ArticlePDF Available

Barriers to Farm Profitability in India: Mechanization, Scale and Credit Markets

Authors:
Preliminary.
Barriers to Farm Profitability in India: Mechanization, Scale and Credit Markets
Andrew D. Foster and Mark R. Rosenzweig
September 2010
1
Although the generalization has many important caveats, across the world the most
efficient and productive agriculture is situated in countries in which farms are family-owned,
large-scale and mechanized. However, comparisons of farming productivity across countries of
the world cannot easily identify the essential barriers to augmenting farming productivity, as
countries differ in their property rights regimes, financial systems, labor markets, agroclimatic
conditions and other institutional and environmental features. A vast literature has highlighted,
usually one at a time, various market imperfections as constraining agricultural productivity in
poor countries. These include, for example, credit market barriers, lack of insurance, problems of
worker effort, and labor market transaction costs. However, many of these market problems are
not confined to poor countries. Moral hazard and adverse selection afflict credit markets in all
settings, and farmers do not have unlimited access to capital anywhere in the world. Nor do
family farms in many developed countries use employment schemes that differ importantly from
those used in those low -income settings where family farms also dominate. And most farmers in
high-income countries do not participate in formal crop, income or weather insurance markets. It
is thus unlikely that labor market problems or lack of insurance or even credit constraints can
alone account for the large differences in the efficiency of farms across developed and
developing countries.
In contrast to agriculture in most developed countries where farming is very efficient,
farming in India, while family-run, is neither large -scale nor mechanized. Figure1 provides the
cumulative distribution of land ownership, based on the Indian Census of 2001. Farming in India
is very small scale - 80% of farms are less than two acres in size and 95% are less than five acres
in terms of owned holdings. Mechanization can be examined using data from a new panel survey
of almost 5,000 crop-producing farmers in 242 villages in 17 of the major states of India, which
we describe and employ extensively below. Figure 2 plots the fraction of farms in the survey data
with a tractor, a mechanized plow or a thresher by land ownership size. As can be seen, less than
five percent of farms below two acres own any of these mechanized implements, but
mechanization increases significantly with ownership holdings, with 30% of farms above 10
acres owning a tractor and over 20% owning a mechanized thresher.
Are farms in India too small and under-mechanized? Our survey data also appear to
This presumption also presupposes that family members do not require supervision to
1
work hard. We show that both hired and family labor require the same amount of supervision
time.
2
indicate that small farms in India are substantially less efficient than larger farms. We use as our
measure of efficiency profits per acre, which reflects the resource costs of farming, inclusive of
the value of family labor, supervisory labor, and own equipment use. By this measure, which
does not take into account the likelihood that larger farmers have lower credit costs,
landownership and farm productivity are strongly positively associated. Figure 3 provides a
lowess-smoothed plot of per-acre profits and landownership from the survey data, net and gross
of labor supervision costs. As can be seen, up to about12 acres, per-acre farm profits increase
with land ownership size. The difference in the two profit measures is labor supervision costs.
The plots thus indicate not only that per-acre profits rise but that per-acre supervision costs fall
with owned acreage. Figure 4 shows that not only do per-acre supervision costs decline with
owned farm size, but above 12 acres, total supervision costs decline. These patterns appears to go
against the conventional idea that small farms, which employ mostly family labor, have a cost
advantage over larger farms who employ a higher fraction of hired labor. This presumption
overlooks how mechanization, which evidently rises with farm size, reduces overall labor use.1
Indeed, the data indicate that total labor costs per acre monotonically fall with ownership
holdings, as seen in Figure 5, which plots total labor costs per acre by owned land size.
Of course, Figures 3 through 5 merely describe associations between scale, labor use and
profitability. It is possible that within India larger farms are located where land is higher quality,
where farmers are better-educated, where credit markets operate more effectively, or where
agricultural conditions generally are more favorable to agriculture. Moreover, land holdings are
endogenous, and may reflect differences in property rights regimes or the capability of farmers.
Many prior studies of scale effects and the role of market constraints on farm productivity have
attempted to correct for particular dimensions of heterogeneity. A major shortcoming of the
literature, however, is that there have not been credible methods of dealing with the endogeneity
of machinery and land ownership and most have examined only particular market constraints and
not the interactions between them in terms of their affects on farm productivity.
Our last round of data indicate that land records were computerized in 80% of the
2
villages.
3
In this paper, we examine theoretically and empirically whether farm scale and lack of
mechanization are important proximate and causal barriers to farm productivity and profitability,
with particular attention to both the problem of eliciting labor effort and the role of credit
markets in an environment with stochastic output. We look at this issue in India, where property
rights are reasonably well-established, and where there exists labor, credit and input rental
2
markets, inclusive of those for land and mechanized inputs. We also examine the barriers to
mechanization. In order to illuminate the role of returns to scale in mechanization and access to
credit in a relatively tractable structure we first develop a model in which there are constant
returns to scale in land, total work (work done using labor and/or equipment) and other variable
inputs, and allow the level and composition of work to be determined by the relative productivity
and costs of different sources of work. Farmers are freely able to rent in or rent out capital
equipment and face constant-price variable inputs; however, labor effort depends on the amount
of supervision. The model provides conditions under which larger farmers will be more efficient
and use more machinery while small farmers will use larger amounts of labor and be less
efficient.
We then augment the model in two ways. First, we incorporate credit market
imperfections, allowing for differences in access to credit by small and large farmers based on
their ownership of land. This augmented model provides predictions on how the returns to capital
and variable inputs and the extent of mechanization (owned equipment) vary with land
ownership. Second, we introduce risk and dynamics, incorporating productivity shocks, savings,
and the persistence of soil nutrients across seasons, to identify how both the ownership of
mechanized assets and land affect input returns and profitability in a risky environment.
The model departs from the traditional literature that focuses on scale issues and/or
mechanization in that it distinguishes between the capacity and quantity of mechanized
implements and builds in the realistic property that higher-capacity mechanized implements
require more physical space. For example, a row-crop cultivator that handles eight rows at a time
will be approximately four times as productive as a two-row cultivator but will need a greater
Note also that to the extent that there is a loss at the end of each row one will lose a
3
certain amount of space per row but will not lose space based on the number of rows. In this
case what would matter in terms of scale economies is the difference between capacity and the
length of a row, which in the case of a square plot would be the square root of area.
4
area at the end of each row to turn around. This approach to modeling the relationship between
3
the scale of operation, mechanization and profitability differs in particular from a more
conventional approach based on investment indivisibilities and, implicitly, fixed-capacity
machinery. For example in the presence of indivisibilities a farmer may wish to purchase a given
machine only if he plans to use that item often enough, thus creating a relationship between the
level of use and the productivity of machinery given area. But there are problems with this
alternative approach. At least in its most simple form, an indivisible machinery model does not
reconcile easily with evidence we present below that there are scale economies across plots for a
given farmer at a given time and that investment in machinery tends to rise with scale rather than
jump abruptly and then stay relatively stable. Moreover, machinery is indeed scalable (e.g.,
tractors have different horse power, row-cultivators vary in terms of number of rows). There is
also significant advantage in terms of tractability of working with a model in which scale
economies only arise through land area.
Besides providing a coherent framework for understanding the interactions among the
size of owned landholdings, ownership of mechanized inputs, credit and labor market
imperfections, and agricultural efficiency, the model also provides tests that enable identification
of the distinct roles of technical scale economies and credit barriers in shaping the relationship
between assets returns and per-acre efficiency by own land size. In the absence of a feasible way
of experimentally varying ownership holdings or farm scale, key to the empirical identification of
scale and credit market effects on profitability and mechanization are the ability to control for
unobserved differences across farmers in ability, preferences and in costs (e.g., interest costs and
shadow labor costs) as well as differences in land quality.
We have collected panel data at the plot (across seasons in the same crop year) and at the
farm level (over the span 1999-2008) on inputs, outputs and investments. Variation across plots
for the same farmer, as we show, can identify pure scale effects, net of measured plot
One shortcoming of our methods is that, because they impound all time-invariant
4
characteristics of farmers into a fixed effect, we cannot identify whether, for example, the low
level of schooling of farmers in India is also a barrier to mechanization and farm efficiency.
5
characteristics, because such an analysis controls for all farm-specific costs. Variation over-time
in the effects of lagged farm profits on contemporaneous profits for the same plot, by farm size,
identify the role of ownership holdings on the ability to attain overall efficiency in input use net
of all differences in plot and farm characteristics that do not time vary. To obtain causal effects of
landownership on profitability and investments, we exploit the fact in the nine-year period 1999-
2008, a fraction of households split and/or received inherited land because a parent died. We
follow an individual farmer before and after inheriting land and/or assets and use the inherited
assets as instruments to explain the change in landownership and capital equipment in an
instrumental-variables set-up.4
Our estimates support the existence of scale economies: for a given farmer, per-acre
profits and the use of capital equipment are higher on larger plots compared with smaller plots,
while per-acre use of labor on larger plots is lower. A farmer who experiences an exogenous
change in owned landholdings exhibits an increase in profits per acre and is more likely to invest
in capital equipment in villages where a bank is proximate. Moreover, profits per acre are higher
on a given plot if a given farmer experiences a favorable farm-level profit shock in the prior
period only for farmers with smaller overall landholdings. These latter results indicate that
ownership helps overcome credit constraints on both investment and variable input use.
Consistent with this and with the higher returns to land among larger landowners, we find that
the marginal returns to capital and to fertilizer decline with owned landholdings. Finally, we
show that in our data, consistent with the higher profitability of a larger scale of operation and
with the relaxed credit needs associated with greater owned landholdings, farmers with small
landholdings lease out to farmers with larger landholdings within a village. This reverse tenancy
does not overcome the adverse ownership distribution of land, as only nine percent of farmers
lease in land.
Our results indicate that lack of mechanization is a barrier to greater farm productivity in
India, and that as a consequence of credit market constraints and scale economies, most farms in
We provide empirical evidence consistent with this assumption in the empirical section
5
of this paper.
6
India are too small to exploit the productivity and cost-savings from mechanization. The flip side
of these findings is that there are too many farms and too many people engaged in agriculture.
This suggests that industrialization may not only augment economic growth but also raise
agricultural productivity to the extent that the exit of people from agriculture to industry is
accompanied by land consolidation, as those exiting sell their land. Any growth-augmenting
agenda that has as its aim the achievement of a more productive agricultural sector therefore
needs to focus on barriers to agricultural exit and land consolidation, especially given the
inherent, pervasive and so far intractable problems of credit and insurance markets.
2. Theory
A. Technical scale economies, cultivated land area and mechanization
In order to illuminate the role of returns to scale associated with mechanization in a
relatively tractable structure we develop a model in which there are constant returns to scale in
land and all variable inputs. For ease of exposition, we define the services provided by labor
and/or equipment as work to be done. The model is set up in such a way that scale in terms of
land size affects the relative productivity of different sources of work but, given area, there are
constant returns in terms of the amount of work done. In particular, for a farmer with given scale
of production measured in acres a let output y per acre in a given crop cycle be
(1)
where e denotes work, and f denotes a variable input such as fertilizer. We assume that (I)
manual labor and machinery services are imperfect substitutes in producing work, (ii) that
manual labor, regardless of the hired or family status of that labor, must be supervised in order
5
to be productive, and (iii) that machinery varies by capacity. These assumptions are embodied in
the following function:
(2)
For an interior solution we require and .
6
We consider the own-versus buy decision once we introduce a credit market below.
7
7
where denotes supervisory labor, denotes manual labor, is a constant-returns
labor-services production function, q denotes the capacity of each machine, and k denotes the
number of machines. Note that the advantage of large farms with respect to higher-capacity
equipment is embodied in the function - increasing the capacity of a machine augments
productivity more the larger is a.
We assume that higher-capacity machines are more costly but that machinery cost does
not rise linearly with capacity. In particular, the price per day of a machine with capacity q is
, where í <1. We also assume that there is a perfect rental market for machines. The cost
67
of production is
(3) ,
fm
where k=the number of machines, p =price of fertilizer, w=wage rate of manual labor, and
s
w=wage of supervisory labor. A profit-maximizing farmer maximizes the value of (1) minus
costs (3) subject to (2) .
In solving this problem and to highlight the particular role that land-size plays in this
structure it is helpful to consider first the cost function
(4) .
Solving (4) first in terms of capacity yields an expression for optimal machine capacity
(5) .
Expression (5) indicates that optimal capacity is determined only by area and the parameter and,
in particular, is not sensitive to the required total work. Larger operations will use higher-capacity
equipment, but an increase in the elasticity of the machinery price with respect to capacity ( )
Note that substituting back into the (2) yields a work production function that is
8
analogous to the CES production function with the exception that the share parameter
, where , depends on area.
8
lowers machinery capacity particularly for large farmers.8
The first-order conditions to the cost minimization problem imply that the ratio of
supervisory to manual labor is constant given prices and technologies and that the ratio of
machinery to labor services is constant given area, prices, and technologies. Because of this
proportionality, we can readily distinguish between how scale affects the demand for inputs
conditional on the amount of work and on how scale affects total input demand by increasing
work. We may write the solution to the cost minimization problem as
(6)
and the conditional factor demands as, for example,
(7)
Implicit differentiation yields
(8) ,
which is positive for sufficiently close to one. That is, for a given work level e, when machinery
is sufficiently substitutable for labor the number of machines, of increasing capacity, increase as
scale increases. The ambiguity in terms of quantity of machinery arises for lower even when
machinery and labor are substitutes because higher-capacity machinery can produce more work in
less time.
Supervisory, manual labor, and the shadow price of work, for a given level of work, all
decline with land area because an increasing share of work is supplied by machinery
(9)
9
(10)
(11) ,
where .
To actually determine how much work is done and the total use of labor and machinery we
now embed the cost-minimization problem in a profit-maximizing one. In particular, let
(12)
or, letting the superscript * denote per-acre quantities:
(13)
The envelope condition implies
(14) .
Profits per acre increase with area, because the cost of work per unit area decreases. Larger
operations are more profitable on a per-acre basis. Similarly, larger operational holdings will use
inputs more intensively, as per-acre work increases in unit area
(15)
and fertilizer per unit area increases in area
(16)
if fertilizer and work are complementary, where is the own price elasticity of demand for
10
work and the fertilizer-work cross-price elasticity. If fertilizer and work are substitutes, the
fact that costs of work decline with area will result in substitution away from fertilizer.
The number of machines k per unit area will be increasing in area, for sufficiently close
to 1, because (I) there will be an overall expansion of work (15) and (ii) k is increasing in total
work. In particular,
(17) .
Whether total expenditures on machinery will rise for as land size increases depends on
whether the pricing of machinery is sufficiently elastic to capacity. Regardless of whether the
number of machines used per unit area increases or decreases, whether a farmer owns a machine
of a given capacity or greater is rising in area as indicated by (5).
Will larger operations use less labor per unit area? The effect of an expansion in area on
the amount of manual labor used per acre is ambiguous. There is an increase in work intensity as
the increasing returns associated with machinery lower unit work costs, but there is also a
decrease in the amount of labor per unit work, as shown in (10). If the demand for work is price
inelastic and/or labor and machines are sufficiently good substitutes, however, both manual and
supervisory labor must decline,
(18) .
The expression for supervisory labor is the same except that the subscript m is replaced with an s.
B. Scale effects and credit market imperfections
In the preceding analysis a was any contiguous plot of land used for an agricultural
operation. We have thus ignored the distinctions between the ownership or rental of land, as well
as equipment, and we have also assumed that over the agricultural cycle farmers can freely borrow
In principle, a similar argument may be made for family labor. A farmer with less area
9
for a given family labor may have lower need to finance hired labor inputs given area and thus
borrow less and face a lower interest cost per unit area. Profit estimates that did not remove
variation in borrowing cost might underestimate his relative profitability. The limitation of this
argument is that family labor and dependents of those family workers must be fed throughout the
agricultural cycle, which reduces the liquidity benefits of having a large endowment of family
labor per unit of area farmed. We do not formally model consumption and family size here
except to note that (a) with food shares at 60-80% it is unlikely that the liquidity effects of family
labor will be substantial and (b) loans to smaller farmers may be otherwise more expensive due
to collateral requirements and/or the relatively high transaction costs per rupee loaned from the
perspective of the lender.
11
against harvest revenues at a zero rate of interest. We now allow for the possibility of credit
constraints. In doing so, we assume that farmers own their plots of land and also own capital
equipment. We first take ownership of both assets as given, and then endogenize the ownership of
equipment. To incorporate capital market considerations we assume that farmers borrow per
acre to finance agricultural inputs and repay this debt with interest during the harvest period. We
assume that the interest rate r on this debt is dependent on the amount borrowed per acre as well as
on total owned land area, with farmers who own a small amount of land a obtaining working
capital at a higher interest rate than larger farmers. Formally, the per-acre amount that must be
repaid in the harvest period is given by
(19) ,
where the interest rate r is increasing in b* and decreasing in owned land. The decrease in interest
rates with land ownership might reflect the use of collateral, a requirement of most bank loans in
rural India (Munshi and Rosenzweig, 2009). In this extended model, ownership of both land and
machinery matters. By assumption owned landholdings reduce the cost of capital. But, while we
retain the assumption that there is a perfect rental market for machinery, ownership (versus rental)
of capital assets such as machinery also influences production decisions through its effect on the
amount of debt that must be incurred to finance inputs. In short, if one owns a productive asset one
does not have to finance the relevant rental cost. Or equivalently one can rent the machine to other
9
farmers and then use the cash to finance other inputs. Thus letting o* denote the rental value of
owned assets
12
(20) .
The farmer’s maximization problem with credit market imperfections can thus be restated as
(21)
0
where r is the rate of return on savings and is assumed to be less than r(a,b*) for all positive levels
of borrowing.
Profit maximization then implies that
(22) ,
where . Thus profits per acre rise with the size of owned
landholdings. The existence of credit market imperfections, as modeled here, steepens the gradient
of per-acre profits with respect to owned area relative to cultivated area, for given (or zero) credit
costs as in (14). This is for two reasons. First, there is a negative effect of owned area on interest
rates given input use per acre, . Second, any savings in cost per unit of work associated
with scale lower the amount borrowed, thus further lowering interest costs and raising profitability.
In addition to affecting the input choices of farmers, the presence of credit market
imperfections creates an empirical problem in measuring true profitability because of the difficulty
of accounting for differences in interest rates and thus the true discounted costs of inputs across
households in informal credit market settings. Expression (22) is relevant to the question of
whether land consolidation will improve true profitability per acre. We now consider the empirical
question of whether it is possible to infer correctly the role of credit market constraints in the
relationship between owned landholdings and true per-acre profitability when borrowing costs are
ignored in computing farm profits. We thus consider the comparative statics associated with
estimated profits, which exclude interest costs. The profit function in terms of estimated profits is
given by
(23) ,
These conditions coincide in the case in which the interest rate is proportional to
10
borrowing per acre.
13
where the inputs are determined by programming problem (21). In this case we have
(24)
where and the second term in parentheses is positive. Estimated profits also increase with
owned landholdings. Comparing (24) to (14) indicates that the gradient in estimated profits, as
with that of true profits, is steeper than would be the case in the absence of credit market effects. In
the case in which there are no technical scale economies associated with machinery so ,
(14) would be zero but (22) would be positive if and (24) would be positive if .10
A direct test of credit market constraints can be obtained by examining the returns to owned
capital assets using true or estimated profits. The marginal return to capital in terms of true profits
is given by
(25) ,
while the marginal return to estimated profits is
(26) .
The observed marginal returns to capital assets in the presence of credit constraints evidently differ
depending on how profits are computed. However, it is easily established that when ,
that is when borrowing costs are independent of land ownership and equal to the returns on
savings, the marginal return to capital assets is zero for either measure of profits. This is because
variation in owned machinery at the margin has no effects on the use of production inputs,
(27) .
14
Therefore, the finding that there is a non-zero return, in terms of estimated profits, to owned capital
assets would reject the hypothesis of perfect capital markets. The finding, moreover, that empirical
profits rises less steeply with landholdings when credit costs are held constant than when they are
not, (24) compared with (14), would establish further that the lower per-acre profitability of
smaller compared with larger landowners is due to disadvantages in the credit market, as depicted
in (19).
Thus far we have taken the amount of owned capital assets as given. In practice, farmers
both own and rent machinery, and the model incorporating credit constraints can explain variation
in equipment ownership even in the presence of a perfect rental market. By the assumption of an
effective rental market all farmers face the same equipment rental price. But due to credit market
imperfections farmer with different landholdings face different borrowing costs. Given that the
rental-equivalent price of owning machinery for one agricultural season depends on one’s own cost
of borrowing, individuals with relatively low borrowing cost will be more likely to own machinery
and those with higher borrowing cost will rent it. This suggests that if, as in (19), financial
intermediaries differentially lower the cost of borrowing for larger versus smaller landowners, then
given an active rental market, larger farmers will be more likely than small farmers to purchase
rather than rent machinery following the entry of such intermediaries.
C. Farm size and profit dynamics
In the preceding section we assumed that the amount a farmer borrowed reflected only his
demand for inputs and his ownership of equipment, ignoring own savings as a source of liquid
capital. In this section we consider the role of landholdings in determining profitability in a
dynamic setting in which profits are stochastic and liquid capital, or cash on hand, affects input
allocations when credit market imperfections are in place. In this setting, if there are credit
restrictions a farmer who has particularly high profits in one period may be able to finance more
inputs and thus accrue greater profits in a subsequent period. If he has access to large amounts of
capital at market rates no such effects should be observed. However, there are other reasons why
there may be a correlation in profits across time for a given farmer. For example, it is well-known
15
that fertilizer use increases nutrient levels in the soil that persist over time. This persistence will
influence fertilizer use and thus profitability in a subsequent period. Because past fertilizer use will
augment past profitability, one might observe a negative correlation between past profits and
current fertilizer use. Inattention to dynamic nutrient effects might lead to the false conclusion that
credit constraints are unimportant even if credit imperfections were present.
Addressing these dynamics in a forward-looking model is complicated and thus we
illustrate the basic structure using a simplified production function with one variable input,
fertilizer, and assume that the production function and the cost of borrowing are quadratic in their
respective arguments. In this model, farmers adjust their end-of-season savings based on
unanticipated income shocks and subsequently use this savings to finance fertilizer purchases. We
assume a stationary problem with state variables representing soil nutrition n* and cash on hand
t
h*. Fertilizer levels are chosen prior to the realization of a shock è. We define the value function
recursively as follows:
(28) ,
where is the discount factor and
(29) ,
where denotes the extent to which unanticipated shocks are saved. For unanticipated
shocks are fully saved as in the permanent income hypothesis and for cash on hand is just a
constant. Soil nutrients depend positively on both the previous period’s stock of nutrients and
fertilizer use and negatively on the output shock ,
(30)
The idea is that more rapidly growing plants, for example, will deplete the soil of nutrients
t
relatively quickly. For example, if è is rainfall, more nutrients are used if rainfall and soil nutrients
are complements. The production and credit functions are
(31)
16
and
(32) ,
where x are the respective arguments in (28) and for notational simplicity we set the fertilizer price
to one. All of the parameters in (31) and (32) are positive; that is we assume that production is
characterized by diminishing returns but the cost of credit r increases at a higher rate with the
amount of credit.
Estimated profits in this model (again, profits that do not account for borrowing costs are):
(33)
Farmers optimally choose their level of savings and use of fertilizer. Given the soil dynamics and
savings behavior, the effect of a previous period shock on next-period’s profits is thus
(34) ,
where is the second derivative from the value function, with and for an
interior maximum.
The two key parameters in (34) are and ë, reflecting the influence of the dynamic nutrient
and savings functions. If so that liquidity h does not depend on unanticipated income shocks
the lagged profit shock only influences profits in the next period because of nutrient depletion. A
positive shock in period 0 in that case leads to greater nutrient depletion and therefore reduces
profitability in period 1. Conversely, if there is no nutrient carryover so that there is only a
liquidity effect: a positive shock in period 0 induces higher savings and thus more cash on hand in
the next period so that less credit is needed for fertilizer. The lower cost of borrowing increases
2
fertilizer use and thus increase profitability in the current period. This effect vanishes if r=0, that
is, if borrowing costs do not rise as the demand for credit increases.
The model thus implies that the finding of a positive lagged profit shock effect on
(estimated) profits is indicative of the presence of liquidity effects. However, it also suggests that
the liquidity effect may be obscured even in the presence of credit market failures due to soil
nutrient dynamics. We show below that th nutrient depletion and credit-market effects can be
17
separately identified using plot-specific information over time for farmers with multiple plots. The
idea is that the nutrient effect only operates for a given plot but that the liquidity effect arises from
an aggregate farm-level shock.
3. Data
Our empirical investigation of the relationships among scale, credit markets, labor use, and
profitability uses four types of data from two surveys that form a panel. The main data sets are the
2007-8 Rural Economic Development Survey (REDS 2007-8) and the 1999 REDS both carried out
by the National Council of Applied Economic Research (NCAER). The surveys were administered
in 17 of the major states of India, with Assam and Jammu and Kasmir the only major states
excluded. The two surveys are the fifth and sixth rounds of a panel survey begun in the 1968-69
crop year. The original sample frame was meant to be representative of the entire rural population
of India at that time. To obtain nationally-representative statistic from the first round data,
sampling weights must be used because a stratified sampling procedure was employed to draw the
sample. This included the oversampling of high-income households within villages and selecting
districts in areas particularly suitable for green revolution crops. By the sixth round, 40 years later,
given household splits, the creation of new towns and villages, and out-migration, the original
sampling weights no longer enable the creation of nationally-representative statistics from the
later-round data. The oversampling of high-income households, however, is an advantage for this
study, given our focus on the relationships among scale, productivity and mechanization, because
there is more variation in own landholdings at the upper tail where mechanization is prevalent.
The 2007-8 survey includes a listing, carried out in 2006, of all of the households in each of
the original 142 villages in the panel survey. Appendix Figure A provides the distribution of own
landholdings in the set of sampled villages in comparison to that from the Census of 2001 reported
in Figure 1. The figure shows that landholding distribution in the sample villages is skewed to the
right relative to the national figures. This is not due to the oversampling of high-income
households, but reflects the geographical sampling. The listing data, which included almost
120,000 households, will be used in the final section to examine land leasing patterns within
18
villages.
The survey of households in the 2007-8 REDS, took place over the period 2007-2009, and
includes 4,961 crop cultivators who own land. The sample of farmers include all farmers who were
members of households interviewed in the 1999 round of the survey plus an additional random
sample of households. The panel households include both household heads who were heads in
1999 and new heads who split from the 1999 households. There are 2,848 panel households for
whom there is information from both the 1999 and 2007-8 survey rounds. The 2007-8 survey is
unique among the surveys in the NCAER long-term panel in that information on all inputs and
outputs associated with farm production was collected at the plot level for each of the three seasons
in the crop year prior to the survey interviews. There is input-output information for 10,947 plots,
with about two-thirds of the plots observed at least twice (two seasons or more). The plot/season
data enable us to carry out the analysis across plots in a given season, thus controlling for all
characteristics of the farmer as well as all input and output prices. Cross-plot, within-farmer
estimates can be biased if plots vary by characteristics that affect productivity. The survey includes
seven plot characteristics. These include depth, salinity, percolation, drainage, color (red, black,
grey, yellow, brown, off-white), type (gravel, sandy, loam, clay, and hard clay) and distance from
the farmer homestead. The multiple season information by plot enables us to obtain estimates that
control perfectly for plot characteristics as well as time-invariant farmer characteristics, as
discussed below. Thus, in addition to the panel of households over the 1999-2007-9 interval, there
is within-crop year panel data of up to three periods on farm plots.
The detail on inputs, outputs and costs enables the computation of farm profits at the plot
and farm level for the 2007-8 survey round and at the household level for the 1999 round.
Information is provided on the use of family, hired and supervisory labor by operation and by age
and gender, along with own use of implements by type and the rental of implements, by type. Other
inputs include pesticides, fertilizer (by type), and water. We subtract out the total costs of all of
these inputs from the value of output using farm gate prices. Thus, our profit measure corresponds
to ‘empirical’ profits in the model as it does not include interest costs associated with using credit
to pay for inputs. Maintenance costs for own equipment is subtracted from gross income, but not
maintenance costs (meals, clothing, shelter) for family labor.
19
The 2007-8 survey also includes retrospective information for each household head on
investments in land and equipment, by type, since 1999. This includes information on land and
equipment that is sold, purchased, destroyed, transferred or inherited. This information will be used
to estimate the determinants of farm mechanization. The acquisition of land is primarily via
inheritance that results from family division - less than 3 percent of farmers bought or sold land
over the entire nine-year period. Division most often occurs when a head dies and the adult sons
then farm their inherited land. Division sometimes occurs prior to the death of a head, which may
result from disputes among family members (Foster and Rosenzweig, 2003). The time variation in
the state variables owned landholdings and equipment thus principally stems from household
splits.
Two key assumptions of the model are (I) that the rental of land does not overcome the
limitations of scale associated with owned plots and (ii) that family labor does not have a cost
advantage over hired labor in terms of the need for supervision. With respect to the first
assumption, the 2007-8 data indicate that only 4.6 percent of cultivated plots, over the three
seasons, are rented (4.9 percent of area). Moreover, the data indicate that in all states of India,
except West Bengal, land is primarily leased from immediate family members (parents and
siblings). This is not unexpected, given the presumed efficiency of cultivating contiguous land area
along with possible moral hazard issues that might arise in terms of farm maintenance. Figure 6
provides the fraction of land leased from immediate relatives, others and ‘landlord.’ Outside of
West Bengal, 72% of land is rented from family members, and 28% from others. Only a negligible
fraction of households report renting land from a landlord. In contrast, in West Bengal, 26% of
farmers rent from landlords, and only 7% from family. These data suggest that, with perhaps the
exception of West Bengal, reform of tenancy laws may not play a major role in overcoming the
disadvantages of small farms.
A key feature of the 2007-8 data, as noted, is that it includes information on supervision
time at the plot level. We can thus directly test an assumption of the model that all manual labor,
whether family or hired, requires supervision. We estimate a supervision cost equation across plots
for the same farmer in a given season:
We could include family size in the specification, but family size may be endogenous;
11
on farms where supervision is particularly advantageous, or hiring supervisiory labor is difficult,
more family members may be in place.
20
sijt 0jt ffijt hhijt Aijt x ijt
ij
(35) l =a + a l + a l + a A + Xa+ e,
sfh
where for plot I of farmer j in season t l=supervisory labor costs, l=family labor costs, l=hired
0j ij
labor costs, A=plot area, a = farmer/season fixed effect, X is a vector of plot characteristics, and e
is an error term, where costs are simply days times the relevant wage per day. Our assumption is
fh
that a = a > 0, that an increase in hired or family labor equally increases supervision time. Note
that thefarmer/season fixed effect picks up all market prices and all farm-level characteristics in a
given season.
f
It is important to control for farmer characteristics to obtain an unbiased estimated of a
h
and a . Supervision is typically carried out by family members, presumably because family
members benefit directly from profitability. This is one of the advantages of family farming.
Supervision time thus may depend on family size if supervision is carried out less efficiently using
hired labor. Farm households that have a greater number of family members thus may both use
family labor more intensively and spend more time supervising. This would create a positive bias
in the coefficient on family labor coefficient in (35). The number of family members is
11
impounded in the farmer/season fixed effect; the a-coefficients estimated with the farmer/season
fixed effect included thus pick up how supervision costs vary across plots according to the
allocation family and hired labor, for given family size.
Table 1 reports the estimates of (35). In the first column, only village and season dummy
variables are included - there is no control for farm characteristics, including family size. The first-
column estimates indicate that an increase in non-supervisory family labor use increases
fh
supervsion costs more than an increase in hired labor, a > a . This result is robust to the inclusion
of the set of detailed plot characteristics. When we control for farmer and season and thus family
size, however, the coefficient estimates for family and hired labor are essentially identical - we
cannot reject the hypothesis that increasing family or hired labor use increases supervsion costs
equally. This result is also robust to the inclusion of plot characteristics. The difference in the
21
coefficient estimates across columns 2 and 4 do suggest that larger families may have an advantage
in supervision, for given scale, but not because family manual laborers require less supervision
than do hired laborers. Instead, as the model suggests and as the descriptive statistics confirm,
mechanization also reduces the need for manual labor and thus supervision costs.
4. Identifying Scale Effects
As indicated in the model, larger landholdings potentially increase profitability by allowing
the use of a higher-capacity (or any) mechanized inputs and also by lowering credit costs. In this
section, we identify the effects of scale, net of credit cost effects, by estimating how variation in the
size of plots for a given farmer affects plot-specific per-acre profitabilty, the likelihood of tractor
use and per-acre labor use. By using farmer fixed effects we are holding constant owned
landholdings, access to credit (and family size) so that variation in area reflects only scale effects.
We also examine the role of fragmentation. We estimate the equation
ijt 0j Aijt -A -ijt N ijt x ijt
ij
(36) ð = b + b A+ bA+ bN + Xa+ u,
ijt 0j ijt -
where ð=profits per acre on plot I for farm j, b =farmer fixed effect, A=plot area (acres), A
ijt ijt ijt
=total area of all other plots, N=total number of cultivated plots, and u is an iid error. The
equation also includes season/state fixed effects to control for input and output prices. The
A
interpretation of the coefficient on plot area b is straightforward, it is the effect of scale on profits.
-ijt ijt
For given total size of the other plots A, an increase in the number of plots N is interpreted as a
decrease in the average size of the other plots. If other plots are smaller, use of mechanized inputs
is less likely so that more resources may be allocated to the larger plot because inputs will have a
N
higher return. The coefficient b would then be positive. An increase in the total size of all plots
might make the rental or ownership of higher-capacity equipment more profitable for the farm,
-A
thus also increasing profits on all plots (b >0).
The first column of Table 2 reports the estimates of equation (36) without the inclusion of
the seven plot characteristics. The second column reports estimates with the plot characteristics
22
included. In both specifications, the estimates are consistent with the operation of scale economies
- the larger the size of the plot, given the farmer’s ownership holdings, capabilities, preferences,
and family size, the higher are profits per acre. And, if other plots are on average smaller or total
cultivated area on the farm is greater, the plot is also more profitable.
Are these profit estimates consistent with scale effects associated with mechanization? The
third and fourth columns reports estimates of equation (37) with per-acre profits replaced by a
dummy variable taking on the value of one if a tractor is used on the plot. The estimates, with and
without plot characteristics included, indicate that, consistent with the theory, a tractor is more
likely to be used on a larger plot and if the total amount of cultivated area is larger (for given
owned area), but if the farm has smaller plots on average, a tractor is less likely to be used. In
columns five and six we see that total labor costs per acre mirror the effects of scale on plot-
specific tractor use - larger plots use less labor per acre and less labor is used per-acre, given plot
size, the larger is the total cultivated area of the farm. But, the smaller are farm plots overall, the
higher are per-acre labor costs on any plot.
5. Owned Landholdings, Efficient Input Use and Profitability
In the preceding analysis we examined at how profitability varied across plots for fixed land
ownership. To explore the overall effects of total farm size on input efficiency we exploit the plot
level data to estimate the marginal returns to a variable input by farm size. Profit-maximization
implies that the marginal return to an additional rupee spent on a variable input should be zero. We
estimate the marginal returns to fertilizer expenditures based on variation across plots in fertilizer
use for a given farmer, stratifying the sample by the size of owned landholdings. If farmers with
small landholdings face higher borrowing costs and are unable to finance the efficient use of
fertilizer, we should find that the marginal returns to fertilizer expenditure are positive for small
farmers but decline as farm size increases. The equation we estimate includes a farmer/season
fixed effect so that only plot-specific characteristics enter the specification. These again include
plot size and soil characteristics. The fixed effect thus absorbs farmer characteristics and input
prices faced by the farm:
Recall that our profit measure does not include credit cost. If costs of credit are high
12
then we are in effect underpricing fertilizer in the computation of profits.
23
ijt 0jt A ijt f ijt x ijt
ij
(37) ð = c + c A+ cf + Xa+ ò,
ijt ijt
where f=plot-specific fertilizer expenditures per acre and ò is an iid error. Profit-maximization
f
implies that c=0.
Table 3 presents estimates of (37) for farmers who own less than four acres of land, farmers
with landholdings above four and less than 10 and for farmers who own more than 10 acres. Again
estimates are shown with and without the plot characteristics. We can reject the hypothesis that
farmers below 10 acres are using sufficient fertilizer; the coefficient on fertilizer is statistically
significantly different than zero for both the <4 and 4-10 acre farmers. Indeed, the estimates imply
that an additional rupee of expenditure on fertilizer yields more than a rupee of profit. For farmers
12
owning 10 or more acres of land, however, the marginal return is effectively zero on average; such
farmers are evidently unconstrained in fertilizer use. The partition of farmers into three groups is
somewhat arbitrary. Figure 7 presents smoothed local-area estimates of the effects of fertilizer on
profits by land ownership from (), along with one-standard deviation bands, for farms up to 50
acres in size. The pattern of estimates indicate that the marginal returns to fertilizer fall
monotonically as landholdings increase and fertilizer is underutilized, given the prices that farmers
face, for farms up to about 40 acres..
To estimate the direct effect of owned land and machinery on per-acre profitability we need
to allow for the possibility that landownership is correlated with unmeasured attributes of farmers.
We use the 1999-2007-8 panel data to estimate the causal impact of landownership on profits and
on investment at the farm level. Prior studies have exploited panel data to eliminate time-invariant
fixed farmer and land characteristics such as risk aversion or ability. However, this is not sufficient
to identify the effect of variation in a capital asset. The equation we seek to estimate is
jt 0t A jt k jt j ijt
(38) ð = d + d A+ dk + ì + å,
j
where t is survey year, k=value of all farm machinery, ì=unobservable household fixed effect, and
24
ijt
å=an iid error. Controlling for farm machinery (mechanization), we expect that the coefficient
Ak
d >0 if there are scale effects and also that machinery has a positive marginal return (d >0),
perhaps a higher return for small farms that are unable to finance capital equipment purchases. The
problem is that farmers who are unobservably (to the econometrician) profitable may be better able
to finance land purchases and equipment, leading to a spurious positive relationship between
landholdings, capital equipment and per-acre profits.
Taking differences in (38) across survey years to eliminate the farmer fixed effect, we get
jt 0 A j k j ijt
(39) Äð = Äd + dÄA+ dÄk + Äå,
ijt
where Ä is the intertemporal difference operator. In (39), even if the errors å are iid, investments
in capital assets such as land or equipment will be affected by prior profit shocks in a world in
which credit markets are imperfect. By differencing we thus introduce a negative bias in the land
ijt ijt
and equipment coefficients - positive profit shocks in the first period make ÄA high when Äå is
ijt jt
low. That is, even if the contemporaneous cov(å, A ) = 0, because assets are measured prior to the
ijt Äaj
profit shock, cov(Äå, ) 0. We show below that for most farms (small farms) in India there is
underinvestment in machinery and that past profit shocks affect current variable input use.
Ak
To obtain consistent estimates of d and d we employ an instrumental-variables strategy.
We take advantage of the fact that over the nine-year interval between surveys 19.9% of farms
divided and farmers inherited land. Moreover, for all heads of farm households in 1999, we know
how much of the land and equipment was inherited before the 1999 survey round. The instruments
we use to predict the change in landholdings of a farmer between 1999 and 2007-8 are thus the
value of owned mechanized and non-mechanized assets inherited prior to 1999 and the value of
assets and acreage of land inherited between 1999 and 2007-8. We also add variables that in our
prior study of household division in India (Foster and Rosenzweig, 2003) contributed to predicting
household splits. These include the age of the head in 1999, an indicator of whether the head in
2007-8 had brothers, and a measure of the educational inequality among the claimants to the head’s
land in 1999.
Table 4 contains the estimates of the first-stage equations predicting the change in
25
landholdings and the value of farm equipment between 1999 and 2007-8. The Anderson Rubin
Wald test of jointly weak instruments rejects the null at the .002 level of significance.
Indeed, post-1999 inheritance of land is a significant predictor of the change in landholdings over
the period along with the head’s age in 1999, while inherited assets obtained prior to 1999 and
inequality in claimants statistically and significantly affect the change in the stock of equipment.
We estimate a variants of (39) in which we omit capital equipment in order to estimate the
unconditional relationship between landownership size and profitability gross of mechanization.
The first two columns of Table 5 report estimates of the two variants of the per-acre profit function
(39) but only controlling for village and time effects, where the reported t-ratios are clustered at the
1999 farm level. This estimation procedure roughly, by village area, controls for land quality
heterogeneity and prices, but not individual farm heterogeneity. The estimates indicate that larger
farms are more profitable per acre, consistent with Figure 2, but capital equipment has little or no
return, conditional on farm size. The farmer fixed-effects estimates are reported in the third and
fourth columns of the table. These estimates suggest that there are no scale effects and that larger
farms are not profitable. However, as discussed, these estimates are biased negatively to the extent
that there are credit constraints on capital investments.
The last two columns of Table 5 report the FE-IV estimates that eliminate the bias in the
farmer fixed-effects estimates. These show that an exogenous increase in landholdings gross of
changes in capital equipment significantly increases per-acre profits. A large proportion of this
effect is evidently due to investments in equipment; controlling for farm equipment reduces the
effect of farm size by 36%. And, the marginal return on capital assets is positive and statistically
significant, at 3.5%. The Kelinberger-Paap and Hansen J diagnostic test statistics, reported in the
table, indicate that we cannot reject the null that the second-stage estimates for either specification
are not identified.
Does the data indicate that there is an optimal farm size? Or put differently, is there a farm
size at which additional increases in owned land no longer increase profits per acre? Figure 8
A
reports the locally-weighted FE-IV land coefficient d by land ownership size ranging from 0.1 to
20 acres. The continuous line depicts the estimated coefficient from the specification that excludes
capital equipment; the discontinuous line portrays the coefficient of land size conditional on owned
26
capital equipment. As can be seen, for the entire range of farm sizes, increases in land increase
profits per acre; the positive effects of size actually rise with land size for farms below 5 acres.
That is, for 95% of the farms in India, increasing farm size would raise profits per acre at an
increasing rate.
Figure 9 reports the locally-weighted FE-IV estimates of the marginal return to capital
k
equipment d , along with the associated one standard deviation bands, across the same range of
owned landholdings. As expected if credit costs decline with land size, the marginal returns to
capital decline with land ownership size - for farm sizes at around three acres, the return to capital
is between .04 and .10, while for farms of 10 acres, the return is between two and four percent.
Smaller farms substantially under-invest in capital equipment, and thus employ too much labor,
given our findings from the plot level data that labor and equipment are substitutes.
6. Farm Size and Equipment Investment and Rental
Figure 9 suggests that credit costs fall with landownership, given the underinvestment in
machinery characterizing small farms. In this section we estimate the effects of landholdings on
both equipment investment and rental. The model suggests that farms owning more land will
purchase more capital equipment to take advantage of scale economies and because they face lower
credit costs. For this analysis we use the retrospective information from the 2008-9 REDS that
provides a yearly history of land and capital equipment acquisition from 1999 up to the survey
interview date. In contrast to the panel data based on information from the 1999 and 2007-8 survey
rounds in which the household unit is defined by the households in 1999, 19% of whom split, the
unit is the household in 2007-8. There are two consequences. First, the sample is larger than the
1999-2007-8 panel, because the latest survey round includes a new random sample of households.
Second, if a farmer split from a household after 1999 his owned land and farm assets at the 1999
date is reported as zero if he was not formerly the household head. 25% of the sample farmers in
2007-8 experienced an increase in landholdings since 1999, of whom 79% inherited land due to
household division. Less than 1.2% of farmers were observed to experience a decline in owned
landholdings.
In principle the data can be used to examine the determinants of net land sales.
13
However, less than 2% of farmers sold or purchased land over the 9-year interval. In contrast,
18% of farmers invested in capital equipment.
27
We create a panel data set from the retrospective history by computing any new
investments made in farm machinery within the three-year period prior to the 2007-8 interview
data and within the three year period 1999-2001. We also compute the stock of equipment and
landholdings in 1999 and three years before the interview in the last round. Thus we create two
observations on capital investment, landholdings and equipment stock value for each farmer. We
13
also examine the determinants of equipment rental. Here we must use information on the value of
hired equipment services in 1999 and in 2007-8 from the 1999 and the 2007-8 surveys, so that the
sample size is reduced to the matched 1999-2007-8 panel.
Our model incorporates credit market imperfections as one of the factors that constrain
mechanization, with owned landholdings serving to mitigate credit costs. We thus add to the
household panel information on bank proximity. From the 1999 and 2007-8 village-level data
providing comprehensive information on village institutions and facilities, we created a dummy
variable indicating whether a commercial bank was within ten kilometers of the village in which
the farm household was located. 84% of farmers were within 10 kilometers of a bank in 1999; 84%
in 2007-8. However, banks were not stationary. 25% of the farmers experienced either the exit of a
bank or a newly-proximate bank. To assess whether landownership plays a role in lowering credit
costs, we interact landholdings and bank presence. The equipment purchase and hire equations we
estimate are thus of the form:
kjt 0t A jt k jt B jt BA jt jt j ijt
(40) K = e + e A+ ek + e B+ e A@B+ì + ç,
Ak
where K=equipment purchase or rental and B=bank proximity. We expect that e >0, e <0, and
BA
e >0; that is, where banks are present, large landowners are more able to finance equipment
purchases and/or rent equipment. To eliminate the influence of unobserved time-invariant farm and
j
farmer characteristics (ì), we again difference across the two periods and use instrumental
variables to eliminate the bias discussed in the previous section. Because a little over half of the
28
observations in the retrospective-based panel are from the newly-drawn sample of households in
2007-8, we do not use information on family circumstances in 1999 as instruments, which is only
available for the 199-2007-8 panel. We use as instruments for the change in owned landholdings,
the change in the value of farm equipment and the change in bank presence, the value of farm
assets inherited since 1999, the amount of land inherited since 1999 and bank proximity in 1999.
The estimates of (40) are presented in Table 6; the first-stage estimates are presented in
Table 7. Because here the second-stage estimates are exactly identified, we cannot use the standard
diagnostics tests of identification. However, inherited land is a statistically significant predictor of
the change in owned landholdings, inherited assets are statistically significant predictors of the
change in the value of the stock of machinery, and bank presence in 1999 is a statistically
significant predictor of subsequent bank location.
The first column of Table 6 reports fixed- effects estimates of the determinants of
machinery investment that do not use the instruments and which exclude the interaction term.
While the signs of the coefficients are as expected, the precision of the coefficient estimates is low
for both land and the equipment stock. When instruments are used, however, as reported in the
second column, both the capital equipment and land coefficients increase substantially and become
statistically significant. In particular, an increase in owned landholdings increases equipment
investment, given the existing stock of equipment, while for given landholdings, those farms that
already own equipment invest less. The effect of bank presence is not precisely estimated,
however, in the linear IV specification. When the interaction between bank proximity and
landholdings is added (column three), we see that evidently the advantage of bank proximity is
only captured by larger landholders - the interaction coefficient is positive and statistically
significant while the linear bank and land coefficients become statistically insignificant. These
results are consistent with land having value as collateral for obtaining bank loans to finance
equipment purchases.
The estimates in columns four through six in Table 6 for equipment rental parallel those for
equipment purchases, except that the interaction term is not statistically significant - the fixed-
effects estimates are negatively biased, but once this bias is eliminated using instrumental variables
large landowners are seen to rent more machinery than smaller landowners, for given owned stock.
29
But bank proximity is not a statistically significant determinant of equipment rental in either the
linear or interactive specification. Formal banks thus do not appear to play a major role in
financing variable inputs. These results thus indicate that larger landowners are more likely to use
and own farm machinery, and part of the reason is that they have better access to lower-cost bank
credit for investment.
7. Credit market imperfections, size, and the effects of profit variability
The previous section provided indirect evidence on the role of credit market imperfections
as a source of scale economies in rural agriculture. We used our model to show that a more direct
test is possible of the interaction between credit market imperfections and scale by examining the
sensitivity of profits to income shocks by land ownership. In this section we estimate how lagged
profit shocks affect per-acre profits, taking into account that such shocks not only affect farmer
liquidity but also soil nutrients. As noted in the theory section, assessment of the effects of past
shocks on current profitability is complicated by the fact that past crop shocks may also affect the
nutrient content of the soil, which, in turn, may also affect profitability. Removing farmer and/or
plot specific fixed effects from estimates of a profit equation may remove fixed aspects of soil
quality that affect profit but will not help if nutrient status is responding directly to past shocks.
This analysis makes use of the 2007-8 household data that includes information on three
consecutive planting seasons. We estimate first the relationship between lagged profits and lagged
fertilizer use and current profits at the farm level. We include village and season dummies to
capture variation in wages and prices, which also influence profitability, as well as farmer specific
effects that are constant across seasons. Because we incorporate lagged variables in the profit
equation and a farmer fixed-effect estimation is restricted to the second and third seasons. Table 8
presents the within-farm estimates of the effects of farm-level lagged profits and fertilizer use on
current profitability per acre, stratified by landholdings.
While there is significant evidence of positive scale economies in terms of profitability,
particularly among small landholding households, we see a negative and significant effect of
lagged farm profits on current profits in each of the three landholding groups. For the lowest
Note that estimation of fertilizer effect is complicated by the fact that the we cannot
14
control for nutrient status. But using the fact that fertilizer responds to nutrient status, we may
show that unconditional on soil nutrient status lagged fertilizer use will positively affect current
profits: .
30
landholding group a 1000 Rupee increase in farm profits per acre is associated with a 266 Rupee
decrease in a subsequent period. This negative coefficient is in principle attributable to two
sources, the first is that an increases in production in one period result in greater nutrient extraction
and thus lower productivity and/or higher fertilizer costs, both of which lower profitability in the
second period. The fact that the lagged fertilizer coefficient is positive in each case is also
supportive of this point–greater usage of fertilizer in the past increases current soil nutrition. Of
14
course, another potential source of this negative coefficient is the fact that the difference in profits
between period t and t+1 is negatively correlated with the differenced residual from the profit
equation as argued above. However, the lagged profit coefficient, as noted, also reflects credit
constraints. It is thus interesting that the lagged profit shocks get progressively more negative as
land size increases. This may reflect the fact that credit costs are lower for larger farmers.
To separate out credit effects from dynamic nutrient effects we make use again of the fact
that we have plot-level data for each farmer, which allows us to separate the effect of a crop shock
on liquidity from the effect of the shock on soil nutrients. Essentially one can augment the dynamic
model by letting cash on hand depend on the unanticipated deviations in the across-plot average
shock so that . We thus differentiate between lagged profits and
fertilizer use on a given plot and lagged profits on all other plots. The coefficient on the lagged
profits specific to a plot will capture the combined nutrient and (a small fraction of) liquidity
effects; the coefficient on the lagged profits from other plots will only reflect the liquidity effect.
To achieve identification, we use a subsample of farmers who cultivate at least two plots over three
seasons.
The results, reported in Table 9, are strongly consistent with the notion that liquidity shocks
importantly determine input use and thus affect profitability among small farmers. In particular,
conditioning on the lagged profits and fertilizer use on a given plot, a 1000 Rupee increase in
31
profits per acre on a farmer’s other plots leads to a substantial 140 Rupee increase in profits per
acre among farmers with less that four acres of land. In farmers with 4-10 acres of land the
corresponding figure is substantially less (62.8 Rupees) and in the largest farmers (10+ acres) the
estimate is even smaller (36.6 Rupees), with neither estimate differing significantly from zero. The
corresponding coefficients on the profits on the particular plot also decline, consistent with the idea
that the own profit effect combines both a technological effect (in this case nutrient depletion) that
is constant across landholding and a declining liquidity effect.
9. Lease markets
The preceding results suggest that there are substantial unrealized returns to profitability in
rural India that are a consequence of current small farm sizes. If this is indeed the case, then even
in a setting in which there are important barriers to wide-scale land consolidation, one should
expect to see transfers of land in the form of leasing and or sales that on net move from smaller to
larger farms. Smaller farmers should be more likely to sell land and reverse tenancy should be the
norm. As we have noted land sales are simply too scarce to characterize patterns. Moreover, our
data suggest that, just as for tenancy, land sales are within family - in our data 95% of land sales
are from parent to child.
Leasing is more prevalent, although some of the identified scale effects, particularly those
associated with credit markets, cannot be exploited through leasing. But, as also noted, leasing is
also quite rare with less than 3 percent of households reporting leasing in or leasing out in a given
year. Given that scale economies arise in part from contiguous land and that most leasing happens
within family, perhaps for reasons of moral hazard associated with land upkeep, the opportunities
for productive trade appear to be small. Nonetheless, the 2006 village listing data gives a large
enough sample size to look at the distribution of this relatively rare event across farms stratified by
ownership size to assess if Indian farmers seek to exploit scale economies.
The relationship between ownership holdings and the probability of leasing in and leasing
out in the 2006 listing data, gross and net of village fixed effects, are plotted in Figures 11 and 12.
The first figure, which does not account for village effects, yields a somewhat confusing picture.
32
One sees, in particular, that although leasing in and leasing out are both quite rare, the leasing in
probability is more than twice as high as the leasing out probability. The problem is that Figure 11
combines both differences in leasing behavior by land size within a village and differences across
villages in average land size that may be correlated with overall levels of leasing. Indeed, our
model suggests that in areas with relatively large plots on average there should less need for leasing
to capture unrealized scale economies and thus the rental market may be inactive. As a result one
might observe a decline in leasing in and/or leasing out with land size even though within a village,
of course, leasing in and leasing out must balance.
Figure 12, which removes village effects, provides a more consistent picture and strongly
supports the hypothesis that leasing goes in the direction of capturing scale economies. In
particular, relative to a household that is 5 acres below the village mean a farmer with 5 acres
above the village mean has .018 (over 50%) higher probability of leasing in and a .014 lower
probability of leasing out. Indian farmers appear to behave as if they also believe that increasing
operational scale is profitable.
10. Conclusion
In this paper we have used new panel data describing Indian farms to examine the question
of whether the size of landownership holdings matters for farm efficiency and profitability,
exploiting both panel information on profits at the plot level and the consequences of division of
landholdings due to households splits. Our empirical results indicate that larger farms are more
efficient, given the resource costs of farming. On farms with larger owned landholdings there is
more mechanization, less labor use per acre inclusive of supervisory labor, and, most importantly,
higher profitability per acre. Our findings suggest that the greater efficiency of larger farms is
partly a scale effect associated with the use of mechanized inputs but also is related to credit
market constraints. Larger landowners appear to have an advantage in the credit market. They face
lower credit costs due to superior collateral and are better protected from income shocks.
Consistent with this we find that larger farms use variable inputs more efficiently, are better
mechanized, and are more insulated against fluctuations in profits associated with weather in terms
33
of input efficiency.
These findings imply that farms in India are too small and under-mechanized.
Consolidation of landholdings would not only raise farm productivity but also lower labor use per
acre as farms adopted mechanized inputs. This in turn implies that industrialization that led to the
exit of many workers from the agricultural sector, if that were accompanied by land consolidation,
would result in a more efficient agricultural sector. The consequences of labor exit and land
consolidation on a large scale, given our estimates, would have to be examined in a general-
equilibrium context in order to appropriately assess these effects.
As with any set of findings that suggest that there are profitable opportunities from altering
the allocation of resources, the question is why farms in India remain small. This question is also
beyond the scope of this paper, but our findings suggest that a serious research program meant to
discover how to improve agricultural efficiency in India, and perhaps other countries where
property rights area reasonably in place, might be directed to examining the question of why there
are so many people in agriculture farming at a small scale and under-exploiting the efficiency of
mechanization in the presence of globally increasing returns to farm size.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0-.5 .5-1 1-2 2-3 3-4 4-5 5-7.5 7.5-10 10-20 20+
Figure 1. The Cumulative Distribution of Owned Landholdings (Acres) in India (2001 Census)
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
<1 12345678910+
Tractor Plow Thresher
Figure 2. Mechanization and Owned Landholdings (Acres), 2007-2008
2000
3000
4000
5000
6000
7000
8000
0 2 4 6 8 10 12 14 16 18 20
Profits per acre, no supervision costs
Profits per acre
Figure 3. Profits per Acre and Profits per Acre less Supervision Costs,
by Owned Landholdings, 2007-8
-0.2
199.8
399.8
599.8
799.8
999.8
1199.8
1399.8
1599.8
1799.8
0 2 4 6 8 10 12 14 16 18 20
Total supervision costs
Supervision costs per acre
Figure 4. Total Supervision Costs and Supervision Costs per Acre by Owned Landholdings Size
100
1100
2100
3100
4100
5100
6100
0 1 2 3 4 5 6 7 8 9 10
Figure 5. Total Labor Costs per Acre, by Owned Landholdings, 2007-8
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Immediate Family Member Other Landlord
India, except West Bengal
West Bengal
Figure 6. Source of Leased-in Land: India and West Bengal, 2006
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0 5 10 15 20 25 30 35 40 45 50
Figure 7. Locally-weighted Within-Farmer and Within-Season Estimates:
The Effects of Plot-Specific Fertilizer on Plot-Specific Profits per Acre (one sd confidence bounds),
by Landholding Size
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0 2 4 6 8 10 12 14 16 18 20
Figure 8. Locally-weighted FE-IV Estimates of the Returns to Capital Equipment Value (and .95 Confidence Intervals),
By Landholding Size
-200
-100
0
100
200
300
400
500
600
700
0 2 4 6 8 10 12 14 16 18 20
Gross of Farm Equipment
Net of Farm Equipment
Figure 9. Locally-weighted FE-IV Estimates of the Effects of Owned Landholdings on Profits per Acre,
Net and Gross of Farm Equipment Owned, by Landholding Size
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0 2 4 6 8 10 12 14 16 18 20
Figure 10. Locally-weighted Within-Plot Estimates:
The Effects of Lagged Farm Profits on Plot-Specific Profits per Acre (one sd confidence bounds), by Landholding Size
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
012345678910
Lease out
Lease in
Figure 11. Relationship Between the Probability of Leasing In and Leasing out Land,
by Ownership Holdings without Village Fixed Effects, 2006 (N=119,349)
-0.01
-0.005
0
0.005
0.01
0.015
0.02
0.025
-5 -3 -1 13579
Lease In
Lease Out
Figure 12. Within-Village Relationship Between the Probability of Leasing In and Leasing out Land,
by Ownership Holdings, 2006 (N=119,349)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0-.5 .5-1 1-2 2-3 3-4 4-5 5-7.5 7.5-10 10-20 20+
Indian Census, 2001
NCAER Listing Data 2006
Appendix Figure A. Cumulative Distribution of Owned Landholdings (Acres), by Data Source
Table 1
Within-Village and Within-Farmer and Season Plot Level Estimates (2007-2008):
Effects of the Use of Hired and Family Labor on Supervision Costs, by Estimation Procedure
Estimation procedure: Village Fixed-Effects Farmer-Season Fixed-Effects
a
Hired labor costs .0402
(3.05) .0383
(2.94) .0387
(3.46) .0399
(3.37)
Family labor costs, less supervision time .134
(4.11) .140
(4.23) .0410
(1.57) .0375
(1.44)
Plot area .00407
(1.91) .00409
(1.92) 3.58
(1.51) 3.35
(1.37)
Owned landholdings 2.74
(0.62) 2.80
(0.62) - -
Plot characteristics included N Y N Y
b
Number of observations 18,484 18,201 18,484 18,201
Number of farmer-seasons 8,685 8,587 8,685 8,587
Absolute value of asymptotic t-ratios in parentheses. Specification includes season dummy variables; clustered t-ratios at the farm
a
level. Plot characteristics include measures of depth, salinity, percolation and drainage; five soil colors (red, black, grey, yellow,
a
brown, off-white); five soil types (gravel, sandy, loam, clay, and hard clay), and distance from the household residence.
Table 2
Within-Farmer, Plot-Level Estimates Across Three Seasons (2007-8):
Effects of Plot Size on Plot-Specific Profits, Labor Costs, and Fertilizer Use per Acre and Use of Tractor Services
Profits per Acre Any Tractor Services Used Total Labor Costs per Acre
Plot area 145.4
(2.34) 157.0
(2.51) .00403
(1.99) .00404
(1.98) -107.2
(2.74) -106.9
(2.73)
Area of other plots 118.7
(1.95) 130.8
(2.14) .00333
(1.77) .00351
(1.76) -55.3
(1.45) -55.4
(1.45)
Total number of plots 482.3
(3.15) 473.8
(3.06) -.0333
(6.68) -.0330
(6.53) 123.2
(1.27) 123.2
(1.27)
Include soil characteristics? N Y N Y N Y
a
Number of plot observations 14,290 14,290 14,290 14,290 14,290 14,290
Number of farmers 4,130 4,130 4,130 4,130 4,130 4,130
Absolute value of asymptotic t-ratios in parentheses. Soil characteristics include measures of depth,
a
salinity, percolation and drainage; five soil colors (red, black, grey, yellow, brown, off-white); and
five soil types (gravel, sandy, loam, clay, and hard clay). All specifications include season*state
dummy variables and plot distance.
Table 3
Within-Farmer, Within-Season Plot-Level Estimates (2007-8):
Effects of Plot-Specific Fertilizer Use on Plot-Specific Profits, by Owned Landholdings
Owned landholdings < 4 acres 4-10 acres 10+ acres
Fertilizer use this season 1.49
(3.73) 1.46
(3.89) 3.23
(2.25) 3.24
(2.28) .182
(0.20) .166
(0.19)
Plot area 29.9
(1.02) 30.4
(0.97) 901.9
(3.03) 747.4
(2.45) 39.3
(0.48) 1.16
(0.02)
Include soil characteristics? N Y N Y N Y
a
Number of plot observations 4,008 4,008 1,935 1,935 1,173 1,173
Number of farmers 1,939 1,939 851 851 464 464
Absolute value of asymptotic t-ratios in parentheses clustered at the village level. Soil characteristics include measures of depth,
a
salinity, percolation and drainage; five soil colors (red, black, grey, yellow, brown, off-white); and five soil types (gravel, sandy,
loam, clay, and hard clay). All specifications include plot distance and fertilizer used in the prior period.
Table 4
Panel Data (1999-2008) First-Stage Farmer FE Estimates: Owned Landholdings and Value of Farm Equipment
Variable: Own Landholdings (acres) Farm equipment x 10-3
Inherited land (acres) after 1999 .193
(2.69) .881
(1.23)
Value of owned inherited mechanized assets in 1999 x 10 -.00458
-3
(0.34) .103
(1.22)
Value of owned inherited non-mechanized assets in 1999 x 10 -.0746
-3
(2.10) 2.86
(3.26)
Value of assets inherited after 1999 x 10 .263
-3
(0.41) -7.45
(1.06)
Standard deviation of the schooling of family claimants in 1999 -.0724
(1.04) 2.245
(2.87)
Head’s age in 1999 -.0247
(2.24) -.280
(1.65)
Whether respondent has brothers -.237
(0.48) 1.40
(0.33)
Number of observations 3,994 3,524
Number of farmers 1,749 1,749
Anderson-Rubin Wald test of weak instruments ÷(7),
2
p-value 22.32
.0022
Absolute value of asymptotic t-ratios in parentheses clustered at the household level.
Table 5
Panel Data Estimates (1999-2008): Effects of Own Landholdings and Own Farm Equipment on Profits per Acre,
by Estimation Procedure
Estimation procedure: Village Fixed-Effects Farmer Fixed-Effects Farmer Fixed-Effects IV
ab
Owned landholdings 13.1
(2.77) 13.1
(2.80) 8.35
(0.48) 3.53
(0.20) 389.6
(1.99) 250.3
(1.37)
Value of farm equipment - .00746
(1.16) - .0114
(3.26) - .0347
(2.14)
Number of observations 3,994 3,994 3,524 3,524 3,524 3,524
Number of farmers 2,138 2,138 1,749 1,749 1,749 1,749
Kleinberger-Paap underidentification test statistic
÷(df), p-value
2(4) 13.4,
.0093 (6) 16.5,
.0113
Hansen J overidentification test statistic
÷(df), p-value
2(3) 0.59
.898 (5) 5.47
.361
Absolute value of asymptotic t-ratios in parentheses. Specification includes year=2008 dummy; clustered t-ratios at the household
a
level. Instruments include land inherited after 1999, assets inherited after 1999, whether the current head has brothers, the standard
b
deviation of the schooling of inheritance claimants, the head’s age in 1999, and owned asset values in 1999.
Table 6
Retrospective Panel Data Estimates (2008): Effects of Own Landholdings and Own Farm Equipment on Investment in Farm
Equipment and Equipment Rental, by Estimation Procedure
Dependent variable Equipment Investment Equipment Hire Expenditure
Estimation procedure FE-Farmer FE-Farmer IV FE-Farmer IV FE-Farmer FE-Farmer IV FE-Farmer IV
bb bb
Owned landholdings 16.3
(0.05) 663.8
(2.15) -400.7
(1.06) 122.7
(1.69) 196.3
(2.01) 201.2
(1.90)
Landholdings x bank - - 1654
(2.50) - - -6.29
(0.04)
Value of owned farm
equipment -.0843
(1.30) -.909
(8.67) -.906
(8.54) -.0187
(3.51) -.0500
(1.16) -.0500
(1.16)
Bank within 10 Km 3524
(2.27) 1820
(0.61) -2920
(0.84) -.319.8
(0.94) 279.6
(0.53) -258.0
(0.29)
Number of farmers 3,522 3,522 3,522 1,833 1,833 1,833
Absolute value of asymptotic t-ratios in parentheses. Specification includes year=2008 dummy; clustered t-ratios at the village
a
level. Instruments include land inherited after 1999, assets inherited after 1999, and the presence of a bank within 10 km in 1999.
b
Table 7
Retrospective Panel Data Estimates (2008) First Stage Fixed-Effects Farmer Estimates: Effects of Own Landholdings and Own
Farm Equipment on Investment in Farm Equipment and Equipment Rental
Dependent variable/
Instrument Owned
Landholdings Own Farm
Equipment Bank < 10 km of
the Village Own Farm Equipment
x Bank Proximity
Inherited landholdings between 1999 and
2008 .938
(18.6) -142.6
(0.40) -.0191
(3.65) .0432
(0.57)
Inherited farm assets between 1999 and
2008 x 10-3 -.00105
(2.61) .546
(5.31) .000973
(2.48) .00229
(1.78)
Bank within 10 km of the village in 1999 -.118
(0.92) -841.8
(0.42) -.918
(10.7) -3.53
(6.92)
Inherited landholdings x bank proximity .04365
(0.85) 1499
(2.22) .0182
(2.92) .904
(10.9)
Number of farmers 3,522 3,522 3,522 3,522
Absolute value of asymptotic t-ratios in parentheses clustered at the farm level. Specifications include season*village dummy
variables.
Table 8
Within-Farmer Estimates Across Three Seasons (2007-8):
Effects of Previous-Period Profit Shocks on Current Profits per Acre, by Owned Landholding Size
Farm size: Owned
Landholdings<4 Owned
Landholdings>=4, <10 Owned
Landholdings>=10
Farm profits per acre, previous season -.266
(4.33) -.503
(6.53) -.797
(5.63)
Fertilizer use, previous season (value per acre) 1.42
(3.62) .104
(1.82) 3.63
(4.18)
Total cultivated area, this season 2494.5
(2.42) 215.5
(0.53) 301.6
(2.17)
Number of farmers 2,176 1,061 580
Absolute value of asymptotic t-ratios in parentheses clustered at the farm level. Specifications include season*village dummy
variables.
Table 9
Within-Plot Estimates Across Three Seasons (2007-8):
Effects of Previous-Period Farm-Level Profit Shocks on Plot-Level Current Profits per Acre, by Owned Landholding Size
Farm size: Owned
Landholdings<4 Owned
Landholdings>=4, <10 Owned
Landholdings>=10
Farm profits per acre, all other plots, previous
season .140
(2.04) .0628
(0.67) .0366
(0.17)
Farm profits per acre, this plot, previous season -.456
(5.60) -.504
(4.78) -.540
(2.88)
Fertilizer use, this plot, previous season (value
per acre) 1.54
(2.61) .789
(1.97) 1.87
(1.22)
Number of plot observations 6,068 3,258 1,919
Number of farmers 1,351 678 311
Absolute value of asymptotic t-ratios in parentheses clustered at the farm level. Specifications include season*village dummy
variables.
Barriers to Farm Profitability in India: Mechanization, Scale and Credit Markets
Andrew D. Foster and Mark R. Rosenzweig
September 2010
Viewed across countries of the world, the most productive agricultural areas are:
1. Family farmed
2. Large scale
3. Mechanized
Indian agriculture is
1. Family farmed, but
2. Very small scale: 80% of farms are less than two acres in size
and 95% are less than five acres (Indian
Census, 2001)
3. Not mechanized: Less than 3% of farms between 2 and 4 acres
own a tractor; but 30% of farms 10+ do (Our
national rural survey, 2008)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0-.5 .5-1 1-2 2-3 3-4 4-5 5-7.5 7.5-10 10-20 20+
Figure 1. The Cumulative Distribution of Owned Landholdings (Acres) in India (2001 Census)
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
<112345678910+
Tractor Plow Thres her
Figure 2. Mechanization and Owned Landholdings (Acres), 2007-2008
2000
3000
4000
5000
6000
7000
8000
0 2 4 6 810121416 1820
Profit s per a cre, no supe rvis ion cos ts
Profits per acre
Figure 3. Profits per Acre and Profits per Acre less Supervision Costs,
by Owned Landholdings, 2007-8
-0.2
199.8
399.8
599.8
799.8
999.8
1199.8
1399.8
1599.8
1799.8
0 2 4 6 8 101214161820
Total supervision costs
Supe rvision co sts per acre
Figure 4. Total Supervision Costs and Supervision Costs per Acre by Owned Landholdings Size
100
1100
2100
3100
4100
5100
6100
012345678910
Figure 5. Total Labor Costs per Acre, by Owned Landholdings, 2007-8
Descriptively, in India we observe negative relationship between profitability per
acre and landholdings, with and without taking into account supervision costs
Drop in supervision and total labor used per-acre as well
The question posed in this paper is whether increasing the scale and mechanizing Indian
agriculture is key to improving agricultural productivity
Prior literature focusing on barriers to high productivity agriculture in low-income
countries:
1. Many studies on scale economies. Key distinction is between output per acre and
profitability. Mixed evidence, unclear definitions of meaning of scale economies,
unclear about nature of markets, ownership versus leased operational holdings.
Biggest empirical problem: scale is endogenous, correlated with
unobservable farmer and land attributes
Unlikely to carry out RCT varying owned land
2. Focus on market imperfections as barriers to agricultural productivity in low-
income countries:
A. Credit markets B. Insurance markets C. Labor markets
But, these market problems are not confined to poor countries:
All farmers everywhere are credit-constrained;
Few farmers purchase formal insurance in any country;
And farmers in developed countries face the same problems of worker
shirking and employ the same labor arrangements as in poor countries.
Problems of credit, insurance, shirking have not been solved anywhere
3. Focus on technology
But Indian farmers using in many cases the same seeds as in developed
countries (Bt cotton, HYV wheat)
4. Focus on behavioral issues impeding efficient resource use
But the brains of Indian farmers are hard-wired the same as humans elsewhere
5. Property rights
This is a big issue in many countries, but in India, property rights are
reasonably well-established: titling, computerized land records
The key difference between high-productivity agriculture in developed countries and a
low-income country like India would thus appear to be scale and mechanization
Put another way, are there too many people employed in agriculture, in the sense that
removing workers and consolidating landholdings to fewer farmers would increase farm
efficiency?
“Surplus” labor revisited? but not via old and discredited micro mechanisms of
compensatory labor supply or nutrition-efficiency wages
This paper:
1. Set out a model to clarify the interrelationships among mechanization, scale
economies, credit market constraints, agency issues, lack of insurance, and
profitability.
2. Key feature is that scale economies (convexities) arise solely because higher-
capacity, mechanized farm implements require more space to operate effectively
(two-row versus 8-row harvester); farmers can choose optimally number and
capacity of equipment (machines are scalable - not an indivisibility issue).
3. Market setting: Labor requires supervision (shirking); credit costs depend on
amount borrowed and on collateral - owned land. But farmers can freely rent
equipment and purchase and sell labor services at constant prices.
4. Tests established distinguishing operational scale effects and size of owned
landholdings on profitability and role of credit market imperfections.
5. Dynamics introduced to test role relationship between ownership scale and
profitability given lack of insurance and credit constraints
6. Empirical application exploits unique data:
A. Panel data at the plot level on all inputs, outputs and costs (3 seasons)
1. Within-farmer/season cross plots: Permits identification of pure
scale effects on profitability and input use: across plots for the same
farmer, so that credit and input costs are held constant (plot
characteristics matter)
2. Within-plots across seasons: perfectly controls for plot and farmer
characteristics; distinction between plot and farm level shocks on
plot-specific profitability
B. Longer-term panel at the farm level (9-year interval)
Identifies effects of changing owned landholdings: exploit family
division, inheritance of land
For same family, observe what happens when land is broken up at
death of head: IV, FE approach
7. Results: Yes, increasing own landholdings increases significantly profits per acre
and mechanization and lowers labor intensity: scale economies and imperfect credit
2. Theory
A. Technical scale economies, cultivated land area and mechanization
Two inputs to production: Total work to be done e, by either machinery or
labor
Fertilizer f
Output per acre y in a given crop cycle is CRS in acreage a, and given by
(1)
The work technology is:
(2)
where
= manual labor services, a constant-returns function of
supervision
q = machine capacity and k = the number of machines.
Important features of the work technology:
A. Machinery varies by capacity.
B. The advantage of large farms with respect to higher-capacity equipment is
embodied in the function - increasing the capacity of a machine
augments productivity more the larger is a.
C. Manual labor and machinery services are imperfect substitutes
D. Manual labor, regardless of the hired or family status of that labor, must be
supervised in order to be productive, and (iii) that machinery varies by
capacity.
Costs:
A. The price per day of a machine = , where í <1. So the rental price
increases non-linearly in capacity q.
B. Rental markets for machine and labor services are perfect.
The cost function is thus
(3)
fm s
where p =price of fertilizer, w=wage rate of manual labor, and w=wage of
supervisory labor.
A profit-maximizing farmer maximizes the value of (1) minus costs (3) subject to (2) .
Optimal machine capacity depends only on area and the elasticity of the price schedule
with respect to capacity - larger farms use higher capacity machines
(5)
Can solve for the effect of land area, conditional on the amount of “work” to be done
A. More machines (of increasing capacity) per unit of work e, the larger is a.
(8)
which is positive for s ufficiently close to one.
B. Less labor per unit of work and lower costs, conditional on work
(10)
(11)
where .
Embed cost function in profit function, letting the superscript * denote per-acre quantities:
(13)
Results:
A. Larger operations are more profitable on a per-acre basis.
(14)
B. Larger operational holdings will use inputs more intensively
(15)
C. Larger operational holdings will use more machinery per acre
(17)
Will larger operations use less labor per unit area?
A. There is an increase in work intensity as the increasing returns associated
with machinery lower unit work costs.
B. There is also a decrease in the amount of labor per unit work.
If the demand for work is price inelastic and/or labor and machines are sufficiently
good substitutes, however, both manual and supervisory labor must decline:
(18) ,
ec
where å = price elasticity of work
B. Scale effects and credit marketimperfections
Now introduce credit market imperfections.
Distinction between ownership and rental now important.
Farmers borrow b* per acre to finance agricultural inputs and repay this debt with
interest r during the harvest period.
But, the interest rate ron this debt is dependent on the amount borrowed per acre as
well as on total owned land area a.
The per-acre amount ñ that must be repaid in the harvest period is given by
(19) ,
12
ab
where the r<0 (collateral value of land), r >0 so ñ<0, ñ>0.
Advantage of owned land: lower interest (by assumption)
Advantage of owned equipment: save on borrowing costs to finance machine rental
Letting o* denote the rental value of owned assets, amount borrowed is given
by
(20) .
The farmer’s maximization problem with credit market imperfections:
(21)
0
where r (< r(a,b*)) = the rate of return on savings
Profits per acre rise more steeply with land if there are credit market imperfections:
(22) ,
A. There is a negative effect of owned area on interest rates given input use per
acre, .
B. Savings in cost per unit of work associated with scale lower the amount
borrowed b*, thus further lowering interest costs.
Note: if no technical scale economies (c`(a)=0), still rising profitability
and if perfect credit markets, rising profitability if there are scale
economies.
Is there a direct test of credit market imperfections?
The marginal return to capital is given by
(25)
Perfect credit markets:
Implications:
A. When borrowing costs are independent of land ownership and equal to the
returns on savings, the marginal return to capital assets is zero.
Non-zero marginal return to machinery implies imperfect credit markets
B. Marginal returns to capital decline with land size: dð*/do*da < 0
2
because debt b* and interest rates r fall with a
C. Larger farms more likely to purchase equipment: face lower borrowing costs but
same rental price as other farmers
C. Farm size, savings, no insurance and profits
Now, introduce stochastic output, lack of insurance and possibility of savings:
dynamics
Before borrowing determined by demand for inputs
Now, also depends on cash on hand
Cash on hand depends on prior profits
But, prior profits also affect current decisions directly: soil nutrient effects
Two reason for looking at dynamics:
1. In empirical work, the existence of lagged profit effects has implications for
econometrics (why farmer fixed effects estimates are biased)
2. Another direct test of credit market imperfections
3. A third advantage to larger land size via credit markets: immunity to shocks
We assume a stationary problem
State variables: soil nutrition n* cash on hand h*
t
Farmers choose fertilizer levels f* prior to the realization of a shock è
The value function:
(28)
where is the discount factor
(29) ,
where denotes the extent to which unanticipated shocks are saved.
For unanticipated shocks are fully saved as in the permanent income
model
For cash on hand is a constant
Soil nutrient dynamics:
(30)
á reflects the idea is that more rapidly growing plants will deplete the soil of
t
nutrients relatively quickly (if è is rainfall, more nutrients are used if rainfall
and soil nutrients are complements).
The production and credit functions are quadratic
and ,
All of the parameters are positive:
diminishing returns
cost of credit r increases at a higher rate with the amount of credit
Profits gross of credit costs:
(33)
The effect of a previous period shock on next-period’s profits is
(33)
where is the second derivative from the value function, with
for an interior maximum.
Two effects of opposite sign (empirical challenge to separate them):
If , liquidity h does not depend on unanticipated income shocks
then, only a nutrient carryover or depletion effect, which is negative
If , no nutrient carryover,
then a positive shock in period 0 induces higher savings and lowers credit
costs so that profits are higher
Note: there is no effect of lagged profits if borrowing does not effect credit
costs or if borrowing is very small: large farmers
Data
Sources:
NCAER REDS 2007-8 round
NCAER REDS 1999 round
Represents rural households in 242 villages in 17 major states of India
(excludes Assam and J & K)
Fifth and sixth rounds of panel survey going back to 1968-69
Originally meant to be representative of entire rural population of India
Oversampled areas conducive to green revolution success higher-income
households
Provide comprehensive inputs and outputs for farming: inclusive of family labor,
supervisory labor, use of own equipment, all costs
Village-level information on institutions and facilities
Histories of family linkages: trace out break-up of households across generations
Four types of information used in our analysis
A. Panel information at the plot level over three seasons from the REDS
2007-8: 10, 947 plots, 2/3 observed at least twice
B. Retrospective histories of equipment and land acquisitions from survey
date in 2007-8 back to 1999: 4, 961 farmers
C. Panel of farm households from the combined 2007-8 and 1999 rounds:
2,848 panel farm households
D. Listing information in 2006 for all households in all villages: 119, 284
households
Use data, guided by model, to estimate effects of land size, mechanization
on profitability on mechanization on returns to mechanization
Key challenges of empirical work:
A. Heterogeneity in land, farmers
B. Endogeneity of land and equipment ownership
C. Identification of pure scale from credit-market effects
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0-.5 .5-1 1-2 2-3 3-4 4-5 5-7.5 7.5-10 10-20 20+
Indian Census, 2001
NCAER Listing Data 2006
Appendix Figure A. Cumulative Distribution of Owned Landholdings (Acres), by Data Source
Key features of methods:
1. Exploit plot level panel, distinguish plot- from farm-level effects
2. Exploit change in landownership over time based on family break-up:
Less than 3% of farmers purchased (sold) land in 9-year period
Over 19% of farmers inherited land in 9-year period 1999-2008
3. Allow estimated effects to vary by land size, as theory suggests (local
linear approximations by land ownership size)
First, data show that two assumptions of model are not unrealistic:
1. Rental of land unimportant, confined to immediate family members
2. No difference between family and hired labor in terms of demands on
supervision
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Immediate Family Member Other Landlord
India, except West Bengal
West Bengal
Figure 6. Source of Leased-in Land: India and West Bengal, 2006
Table 1
Within-Village and Within-Farmer and Season Plot Level Estimates (2007-2008):
Effects of the Use of Hired and Family Labor on Supervision Costs, by Estimation Procedure
Estimation procedure: Village Fixed-Effects Farmer-Season Fixed-Effects
a
Hired labor costs .0402
(3.05) .0383
(2.94) .0387
(3.46) .0399
(3.37)
Family labor costs, less
supervision time .134
(4.11) .140
(4.23) .0410
(1.57) .0375
(1.44)
Plot area .00407
(1.91) .00409
(1.92) 3.58
(1.51) 3.35
(1.37)
Owned landholdings 2.74
(0.62) 2.80
(0.62) --
Plot characteristics included N Y N Y
b
Number of observations 18,484 18,201 18,484 18,201
Number of farmer-seasons 8,685 8,587 8,685 8,587
Absolute value of asymptotic t-ratios in parentheses. Specification includes season dummy
a
variables; clustered t-ratios at the farm level. Plot characteristics include measures of depth,
a
salinity, percolation and drainage; five soil colors (red, black, grey, yellow, brown, off-white); five
soil types (gravel, sandy, loam, clay, and hard clay), and distance from the household residence.
Road map of empirical applications:
A. Estimates of the pure effects of scale and returns to fertilizer use by owned land:
exploits within-farm plot differences in size (8 measures of plot characteristics)
on profits per acre, use of mechanized inputs, labor intensity (farmer FE)
B. Estimates of mechanization and landownership size on profits per acre:
exploits household panel; uses IV, farmer fixed-effects for causal effects
C. Estimates of the determinants of mechanization and role of credit:
Exploits retrospective history of equipment acquisitions; uses IV, farmer fixed-
effects
D. Estimates of lagged profit shocks, by landholding size, on current profits
distinguishing soil-nutrient from liquidity/credit effects
Exploits panel of plots across three seasons; (plot and farmer fixed effects)
E. Final section: are we as smart as farmers? Does leasing behavior of Indian farmers
reflect the land size effects we find (who rents from whom)
Exploits listing data: observe entire market
Table 2
Within-Farmer, Plot-Level Estimates Across Three Seasons (2007-8):
Effects of Plot Size on Plot-Specific Profits, Labor Costs, and Fertilizer Use per Acre
and Use of Tractor Services
Profits per Acre Any Tractor
Services Used Total Labor Costs
per Acre
Plot area 145.4
(2.34) 157.0
(2.51) .00403
(1.99) .00404
(1.98) -107.2
(2.74) -106.9
(2.73)
Area of other plots 118.7
(1.95) 130.8
(2.14) .00333
(1.77) .00351
(1.76) -55.3
(1.45) -55.4
(1.45)
Total number of plots 482.3
(3.15) 473.8
(3.06) -.0333
(6.68) -.0330
(6.53) 123.2
(1.27) 123.2
(1.27)
Include soil characteristics? N Y N Y N Y
a
Number of plot observations 14,290 14,290 14,290 14,290 14,290 14,290
Number of farmers 4,130 4,130 4,130 4,130 4,130 4,130
Absolute value of asymptotic t-ratios in parentheses. Soil characteristics include measures of
a
depth, salinity, percolation and drainage; five soil colors (red, black, grey, yellow, brown, off-
white); and five soil types (gravel, sandy, loam, clay, and hard clay). All specifications include
season*state dummy variables and plot distance.
Table 2
Within-Farmer, Plot-Level Estimates Across Three Seasons (2007-8):
Effects of Plot Size on Plot-Specific Profits, Labor Costs, and Fertilizer Use per Acre
and Use of Tractor Services
Profits per Acre Any Tractor
Services Used Total Labor Costs
per Acre
Plot area 145.4
(2.34) 157.0
(2.51) .00403
(1.99) .00404
(1.98) -107.2
(2.74) -106.9
(2.73)
Area of other plots 118.7
(1.95) 130.8
(2.14) .00333
(1.77) .00351
(1.76) -55.3
(1.45) -55.4
(1.45)
Total number of plots 482.3
(3.15) 473.8
(3.06) -.0333
(6.68) -.0330
(6.53) 123.2
(1.27) 123.2
(1.27)
Include soil characteristics? N Y N Y N Y
a
Number of plot observations 14,290 14,290 14,290 14,290 14,290 14,290
Number of farmers 4,130 4,130 4,130 4,130 4,130 4,130
Absolute value of asymptotic t-ratios in parentheses. Soil characteristics include measures of
a
depth, salinity, percolation and drainage; five soil colors (red, black, grey, yellow, brown, off-
white); and five soil types (gravel, sandy, loam, clay, and hard clay). All specifications include
season*state dummy variables and plot distance.
Table 3
Within-Farmer, Within-Season Plot-Level Estimates (2007-8):
Effects of Plot-Specific Fertilizer Use on Plot-Specific Profits, by Owned Landholdings
Owned landholdings < 4 acres 4-10 acres 10+ acres
Fertilizer use this season 1.49
(3.73) 1.46
(3.89) 3.23
(2.25) 3.24
(2.28) .182
(0.20) .166
(0.19)
Plot area 29.9
(1.02) 30.4
(0.97) 901.9
(3.03) 747.4
(2.45) 39.3
(0.48) 1.16
(0.02)
Include soil characteristics? N Y N Y N Y
a
Number of plot observations 4,008 4,008 1,935 1,935 1,173 1,173
Number of farmers 1,939 1,939 851 851 464 464
Absolute value of asymptotic t-ratios in parentheses clustered at the village level. Soil
a
characteristics include measures of depth, salinity, percolation and drainage; five soil colors (red,
black, grey, yellow, brown, off-white); and five soil types (gravel, sandy, loam, clay, and hard
clay). All specifications include plot distance and fertilizer used in the prior period.
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0 5 10 15 20 25 30 35 40 45 50
Figure 7. Locally-weighted Within-Farmer and Within-Season Estimates:
The Effects of Plot-Specific Fertilizer on Plot-Specific Profits per Acre (one sd confidence bounds),
by Landholding Size
Landownership, profits and the returns to mechanization
The profit equation we seek to estimate is
jt 0t A jt k jt j ijt
(38) ð = d + d A+ dk + ì + å,
where t is survey year
k=value of all farm machinery
j
ì=unobservable household fixed effect: land quality, farmer ability
ijt
å=iid error
jt j
Problem cov(A,ì) 0
Standard fixup - exploit panel data, farmer fixed effect removed
With two observations (1999, 2007-8):
jt 0 A j k j ijt
(39) Äð = Äd + d ÄA+ k + Äå,
where Ä is the intertemporal difference operator.
But, our model says that lagged profit shocks will affect current profits, thus making
ijt ijt
possible acquisition of land or capital: ÄAhigh when Äå is low.
ijt jt
Even if the contemporaneous cov(å, A ) = 0 (assets are measured prior to the shock)
ijt j
cov(ÄåA) < 0 FE technique introduces a negative bias, where credit is
constrained
Ak
We use IV to obtain consistent estimates of d and d
Over the nine-year interval between surveys 19.9% of farms divided and farmers
inherited land.
The instruments we use to predict the change in landholdings and equipment of a
farmer between 1999 and 2007-8 are:
A. The value of owned mechanized and non-mechanized assets inherited
prior to 1999
B. The value of assets and acreage of land inherited between 1999 and
2007-8.
C. The age of the head in 1999, an indicator of whether the head in
2007-8 had brothers, and a measure of the educational inequality
among the claimants to the head’s land in 1999 (F&R, 2003).
Table 4
Panel Data (1999-2008) First-Stage FE Farmer Estimates:
Owned Landholdings and Value of Farm Equipment
Variable: Own Landholdings (acres) Farm equipment x 10-3
Inherited land (acres) after 1999 .193
(2.69) .881
(1.23)
Value of owned inherited mechanized assets
in 1999 x 10-3 -.00458
(0.34) .103
(1.22)
Value of owned inherited non-mechanized
assets in 1999 x 10-3 -.0746
(2.10) 2.86
(3.26)
Value of assets inherited after 1999 x 10 .263
-3
(0.41) -7.45
(1.06)
Standard deviation of the schooling of family
claimants in 1999 -.0724
(1.04) 2.245
(2.87)
Head’s age in 1999 -.0247
(2.24) -.280
(1.65)
Whether respondent has brothers -.237
(0.48) 1.40
(0.33)
Number of observations 3,994 3,524
Anderson-Rubin Wald test of weak
instruments ÷(7), p-value
222.32
.0022
Table 5
Panel Data Estimates (1999-2008): Effects of Own Landholdings and Own Farm Equipment on
Profits per Acre, by Estimation Procedure
Estimation
procedure: Village Fixed-Effects Farmer Fixed-Effects Farmer Fixed-
a
Effects IVb
Owned landholdings 13.1
(2.77) 13.1
(2.80) 8.35
(0.48) 3.53
(0.20) 389.6
(1.99) 250.3
(1.37)
Value of farm
equipment - .00746
(1.16) - .0114
(3.26) - .0347
(2.14)
Number of
observations 3,994 3,994 3,524 3,524 3,524 3,524
Number of farmers 2,138 2,138 1,749 1,749 1,749 1,749
Kleinberger-Paap underidentification test statistic
÷(df), p-value
2(4) 13.4,
.0093 (6) 16.5,
.0113
Hansen J overidentification test statistic
÷(df), p-value
2(3) 0.59
.898 (5) 5.47
.361
Absolute value of asymptotic t-ratios in parentheses. Specification includes year=2008 dummy;
a
clustered t-ratios at the household level. Instruments include land inherited after 1999, assets
b
inherited after 1999, whether the current head has brothers, the standard deviation of the schooling
of inheritance claimants, the head’s age in 1999, and owned asset values in 1999.
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0 2 4 6 8 10 12 14 16 18 20
Figure 8. Locally-weighted FE-IV Estimates of the Returns to Capital Equipment Value (and .95 Confidence Intervals),
By Landholding Size
-200
-100
0
100
200
300
400
500
600
700
02468101214161820
Gross of Farm Equipme nt
Ne t of Fa rm Equipment
Figure 9. Locally-weighted FE-IV Estimates of the Effects of Owned Landholdings on Profits per Acre,
Net and Gross of Farm Equipment Owned, by Landholding Size
Table 6
Retrospective Panel Data Estimates (2008): Effects of Own Landholdings and Own Farm
Equipment on Investment in Farm Equipment and Equipment Rental, by Estimation Procedure
Dependent
variable Equipment Investment Equipment Hire Expenditure
Estimation
procedure FE-Farmer FE-Farmer
IVbFE-Farmer
IVbFE-
Farmer FE-Farmer
IVbFE-Farmer
IVb
Owned
landholdings 16.3
(0.05) 663.8
(2.15) -400.7
(1.06) 122.7
(1.69) 196.3
(2.01) 201.2
(1.90)
Landholdings x
bank - - 1654
(2.50) ---6.29
(0.04)
Value of owned
farm equipment -.0843
(1.30) -.909
(8.67) -.906
(8.54) -.0187
(3.51) -.0500
(1.16) -.0500
(1.16)
Bank within 10
Km 3524
(2.27) 1820
(0.61) -2920
(0.84) -.319.8
(0.94) 279.6
(0.53) -258.0
(0.29)
Number of farmers 3,522 3,522 3,522 1,833 1,833 1,833
Absolute value of asymptotic t-ratios in parentheses. Specification includes year=2008 dummy;
a
clustered t-ratios at the village level. Instruments include land inherited after 1999, assets inherited
b
after 1999, and the presence of a bank within 10 km in 1999.
Table 7
Retrospective Panel Data Estimates (2008) First Stage Fixed-Effects Farmer Estimates: Effects of
Own Landholdings and Own Farm Equipment on Investment in Farm Equipment and Equipment
Rental
Dependent variable/
Instrument Owned
Landholdings Own Farm
Equipment Bank < 10 km
of the Village
Own Farm
Equipment x
Bank Proximity
Inherited landholdings between
1999 and 2008 .938
(18.6) -142.6
(0.40) -.0191
(3.65) .0432
(0.57)
Inherited farm assets between
1999 and 2008 x 10-3 -.00105
(2.61) .546
(5.31) .000973
(2.48) .00229
(1.78)
Bank within 10 km of the
village in 1999 -.118
(0.92) -841.8
(0.42) -.918
(10.7) -3.53
(6.92)
Inherited landholdings x bank
proximity .04365
(0.85) 1499
(2.22) .0182
(2.92) .904
(10.9)
Number of farmers 3,522 3,522 3,522 3,522
Absolute value of asymptotic t-ratios in parentheses clustered at the farm level. Specifications
include season*village dummy variables.
Table 8
Within-Farmer Estimates Across Three Seasons (2007-8):
Effects of Previous-Period Profit Shocks on Current Profits per Acre, by Owned Landholding Size
Farm size: Owned
Landholdings<4 Owned
Landholdings>=4,
<10
Owned
Landholdings>=10
Farm profits per acre, previous
season -.266
(4.33) -.503
(6.53) -.797
(5.63)
Fertilizer use, previous season
(value per acre) 1.42
(3.62) .104
(1.82) 3.63
(4.18)
Total cultivated area, this season 2494.5
(2.42) 215.5
(0.53) 301.6
(2.17)
Number of farmers 2,176 1,061 580
Absolute value of asymptotic t-ratios in parentheses clustered at the farm level. Specifications
include season*village dummy variables.
Table 9
Within-Plot Estimates Across Three Seasons (2007-8):
Effects of Previous-Period Farm-Level Profit Shocks on Plot-Level Current Profits per Acre,
by Owned Landholding Size
Farm size: Owned
Landholdings<4 Owned
Landholdings>=4,
<10
Owned
Landholdings>=10
Farm profits per acre, all other plots,
previous season .140
(2.04) .0628
(0.67) .0366
(0.17)
Farm profits per acre, this plot,
previous season -.456
(5.60) -.504
(4.78) -.540
(2.88)
Fertilizer use, this plot, previous
season (value per acre) 1.54
(2.61) .789
(1.97) 1.87
(1.22)
Number of plot observations 6,068 3,258 1,919
Number of farmers 1,351 678 311
Absolute value of asymptotic t-ratios in parentheses clustered at the farm level. Specifications
include season*village dummy variables.
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0 2 4 6 8 10 12 14 16 18 20
Figure 10. Locally-weighted Within-Plot Estimates:
The Effects of Lagged Farm Profits on Plot-Specific Profits per Acre (one sd confidence bounds), by Landholding Size
-0.01
-0.005
0
0.005
0.01
0.015
0.02
0.025
-5-3-113579
Lea se In
Lease Out
Figure 12. Within-Village Relationship Between the Probability of Leasing In and Leasing out Land,
by Ownership Holdings, 2006 (N=119,349)
Conclusion
A. The principal proximate barriers to profitable, efficient agriculture in India is the
small size of owned landholdings and lack of mechanization
Larger farms are more mechanized, have higher profits per acre, use labor less,
more efficiently use fertilizer and equipment, and their incomes are less
susceptible to profit shocks
B. Small landholdings lead to under-mechanization and higher labor costs due to
technical scale economies, credit market imperfections and lack of insurance
C. Indian farmers attempt to increase scale through the land rental market, but that
market is very limited (contiguity important) and does not mitigate credit issue
D. First step: key question is why are farms so small? Why are there too many farms?
E. Possible that attraction of people out of agriculture could make farming more
efficient and productive if leads to land consolidation
“Out of agriculture” policy may be a pro agriculture policy
... As a result, it may lead to a greater stabilization of revenues [35]. The research also points to the fact that the size of farms has a statistically insignificant impact on the variability of income in low-developed countries [36][37][38]. For developed countries, including the countries of the European Union, such as those from the former Soviet bloc, numerous studies show that, with the increase in the size of farms, land and labor productivity and income per working person usually increase [34,[39][40][41][42]. ...
... The labor pull effect of people working in agriculture largely depends on macroeconomic factors, especially the economic development of the country and the level of unemployment. With a low unemployment rate, the labor pull effect will be strengthened through the labor push effect [38,[78][79][80]. Despite the negative consequences of the technology treadmill, farmers are under pressure to implement technological progress. ...
... Moreover, it should be noted that the level of indebtedness of the analyzed farms was low (Table 1), which did not burden farms with credit costs. Farmers usually reduce the financial risk related to debt, which may inhibit the dynamic development of farms but increase their resistance to financial market disturbances [38,[119][120][121]. ...
Article
Full-text available
Farms in Poland come in a wide variety of sizes, and the effect of farm size on the profitability of labor has not been sufficiently investigated. This paper takes a fresh look at the model for labor profitability determinants of family farms in Poland in relation to their economic size. The purpose of this paper is to analyze the factors that determine the labor profitability index in farms of various economic size classes (classes ES1-ES5). In the analysis of factors shaping the profitability of labor in family farms, a panel analysis was applied. Family farm income expressed per family labor unit was adopted as the dependent variable. The following variables were used as explanatory variables: (1) macroeconomic index of price relations ("price gap"); unemployment; average monthly gross wages and salaries; inflation; (2) technical-agricultural production efficiency index; (3) microeconomic ratio of total assets to agricultural land; technical equipment for work; land-to-labor ratio; debt ratio, subsidy ratio, and investment effort. A diversified influence of selected factors determining the level of profitability of labor in agriculture in particular groups of farms was found. The econometric models developed also indicate different strategies that are adopted by farmers on various farms. There is no single solution here; strategies for improving the profitability of work must take into account the specificity of a given entity. The models estimated indicate the necessity of using other mechanisms and tools of agricultural policy for farms of various economic sizes. It should be expected that, in the future, there will be a dichotomous development of farms. Medium-sized farms will become larger and economically effective, and smaller farms will perform residential functions, with the disappearing function of agricultural production.
... First, greater distances between fragments increase the time needed to move between them and associated set-up costs for equipment (Bentley, 1987). Second, if fragment size falls below a critical threshold, mechanization will be impossible, thus precluding the use of labor-saving machinery altogether and increasing production costs (Foster & Rosenzweig, 2010). Empirically, fragment size and distance between fragments constitute two distinct aspects of fragmentation that can be considered separately from each other and that may differentially affect farms of different sizes. ...
... While parental bequest decisions are complex and descendants may try to influence them pre-mortem (Goetghebuer & Platteau, 2010), the amount and location of land inherited and thus a holding's level of fragmentation is no longer a decision variable for recipients post mortem. In fact, studies use amount or value of land inherited or brought into marriage as an instrument for outcomes in terms of consumption (Estudillo et al., 2001), intra-household bargaining (Anderson & Eswaran, 2009), production (Foster & Rosenzweig, 2010) and decisions about new household formation (Foster & Rosenzweig, 2002). The only way for households to adjust their holding's level of fragmentation following inheritance is via participation in land markets. ...
... Class means type of animals, that is, goats, sheep, buffalos, cows, camels, poultry.6 See Appendix I.7 The probit model results inTable 2are inconsistent with the results from the duration model(Table 1), as they do not account for time-varying information(Abdulai & Huffman, 2005;Foster & Rosenzweig, 2010). Notably, the factors that are significant in the selection stage of the ESR model (seeTable 2) differ from the hazard model (seeTable 1) due to the model assumptions and the set of variables used. ...
Article
This paper investigates the factors that influence the adoption of laser land leveler and its impact on groundwater usage in the Punjab province of Pakistan. A farm household survey of 504 agriculture producers was conducted in 2019. A discrete‐time duration model is used to investigate factors influencing the speed of adoption and an endogenous switching regression (ESR) model is used to evaluate its impact on groundwater usage. About 70% of the surveyed households adopted the technology, and the average time to adoption was 9 years. Key factors accelerating the speed of adoption include strong legal land rights, access to information about the technology, and exposure to the technology. In contrast, long distance to rental market deaccelerates the speed of adoption. The adoption of laser land leveler reduced groundwater use by about 23%. The results imply that institutional arrangements, such as improving access to extension services, exposure to innovation, and legal land rights, can enhance the adoption and diffusion of the technology and conserve groundwater.
... The inverse farm size-productivity relationship is not universal. It is controversial to conclude that smalls are more productive because some studies find different results (Kumar et al., 2020;Otsuka et al., 2013;Foster and Rosenzweig, 2010) since it depends on the input use, owner's personal attributes, managerial capability, production technology, etc. The reason for the rising average per-unit cost of small farms can be made visible by looking into Table 5. ...
Article
Carp polyculture emerged as a promising practice for the development of aquaculture in Bangladesh in recent years. However, the economic potentiality of carp polyculture relied on the profitability and inputs use pattern. Therefore, this study investigated the economic viability of carp polyculture farming, particularly focused on profitability, profit sensitivity, economies of scale, and liquidity. A survey was conducted using a structured questionnaire to gather cross-sectional data, which included 183 randomly selected carp polyculture farmers from five districts in Bangladesh. Descriptive and econometric analysis was adopted to evaluate the economic performance of the carp polyculture system. The results showed that feed was the most important production input since it accounted for a major share of operating costs (57.3%). Freshwater carp polyculture farming was generally found profitable, although the profit was considerably sensitive with the alteration of feed price and fish price. Economies of scale were evident in this study, where the farm operated in economies of scale had lower productivity but greater cost efficiency. However, some inputs were inefficiently used by the diseconomies scale of farms, which led them to loss of return even though they had higher productivity. High cash flow coverage ratio due to higher yield strategies and low debt-servicing ratios were found, indicating lower liquidity risk for carp polyculture farming. Based on the present findings, it is suggested that farmers should pay more attention to the efficient use of inputs to facilitate a greater cost efficiency, continual cash flow, and sustainable increase in fish production.
Conference Paper
Full-text available
Article
Full-text available
Farmers in Nepal face many of the same global challenges associated with initiating and scaling poultry husbandry as many other developing countries. These include, access to innovative approaches in finance, credit, coop design, marketing and sales. As with most low-income countries, Nepalese poultry farmers also lack adequate training in poultry husbandry including biosecurity. In this paper, we describe a collaborative workshop-subsidy approach to addressing these challenges conducted by a partnership with the UC Davis School of Veterinary Medicine, the College of Engineering, the School of Education and a farming co-operative based in the semi-rural area of Bhaktapur, Nepal. The program included two workshops covering aspects of poultry rearing including coop construction, chick rearing, biosecurity and husbandry. Both workshops were a combination of lectures and hands-on learning. Following completion of the workshops, each farmer received subsidized materials for coop construction and poultry rearing. The co-operative provided training facilities and a market for selling eggs. Despite an outbreak of Highly Pathogenic Avian Influenza (HPAI), which affected the scale of program implementation, our results suggest that the workshop subsidy collaborative approach can be successful in reducing market entry barriers. Our 6-month post-workshop survey showed that two-thirds of the workshop participants ultimately built their own coop and raised chicks. Half of these participants reported market available egg production and a doubling of egg consumption at home.
Article
Full-text available
Small farm size and fragmented land are considered constraining agricultural development. This study uses the Vietnam Household Living Standard Survey 2016 (VHLSS 2016) dataset to measure the technical efficiency of rice smallholders and its determinants, including farm size, in the Mekong Delta. Data envelopment analysis was employed to examine efficiency scores in the first stage based on data of 506 paddy farms. The overall efficiency calculated through slack-based measure was low at 0.59 and the input slacks were quite large. This indicated that local farmers have not been using their resources efficiently in producing paddy. Further, farms smaller than 2 ha faced low overall efficiency at 54% and higher slacks in terms of all input types. The second-stage Tobit result showed that all types of efficiency could be improved if farmers expanded their farm size and reduced the over-use of inputs. Thus, enabling small farms to achieve economies of scale through collective farming in the Large Field Model (LFM) will be critical for upgrading production efficiency and reducing slacks as labor costs rise and natural resources are constrained. It is recommended that farmers should follow strictly eco-friendly farming packages in order to reduce their current excessive usage of seed cost by 28 USD/ha, pesticides by 61 USD/ha, and fertilizers by 155 kg/ha to reach efficient production frontier. The government needs to take measures to replicate and closely monitor climate smart agriculture programs in large-scale production to improve the overall efficiency of paddy sector, in addition to the important goal of protecting the environment and natural resources of the region. Availability of data and materials The VHLSS 2016 questionnaire section and dataset analyzed in this study are available from the corresponding author on reasonable request.
Article
Prior studies emphasize more on the importance of land titling, but do not explored the institutional effects of a specific land titling policy. Then, this paper explores the effect of land consolidation titling policy on Chinese farmers’ fertiliser use. Based on a quasi-natural experiment data, the difference-in-differences estimation shows that land consolidation titling inspires farmers to reduce chemical fertiliser use by 63.52 yuan per mu. Additionally, land consolidation titling significantly motivates farmers to use organic fertiliser by 24.29 jin per mu. Further, the mediating effect analysis shows that the change in farmers’ fertiliser use can be explained by improving the resource attributes’ value and enhancing the security of property rights, respectively. These findings reveal the importance of the policy combination of land consolidation and land reallocation, which can effectively support soil health and agricultural green development.
ResearchGate has not been able to resolve any references for this publication.