THE  WORLD  BANK  ECONOMIC  REVIEW, VOL.  1, NO. 1:  103-148
SnipL. Iqg,
Modeling the Impact of Agricultural Growth and
Government Policy on Income Distribution in India
Jaime Quiz6n and Hans Binswanger
This article uses a limited general equilibrium model to investigate the growth and
equity effects of a variety of economic and technical changes and selected agricultural
policies in India. It explores how changes in food prices, rural wages, and farm profits
associated with the Green Revolution period affected income distribution between net
buyers and sellers of food. The model shows that income gains from the Green Revolu-
tion initially accrued to the wealthier rural groups but that after 1972-73 they were
transferred to urban consumers and that by 1980-81 the per capita incomes of poor
and wealthier rural groups alike were barely above their respective 1960-61 levels. The
model is also used in counterfactual analysis of the impact of changes in technological,
demographic, investment, taxation, and income redistribution variables. Its findings
indicate the importance of trade policies for the nature of the equity outcomes from
agricultural growth and suggest that a reduction in population growth and an increase
in nonagricultural employment and income are required to convert agricultural growth
into reduced rural poverty.
As a result of the Green Revolution, agricultural productivity in India has risen
sharply over the last two decades and India has become a self-sufficient producer
of basic food grains. While there is no dispute about the rapid increase in
production, economists have not had available a similarly compelling analysis of
who has benefited from this growth.
Debates about the effects of the Green Revolution and Indian agricultural
policies on the distribution of income have, almost without exception, been
limited to the question of how income is distributed across small and large farms
and between landowners and workers, rather than between producers and con-
sumers of food. Typical subjects of study have been the differences in adoption
behavior of small and large farms, the distributional impact of their differential
access to credit, and the direct labor-use effects of high-yielding varieties or of
irrigation.
The determination of these direct first effects of changes in agricultural tech-
Jaime Quiz6n is at Chase Econometrics, Philadelphia. Hans Binswanger is at the World Bank.
Copyright © 1986 by the International Bank for Reconstruction and Development / THE WORLD BANK.
103               . -     .



104 THEWORLDBANKECONOMICREVIEW,VOL. 1,NO. 1
niqces and policies is, of course, necessary and important. Meanwhile, however,
the longer-term macroeconomic effects of changes in agricultural technology and
agriculture-related policies have not received sufficient attention. This paper
presents a limited general equilibrium model which incorporates most of the
relevant macroeconomic factors needed to determine t!he distributional impact
of the Green Revolution. The model also allows assessments of other trends and
policies that may be determinants of income distribution. This analysis is di-
rected to the following objectives:
* To trace changes in income distribution between rural and urban groups
and between different income groups
* To determine the equity effects of the Green Revolution
* To suggest how changes in economic, demographic, and technical trends
would be likely to influence income distribution
* To indicate the effects of alternative government policies on equity and
poverty.
In order to address these objectives, we developed a limited general equilib-
rium model that is capable of accounting for changes in rural and urban income
induced by changes in agricultural commodity supply and demand. This model
is described in section I of this paper and is presented in mathematical terms in
appendix A. The major elements of the model are:
* The demand and supply of four agricultural outputs
* The demand and supply of three agricultural inputs
• Real incomes of rural and urban inhabitants at different income levels.
The key feature of the model is that prices and quantities of agricultural
output and variable inputs are endogenous. The model cliffers from an economy-
wide model, however, in that nonagricultural income and production are treated
as exogenous.
Succeeding sections of the paper discuss several of the applications of the
model.
Section II briefly describes a standard exercise which we carried out to com-
pare the model's endogenous ex post predictions of the quantities and prices of
agricultural inputs and outputs with the actual paths of such quantities and
prices as shown by macroeconomic data. Section III disczusses the use of selected
equations from the model to account for changes in rural and urban income in
India during the period from 1961 to 1981. In section IV, the model is used in
counterfactual analysis to determine how income and agricultural variables
would have changed under various hypothetical scenarios.
The model's findings are summarized in section V. Among other things, our
investigations show that India's progress in agriculture during the twenty-year
period in question apparently had little net positive effect on the incomes of



Quiz6n and Binswanger 105
either the rural well-to-do (the landowners) or the rural poor. The chief benefi-
ciaries of increased agricultural output (which was accompanied by government
policies that caused a relative decline in food prices as compared with manufac-
tured goods prices) were urban residents. Our findings suggest that the incomes
of the rural poor in India would be more likely to improve as a result of demo-
graphic changes and increases in nonagricultural employment than as a result of
technological improvements in agriculture.
These conclusions, it should be understood, were arrived at through an ambi-
tious attempt to try to understand an exceedingly complex reality. Our efforts to
do so are subject to various limitations, many of which stem from a lack of
complete data. In order to construct and utilize our model, many assumptions
had to be made, and readers will find many caveats scattered throughout this
article. The strength of the model, however, arises from our econometric estima-
tion of parameter values which are based on the very large amounts of data
compiled and incorporated into it. Despite its limitations, we hope that this
paper can further the evolution of analysis of important issues in economic
development.
I. A SUMMARY OF THE MODEL
The limited general equilibrium model for our investigation determines quan-
tities and prices in seven markets: three input markets, labor, draft power, and
fertilizers; and four agricultural output markets, rice, wheat, coarse cereals, and
other crops. It also determines residual farm profits. Given these prices and
quantities, it then determines the real incomes of four rural and four urban
income quartiles (Rl, R2, R3, and R4 and Ul, U2, U3, and U4, respectively, in
the appendixes).
The supply of the four agricultural commodities and the demand for the three
factors of production are modeled as a jointly estimated system of output supply
and factor demand equations.' Output supply and factor demand shift in re-
sponse to changes in exogenous endowment and technology variables: land
(cultivable area), annual rainfall, irrigation, high-yielding varieties, roads, farm
capital (animals and implements), regulated markets, and technological change.
The supply of labor is responsive to the real rural wage. Agricultural labor is
supplied by rural groups and also by some urban emigration, which is responsive
to the rural wage.
The supply of draft power is responsive to the real rental rate for draft animals
and is supplied by each of the rural groups.
The fertilizer supply is treated as an aggregate of nutrient tons, which is
responsive to the price of fertilizer relative to nonagricultural goods prices.
1. Separate systems were estimated for each of four agroclimate zones. These systems were then
aggregated to the national level. A flexible functional form was used to allow for cross-price effects
among all seven outputs and factors.



106 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. 1
The supply of land is exogenously given as the cultivated area. This is appro-
priate, because area expansion in Indian agriculture has virtually stopped since
the mid-1960s. However, this treatment still allows cropped area to vary endog-
enously via changes in the extent of double and triple cropping. And, of course,
the area allocated to different crops can vary.2 While the supply of land is
exogenous, net returns to land (the residual farm profits after variable factors
have been paid) are determined endogenously.
Consumer demand is responsive to the prices of commodities and the real
income of each of the eight income groups. Poorer groups have higher income
elasticities than richer groups. Each income group's demand must therefore be
modeled separately. Demand was estimated econometrically; a flexible func-
tional form was used, so that all (compensated) cross-price elasticities were
directly estimated. Aggregate demand is the sum of the demands of all the
income groups.
Nominal income is computed as each group's supply of agricultural produc-
tion factors multiplied by the factor prices, plus an exogenously given compo-
nent for nonagricultural income. Real income is calculated for each of the eight
groups as their nominal income deflated by an endogenous consumer price index
that is specific to that group's consumption patterns and reflects all endogenous
changes in food prices.
Prices and quantities of commodities and factors of production are determined
as those which equate aggregate supply and demand in each of the seven mar-
kets. The government can influence agricultural prices through the use of tariffs,
food imports and exports, food grain storage, forced procurement at fixed
prices, and consumer ration shop sales at nonequilibrium  prices.3 The model
solves simultaneously for changes in endogenous prices and quantities and thus
determines for each income group the change in its nominal income, price defla-
tor, real income, labor supply, draft power supply, and level of consumption.
Nonagricultural prices are given exogenously and are used as the numeraire of
the model. Because nonagricultural income is also given, nonagricultural pro-
duction is exogenous and consumption of this output must adjust via trade.
The base year used in constructing the model is 1973-74. Initial values are
computed largely from an extensive rural household survey by the National
Council for Applied Economic Research.4 The entire model is written in loga-
rithmically linear equation form.
There are several important characteristics of the model which must be kept in
mind while interpreting our findings.
First, it is well known that the distributional outcomes from general equilib-
2. Neither the total cropped area nor the area under different crops is explicitly traced in the model
because the supply equations do not distinguish between area and yield supply.
3. Although we deal mainly with food trade in this paper, forced procurement and food subsidies are
discussed in Binswanger and Quiz6n (1986).
4. For a fuller discussion of data sources and estimation of parameter values, see appendix B and Pal
and Quiz6n (1983).



Quiz6n and Binswanger 107
rium models depend crucially on labor market assumptions (Taylor 1979). We
model the real rural wage by equating supply and demand for labor; that is, it is
a full employment model. This treatment is consistent with the empirical evi-
dence that there is little year-round unemployment in rural areas and that most
unemployment is seasonal (Krishna 1976). Moreover, real wages are variable
both within and across years; that is, no model of constant nominal or real
wages is consistent with the data. Econometric studies of labor demand (Even-
son and Binswanger 1984) and supply (Bardhan 1984; Rosenzweig 1984) are
also consistent with our neoclassical treatment of the rural labor market.
In spite of this evidence in favor of a neoclassical approach, we are keenly
aware that there is considerable friction in rural labor markets. For example,
there are substantial and persistent interregional wage differentials, and seasonal
unemployment is clearly present. But our model is not regional and does not deal
with intrayear wage determination.
Similarly, because the model aggregates across different regions, it is not able
to account for regional concentration of the Green Revolution. Because, in the
longer term, increased production led to a decline in agricultural prices, farmers
who had not adopted the Green Revolution technology-and whose yields had
not increased-were harmed. Thus our simulation obscures both the more radi-
cal income gains in beneficiary areas and the declines in the nonadopting re-
gions.
The model treats nonagricultural incomes (and implicitly urban wages and
nonagricultural output) as exogenously determined. The purchasing power of
the nonagricultural incomes, however, depends on agricultural prices. When
these prices rise, urban agricultural demand will fall because of both price and
income effects. But other feedbacks from agricultural activity to the nonagricul-
tural sector are not allowed for in the model. One consequence of our treatment
of the nonagricultural sector is that changes in food prices have no effect on the
nominal urban wage; that is, reductions in food prices benefit urban wage
earners and are not passed along to employers in the form of lower wages.
Although the model determines what happens to real farm profits and the
incomes of the rural income groups, it does not treat endogenously what subse-
quently happens to private savings and private agricultural investments brought
about by the changing fortunes of farmers. Thus our model is not a very long-
run model. The reason for this treatment is that no econometric studies exist
which quantify the link between farm profits and farm investment.
Because there is no adequate empirical evidence for the actual changes in
factor or asset endowments, we have not attempted to track these changes in our
analysis of income distribution trends and we do not have endogenous endow-
ment changes in our simulations. For such an analysis, one would need either to
get comprehensive and accurate data or to be able to model investment processes
in land and other factors of production for each of the four rural income groups.
At the present time, the absence of such empirical knowledge makes the model-
ing of endowment changes a distant goal.



108 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. 1
Finally, the model leaves out the effects of the market f'or foreign exchange on
agricultural performance, and vice versa. India is modeled as a state-trading
economy in which decisions to export or to import agricultural commodities rest
solely with the government. These decisions are exogenous to the model.
II. COMPARING MODEL PREDICTIONS WITH ACTUAL CHANGES
A set of experiments was performed to compare the model's predictions of
agricultural prices and quantities with the actual prices and quantities reported.
Ideally, one would want to compare the model's predictions of income distribu-
tion with actual patterns. Unfortunately, the data needed for such a comparison
do not exist. Changes in exogenous variables (such as population, agricultural
technology, capital and inputs, and nonagricultural prices and income) were
introduced into the model for the five-year periods between 1960-61 and 1980-
81, and the model's calculated production and prices were compared with the
actual quantity and price data reported for those periods (see part B of appendix
table 11). Difficulties encountered in compiling actual data for the comparison
are discussed in appendix B.
In table 1, we compare indexes of actual and predicted values for six years and
give the ratios of predicted to actual levels for each variable (with 1973 as the
base). As can be seen, the fit between predicted and actual values is generally
close despite the substantial changes that occurred in many actual values during
the period. Of 65 predictions, 28 differ from the actual figure by 10 percent or
more and only 10 by 20 percent or more. The poorest predictions are for the
extreme years 1960-61 and 1980-81.
Although during the period as a whole we overpredicted the growth rate in
agricultural output by only about 0.5 percent per year, our quantity predictions
are better than our price predictions. On the price side, the most serious problem
is the overprediction of the rate of growth in agricultural prices from 1975-76 to
1980-81. Figure 1 shows that actual terms of trade imoved rapidly against
agriculture during that period, but our model does not filly capture this down-
ward trend, apparently because our model exaggerates the growth of demand.
Notwithstanding these difficulties, the results show that our model is able to
replicate reasonably actual agricultural conditions for the period.
Among the individual variables, fertilizer consumption in the pre-Green Rev-
olution period is the one tracked least accurately. We overpredict fertilizer con-
sumption in those early years by a factor of 200 percent. This error is partly due
to an extremely low base-year value. We also underestimate the rapid growth in
fertilizer demand in the 1975-76 to 1980-81 period. This may be partly because
we are not able to account for the rapid growth in the fertilizer subsidy in our
simulations.



Table 1. Comparative Indexes of Production, Employment, Wages, and Prices
Agricultural year
Variable          1960-61  1965-66  1970-71  1975-76  1980-81
All crop production
Actual value                 78.46    79.95   101.02   108.23   122.16
Predicted value              74.83    78.28    98.62   108.49   130.28
Ratio of predicted to actual value    0.95  0.98  0.98   1.00     1.07
Rice production
Actual value                 82.82    81.49   101.91   106.01   121.37
Predicted value              82.65    83.81   100.05   107.68   125.95
Ratio of predicted to actual value    1.00  1.03  0.98   1.02     1.04
Wheat production
Actual value                 47.32    48.24    99.69   116.10   149.54
Predicted value              41.39    52.83    95.59   117.51   162.49
Ratio of predicted to actual value    0.87  1.10  0.96   1.01     1.09
Coarse cereal production
Actual value                 89.19    90.96   106.45   108.52   110.81
Predicted value              82.12    80.96    99.35   108.52   119.59
Ratio of predicted to actual value    0.92  0.89  0.93   1.00     1.08
Other crop production
Actual value                 83.05    86.86    99.07   107.05   116.10
Predicted value              75.57    80.79    97.62   105.26   127.27
Ratio of predicted to actual value    0.91  0.93  0.99   0.98     1.10
Fertilizer consumption
Actual value                 11.45    32.50    84.30   108.53   205.85
Predicted value              35.21    58.75    74.44   114.46   182.02
Ratio of predicted to actual value    3.08  1.81  0.88   1.05     0.88
Employment
Actual value                 85.17    90.62    96.07   102.62   109.17
Predicted value              81.54    86.60    95.49   103.36   111.74
Ratioofpredictedto actual value    0.96  0.96   0.99     1.01     1.02
Rice prices
Actual value                 92.89    93.68    97.15   101.78    89.97
Predicted value              92.98   116.50    81.19   107.88   120.87
Ratio of predicted to actual value    1.00  1.24  0.84   1.06     1.34
Wheat prices
Actual value                100.57   109.06   108.30   106.36    85.35
Predicted value             120.20   134.52    86.33   108.54   103.66
Ratio of predicted to actual value    1.20  1.23  0.80   1.02     1.21
Coarse cereal prices
Actual value                 93.13   106.49    86.09    90.70    74.70
Predicted value             102.38   116.84    85.77    95.20   101.38
Ratio of predicted to actual value    1.10  1.10  1.00   1.05     1.36
Other crop prices
Actual value                100.74    99.20   103.56    95.59   101.66
Predicted value              93.22   114.29    87.44   105.41   126.31
Ratio ofpredictedto actual value    0.93  1.15  0.84     1.10     1.24
Labor wages
Actual value                102.57   104.85   109.57    97.69    98.40
Predicted value             116.22   121.74    93.05   102.00   105.57
Ratioofpredictedto actualvalue    1.13  1.16    0.85     1.04     1.07
Prices of all commodities
Actual value                100.00   100.00   100.00   100.00   100.00
Predicted value             100.11   113.49    88.80   105.30   119.08
Ratio of predicted to actual value    1.00  1.13  0.89   1.05     1.19
Source: World Bank data; see appendix table 11.
109



110 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. 1
Figure 1. Agricultural/Nonagricultural Terms of Trade for India,
1960-61 to 1980-81 (Actual Data; 1973-74 = 100)
X,  110
C
C
0
-    100
u
E @  90
0 \:
o *
E
80
<Iu
1960-61 1965-66  1970-71 1973-74 1975-76  1980-81    Year
Source: Appendix table 11, part A.
III. ACCOUNTING FOR CHANGES IN INCOME DISTRIBUTION
In this section we compute a reference path of the real incomes of each of the
rural and urban income groups during the period from 1960-61 to 1980-81.
We generate the implied distribution of income among the eight groups by using
actual estimates of agricultural output, agricultural prices, wages, and fertilizer
consumption, as well as the exogenous variables that affect the income and
factor market equations in the model. The numbers in table 2 are indexes of the
predicted levels and are calibrated so that the predicted level of each variable is
equal to 100 for 1970-71, the end of the first phase of the Green Revolution.
We assumed that during the twenty-year period, the across-quartile shares in
ownership of factor inputs and within-quartile shares of nonagricultural and
factor incomes in total income remained equal to their respective base-year
(1973-74) values (see discussion in section I). We also aissumed that the rates of
growth in the population, in the agricultural capital stock, and in the nonagri-
cultural income of each quartile were the same across the groups. But there may
have been other causes of change in actual incomes that we were unable to
account for, such as changes in taxation, in investment behavior, in people's
occupations, and in food subsidies. Table 2 shows what would have happened to
real income as a result of changes in agricultural production and technology,
agricultural output and input prices, nonagricultural incomes and prices, and
population. Although the total endowments of the various groups change over



Quiz6n and Binswanger III
Table 2. Simulated Indexes of Income Distribution and Income Sources, India,
1960-61 to 1980-81
(1970-71 = 100)
Agricultural year
Endogenous variables    1960-61  1965--66  1970-71  1973-74  1975-76  1980-81
Real per capita income
(actual)
National                    92.0     95.0      100      95.1      95.4    105.9
Rural, by quartile
First (poorest)          101.0     99.0      100      95.9      97.4    107.0
Second                    96.9     95.8      100      94.6      94.8      99.9
Third                     93.8     93.5      100      93.8      93.3      96.3
Fourth (richest)          88.5     88.6      100      92.4      90.7      88.8
Aggregate,                92.9     92.4      100      93.6      92.9      94.9
Urban, by quartile
First (poorest)           91.9    100.4      100      98.1    100.7    136.0
Second                    90.9    102.8      100      99.3    102.6    141.9
Third                     90.2    102.7      100      99.7    102.5    139.3
Fourth (richest)          87.6    102.3      100      99.8    102.2    133.5
Aggregatea                89.4    102.3      100      99.4    102.2    136.7
Agricultural employment      98.2    100.1       100     112.3    118.8    118.5
Real agricultural wage bill  91.2      95.3      100     101.4    104.9    105.4
Realresidualfarmprofits      64.2      67.9      100      86.0      85.1      76.4
Nonagricultural income       71.9      93.6      100     111.3    121.8    182.7
Real per capita disposable
income                    92.4     94.5      100      96.7      97.8    113.6
Total actual agricultural
output                    79.3     81.2      100      99.4    107.1    119.6
Actual prices = agricultural/
nonagricultural goods     89.8     97.2      100      97.7      91.6      76.3
a. These estimates of per capita income are computed as in equation 16 of appendix A, in which the
subscript k now refers to either the rural quartiles (Rl to R4) or the urban quartiles (Ut to U4) only.
time, the relative endowment position of each group was assumed to remain the
same.
The last two rows of table 2 show the actual growth of total agricultural
output and the change in agricultural terms of trade. Agricultural production
grew rapidly during the early Green Revolution period (1965-66 to 1970-71)
and again from 1973-74 onward, while agricultural terms of trade rose prior to
the Green Revolution, stayed fairly constant until 1973-74, and then dropped
substantially by 1980-81.
These changes in quantity and price explain the changes in farm profits. Farm
profits were seriously depressed in 1960-61 and in 1965-66 but then moved
dramatically upward by 1970-71. By 1973-74 they had declined to 85 percent
of their 1970-71 level, and by 1980-81 to 76 percent of the 1970-71 level. In
these years, declines in output prices outweighed rapid growth in agricultural
output.
Employment in agriculture (estimated in our model) grew by about 20 percent



112 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. I
during the twenty-year period. Because real wages declined by about 5 percent,
the total real wage bill for the period rose by about 15 percent.
Nonagricultural real income more than doubled during the period, with the
most rapid increases occurring just prior to the Green Revolution and between
1975-76 and 1980-81. This latter gain occurred partly because the numeraire
by which nonagricultural income is deflated gives a large weight to agricultural
commodities, the prices of which had declined.
The trends in output and factor prices, and in agricultural and nonagricultural
income, suggest that real aggregate per capita income among rural people grew
by only about 8 percent during the early Green Revolution, after which it
declined and stagnated. Despite a drastic shift in the disitribution of rural income
from wages to profits in the early period, rural income distribution was remark-
ably stable for the period as a whole. The effect of adverse wage trends on the
rural poor was partially alleviated because agriculturail employment increased
somewhat and because the poor participated to a small extent in the growth of
farm profits. About 11 percent of their income was derived from such profits.
They also had substantial gains in nonagricultural incomes,5 and as consumers
they benefited from the decline in agricultural prices during the last five years of
the twenty-year period.
The first period of the Green Revolution was one of substantial gains in farm
profits. But the rapid gains in production during the late 1970s did not translate
into further advances in income because the prices of agricultural products fell.
The production gains from the early Green Revolution period were associated
with rising prices because the government used the gains largely to replace
imports. But once self-sufficiency in food grain production was more or less
assured, the surplus grain production had to be absorbed domestically. This was
a classic example of the process in which productivity gains in agriculture were
transmitted to consumers (both rural and urban) by way of declining prices.
It is therefore not surprising to find that urban groups showed the largest gains
in income, although their gains were largely a phenomenon of the last five years
of the twenty-year period. They appear to have gained during the first quinquen-
nium as well, but that gain was made despite rises in agricultural prices. In the
late 1970s the combination of rapid nonagricultural growth and declining agri-
cultural terms of trade greatly benefited the urban groups, with the biggest
beneficiaries being the urban poor, since they spend a larger share of their
incomes on food.
IV. SIMULATIONS OF ALTERNATIVE POLICIES AND TRENDS
The previous section offers only partial explanations of why wages, farm
profits, and income distribution evolved the way they did. An assessment of how
5. Nonagricultural sources provided 21 to 26 percent of the nominal per capita incomes of rural
groups.



Quiz6n and Binswanger 113
each individual change in a policy or a trend affects the model's outcomes is
required to separate out the influence of different factors.
We do that in this section by comparing the results of a simulated change in
selected trends or policies with the "base case,' which reflects what actually
occurred in India during this period. Thus we simulate a change in a specific
exogenous demographic variable, for instance, and trace the effect of the change
on production, prices, employment, and farm profits and through them see how
income distribution would have been affected. Figure 2 illustrates this process.
UoU1 is the path of the specific endogenous variable given actual policy trends
and events, and a is the value of U at Ul. The variable U could be any of the
twenty-eight shown in the left-hand column of tables 3 through 6. UOU2 is the
simulation path of U if an exogenous change or intervention occurred, such as
any of those shown in the column headings of tables 3 through 6. The value b is
the difference between U2 and Ul, the induced change in U at T1; and C is the
percentage change in U, or bla. These percentage changes are the values re-
ported in tables 3 through 6. In table 3, To to T1 is a ten-year period, while in
tables 4, 5, and 6, T1 is perhaps three to five years, sufficient time for farmers to
respond to changes in technology, policies, and prices by adjusting their produc-
tion patterns.
In order to explicitly track the changes in the terms of trade between agricul-
Figure 2. Simulated Changes in Trends and Policies: Derivation of Values,
Tables 3 through 6
L    I                                                 U2
'5~~~~~~~~~~~~~~~~~~~~~~U
5 U0~~~~~~~~~~~~~~~~U
U~
-r   II
TO                                 T,        ~~~~~~~~~~~~~~Time
Note: C = percentage change in simulated value of U from its "base" trend value, as the result of a
simulated change in one or more exogenous variables, C values are those shown in tables 3 through 6.



114 THEWORLDBANKECONOMICREVIEW,VOL. 1, NO.1
ture and nonagriculture, we use nonagriculture commodities as the numeraire.
The change in the gross national product (GNP) deflator shown below is there-
fore a direct function of the change in terms of trade, not of inflation.
Demographic and Urban Growth Scenarios
In demographic scenarios 1.1 a and 1.1 b (see table 3), the assumption is made
that population growth in India (both rural and urban) is reduced by 10 percent
during a decade. Total nominal nonagricultural income is reduced by 10 percent
as well and is therefore unaffected on a per capita basis. In scenario 1.1a the
labor force continues to grow at the same rate as before-that is, this scenario is
a stylized representation of a reduction in fertility alone, which would not affect
the size of the labor force. In scenario 1.1b we assume that the reduction in
fertility has caused a (long-run) decline in labor force growth and shows, in a
stylized way, the effects of these reductions in fertility and labor force growth
during a ten-year period.
In the first row of table 3 we see that the postulated fertility decline leads to a
substantial gain in national income of about 5.6 percent and 5.2 percent in the
two cases. In the second row we see that output declines somewhat more sharply
(-1.2 percent) when labor force growth is also reduced than it does initially
when the labor force still continues to grow (-0.6 percent). Aggregate prices,
reflected in the GNP deflator, decline sharply (-19.4 percent and -18.1 percent
respectively). The main difference between the two scenarios is in wages. Real
wages decline by about 3 percent during the first decade because of reduced
demand for agricultural output, whereas they increase by about 10 percent in
our (stylized) long term. This is because long-run agricultural labor employment
declines by about 5.5 percent in the latter case, so wages must rise.
In scenarios 1.1a and 1.1b there is a progressive irnpact on rural and urban
income. The rich rural group (landowners) loses only a little in real income in
the first decade (-3 percent), but in the long run this group loses more (-7.4
percent) as its members must begin to pay higher real wages. The increase in
wages, coupled with declining demand, leads to a sharp reduction of 45 percent
in residual farm profits. Under both scenarios, the poor in both rural and urban
areas gain from the substantial decline in food prices. They gain a little more so
in the long run because they also benefit from increase,d scarcity of labor (+ 15.5
percent and +19.7 percent for the rural and urban poor, respectively). The
urban group gains the most since it benefits not only from lower food prices but
also encounters less erosion of income as rural-to-urban migration is reduced.
Nutrition, measured here as cereal consumption, improved in all groups ex-
cept in the rich urban and rural groups. For the lowest income groups the
improvement is somewhat smaller than the change in real income, whereas for
the richest groups the change in nutrition is much smaller than the change in
income. This reflects the fact that richer groups have lower income elasticities.



Quiz6n and Binswanger 115
Simulated scenario 1.2 in table 3 is one of urbanization. Rural population is
assumed to decline by 10 percent, while urban population increases by 40.2
percent, enough to absorb the rural population. (Nominal urban income is
increased by 40.2 percent in order to hold nominal per capita income constant.)
In the absence of a continued rural-urban income differential, no migration
would occur and the scenario would be unrealistic.
The main feature of scenario 1.2 is the assumed reduction in the number of
agricultural producers while the number of consumers remains constant. There-
fore, agricultural terms of trade rise sharply, which drives the GNP deflator up by
32 percent.
The reduction in agricultural population leads to a real wage increase of 9.4
percent, while the sharp increase in agricultural prices allows residual farm
profits to rise by 58 percent. These effects drive the income distribution effects.
Large farmers gain by nearly 30 percent, while the rural poor benefit both from
increased wage income and increased farm profits, which account for 11.32
percent of their real income (appendix table 8). These gains more than offset
their losses as consumers, and their incomes rise by a modest 3 percent. For the
urban poor, however, the outcome is a fall of 21 percent in income as food prices
rise. The losses of the second urban quartile are somewhat higher than the losses
of the first (poorest) quartile. This occurs because the poorest urban quartile
supplies some labor to the agricultural sector, whereas the second quartile does
not. The urban rich lose less (-15 percent) than other urban groups because
they spend a smaller part of their income on food.
Simulated scenario 1.3 combines scenarios 1.1b and 1.2. Overall population
and labor force growth rates both decline by 10 percent, but the decline is
accompanied by a rural-to-urban migration. The net effects are a decrease of 20
percent in the rural population and an increase of 30.2 percent in the urban
population. The result is a large gain (14.3 percent) in real national per capita
income. Meanwhile, agricultural prices increase, which leads to a modest in-
crease in residual farm profits (13 percent). Therefore, all rural income groups
experience real income gains of about 20 percent. Urban groups, however, lose.
We should note here that this scenario would not last indefinitely. People
would not continue to move to urban areas in the face of a substantial decline in
real wage incomes. Nominal incomes must rise if urbanization is to continue. We
investigate the effect of such a rise alone in scenario 2.1 and then combine it with
the demographic cum urbanization scenarios in scenario 2.2.
Scenario 2.1 lets the exogenous component of urban income increase by about
19 percent. Aggregate agricultural output increases only slightly (0.7 percent) in
response to the increased demand for food because output is quite inelastic.
Instead, the increased demand for food resulting from the rise in urban income
translates into a substantial increase in the aggregate price level of food (16.6
percent). Thus, much of the increase in urban consumption of food must come
from reduced consumption among the poorer two rural groups.



Table 3. Simulated Effects of Demographic Changes
(percentage change)
Scenarios for changes in exogenous variables
Reduced
population growth             Urbanization
Laborforce                Population and
growth rate                laborgrowth              Rise in
Steady       Reduced        Steady       Reduced       urban income       Combined scenarios:
C,\       Endogenous variables         (sl.la)        (sl.lb)       (sl.2)        (sl.3)          (s2.1)        1.1b + 1.2 + 2.1 (s2.2)
National income per capita         5.63           5.20         9.08         14.29           6.49                  20.78
Output
Total                           -0.64         -1.23          0.91        -0.32            0.74                   0.42
Rice                            -0.39         -1.97          2.32          0.35           1.25                   1.60
Wheat                           -4.58         -5.70         10.95          .5.25          5     66              10.92
Coarse cereals                  -2.43         -3.37        -17.12       -20.48          -5.61                 -26.10
Other crops                      0.44           0.56         2.04          2.60           0.84                   3.43
GNP deflator                     -19.44        -18.12         31.76         13.64          16.64                  30.28
Prices
Rice                           -27.40        -25.51         45.02         19.51          23.04                  42.55
Wheat                         -34.83         -32.68         58.14         25.47          30.09                  55.56
Coarse cereals                -32.83         -25.81         28.38          2.57          16.58                  19.14
Other crops                    -21.41        -21.00         40.12         19.12          21.08                  40.20
Wage rate                         -2.88          10.14         9.42         19.56         -0.35                   19.22
Employment                       -0.71          -5.46        -5.92        -11.38          -0.47                 -11.85
Wage Bill                         -3.59           4.68         3.50          8.18         -0.82                    7.37
Profits                         -36.01         -45.37         58.46         13.09          35.14                  48.23



Income per capita (by quartile)
Rural
Poorest                         14.91          15.53          3.13         18.56          -3.89                    14.67
Second                           8.61           7.72         11.62         19.33            1.51                   20.84
Third                            4.64           2.46         17.31         19.77            5.13                   24.90
Richest                        -3.34          -7.44          29.86         22.43           12.59                   35.02
Urban
Poorest                         14.99          19.67       -20.95         -1.28             4.10                    2.82
Second                          16.55          21.02       -26.86         -5.84             2.73                   -3.11
Third                           14.22          19.02       -23.29         -4.27             4.47                    0.20
Richest                          8.44          13.62       -14.67         -1.05             9.12                    8.07
Per capita cereal consumption
Rural
Poorestquartile                 13.78          12.61          1.24         13.85          -2.95                    10.91
Richest quartile               -0.54          -2.24          10.34          8.09            4.60                   12.70
Urban
Poorest quartile                14.64          16.93       -17.04         -0.11             1.96                    1.85
Richest quartile               -0.58            0.55          3.15          3.70            6.54                   10.25
Aggregate per capita cereal
consumption                        8.03           6.80          0.53          7.33            0.95                    8.28
Note: The values shown are the percentage changes in the endogenous variables from their base case level, as the result of the induced change in the exogenous
variables shown in the column headings.
All income, output, price, wage, profit, and consumption variables are real values.
The income groups shown are expenditure quartiles for rural and urban populations separately. The four rural expenditure quartiles together constitute 0.8009 of
the total population, while the four urban quartiles include 0.1991 of the total population.



118 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. I
Since agricultural output rises only slightly, real wages are largely unaffected,
but residual farm profits rise by 35 percent because of higher food prices. The
rural poor lose 3.9 percent of their real income, but large farmers gain by more
than 12 percent. Since urban groups must share their initial income gain of 19
percent with large farmers, the urban gain is reduced to about 9 percent for the
urban rich and only 3 percent to 4 percent for the urban poor.
Scenario 2.2 groups the combined effects of the slowdown in population
growth and labor force growth (1.lb), faster urbanization (1.2), and urban
income growth (2.1). The effects on the endogenous variables are largely addi-
tive. Real urban incomes stay about constant, except fo:r those of the urban rich,
which rise by 8 percent. The incomes of the rural poor rise by about 14.7
percent, while those of the rural rich increase 35 percent. As long as the reduc-
tion in the number of agricultural producers and the rise in nominal urban
income are not accommodated by more imports, the rural groups are the main
beneficiaries. It is important, however, to realize that if increases in food prices
were precluded by additional imports, the distributional outcome would be
more favorable to the urban groups and less favorable to the rural rich.
Technical Change Scenarios
In simulated scenarios 4.1 and 4.4 (see table 4), yields of each individual crop
or crop group are assumed to rise by 20 percent, a change corresponding to a
major varietal shift like the Green Revolution. In scenario 4.5, the yield gain is
smaller (10 percent) but is distributed evenly across all crops.
We present two versions of each of these scenarios. In the first versions (sce-
narios 4.1a, 4.2a, 4.3a, and 4.4a), the economy is considered closed and addi-
tional production is consumed in India. But in the second versions (scenarios
4.lb, 4.2b, 4.3b, and 4.4b), the extra yield is either exiported or used to reduce
imports of the commodity in question. The exported quantities (or the reduction
in imports) were considered to be base-year domestic production multiplied by
20 percent. It is only after this initial increased quantity of output is exported
that our simulation allows for farmer adjustment of crop mix into the more
profitable crops as yields and prices change.6
Note that the b scenarios correspond to an assumption of state trading; it is
not an open economy model with trade in many commodities. The exports (or
reduced imports) of the b scenarios are an extreme assumption which the gov-
ernment is unlikely to carry out. It would probably alter exports (or imports) by
a magnitude ranging from full domestic absorption of the surplus, as in the a
scenarios, to full export of the increase in the b scenarios. Any desired intermedi-
ate point can be obtained by computing the appropriate linear combination of
the impacts of the a and b scenarios.
When an increase of 20 percent in rice yields has to be absorbed domestically,
6. For a detailed discussion of how technical change is introduced in the model, see appendix III of
Quiz6n and Binswanger (1984).



Quiz6n and Binswanger 119
the result is a sharp decline in the price of rice (-31 percent) and in the price of
its closest substitute, wheat (-15 percent). Rice production increases by about
20 percent, while wheat production declines by about 6 percent. Prices of the
other agricultural commodities also decline by about 6 to 8 percent. The GNP
deflator therefore declines by about 12 percent, while total agricultural output
increases by about 5 percent. The price decline and the increase in agricultural
output imply a real national income gain of about 4 percent.
The increased agricultural output requires only moderately larger labor inputs
(1.1 percent), and the increased demand for labor results in modestly higher (1.5
percent) real agricultural wages.
The declines in agricultural prices, combined with the rise in wages, lead to a
reduction in residual farm profits despite the increase in agricultural productiv-
ity. The price, farm profit, and wage effects largely explain the distributional
outcome. Net buyers of food gain, and the more so the larger is their share of
income spent on food. The urban poor gain the most (12 percent and 13.6
percent). The rural poor also benefit, since they too spend most of their income
on food. Moreover, they benefit from the slight rise in wage levels. Since reduc-
tion in farm profits affects them only slightly, they end up with a net gain in real
income of 7.5 percent. The rural rich, however, derive much income from farm
profits, and their gain as consumers is not sufficient to offset their loss in profits.
Their real income therefore falls by 1.4 percent.
A decision to export all the initial increase in rice production would sharply
alter the distributional outcome. Since national income would rise by 5.7 per-
cent, domestic demand would increase, which would lead to a rise of 10 percent
in the domestic price level. Meanwhile, aggregate agricultural output would also
rise because of the additional incentive to export. Rice production alone in-
creases by 28 percent, which is 8 percent more than the increase caused by the
technical change. Increased profitability, in other words, leads to extra resources
being allocated to rice.
In this scenario, employment and real wages increase modestly. But price
increases, combined with improved efficiency in production, lead to a rise in
residual farm profits of 36 percent. These price, wage, and profit changes com-
bine to produce a regressive distributional impact. All urban consumer groups
lose, with the poor being hardest hit, and cereal consumption declines.7 Mean-
while, the losses of the rural poor on the consumption side reduce their income
gain to a mere 1.3 percent, while the rural rich experience a major gain in
income of 14.8 percent.
The sharp effects of trade on income distribution are also evident in the other
technical change scenarios, although magnitudes and other details differ signifi-
cantly by commodity. Except in the case of coarse cereals, the gains of the urban
groups are larger than those of any rural group when the extra output caused by
7. The losses of the poorest urban group are slightly less than those of the second quartile because of
the greater direct participation of the poorest in the agricultural labor market.



Table 4. Simulated Effects of Technical Change and Increased Exports
(percentage change)
Scenarios for changes in exogenous variables
20 percent increase in yield                       10 percent increase in
Rice              Wheat            Coarse cereal       Other crops        all crop yields
Closed   Exports   Closed   Exports   Closed   Exports   Closed   Exports   Closed   Exports
Endogenous variables    (s4.la)   (s4.1b)   (s4.2a)   (s4.2b)   (s4.3a)   (s4.3b)   (s4.4a)   (s4.4b)   (s4.Sa)   (s4.5b)
National income per capita  4.10     5.65      1.98     0.02       1.01     2.08      7.31     10.58     7.20     10.20
Output
Total                     5.35     6.19      2.11     2.13      2.26      2.48     10.34     12.18    10.03     11.49
Rice                    20.09    27.90    -1.67    -1.35         2.82     0.51    -0.55       0.64    10.35     13.85
Wheat                   -5.94      1.78     18.33    29.27       1.46     1.82      1.14      9.35     7.49     21.11
Coarse cereals            4.86    -5.42      1.49   -3.57      12.75    22.21    -4.55   -14.02        7.28    -0.40
Othercrops               0.16    -1.62       0.81   -0.49      -0.16    -0.69      21.20    24.52     11.00     10.86
GNPdeflator              -11.96    10.13    -6.59       6.14    -4.91       5.59   -12.78     22.24   -18.13      22.05
Prices
Rice                   -30.87      9.10    -8.86      9.60    -1.81    10.85   -10.36        35.29   -25.95     32.42
Wheat                  -21.13    19.62   -29.26       8.04    -4.05    13.55    -8.04        47.72   -31.24     44.46
Coarse cereals          -6.74    13.92    -6.07       7.24   -35.50    -6.24    -17.89       18.84   -33.10     16.88
Othercrops              -8.12    14.45    -3.47       7.79    -3.73       6.89   -23.78     24.05   -19.55      26.59
Wage rate                   1.54     2.39      0.14     0.82    -2.08       2.10      0.21    -0.87    -0.10       2.22
Employment                  1.14     0.81      0.23     0.17     -0.37      0.83      0.18    -0.88      0.59      0.46
Wagebill                    2.68     3.20      0.37     0.99    -2.45       2.93      0.39    -1.75      0.50      2.69
Profits                   -7.54    36.07    -4.11    17.63    -1.08    15.03          4.11     80.75    -4.31     74.74



Income per capita (by quartile)
Rural
Poorest                 7.52     1.33      2.6S    -0.62       3.98     1.48      5.74    -2.05       9.95     0.07
Second                  5.06     6.06      1.89     2.02       1.84     2.69      5.72      9.64      7.26    10.20
Third                   3.55     9.22      1.23     3.71       0.75     3.54      5.82     17.56      5.67    17.02
Richest               -1.37    14.85    -0.04       7.21     -0.12      6.15      5.20     32.86      1.84    30.53
Urban
Poorest                12.02    -6.90      7.47    -4.45       3.28    -3.92     12.79   -18.02      17.78   -16.65
Second                 13.61    -7.69      6.15    -5.43       1.78   -5.22      13.47   -20.66      17.51   -19.50
Third                  11.09    -7.07      5.46    -4.69       1.24   -4.53      12.60   -18.12      15.19   -17.20
Richest                 5.74    -4.77      2.79    -2.91      0.09    -2.63      11.04   -10.69       9.83   -10.50
Per capita cereal consumption
Rural
Poorest quartile       11.21     2.03      2.62    -1.27       7.83     2.06      0.63    -3.88      11.14    -0.53
Richest quartile        4.27     6.46      3.03     2.59    -0.68       1.20    -4.27       8.24      1.18     9.25
Urban
Poorest quartile       13.07    -5.01      9.48    -3.49       5.63   -2.78       5.27   -16.96      16.72   -14.12
Richest quartile        6.75     3.50      0.81     0.20    -1.36    -0.35      -5.43    -0.81        0.39     1.27
Aggregate per capita cereal
consumption              10.14     3.70      3.57       .00     4.16      1.02    -0.93    -0.12        8.47     2.29
Note: The values shown are the percentage changes in the endogenous variables from their base case level, as the result of the induced change in the exogenous
variables shown in the column headings.
All income, output, price, wage, profit, and consumption variables are real values.
The income groups shown are expenditure quartiles for rural and urban populations separately. The four rural expenditure quartiles together constitute 0.8009 of
the total population, while the four urban quartiles include 0.1991 of the total population.



122 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. 1
technical change is absorbed domestically. (In the case of coarse cereals, the
gains of the rural poor exceed those of the urban group because urban con-
sumers buy very little coarse cereals.) It is clear that the export of the initial gain
caused by the technical change always leads to losses for urban consumers and is
associated with a sharply regressive distribution of its benefits in rural areas.
Differences in the magnitude of the effects associated with technical change
are partly a reflection of each commodity classification's share of agricultural
output. Rice and other commodities have the largest shares, 26.7 percent and
51.3 percent respectively (see appendix table 9). Technical changes that affect
the production of these commodities therefore contribute more to national in-
come. The shares of coarse cereals and wheat are roughly 10.7 percent and 11.3
percent respectively, so their national income contributions are more modest.
However, final demand elasticities matter as well.. Other commodities have
the highest income elasticity (see appendix table 4). In the no-trade scenario,
therefore, the decline in the own price of other commodities (-23.8 percent) is
smaller than the decline in the own price of any other crop following an equal
technical change. Coarse cereals are at the other extreme. A 20 percent increase
in yield leads to a 35.5 percent decline in prices.
The income distribution impacts of the trade ancl no-trade scenarios differ
accordingly. Technical change that affects other commodities benefits urban
groups fairly evenly, and the disparities among rural groups are also modest. But
technical change that affects coarse cereals clearly benefits the poor urban and
poor rural groups, while neither the urban nor the rural rich gain.
As the all-crop scenarios (4.5a and b) illustrate, trade policy is the major
determinant of the distributional outcome of technical change. The gains for the
urban poor can vary from a high of 17.8 without trade to a low of -16.6
percent in the open economy scenario, depending on how much of the gain in
yield is exported or used to reduce imports. For the rural rich, gains can vary
from 1.8 percent to 30.5 percent while the impact oII the urban rich can range
from a gain of 9.8 percent to a loss of 10.5 percent.
When technical change affects all crops positively and there is no expansion of
trade, the poorest rural group gains 9 percent from drops in prices but virtually
nothing from wage rises. When, however, the full gains from technical change
are exported, the poorest rural group sees no fall in prices. But its wages rise,
and to a small extent this group also benefits from the massive rise in farm
profits. On balance, both of the poorest rural quartiles would still be much
better off without an increase in trade. The situation is reversed for the second
income quartile. The positive farm profit effects outweigh the negative food
price effects. When exports increase as a consequence of technical change, the
income gain of the second rural quartile is 10.2 percent; without exports the
gain drops to 7.3 percent.
A remarkable feature of technical change scenarios is the modest impact they
have on real rural wages, regardless of what happens in trade. The largest
absolute change in the real wage bill is an increase of 3.2 percent in scenario



Quiz6n and Binswanger 123
4.1b. This very small wage response is not caused by the elasticity of labor
supply, which is instead very inelastic. The total supply elasticity of rural labor,
including the migration response, is less than 0.5. Demand for labor is also
inelastic (-0.48) and thus cannot account for the limited wage response. In-
deed, when labor is withdrawn from rural areas, either because of reduced
fertility (scenario 1.1 b) or rural-to-urban migration (scenario 1.2), real rural
wages increase sharply.
Real wages remain stable despite technical change because technical change
has contradictory effects on the demand for labor. As yields increase, less labor is
needed to produce any given level of output, and thus labor demand is de-
pressed. At the same time, however, the technical change has increased real
incomes and reduced the relative prices of agricultural commodities, so that
demand for the products increases. Thus, while the per unit labor requirement
has declined, the impact on total labor demand is offset by the induced rise in
demand for and production of agricultural commodities. It is the balance be-
tween these offsetting forces which determines the final effect on the demand for
labor.
The findings shown in tables 3 and 4 should dispel the notion that technical
change is responsible for the slight wage decline observed in section II. This
decline must instead be the result of inadequate growth in labor demand in the
nonagricultural sector.
Investment Scenarios and Fertilizer Subsidies
In this subsection we present only closed economy scenarios.
Under scenario 5.1 (see table 5), the assumption is that investment in irriga-
tion is accelerated enough to increase the percentage of area irrigated by 10
percent. This leads to an increase in aggregate output of 2.7 percent and a drop
in the aggregate price level of 5.8 percent. Because irrigation requires labor,
labor employment and real wages rise slightly. Residual farm profits, however,
decline by 4.8 percent as a consequence of slightly higher labor costs and lower
output prices. The income distributional outcomes follow from these price and
profit changes. The landless gain modestly (2.9 percent), while large farmers
lose (-0.7 percent). All urban households gain substantially, with the poorest
showing the largest gain (6 percent).
In the aggregate, real per capita income rises modestly (1.7 percent). Changes
in the yield and price of individual commodities thus reflect shifts in income
distribution rather than aggregate income growth. Wheat shows the biggest
production increase and the largest price drop.
Scenario 5.2 focuses on expanding such capital inputs as tractors, other imple-
ments or machines, and livestock, and on improving the marketing infrastruc-
ture. Both of these are accelerated by 10 percent. Real per capita income then
decreases slightly (0.2 percent) as a consequence of producer losses. Aggregate
agricultural output increases by about 0.8 percent, and the price index drops by
3 percent. The effects on income distribution are similar to those of increased



Table 5. Simulated effects of increased agricultural investment and inputs
(percentage change)
Scenarios for change in exogenous variables
1 0 percent
rise in     Irrigated land up
10 percent   capital inputs   10 percent plus    20 percent
increase in   and marketing    5 percent rise in    fertilizer
irrigated area  infrastructure  inputs and marketing   subsidy
Endogenous variables        (sS. 1)       (sS.2)             (sS.3)          (sS.4)
National income per capita        1.71       -0.20               1.61             1.30
Output
Total                          2.72          0.81              3.12             1.26
Rice                           0.64        -1.24               0.02             0.34
Wheat                          5.14          1.34              5.81             1.29
Coarse cereals                  1.88         2.12              2.94           -2.13
Other crops                    3.48          1.47              4.22             2.47
GNP deflator                    -5.76        -2.95             -7.23            -1.13
Prices
Rice                          -6.93        -2.11             -7.98            -1.76
Wheat                        --12.77       -5.34            -15.44            -1.78
Coarse cereals                -9.39        -7.19            -12.98              1.25
Other crops                   -6.38        -3.93             -8.34            -1.94
Wage rate                        0.71          0.42              0.92           -1.90
Employment                       0.44          0.25              0.56           -0.77
Wage bill                         1.14         0.67              1.48           -2.67
Profits                         -4.79        -8.20             -8.89              5.58
Income per capita (by quartile)
Rural
Poorest                      2.92          1.64              3.74           -0.35
Second                       1.71          0.03              1.73             0.75
Third                        0.90        -1.02               0.39             1.54
Richest                    -0.67         -2.39             -1.87              2.54
Urban
Poorest                      6.04          2.66              7.37             0.60
Second                       5.73          2.56              7.01             0.74
Third                        5.15          2.38              6.35             0.60
Richest                      3.50          1.73              4.37             0.40
Per capita cereal consumption
Rural
Poorest quartile             2.57          1.21              3.18           -0.74
Richest quartile            -0.07        -1.30             -0.72              0.60
Urban
Poorest quartile             5.59          2.25              6.72             0.06
Richest quartile            -0.38        -0.95             -0.86            -0.33
Aggregate per capita cereal
consumption                    1.84          0.04              1.86             0.07
Note: The values shown are the percentage changes in the endogenoas variables from their base case
level, as the result of the induced change in the exogenous variables shown in the column headings.
All income, output, price, wage, profit, and consumption variables are real values.
The income groups shown are expenditure quartiles for rural and urban populations separately. The
four rural expenditure quartiles together constitute 0.8009 of the total population, while the four urban
quartiles include 0.1991 of the total population.
124



Quiz6n and Binswanger 125
irrigation, but the disparities among the income groups are substantially less
because the effect on output is smaller.
Scenario 5.3 combines the two previous scenarios, except that investment in
irrigation is accelerated twice as much (10 percent) as investment in capital and
marketing (5 percent). Since the distributional effects of scenarios 5.1 and 5.2
are so similar, their combined effects are largely a matter of increased magni-
tudes.
Scenario 5.4 portrays a simple fertilizer subsidy scheme in which government
pays 20 percent of the actual cost of fertilizers.8 Since we assume that the supply
elasticity of fertilizers is high (4.0), this scheme results in a considerable shift of
the supply curve. We also assume that fertilizer is not rationed or is sold in the
black market at higher prices. In other words, the subsidy actually reaches the
farmers and alters their fertilizer use.
Under this scenario, aggregate agricultural output increases by about 1.3
percent with the output gains concentrated in wheat and in other commodities.
Output of coarse cereals, which are not fertilizer-responsive, declines by 2.1
percent. The GNP deflator declines by 1.1 percent, which leads to gains for
urban consumers. Fertilizer is substituted in part for labor, and the real wage bill
declines by 2.7 percent. Thus the rural poor lose, since their losses in wages
outweigh their gains as consumers. The rural rich, however, obtain a gain in
income because of higher farm profits. For them, the fertilizer subsidy and lower
wages more than offset the negative effect of smaller output prices.
Note, too, that the rural poor's cereal consumption declines by more than
their real income loss because the prices of coarse cereals rise.
Taxation and Redistribution Scenarios
Scenarios 6.1 and 6.3 (see table 6) postulate various forms of taxation each of
which raises Rsl2 billion (billion is 1,000 million). It is assumed that the money
is used for purposes which do not affect agricultural demand or supply.
In scenario 6.1 a land tax of 10 percent of residual farm profits is levied.
Residual farm profits are considered a proxy for land rents.
In scenario 6.2 the assumption is that a progressive income tax is levied on
rural residents alone. In order to raise Rs12 billion at 1973-74 prices, rates of
3.1 and 6.2 percent on the nominal incomes of the upper two rural quartiles are
required. The two poorer quartiles are untaxed.
Scenario 6.3 involves imposing an excise tax on nonagricultural goods. This is
achieved by means of an exogenous increase of 9.7 percent in the price index for
nonagricultural goods. Unlike the land tax or the income tax, the excise tax
would also fall on the urban income groups.
Scenarios 7.1 to 7.3 are income redistribution schemes that assume an in-
crease of 30 percent in the nominal per capita income of the poorest rural
8. Indian fertilizer subsidy policy is complex, and a more detailed analysis is required to assess its exact
impact.



Table 6. Simulated effects of changes in taxation and income redistribution policies
(percentage change)
Scenarios for change in exogenous variables
30 percent income redistribution
From                   From
10 percent   Rural   Nonagricultural 10 percent From rural nonagricultural From land
land tax  income tax    excise tax  land tax  income tax    excise tax  transfer
Endogenous variables    (s6.1)     (s6.2)       (s6.3)       (s7.1)     (s7.2)      (s7.3)       (s8. 1)
National income per capita  -3.14     -2.91        -2.67         0.32       0.55         0.79        0.57
Output
Total                    -0.29      -0.25        -0.02         0.12       0.17         0.39        0.18
Rice                     -1.00      -0.81        -0.27         0.68       0.87         1.41        1.01
Wheat                    -3.17      -2.84        -2.06         0.86       1.20         1.97        1.35
Coarse cereals             3.59       3.27        2.24        -0.12      -0.44        -1.47       -0.60
Other crops              -0.16      -0.18         0.03        -0.26     -0.28         -0.07       -0.32
GNPdeflator                -7.11      -6.05         2.19         3.68       4.74        12.98        5.19
Prices
Rice                    -10.47      -8.80        -0.39         6.01       7.67        16.09        8.45
Wheat                   -14.07    -12.15         -3.51         6.46       8.38        17.03        9.21
Coarse cereals           -5.61      -4.44         3.25         5.61       6.78        14.46        7.25
Other crops              -8.60      -7.41        -0.32         3.75       4.95        12.03        5.41
Wage rate                    0.36       0.40       -0.54         0.57       0.61       -0.34         0.63
Employment                   0.28       0.27       -0.09         0.10       0.09       -0.28         0.08
Wage bill                    0.64       0.67       -0.63         0.66       0.69       -0.61         0.72
Profits                   -25.10    -12.84         -5.03        -3.37       8.89        16.70        9.76



Income per capita (by quartile)
Rural
Poorest                    0.58        1.46       -1.00         27.87       28.74         26.28        28.62
Second                   -2.99       -0.46        -2.56         -2.25        0.28         -0.82         0.31
Third                    -5.41       -4.88        -2.12         -2.39       -1.86          0.90         1.37
Richest                 -10.10      -10.76        -3.51         -2.24       -2.89          4.35        -4.31
Urban
Poorest                    5.82       5.59        -1.19         -3.62       -3.85        -10.63        -4.23
Second                     6.72        6.09       -1.81         -3.57       -4.20        -12.10        -4.63
Third                      5.98        5.44       -2.57         -3.03       -3.57        -11.58        -3.93
Richest                    3.82        3.66       -4.54         -1.88       -2.03        -10.24        -2.24
Per capita cereal consumption
Rural
Poorest quartile           0.59        0.98       -0.37          16.52      16.91         15.57        16.73
Richest quartile         -3.10       -3.36          0.62        -1.06       -1.32          2.66        -1.82
Urban
Poorestquartile            4.86        4.54       -0.52         -3.47       -3.78         -8.84        -4.15
Richest quartile         -0.29       -0.21        -0.04         -0.38       -0.31         -0.13        -0.36
Aggregate per capita cereal
consumption                -0.59       -0.46        -0.20           0.58       0.71          0.97         0.78
Note: The values shown are the percentage changes in the endogenous variables from their base case level, as the result of the induced change in the exogenous
variables shown in the column heading.
All income, output, price, wage, profit, and consumption variables are real values.
The income groups shown are expenditure quartiles for rural and urban populations separately. The four rural expenditure quartiles together constitute 0.8009 of
the total population, while the four urban quartiles include 0.1991 of the total population.



128 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. 1
quartile. In scenario 7.1 the source of the added 30 percent is the land tax
discussed above, which is just sufficient to finance the increase. In scenario 7.2
the progressive income tax is used to finance the increase. In scenario 7.3 the
source is the excise tax at the rate of 9.7 percent.
The land tax translates into substantial price drops (the GNP deflator is -7
percent) but only a minimal decline in aggregate agricultural output (-0.3
percent). The decline in residual farm profits is 25 percent. Urban groups gain
from the price decline, while rural groups (except for the poorest) suffer a real
income loss, the more so the richer they are. This income redistribution leads to
a redirection of production toward coarse cereals (3.6 percent) and away from
rice and wheat (-1 and -3.2 percent respectively).
Qualitatively, the effects of the income tax are very similar to those of the land
tax. The only exception is the much smaller effect of the income tax on farm
profits, an effect that is the result of reduced final dernand rather than a direct
tax effect. Because the rural poor escape direct taxation and the rural rich carry
the entire tax burden, the income distribution effect of the income tax is more
progressive than that of the land tax.
The excise tax has a more even incidence than the other taxes, since it also
falls on the urban groups. The urban rich and the rural rich are the groups
whose incomes are reduced the most, -3.5 and -4.5 percent respectively.
Depending on how taxes are levied, the planned gain of 30 percent in income
for the rural poor is somewhat eroded. The gain from the excise tax falls to 26.3
percent, the gain from the land tax falls to 27.8 percent, and the gain from the
income tax falls to 28.7 percent. This happens because the rural poor have a
higher propensity to consume food than the rich, and their increased demand for
food causes food prices to rise. In the land tax and income tax scenarios, the
richer rural groups lose very little because rising food prices increase farm
profits. In fact, the rich rural group loses only about as much as the untaxed rich
urban groups, whose loss stems from rising food prices. These higher prices
mean, in effect, that the urban groups end up paying part of the income transfer
to the rural poor. In the excise tax scenario, the price level rises particularly fast
(13 percent) because increases in food prices and the excise tax both affect the
price level. Therefore, almost the entire burden of an excise tax would be placed
on the urban groups. Large farmers would show a net gain as a consequence of
the increased food demand of the rural poor.
In scenario 8.1, sufficient land is transferred from the fourth rural quartile
(the richest) to the first quartile to give the rural poor an initial income boost of
nearly 30 percent. The effects are very similar to the land tax and income tax
scenarios, although the rural rich lose a bit more, since they are the only taxed
group.
The taxation and income redistribution scenarios show that a substantial
increase in the income of the rural poor could be achieved that would cause only
small losses-or, in some cases, even a net gain-for the rural rich. This is an
important, and initially counterintuitive, result that is again critically dependent



Quiz6n and Binswanger 129
on letting food prices rise. If, for example, the government decided to accommo-
date the increased food demand via imports, large farmers would inevitably lose
as their profits would not increase.9
V. CONCLUSION
During the past two decades, Indian agricultural output has grown at an
annual rate of 2.7 percent, which is extremely high by international standards.
Production is now at a level that would be sufficient to feed India's population,
which has increased by 2.2 percent annually. The technical changes associated
with the Green Revolution have been an important part of this increased output,
and there is no question that, had they not occurred, India would be far worse
off today than it is. During the early Green Revolution period, the real per capita
income of the rural population of India rose by about 8 percent. However, these
gains were rapidly eroded. The sobering point is that in 1980-81 real rural per
capita income appears to have been only about 2 percent higher than in 1960-
61.
The early productivity gains of the Green Revolution were retained by the
agricultural sector because Indian policymakers used these gains to reduce im-
ports of foods. Food prices therefore continued to rise slightly. But when near
self-sufficiency was reached, all the extra output had to be absorbed domesti-
cally, food grain prices declined, and terms of trade moved substantially against
agriculture. The benefits of the productivity gains were thereby transferred to
consumers, a classic case of the agricultural treadmill.
The early Green Revolution period was associated with a sharp rise in residual
farm profits, while the real wage bill rose much more modestly. The real income
gain of that period was distributed regressively; large farmers gained the most
while the rural poor gained very little. However, the subsequent rapid drop of
about 25 percent in residual farm profits reduced the per capita incomes of the
rural rich to their 1960-61 levels. By 1980-81, both the absolute level of real
rural per capita income and its distribution appear to have returned to about
what they were in 1960-61.
Real rural wages (as measured by actual data) appear to have risen somewhat
during the early Green Revolution but then dropped back so that by 1980-81
they were barely above the 1960-61 level. Agricultural employment (as mea-
sured by the model) rose substantially but at a rate slower than rural labor force
growth. The simulations show clearly that the main reasons for these trends are
the adverse demographic trends and the insufficient growth in nonagricultural
employment. The rural poor did not lose too much only because they shared
9. Note that the government can also alter the income distribution by direct food distribution at
subsidized prices to rural or urban groups. These issues are discussed in a separate paper (Binswanger and
Quiz6n 1986).



130 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. 1
somewhat in farm profit growth, in nonagricultural income growth, and in the
consumer benefits from declining agricultural terms of trade.
There were probably subgroups of the rural poor, including the totally land-
less agricultural workers, who suffered more severely than indicated. In addi-
tion, several areas of the country did not participate in the adoption of Green
Revolution technology and their farmers must have been hurt by the decline in
agricultural prices. We are currently investigating interregional issues, but it is
evident that the simulations presented here have obscured the sharper declines in
income experienced by these subgroups and regions.
A further sobering point is that some intuitive notions of how agricultural
development trends or policies affect income distribution are likely to be wrong.
Much of the debate on the distributional impact of technical change has concen-
trated excessively on the nature of the technical changes. Our analysis suggests
that trade policy is a far more important determinant of income distribution. All
growth-oriented policies or technical changes tend to benefit net buyers of food
if the additional agricultural output is absorbed domestically rather than used to
reduce imports or increase exports. Forcing domestic absorption has been the
policy customarily pursued by India except during the early part of the Green
Revolution, when agriculture production gains were used largely to reduce im-
ports. Since the gains that occurred later were not used to expand exports, net
food buyers in both rural and urban areas benefited.
This evidence, however, does not imply the general superiority of policies
prohibiting external food trade for equity objectives. Given a shortfall in food
production, rather than the sharp increases examined here, use of food imports
to avert shortages and a marked rise in prices will serve equity goals.
There are, of course, differences in the impact of different technical changes or
investment policies. As intuition would suggest, technical changes resulting in
greater production of coarse cereals benefit poor rural consumers more than
technical change directed toward improving production of other types of food.
And investment in irrigation has a greater effect on the demand for labor than a
fertilizer subsidy, which tends to encourage the substitution of fertilizers for
labor.
But none of the agricultural development measures can affect rural labor
demand or wages nearly as much as changes in population growth or growth in
the demand for nonagricultural labor. This does not mean, of course, that we
should not pay attention to the selection of agricultural techniques or other
employment determinants. It simply suggests that maojor improvements in the
incomes of the rural poor must eventually come mainly from demographic
changes and growth in nonagricultural employment.
The only other avenue for substantially affecting incomes of the poorest
group, the rural poor, are direct income transfers or land redistribution. Such
transfers increase food demand and under closed economy conditions lead to
rising food prices, which thereby erode the gains to the rural poor to roughly 90
percent of the initial nominal transfer. More irnportant than this erosion of



Quiz6n and Binswanger 131
benefits, however, is the effect these price changes have on shifting the burden of
the tax required to finance the income or asset transfers. If the rural rich are the
source of taxes (or land) to finance the income (or land) transfer, the rise in food
prices and therefore in farm profits drastically reduces the real incidence of the
tax on them and shifts it to the urban groups. In the case of more broad-based
taxes such as an excise tax, the rural rich can even become net beneficiaries from
the income distribution because the increase in farm profits exceeds their tax
burden. Note that these conclusions would not hold in a free trade situation in
which food prices remained at a constant level. Again, the crucial nature of the
trade decision on the real outcome of government policies must be emphasized.
APPENDIX A. THE MODEL IN MATHEMATICAL TERMS
The model is an extension of the theoretical model described in Quiz6n and
Binswanger (1983). Producer behavior is represented by a system of output
supply and factor demand equations called the producer core. Analytically, the
producer core is derived from a variable profit function [I  = II*(V,Z,T), where
II* is maximized variable profits, V = (P, W) is the vector of prices of outputs (P)
and variable inputs (W), Z is a vector of fixed inputs, and r is a technology index
(appendix table 1 lists definitions of the symbols used in the model). The output
supply and factor demand curves are derived from II- via Shepard's lemma; that
is, the vector of outputs (Y) and (negative) variable inputs (-X) is written as
Q = [Y, -X]  i
In terms of rates of changes, they are written as
(1)                   Q'= E I3iV  + E f3gZ; + Es'        ieO, VI
0 is the set of outputs and VI the set of variable inputs. The prime notation X'
of a variable X indicates the total rate of change over time of variable X. The
star notation X* refers to the rate of change of an exogenous variable or to the
exogenous component of an endogenous variable. Oij are the elasticities of sup-
ply (or demand) of an output i (or factor i) with respect to a price j. The Zs are
exogenous variables and fixed inputs affecting producer behavior, and the (3ig are
supply or demand elasticities with respect to those fixed inputs. Some of the Z
variables are subject to government policy. E, = aQi/l3T l/Qi a8r/O- are the
technology shifters of the supply and factor demand equations if fixed inputs are
held constant. (For a detailed discussion of these technology concepts used, see
Quiz6n and Binswanger 1984.)
Output demand is treated in a more disaggregated fashion. Let k = 1, . . . , K
refer to income groups. Then total final demand is defined as
2Yi  E Yk          ieO



Appendix Table 1. Symbols Used in the Model
dQ, 1  aT
Ej'    =   dt  0- dt    technology shifters, that is, shifts in output supplies and factor
demands for given fixed input levels. These are profit function definitions.
G      =  square matrix of elasticities and shares
K      =  column vector of exogenous shifter variables
L      =  labor services
M      =  total nominal income
m      =  real per capita income
MN    =  nonagricultural income
N      =  population
NA     =  nonagricultural commodities
P      =  output prices
P      =  output price indexes
Q      =  [Y, -X] = vector of outputs and (negative) variable inputs
S      =  rent to fixed factor
Si     =  share of output i in total revenue or share of factor i in mnput prices
t      =  time
U'    =  column vector of endogenous variables
V      =  [P, WI = vector of outpat and variable input prices
W      =  wage rate or variable input prices
w      =  real wage rate
X      =  variable inputs
Y      =  outputs
y      =  per capita output
Z      =  fixed factors; Z, refers to land
T      =  technology index
=  price elasticity of supply or demand in production (outputs and inputs respectively)
Tr     =  variable profit function
isik   =  proportion of i from (by) group k
e      =  price elasticity of factor supply
e      =  labor supply per person
4)     =  variable profit shares
Pik    =  share of group k's consumer expenditures spent on commodity i
Pk     =  share of real income accruing to group k
Modifiers of variables unless already defined above:
X      =  level
X, XT  =  a column and a row vector of the X variables, respectively
dX 1
XI    =  dt X  =  total rate of change (n growth rate) of variable X with respect to time
X'-    =  exogenous component of the rate of change of a variable (except that 7r* stands for
maximized variable profits)
Indexes and sets of inputs and outputs
g      =  shifter variables
i      =  commodities (outputs, inputs)
j      =  commodities (outputs, inputs)
k      =  income groups
K      =  set of income groups or their total number
O      =  set of outputs or total number of outputs
I      =  set of inputs or total number of inputs
VI    =  set of variable inputs or their total number
132



Quiz6n and Binswanger 133
where Yik is the total demand of consumer group k. Rewriting equation 2 in
terms of changes,
(3)                         yi = E Xik Yk            ieO
k
where Xik = Yzk / Yi is the proportion of commodity i consumed by income group
k. The consumption of each income group is described by an income-group-
specific consumer demand system:
(4)                        Yk = Nkyk(P,mk)
where the underbars denote a column vector of the variable; for example, Yk
(Y1, Y2, . . ., YO), Nk is the population in income group k, and Yik is the per
capita demand which depends on output prices and the per capita income of the
income group. Transforming each equation into rates of changes leads to
(5)           Y,k = Yi'k + Nk =   UcijkPJ +  iXimkmk + Yik + Nk  i,je O
Here, yik is an exogenous change in per capita demand of income group k, and
aoj and aiim are the price and income elasticities of final demand.
We assume that the population in each income group grows at an exogenous
rate Nk4- But the rural population grows via immigration or via diminished
emigration and vice versa for the urban population. We assume this migration to
be responsive to the real rural wage rate. Differentiating Nk with respect to time
and the real wage and converting to rates of changes leads to
(6)              Nk = Emk W' + Nk = Emk (WL - Pk) + Nk
where Emk is the migration elasticity into (or for the urban group, out of) the
speific income group with respect to the real rural wage and Pk is an income-
group-specific price deflator defined below. Let the rural income groups be
indexed k = 1, . . . 4 and the urban group by k = 5,.. . 8. Total labor supply to
agriculture is
8
L = L Lk,
k=1
or in rates of changes
8
(7)                         L     - =     XLkLk
k=
where XLk = Lk/L is the proportion of labor supplied to agriculture by income
group k. Labor supply of income group k is Lk = ekNk where 4k is total labor
supply per person. Differentiating with respect to the real wage and time and
converting to rates of change, we find
(8)          Lk = efkw' + f + Nk = ek (WL - A) + k + N'



134 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. 1
where eek is the total labor supply elasticities of income group k and f£- is an
exogenous shifter in the labor supply to agriculture of income group k.
The supply of bullocks is similarly aggregated as
8
Xi = 2 Xik
k=1
Rates of change are aggregated as usual; that is,
8
(9)                             Xi,= kEXkXik              i = bullocks
k=1
The supply of each input is only dependent on its own price, Wj; therefore,
(10)                       Xik = Eik(Wi- Pk) + Xi = bullocks
While the model contains many fixed input (Z) variiables, such as irrigation
and rainfall (see appendix table 10), we treat land (indexed as ZI) as the only
fixed factor which is a recipient of residual farm profits. The change in residual
farm profits per unit of land (rental rate of land) S' is derived residually from the
profit function, a derivation given in detail in Quiz6n and Binswanger (1986).
(11)                    S' =  E oyV' + ±   OiQi' -Z               ieO, VI
where
O=  QiVi
II*;
are variable profit shares, which are positive for outputs and negative for inputs.
Changes in income-group-specific consumer price levels Pk can be related to
the endogenous changes in agricultural output prices as follows:
(12)                      Pk = S PikPi + E PNAkPNA
ZrO      icO
where ltik is the share of total consumer expenditures spent on commodity i by
income group k. The subscript NA refers to nonagricultural commodities. The
GNP deflator P is derived in the same way by dropping the k subscripts.
The nominal income of income group k is Mk and is defined as the sum of all
net factor incomes accruing to the group plus nonagricultural incomes MN. 10
(13)                    Mk = E XikWi + Z,kS + MNk
jE VI
where Zlk is land supplied by group k.
10. Note that we treat all rural labor supply as "agricultural" labor because we assume wage equaliza-
tion between the agricultural and nonagricultural rural labor markets.



Quiz6n and Binswanger 135
Real per capita income is derived by dividing by the number of people and the
consumer price index: that is,
(14)                        mk = Mk/PkNk
Differentiating 13 and 14 totally and converting to rates of changes leads to
(15)   m' = E 6ik(Wi + Xik) + bsk(S + Zk) - h - N' + 6MNkMNk
iEVI
where the 6i are the shares of net income arising from the respective source.
The real per capita income of India's rural and urban population is defined as
(16)                        m = r.XNkmk
k
where XNk = Nk/;Nk are the initial shares of group k in the total population.
Differentiating and converting into rates of changes leads to
(17)                m' =  SXNk -Mk =  vkMk
k     m        k
where vk is the proportion of real income accruing to group k.
Note that m' is not equal to the conventional definition of a change in real per
capita income, which would be
(18)                      ml = M'-N'-P'
where P' is computed from the equivalent of equation 12 but the k subscript is
dropped, and M is defined as in equation 13 but again the k subscript is
dropped. The difference between equations 17 and 18 is that 17 utilizes real
income weights, for which each group's real income is deflated by a group-
specific price deflator. Thus 17 is closer to a measure of a change in real per
capita welfare than is 18.
The model treats India as a closed economy with respect to agricultural com-
modities except that trade by the government is allowed and is easily treated as a
fixed addition to and reduction from domestic supply. The full model consists of
equations 1, 3, 5, 6, 7, 8, 9, 10,11, 12, 15, and 17.
The equation system can be exhibited in matrix form:
(19)                          GU' = K*
where G is a square matrix of elasticities and shares, U' is the column vector of
endogenous variables, and K'- is a column vector of exogenous shifter variables.
(For simpler examples of such full systems, see Quiz6n and Binswanger 1983,
1986.) The effect of a shift in an exogenous variable on the endogenous vari-
ables in the system can be solved as
(20)                         U' = G`K`
which exists so long as the matrix G is nonsingular.



136 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. 1
APPENDIX B. DATA AND PARAMETER VALUES
The data used to compile our core (or G) matrix come from a variety of
sources.11 The agricultural commodities, rice, wheat, inferior cereals, and other
crops, are exhaustive in that they account for all crop production in the agricul-
tural sector. Livestock products are aggregated into "other agricultural commod-
ities."
The commodity-specific output supply and the fertilizer and labor demand
elasticities for the semiarid tropics (SAT) are from 13apna, Binswanger, and
Quiz6n (1984); for North India and the eastern rice region (ER) from Evenson
(1981); and for the coastal rice region (CR) of South India from unpublished
estimates. Estimation equations were derived from a normalized quadratic
profit function on which all regularity conditions were imposed, except for the
condition of convexity of the resultant Hessian matrices. The estimates were
therefore adjusted in an ex post manner in order to satisfy this convexity con-
straint, following trial and error procedures described in Quiz6n and
Binswanger (1984).
The bullock power demand elasticities have been estimated in Evenson and
Binswanger (1984). Only the own-price elasticity and the cross-price elasticity
with respect to labor are available for bullock power demand.
The output demand elasticities are from Binswanger, Quiz6n, and Swamy
(BQS, 1984) and are averages for all India. Original price coefficient estimates in
the reported demand equations were first adjusted following trial and error
procedures to satisfy convexity restrictions. Then, from the twenty-eighth round
of the National Sample Survey (28 NSS) Tables on Consumer Expenditures, the
average commodity prices, the per capita quantities consumed, and the real per
capita expenditures and incoimes of each of our defined expenditure quartiles
were computed. These were used with the adjusted convex price coefficient
estimates and the income coefficient estimates from BQS to obtain all expendi-
ture-quartile-specific output price and income elasticiities of demand.12 Total
consumption by commodity and by group, computed straightforwardly from
the 28 NSS, was used to obtain the Xik output consumption weights in equation 3
and the Iik weights in equation 12.
The 1970-71 National Council for Applied Economic Research (NCAER) Ad-
ditional Rural Income Survey (ARIS) is a national rural household survey that
contains a wealth of data, including information on household ownership of
different agricultural factors of production, household incomes by income
source, and costs of agricultural production by factor of production. The survey
does contain data on hired labor but not, however, oII family labor input. We
therefore used data on family labor input by farm size group from the Farm
11. All notations used in this appendix are defined in appendix A.
12. For the highest income group, urban 4, the estimated elasticities for coarse cereals had high
negative values. These values were reduced to a minimum of -1.



Quiz6n and Binswanger 137
Management (FM) studies and matched each household in the NCAER survey with
the corresponding farm size group in the FM study which most closely resembled
the agroclimatic features of the district in which the NCAER household resided.
For a number of semi-arid districts, we used family labor data from the more
recent village studies of the International Crops Research Institute for the Semi-
Arid Tropics (ICRISAT). With this addition, the NCAER survey enables computa-
tion of the Xik input supply weights in equations 7 and 9, the xit profit shares in
equation 11, and the bik income weights in equation 15. Pal and Quiz6n (1983)
describe in detail how all these shares are computed from the NCAER-ARIS survey.
The only unaccounted parameters in our G matrix thus far are the input
supply elasticities, that is, e,k in equations 6, 8, and 10. On the basis of Ro-
senzweig's (1980) econometric estimates, we assume Els to be equal to 0.3. The
migration elasticities, Emk, are computed from Dhar (1980) and are equal to
0.1083 for the rural groups and -0.4356 for the urban groups.13 For bullock
labor, the own-price elasticity is assumed to be equal to 0.4993, that is, the
average of the value weighted sums of the own-price elasticities of supply for
agricultural outputs in each agroclimatic region. This follows from the notion
that bullocks are reproducible out of agricultural output. Finally, the fertilizer
supply elasticity is set at 4.0, a high value which reflects opportunities for
international trade.
The most important elements of our G matrix are given in appendix tables 3
through 9. All parameters pertaining to cost, income, and factor supply shares
are listed in Pal and Quiz6n (1983). The f,lg matrix (equation 1) of shifter
variables is listed in appendix table 10. These elasticity estimates are from the
same estimation equations used to construct the matrix of output supply and
variable input demand elasticities. The complex K" vector of exogenous shifter
variables (equation 19) can be reconstructed from the appendix tables and other
data sources that already have been mentioned.
13. Quiz6n and Binswanger (1984) explain how these migration elasticities were computed from
Dhar's study.



Appendix Table 2. Agroclimatological Regions of India Showing the States,
Union Territories, and Districts that Comprise Them
Agroclimatological   State or union
region            territory                        Districts,
Semiarid tropics (SAT)  Andhra Pradesh    Adilabad, Nizamabad, Karimnagar, Medak,
Warangal, Mahbubnagar, Hyderabad, Nalgonda,
Khammam, Kurnocl, Guntur, Vishakapatnam,
Anantapur, Cuddapah, Ongole, Nellore, Chitoor
Gujarat           All
Karnataka         Bidar, Gulbarga, Bijapur, Belgaum, Dharwar,
Raichur, Shimoga, Bellary, Chikmagalur,
Chitradurga, Hassan, Tumkur, Mandya, Mysore,
Bangalore, Kolar
Madhya Pradesh   All
Maharashtra       All
Rajasthan         All
Tamil Nadu        Dharmapuri, The Nilgiris, Coimbatore, Salem,
Tiruchirapalli, Pudukkottai, Madurai,
Ramanathapuram, Tirunelveli
Dadra and Nagar
Haveli
Eastern rice (ER)    Arunachal Pradesh  All
Assam             All
Bihar             All
Manipur           All
Meghalaya         All
Mizoram           All
Nagaland          All
Orissa            All
Tripura           All
Uttar Pradesh     Jalaun, jhansi, Hamirpur, Banda, Fatehpur, Rae
Bareli, Sultanpur, Faizabad, Basti, Allahabad,
Pratapgarh, Jaunpur, Azamgarh, Gorakhpur,
Mirzapur, Varanasi, Ghazipur, Ballia, Deoria
West Bengal       All
Coastal Rice (CR)    Andhra Pradesh    Srikakulam, East Godavari, West Godavari, Krishna
Goa
Karnataka         North Kanara, South ]Kanara, Coorg
Kerala            All
Pondicherry
Tamil Nadu        Chingliput, North Arcot, South Arcot, Thanjavur,
Kanyakumari
Northern wheat (NW)  Chandigarh
Delhi
Haryana           All
Himachal Pradesh  All
Jammu and         All
Kashmir
Punjab            All
Uttar Pradesh     Dehradun, Saharanpur, Bijnor, Nainital,
Muzaffarnagar, Meerut, Moradabad, Rampur,
Bulandshahr, Budauin, Bareilly, Pilibhit, Mathura,
Aligarh, Agra, Etah, Mainpuri, Farukhabad,
Shahiahanpu, Kheri, Etawah, Hardoi, Sitapur,
Kanpur, Unnao, Lucknow, Barabanki, Bahraich,
Gonda
- Not applicable.
a. For those states that fall into two agroclimatological regions, districts are allocated and identified
individually.
138



Appendix Table 3. Price Elasticities of the Producer Corefor All India
Prices
Coarse    Other
Quantities    Rice     Wheat   cereals    crops   Fertilizer   Labor   Bullocks
Rice           0.5531  -0.1280  -0.1271  -0.1834  -0.0206  -0.0940    0.0000
Wheat         -0.0900    0.4454  -0.1583  -0.0879  -0.0614  -0.0479    0.0000
Coarse cereals  -0.2280  -0.1088    0.7554  -0.2039    0.1791  -0.3986    0.0000
Othercrops    -0.1632  -0.0320  -0.0652    0.2955  -0.1011    0.0663    0.0000
Fertilizer      0.0026    0.1203  -0.4635    0.7525  -0.8355    0.4278    0.0000
Labor           0.1019    0.0228    0.2045  -0.0489    0.0753  -0.4782    0.1225
Bullocks       0.0000    0.0000    0.0000    0.0000    0.0000    0.1335  -0.4041
Note: Elasticities are computed at base year 1973-74 prices and quantities. Estimates are aggregated
from Evenson (1981), Evenson and Binswanger (1984), and Bapna, Binswanger, and Quiz6n (1986).
139



Appendix Table 4. Own-Price and Expenditure Elasticities of Demand for All India,
by Commodity and by Expenditure Quartile
Own-price elasticities of demand                Income elasticities of demand
For       For                                  For      For
For       For      coarse    other      For      For    For    coarse    other    For
Expenditure quartile    rice   wheat    cereals    food    Nonfood   rice    wheat   cereals   food    nonfood
Rural 1       -0.7752  -0.7172  -0.6153  -0.8857  -0.5431  0.8196  1.1325    0.0653  1.2734   1.4408
Rural2        -0.8311  -0.7247  -0.5507  -0.8077  -0.5529  0.7436  1.039   -0.3328  1.1540   1.5724
Rural3        -0.8735  -0.7217  -0.4544  -0.7878  -0.5530  0.6825  1.0011  -0.5889  1.1158   1.5760
Rural 4       -1.0363  -0.7218  -0.0000  -0.7255  -0.5258  0.3768  0.8966  -1.0000  1.02154  1.5504
Urban 1       -0.8088  -0.7233  -0.5883  -0.8426  -0.5484  0.7771  1.0760  -0.1302  1.2040   1.5211
Urban 2       -0.8425  -0.7241  -0.5255  -0.8009  -0.5532  0.7271  1.0279  -0.4055  1.1420   1.5761
Urban 3       -0.9420  -0.7195  -0.2758  -0.7604  -0.5463  0.5672  0.9505  -0.9449  1.0682   1.5698
Urban4        -1.1286  -0.6786  -0.0000  -0.7123  -0.4943  0.0844  0.7698  -1.0000  0.9827   1.5350
Note: Elasticities are computed at base year 1973-74 prices and quantities. Estimates are from Binswanger, Quiz6n, and Swamy (1984).



Appendix Table 5. Shares of Commodities in Consumption of Each
Expenditure Quartile
Commodity
Expenditure                     Course    Other
quartile    Rice     Wheat   cereals   food    Nonfood    Total
Rural 1    0.3152   0.0847   0.1792   0.2643    0.1565    1.0000
Rural2     0.2789   0.1021    0.1215   0.3128     0.1848    1.0000
Rural 3    0.2611    0.1029   0.0870   0.3364    0.2126    1.0000
Rural 4    0.1389   0.1254   0.0613   0.3661      0.3082    1.0000
Urban 1    0.2004   0.1525    0.0744   0.3875     0.1851    1.0000
Urban 2    0.2503   0.0935   0.0346   0.4004    0.2212    1.0000
Urban 3    0.1923   0.0955   0.0265   0.4073    0.2784    1.0000
Urban 4    0.0926   0.0634   0.0101   0.4261      0.4077    1.0000
Source: 28 NSS, Tables on Consumer Expenditures, 1973-74.
Appendix Table 6. Share of Each Expenditure Quartile for All India,
Total Consumption by Commodity
Commodity
Expenditure                      Coarse    Other
quartile    Rice     Wheat   cereals   food    Nonfood
Rural 1    0.1405   0.0672   0.2125   0.0530    0.0505
Rural 2    0.2027   0.1293    0.2432   0.0998    0.0935
Rural 3    0.1938   0.1606   0.1877   0.1146    0.1152
Rural 4    0.1857   0.3131   0.2262   0.2186    0.2946
Urban 1    0.0476   0.0812   0.0507   0.0766    0.0394
Urban 2    0.0806   0.0616   0.0310   0.0910    0.0530
Urban 3    0.0837   0.0885    0.0308    0.1361    0.1097
Urban 4    0.0655   0.0985    0.0179    0.2103    0.2441
Total      1.0000    1.0000    1.0000   1.0000    1.0000
Source: 28 NSS, Tables on Consumer Expenditures, 1973-74.
Appendix Table 7. Share of Each Expenditure Quartile in the Total Supply of
Agricultural Inputs for All India
Agricultural input
Expenditure   Agricultural           Agricultural
quartile      labor      Bullocks   land owned
Rural 1      0.2380      0.0973      0.1137
Rural 2      0.2651      0.1514      0.1625
Rural 3      0.2466      0.2327      0.2553
Rural 4      0.2315      0.4402      0.4685
Urban 1      0.0118      0.0063      0.0669
Urban 2      0.0014      0.0097      0.0359
Urban 3      0.0051      0.0260      0.0320
Urban 4      0.0005      0.0358      0.0377
Total        1.0000      1.0000      1.0000
Sources: 1970-71 NCAER-ARIS survey; 26 NSS, Tables on Landholdings, All India, 1981-82.
141



Appendix Table 8. Share in Total Income from Agricultural Inputs for
All India by Expenditure Quartile
Agricultural input
Agricultural   Total
Expenditure Agricultural         Residual  implements and  agricultural
quartile    labor    Bullocks farm profits,    machinery  income
Rural 1     0.5283    0.0256     0.1132        0.0737       0.7408
Rural 2     0.4278    0.0254     0.2422        0.0567       0.7521
Rural 3     0.3383    0.0258     0.3298        0.0498       0.7437
Rural 4     0.2215    0.0241     0.4726       0.0646        0.7828
Urban 1     0.0713    0.0201     0.0069       0.0000        0.0983
Urban 2     0.0028    0.0099    0.0359         0.0000       0.0486
Urban 3     0.0061    0.0169    0.0320        0.0000       0.0550
Urban 4     0.0002    0.0093     0.0377       0.0000       0.0472
All Groups    0.2042    0.0157    0.2213       0.0380       0.4792
a. Defined as net returns to land.
Source: 1970-71 NCAER-ARIS survey.
Appendix Table 9. Summaries of Agricultural and Demographic Data Used
in the Model
Share in total real income by expenditure quartile for all India
Rural 1    0.0898    Urban 1    0.0345
Rural 2    0.1425    Urban 2    0.0456
Rural 3    0.1752    Urban 3    0.0696
Rural 4    0.3185    Urban 4    0.1244
Total rural 0.7260   Total urban 0.2740
Share in the population by expenditure quartile for all India
Rural 1    0.2002    Urban 1    0.0498
Rural 2    0.2002    Urban 2    0.0498
Rural 3    0.2002    Urban 3    0,0498
Rural4    0.2002     Urban 4    0.0498
Totalrural 0.8009    Total urban 0.1991
Share of agricultural commodities in the value of total agricultural output for all Indiaa
Rice         0.2666
Wheat        0.1073
Coarse cereals 0.1128
Other crops   0.5133
Share of agricultural inputs in the total cost of agricultural production for all Indiab
Agricultural labor               0.3258
Bullocks                         0.1086
Residual farm profits            0.3088
Fertilizer                       0.0331
Agricultural implements and machinery 0.0979
a. Data are from the national accounts.
b. Data are from the 1970-71 NCAER-ARIS survey.
142



Appendix Table 10. Output Supply Elasticities with Respect to Exogenous
Shifter Variables
Variable
High-yield Irrigated
Commodity    Rain    varieties    area    Roads    Land    Capital
Rice           0.3563    0.27s5   0.0011  -0.2116  0.4801   -0.0458
Wheat          0.2178    0.3764   0.7965  -0.0488  0.2871      0.2566
Coarsecereals  -0.0575  -0.1931   0.2547    0.3207  0.2000    0.1025
Other crops    0.0750    0.0340   0.3629    0.0911  0.3056    0.1155
Fertilizer     0.1558    0.5606   0.6370    0.5422  0.0000    0.0000
Labor          0.0557    0.0526   0.0917  -0.0027  0.61291    0.0761
Bullocks       0.0578    0.0441   0.1022    0.0018  0.8882   -0.0183
Note: Elasticities are computed at base year 1973-74 quantities. Estimates are aggregated from
Evenson (1981), Bapna, Binswanger, and Quiz6n (1984), and Evenson and Binswanger (1984).
143



Appendix Table 11. Data Description and Data Sources for Counterfactual Analysis
Variable or
exogenous shock                        Additional Description,                                    Data sourceb
A. For actual price and quantity levels
1. Rice production        Production index for rice                                Index Numbers of Area, Production, and Yield of
Principal Crops in India-Cropwise, Directorate of
Economics and Statistics, Ministry of Agriculture
2. Wheat production       Production index for wheat                               Same as 1 above
3. Coarse cereal production  Production index for coarse cereals                   Same as 1 above
4. Other crop production    Production index for pulses and nonfood grains         Same as 1 above
5. All crop production    Production index for all crops                           Same as 1 above
6. Fertilizer consumption    Computed as the sum of fertilizer production, net fertilizer  Produciion, Imports, Distribution and Consumption of
imports, and net withdrawals from fertilizer stock      Fertilizers, Fertilizer Association of India
7. Employment             Index of employed male rural workers                      Union Primary Census Abstract, Census of India
8. Rice prices,           Wholesale price index for rice                           Index Numbers of Wholesale Prices in India, Economic
and Scientific Research Foundation
-4     99. Wheat prices,           Wholesale price index for wheat                          Same as 8 above
10. Coarse cereal pricesc  Wholesale price index for coarse cereals                 Same as 8 above
11. Other crop prices,      Production-weighted wholesale price index for other crops  Same as 8 above
12. Labor wagesc           Index of money wage rate of agricultural laborers as defined in  Jose (1974) and Agricultural Wages in India, Directorate
Jose (1974)                                              of Economics and Statistics, Ministry of Agriculture
13. Prices of all commoditiesc  Consumption-weighted price index for all commodities (food    Same as 8 above
and nonfood articles)
14. Real per capita income    Personal disposable income at constant prices divided by the  Macroeconomic Aggregate and Population, Central
population; this was used as an exogenous shock in section II   Statistical Office, Department of Statistics, Ministry of
Planning
B. For the Exogenous Shifters
1. Rate of change in net                                                           Area under Principal Crops in India, Directorate of
cropped area                                                                      Economics and Statistics, Ministry of Agriculture
2. Rate of change in      Assumed to be equal across expenditure groups for given  Population by Sex, Sex Ratio, Percentage Decadal
population               period of time                                            Variation of Population and Urban Population as a
Percentage of Total Population, 1901-1981, Census
of India



3. Rate of change in capital  Assumed to be equal across expenditure groups for given  Gross Domestic Capital Formation by Industry of Use,
used in agricultural       period of time. "Capital" refers to the value of           Central Statistical Office, Department of Statistics,
production                 household-owned livestock and machinery and implements     Ministry of Planning
used in agricultural production. How this variable was   Proportion of Households Reporting and Average Value
constructed is explained in the text.                      per Household of Individual Items of Assets and
Liabilities as of 30th June 1971 according to Asset
Groups, Reserve Bank of India
4. Rate of change in the    Assumed equal to the rate of growth in the total number of  Total Buffalos and Cattle Used for Work, Directorate of
supply of draft animals in    buffalo and cattle used for work only and assumed to be  Economics and Statistics, Ministry of Agriculture
agricultural production    equal across expenditure groups for given period of time
5. Rate of change in the    Other inputs are taken to be all inputs other than land, labor,    Value of Output from Agriculture, Central Statistical
supply of other inputs in  draft power, capital (as defined above), and fertilizer. Its rate  Office, Department of Statistics, Ministry of Planning
agricultural production    of change is assumed to be equal to the rate of growth in the
value of total agricultural production (at constant prices)
6. Rate of change in        Nonagricultural prices are weighted averages of wholesale  Index of Wholesale Prices, Economic and Scientific
nonagricultural pricesc    prices of nonfood articles                                 Research Foundation
7. Rate of change in the    Assumed equal to 6 above                                   Same as 6 above
-4        price of capital services
and other inputs used in
agricultural productionc
8. Rate of change in        Assumed to be equal across expenditure groups for given    Net Domestic Production at Factor Cost by Industry of
nonagricultural incomec    period of time. Initial estimates of this exogenous shock were    Origin, Central Statistical Office, Department of
obtained from two independently computed indexes of        Statistics, Ministry of Planning
nonagricultural incomes. However, this variable was treated   Macroeconomic Aggregates and Population, Central
as endogenous in section 11 (see text and appendix A)      Statistical Office, Department of Statistics, Ministry of
Planning
9. Rate of change in the    Defined as the rate of change in the net available rice supply,    Availability of Food Grains in India, Directorate of
domestic availability of   that is, net imports of rice plus net releases from        Economics and Statistics, Ministry of Agriculture
rice caused by trade and   government-held stocks of rice
buffer stock operations
10. Rate of change in the    Same as 9 above, but for wheat                             Same as 9 above
domestic availability of
wheat caused by trade
and buffer stock
operations                                                                                           (Table         on the following page.)



Appendix Table 11. Continued
Variable or
exogenous shock                           Additional Description,                                      Data sourceb
11. Rate of change in the    Same as 9 above, but for coarse cereals                    Same as 9 above
domestic availability of
coarse cereals caused by
trade and buffer stock
operations
12. Rate of change in the    Same as 9 above, but for fertilizer                        Production, Imports, Distribution, and Consumption of
domestic availability of                                                              Fertilizers, Fertilizer Association of India
fertilizer caused by trade
and buffer stock
operations
13. Technical change in rice   For 1960-61 to 1980-81 assumed to be equal to 75 percent of  Index Numbers of Area, Production, and Yield of
production                 the rate of change in rice yield per hectare               Principal Crops in India-Cropwise, Directorate of
Economics and Statistics, Ministry of Agriculture
14. Technical change in      Same as 13 above, but for wheat                            Same as 13 above
011        wheat production
15. Technical change in      Same as 13 above, but for coarse cereals                   Same as 13 above
coarse cereal production
16. Technical change in other  Same as 13 above, but for other crops                    Same as 13 above
crops production
17. Rate of change in the    Percentage irrigated area: the ratio of gross area under   Gross Area under Irrigation by Crops, Directorate of
percentage of irrigated   irrigation to gross cropped area                            Economics and Statistics, Ministry of Agriculture
area                                                                                Area under Principal Crops in India, Directorate of
Economics and Statistics, Ministry of Agriculture
18. Rate of change in road   Road density is defined as the ratio of road length (in    Extra-Municipal Roads (Classified According to
density                    kilometers) to total geographic area (in kilometers2).     Surface) Including National Highways Maintained by
"Roads" refers to all surfaced and motorable unsurfaced    P. W D. and Local Bodies and Roads Constructed in
roads                                                      C.D. and N.E.S. Blocks, Ministry of Shipping and
Transport
a. Data values are three-year averages. They are indexed such that the average for the three years 1972-73 to 1974-75 is equal to 100.
b. Only table headings and the agency which reports them are listed here.
c. These variables are deflated by P. See text of appendix A for further explanation.



Quiz6n and Binswanger 147
REFEREN CES
Bapna, Shanti, Hans P. Binswanger, and Jaime B. Quiz6n. 1984. "Systems of Output
Supply and Factor Demand Equations for Semi-Arid Tropical India." Indian Journal of
Agricultural Economics 34, no. 2 (April -June).
Bardhan, Pranab. 1984. "Determinants of Supply and Demand for Latin in a Poor
Agrarian Economy: An Analysis of Household Survey Data in Rural West Bengal." In
Hans P. Binswanger and Mark R. Rosenzweig, eds., Contractural Arrangements,
Employment, and Wages in Rural Labor Markets in Asia. New Haven, Conn.: Yale
University Press.
Binswanger, Hans P., and Jaime B. Quiz6n. 1986. "Distributional Consequences of
Alternative Food Policies in India." In Per Pinshrup-Anderson, ed., Consumer-Ori-
ented Food Subsidies: Costs, Benefits, and Policy Options for Developing Countries.
Baltimore, Md.: Johns Hopkins University Press. 1986.
Binswanger, Hans P., Jaime B. Quiz6n, and Gurushri Swamy. 1984. "The Demand for
Food and Foodgrain Quality in India." Indian Economic Journal 31, no. 4 (April-
June).
Dhar, Sanjay. 1980. An Analysis of Internal Migration in India. Ph.D. diss., Yale Univer-
sity.
Evenson, Robert E. 1981. "Green Revolution in North Indian Agriculture: An Ex-Post
Assessment of Economic Effects." Economic Growth Center, Yale University. Pro-
cessed.
Evenson, Robert E., and Hans P. Binswanger. 1984. "Estimating Labor Demand Func-
tion for Indian Agriculture." In Hans P. Binswanger and Mark R. Rosenzweig, eds.,
Contractural Arrangements, Employment, and Wages in Rural Labor Markets in Asia.
New Haven, Conn.: Yale University Press.
Jose, A. V. 1974. "Trends in Real Wage Rates of Agricultural Laborers." Economic and
Political Weekly 9.
Krishna, Raj. 1976. Rural Unemployment: A Survey of Concepts and Estimates for
India. World Bank Staff Working Paper 234. Washington, D.C.
Krishna, Raj, and G. S. Raychandhuri. 1980. Trends in Rural Savings and Private
Capital Formation in India. World Bank Staff Working Paper 382. Washington, D.C.
Pal, Ranjan, and Jaime B. Quiz6n. 1983. "Factor Cost, Income, and Supply Shares in
Indian Agriculture." World Bank Agriculture and Rural Development Department Re-
search Unit Discussion Paper 15. December. Available from the author upon request.
Quiz6n, Jaime B., and Hans P. Binswanger. 1983. "Income Distribution in Agriculture:
A Unified Approach." American Journal of Agricultural Economics 60, no. 3 (Au-
gust).
Quiz6n,Jaime B., and Hans P. Binswanger. 1984. "Factor Gains and Losses in the Indian
Semi-Arid Tropics: A Didactic Approach to Modelling the Agricultural Sector." World
Bank Agriculture and Rural Development Department Research Unit Discussion Paper
9 (revised). May. Available from the author upon request.
Rosenzweig, Mark R. 1980. "Neoclassical Theory and the Optimizing Peasant: An
Econometric Analysis of Market Family Labor Supply in a Developing Country."
Quarterly Journal of Economics 95, no. 1 (February).



148 THE WORLD BANK ECONOMIC REVIEW, VOL. 1, NO. 1
1984. "Determinants of Wage Rates and Labor Supply Behavior in the Rural
Sector of Developing Country." In Hans P. Binswanger and Mark Rosenzweig, eds.,
Contractural Arrangements, Employment, and Wages in Rural Labor Markets in Asia.
New Haven, Conn.: Yale University Press.
Taylor, Lance. 1979. Macro Models for Developing Countries. New York: McGraw-
Hill.