty-line income of Rs. 405 for All-India (in 1972/73 prices) discussed earlier. ~. ........ -8- 4.1 Estimation Procedures The poverty of rural Bihar is accentuated by the skewed distri- bution of land and, therefore, incomes from land in the state. Our first order of business is to attempt to approximate the current relationship between the distribution of holdings and that of the incomes from cropping activities among various size farms in Bihar. For estimating cropping incomes per holding two different sets of assumptions were adopted. The first set calls for the following steps and assumptions: a) On the basis of a table that gives area of principal crops by size distribution of holdings for the State of Bihar, to be found in the 1970/71 Agricultural Census, we determine the cropped areas for seven most important crops in the state (namely, paddy, maize, wheat, ragi, gram, arhar, and sugarcane). It is to be remembered that the computations are thus based on the cropping patterns as they were reported in the census. Because of the nature of data on which the census figures were compiled, there is a downward bias in the cropping acreage data of unknown magnitude. (It is known that the cropping intensities reported in the agricultural census are lower than those reported by the Department of Agriculture.) b) For the first three foodgrains, paddy, maize, and wheat, we distinguish between irrigated acreage and unirrigated acreage (both of which are reported in the table mentioned above). The output performance (and therefore the "net income", as explained below) of the HYV's was assumed for the irrigated areas and those of the local, improved varieties for the unirrigated cropped areas (with appropriate adjustments for differences in input patterns). For the rest of the selected crops (which are mostly unirrigated, except for sugarcane in some districts) no distinction was made between irrigated and unirrigated acreage. c) It was assumed that there are no economies of scale (nor dis- economies) in production across holding size that affect yield and input patterns per hectare. This assumption means that for any given crop the per acre yield and input requirements are the same regardless of the size of holdings and that only the cropping patterns would differ from one holding size to another. The data with respect to yield and input requirements for the designated crops were taken from a background paper prepared by GOB for the mission (Farm Budget Plans with Profitability, Employment Potential and Income Distribution Possibility). These figures tend to represent situations much more favorable than those prevailing in Bihar as a whole. d) The "net income" is defined as the difference between gross income per hectare (including by-products) and the per-hectare costs which include material inputs (seeds, organic and inorganic fertilizers, irrigation charges, and plant protection charges), labor inputs (hired labor only, except that upto and including the size group with holdings between one and two hec- tares--averaging 1.4 hectare per holding for the size group--there were no provisions made for hired labor), bullock power inputs and interest charges on credits where applicable, The prices used are those of 1974/75. -9- e) The "net income" from each of'the seven crops was summed for each size class. On a further assumption that the remaining cropped acreage (about 22 percent of the total cropped acreage) was as "profitably" used as those included, values were imputed to crops other than the seven explicitly dealt with. The total monetary value of crops income per holding is thus the sum of the incomes from the seven major crops and thoseimputed to the remainder. The computation of cropping incomes based on a second set of assumptions follows the same steps as the first one except for two important exceptions. Instead of assumption (b) above, no distinction was made between irrigated and unirrigated acreages for the three principal crops of paddy, maize, and wheat. In contrast to assumption (c) above, the second set relied on the crop budget data prepared by the mission. Moreover, this set of assumptions include only five crops. In addition to the three principal crops only gram and sugarcane are explicitly dealt with. The combination of the assumptions in this set is perhaps biased towards the lower end of the actual situation in Bihar, although by no means we can be sure of the magni- tude of such a bias. The following tabulation contrasts the estimated net incomes from cropping activities, based on either set of assumptions, for the smallest size groups holding less than 2 hectares: INCOMES FROM CROPPING ACTIVITIES (Rs./Holding) Size Groups (Ha.) Assumption Set I Assumption Set II Less than 0.5 379 194 0.5 - 1.0 1,163 591 1.0 - 2.0 2,082 1,122 4.2 The Incidence of Rural Poverty GOB estimates that some 75 percent of the population of Bihar are below the official poverty line. Table 3, which presents cropping incomes for operating farm households by size groups, tends to confirm this estimate. Even though incomes from allied activities (animal husbandry, forestry, and fishing) would boost the total per holding, say, by 20 percent or so, if we would take account of agricultural laborers, the GOB estimate may very well be an underestimate of the incidence of poverty, Secondly, Table 3 starkly reveals that most of the rural poor are literally on the margin of subsistence with little capacity to save (or to take risks). Severe flood or drought can easily wipe out what little inventories they accumulate over time and push many of them below the margin, often into deeper indebteeriess. In fact, operating farm households in the smallest size groun, operating below half a hectare (and on average cropping only .26 hectare, on net cultivated area of .21 hectare, or a little above half an acre), would be making from crops less than the official poverty-line income per capita. Indeed, they cannot be much better off than landless agricultural laborers and their families. -10- Table 3: ESTIMATED NET INCOME FROM CROPPING ACTIVITIES BY SIZE OF HOLDINGS AVERAGE FOR BIHAR STATE Operational -ropped Area Net Income Size group Percht of Holding Size per Holding from crops total as /1 total number (hectares) (hectare) per holding (Rs) of Statetae of holdings Less than 0.5 .26 379 11.9 46.55 0.5 - 1.0 .80 1,163 1.0 17.79 1.0 - 2.0 1.51 2,082 20.6 14.64 2.0 - 3.0 2.49 2,3382 12.0- 7.59 3.0 - 4.0 3.47 2,679 8.1 4.49 4.0 - 5.0 4.48 3,426 6.8 2.95 5.0 - 10.0 6.28 4,653 13.3 4.22 10.0 - 20.0 11.59 8,397 7.9 1.38 20.0 - 30.0 19.52 14,231 2.L .25 30.0 - 40.0 27.52 19,751 1.0 .08 4o.0 - 50.0 37.15 28,345 .6 .03 50.0 and above 61.54 53,175 1.2 .03 All Sizes- 1.49 1,479 100.0 100.00 The total for the size group is a product of "net income from crops per holding" and the number of holdings in that size group. 2 Note that the underlying assumption with regards to treatment of labor inputs changes between size group, 1.0 - 2.0, and size group, 2.0 - 3.0. The former group's income is computed on the assumption that all labor input is family labor (not costed), while the latter assumes that some input of labor is hired and therefore is to be couted. 3/ The average for the state weighted by the number of holdings in each size group. S- 11 - The agro-climatic and geological features of Bihar, together with the socio-economic characteristics of its agrarian structure *(distribution of holdings and tenurial relations as well as limited resources available) :nd the needs of local subsistence, determine the cropping patterns as observed in the state. The poor cropping systems resulting from these factors have limited the productivity of land per acre per year. Together with the unfavorable land/man ratio, the low land productivity has meant that the productivity of labor has remained low. Given that in the decades to come the state's land/man ratio is to deteriorate further, the only way open for Bihar's agriculture to increase the productivity of labor (and farmers' income) is to increase the productivity of land per acre per year, that is to say, to intensify the use of land (multiple cropping as well as higher yields per cropped area). The process of intensification involves, inter alia, (i) eliminating fallowing, (ii) growing more valuable crops, and (iii) evolving the farm systems that increase the efficiency of farm resources as well as (iv) increasing yields per crop (including post-harvest handling of crop). The scope for improvement is great. As with other states on the eastern Gangetic Plain, Bihar has one of the greatest potentials for growth in its still largely untapped groundwater. Control of water, in the sense of irrigation and drainage as well as flood control, is perhaps the most important single enabling factor in this process of intensification of land use. AccordIng to the Agricultural Census of 1970/71, the cropping intensity in the state averages only 112 percent, when the concept is defined in terms of total cropped area over net cultivated area. Of course, this simple indicator of the land use in Bihar differs from one area to another and according to the size of operational holdings. The cropping intensity is higher in such districts as (old) Patna and Champaran where the irrigation facilities are better developed than in Dharbanga and Purnea, to say nothing of the plateau area typified by Ranchi. Again, the cropping intensity is higher for the smaller holdings than the large. 5. Economic Implications Given the reality of Bihar's agriculture, as outlined above, there are some inter-related issues of particular significance that need general attention. First, it is clear that food production in Bihar must be quickly and substantially increased if her rising population is to be fed in the years to come even at present inadequate levels of nutrition. Secondly, where non-agricultural sectors are so small relative to agriculture in terms of employment, as in Bihar, generating jobs outside agriculture at even twice or three times the rate of labor force growth will not absolve the farming sector from having to absorb the "residual" increase in the labor force. Thirdly, in order to provide jobs and income to an increasing number, it is clear that farming must be intensified across all farm sizes by removing the impediments to the adoption and spread of improved technology -12- and farming systems. Fourth, the twin goals of output growth and employ- ment generation in agriculture must be achieved because the broadly-based agricultural growth is essential for economic reasons. It is to be remembered that agriculture is the largest sector and that the incomes generated and broadly-shared in that sector will heavily influence the development of other sectors by way of the effective (domestic) demand for their products. It is to be remembered also that increasing sub- stantially the levels of living of the rural mass (by way of increase in the productivity of labor in rural areas) is perhaps the only way by which the rapid rise in population can be effectively checked. Nonetheless, in the fifth place, the fact remains that these goals must be pursued in one of the most difficult socio-economic circum- stances characterizing agriculture in Bihar. The environment of agriculture at issue here is the grossly unequal distribution of land holdings and the consequent unequal access to economic and public service inputs (e.g., water, power, extension and credit services, etc.)among different groups of farmers (landlords, cultivators of various sizes, tenants, and landless workers). Although many policy measures and projects implemented in Bihar theoretically touched small/marginal farmers as well as larger farmers, and some of the benefits accrued to the landless, the records in Bihar and elsewhere indicate strongly that the theoretically correct assumption of scale-neutrality of most of these measures broke down in practice when faced with the reality of the agrarian structure. If the apparently widening income gap between the large and small/marginal farmers is to be reduced and the scope of participation in growth enlarged in Bihar's agriculture, the less privileged must be assisted in overcoming their particular, structurally-imposed, impediments and in securing access to economic and public service inputs. If institutional innovations could be devised that encourage the introduction and propagation of high-yielding practices successfully in Bihar's agrarian structure, income from, and employment in, agriculture as well as its output would increase. Furthermore, additional income and investment in agriculture, which involves 80 percent of the Biharis, would have to result in larger incomes and employment not only in ancillary activities in agriculture but also in other sectors of the economy.