THE

ECONOMIC

ANNUAL NUMBER FEBRUARY 1962

WEEKLY

An Aspect of Indian Agriculture

Amartya Kumar Sen
The agricultural sector of our economy has recently been subjected to a great deal of empirical
investigatioriy and while much of the picture continues to remain obscure, some of it looks a whole. lot clearer.
This note is concerned with the interpretation of certain broad relationships that have been observed
in Indian agriculture, mainly in the recent Studies in Economics of Farm Management produced by the
Ministry of Food and Agriculture, for ''six typical regions in the country."
We shall draw two types of conclusions ;
(a) policy implications for Indian agriculture, and
(b)

implications about the use of certain concepts in Indian agricultural studies.

A OBSERVATIONS
WE start w i t h three specific results
of observation w h i c h are f o u n d
to be broadly v a l i d in I n d i a n agriculture. We formalise them below.
Observation I :
When
family
labour employed
in
agriculture is
given an "imputed value" in terms of
the ruling wage ratet much of Indian
agriculture
seems
unremunerative.
In fact in Bombay for the two districts studied, taken together, "we
get a loss in crop p r o d u c t i o n to the
tune of Rs 5.50 per acre '' (Bombay
Studies. 1954-55, p 6 5 ) . This was
made up of a Joss in Ahmednagar
and a profit in Nasik. The loss in
the former is reported also in the
second study for Bombay.
(Bombay
Studies, 1955-56, p 1 0 1 ) .
Similar,
if less extreme, losses are found in
all other States as w e l l , even though
they are confined to given areas or
given size groups. In Madras, however, t h e i r was an overall loss b o t h
according to the Cost Accounting
M e t h o d as well as the Survey Method for 1054-55, though the former
showed a profit of Rs 4.9 per acre
and the latter a loss of Rs 30,8 per
acre for 1955-56. (Madras Studies,
1955-56. pp 4 7 - 4 8 ) . To quote a
few more results in Punjab, "there
has been a loss per cropped acre
both in i r r i g a t e d and u n i r r i g a t e d
areas in the Survey Sample though
some profits are realised on bigger
holdings in i r r i g a t e d areas in the
Cost A c c o u n t i n g Sample''
(Punjab
Studies. 1955-56, p
242)
and in
West Bengal, " o n l y 60 per cent of
the f a r m ' have earned profits and
the r e m a i n i n g 40 per cent have i n curred losses f r o m farming.'" (West
Bengal Studies,
1955-56. p 1 4 3 ) .
There is no area studied where a
substantial
number of farms have

not made "losses" after i m p u t i n g

a wage value to f a m i l y labour.
Observation II : By and forge, the
"profitability"
of
agriculture
increases with
the
size of
Iwlding,
"profitability"
being
measured
by
the surplus (or deficit) of output
over costs
including the
imputed
value of labour. T h i s is not universally v a l i d , but w i d e l y found.
In
some areas the relationship is quite
striking, e g, in U P "the p r o f i t
per acre is highest on holdings of
20 acres and above- It is correlated
directly w i t h the size of h o l d i n g "
U P 'Studies, 1954-55. p 5 2 1 . S i m i larly in Punjab, '"the losses show a
tendency to taper off w i t h every
increase' in the size of h o l d i n g " .
(Punjab Studies, 1955-56. pp 241-2).
Similar, though less strong, correlations are seen in most other areas as
well.
Observation I I I : By and large,
productivity per acre decreases with
the size of holding. This trend w i t h
gross output per acre is observed,
more or less strongly, in practically
all the regions studied. " V a l u e added'' figures are not available, but a
rough calculation indicates a similar
trend in most of the areas studied
for "value added" per acre as welt.
B

INTERPRETATION

There are some attempts in the

Studies to explain these rather remarkable results. In general, observation 1 is regarded as a sign of
poverty of I n d i a n agriculture, though
1

It must he remembered that the

value of " bullock labour" includes
the value of human labour in its
up-keep, viz, wages of hired labour
as well as imputed value of family
labour. The same is true of items
like manure produced on the farm,
or feed cost of bullocks fed on
farm byproducts.
243

in some of the Studies the result is

just mentioned without any interpretation. Perhaps the greatest concern is expressed in the Madras
Studies. " T h i s is an a l a r m i n g situation, for if 50 per cent or more of
the fanners are c a r r y i n g on the
business at loss, the f a r m i n g comm u n i t y cannot be considered to be
comfortably placed in any sense of
the t e r m " . (Madras Studies. 195556, p 1 4 6 ) . In answering the quest i o n as to why these 50 per cent or
more are c a r r y i n g on in spite of the
loss, the answer that is given is " s t i l l
they carry on the business for the
meagre subsistence it gives, l o o k i n g
at it as a way of l i v i n g and not as a
business proposition at all.'" vop cit
p 146).
C o m i n g t o Observation I I . w e f i n d
a great variety of explanations.
Punjab gets a technical answer.
" T h i s f a l l in the losses on higher
holdings may be a t t r i b u t e d to proportionately more area under American cotton than on smaller holdings. American cotton is a very
remunerative crop and yields positive net returns''.
(Punjab Studies.
1955-56 p 6 0 ) . This treats the
phenomenon as a special feature of
Punjab agriculture and explains it
in terms of a consideration that is
quite regional, viz. American cotton
which is grown in P u n j a b l a n d
w h i c h , we also gather, is so " v e r y
remunerative'' that it yields- "posit i v e " net r e t u r n s ! ) . The U P Report takes a broader line, and refers to the i n d i v i s i b i l i t y of certain
capital goods and the consequent
economy of large scale, in t r y i n g to
explain the h i g h cost of production
per acre of smaller farms. " T h e
reason may be that the smaller sized
holdings have to m a i n t a i n at bast a
p a i r of bullocks and a set of imple-

A N N U A L NUMBER FEBRUARY 1962

THE

244

ECONOMlC

WEEKLY

A N N U A L NUMBER

THE ECONOMIC WEEKLY

menta necessary f o r f a r m operations
a n d hence the expenses per acre are
highest in this g r o u p . " (U P Studies,
1954-55, p 5 1 ) . Madras, however,
gets a different explanation. "There
is a persistent tendency to raise the
q u a n t u m of inputs, especially in the
smaller h o l d i n g s and i n the i r r i g a t e d
holdings in the hope that output also
w i l l increase, i f not proportionately,
at least absolutely, so that product i o n o n w h i c h c u l t i v a t o r s i n general
w i l l concentrate in a system of subsistence f a r m i n g , m a y be the m a x i mum.
But the response of output
is hardly within their control, depending as it does largely on favourable n a t u r a l c o n d i t i o n s . '
(Madras
Studies, 1955-56 p 56, p 146; italics
m i n e ) . O r i n other words, the
farmers misjudge w h a t they can
achieve by p u t t i n g more and more
inputs and end up in losses and this
tendency to misjudge is stronger for
smaller holdings.
Observation I I I does not seem to
puzzle anybody, p a r t i c u l a r l y since it
is seen that increased o u t p u t is related to increased inputs i n c l u d i n g
h u m a n labour.
T h e smaller farms
get more i n p u t s per acre (at i m puted m a r k e t value) and therefore
get more o u t p u t as w e l l . This answers the question by postponing i t ,
and the question w h y i n p u t per
acre falls as the size of the f a r m i n creases, remains.
Of course, there
are some special explanations, e g
the U P argument about i n d i v i s i b i l i t y of bullocks, etc, m a k i n g for a
larger i n p u t of such goods per acre
for smaller farms. B u t this does not
e x p l a i n w h y other i n p u t s per acre
also rise when the size falls, p a r t i cularly the a m o u n t of h u m a n labour
applied. T h e total a m o u n t o f f a m i l y
labour applied per acre goes up rem a r k a b l y as the size falls, so that
i n spite o f the fact that i n some
areas the amount of wage labour
applied falls as the size gets smaller,
the total amount of labour per acre
is inversely correlated w i t h the size
of the f a r m .
Before it is attempted to p r o v i d e
an interrelated set of explanations
to
all the three observations, it
m i g h t be useful to consider an extremely simple economic p i c t u r e .
I n the D i a g r a m
curve M P represents the schedule of net marg i n a l p r o d u c t o f l a b o u r applied
to a given acre of l a n d . If cultivat i o n is family-based, and if the per

acre supply o f labour i n the f a m i l y

equals or exceeds OP, we can expect
t h a t O P amount o f labour w i l l b e
applied, p r o v i d e d labour has no outside o p p o r t u n i t y of employment and
provided there is no significant disu t i l i t y of w o r k in the relevant range
of effort.
Next imagine that the
wage rate r u l i n g in the area is O W .
If we i m p u t e OW wage to each u n i t
of OP labour, O W A P is the total
"cost" of labour.

Lobour

Units

N o w let us look at the observations needing interpretation. The

area O M P (net o u t p u t ) could be
more or less t h a n area O W A P
(labour c o s t ) . I f i t i s more, there
is a p r o f i t ; if less, a loss. There is
no p a r t i c u l a r reason w h y it should
be more, and so there is n o t h i n g in
the least s u r p r i s i n g i n Observation I .
It is a natural result of an economy
w i t h surplus labour w h i c h has
f a m i l y based non-wage c u l t i v a t i o n .
There is n o t h i n g p a r t i c u l a r l y " a l a r m i n g " i n this, and i n fact, i f product i o n were restricted to OC, where a
p r o f i t must necessarily be earned,
the overall position w o u l d have been
m u c h worse.
Next, i t should b e noted that i n
practically a l l the areas studied, the
p r o p o r t i o n o f f a m i l y labour t o h i r e d
labour falls w i t h the size of farms,
as indeed one w o u l d expect.
The
tables on this field in all the Studies
indicate a weak or a strong tendency towards this relationship. I n cidentally, the U P Studies f o r 195455 states, rather curiously, that "the
use of f a m i l y labour in various size
groups differs w i d e l y and does not
follow any pattern i n relation t o
size". (U P Studies, 1 9 5 4 5 5 p 5 1 ) .
It is stated elsewhere that " t h e cont r i b u t i o n of f a m i l y labour to the
total f a r m labour is the highest in the
lower size groups", (op cit p 3 7 ) .
246

FEBRUARY 1962

A look at T a b l e 4.10 on page 55

dealing w i t h this data confirms the
latter statement completely, and also
shows that the p r o p o r t i o n of f a m i l y
labour falls monotonically ( w i t h one
small exception in the Cost Accounti n g Sample a n d none i n the Survey
Sample) w i t h the increase in size.
Given this relationship, Observations I I arid I I I are the obvious ones
to expect. If the wage rate is O W,
a wage-based f a r m
will
restrict
labour i n p u t to O C, and have more
p r o f i t (area M W B as opposed to
area M W B minus area B A P ) compared w i t h family-based f a r m i n g .
Sinre size is positively correlated
w i t h wage-based f a r m i n g , Observations I I and I I I are immediately explained-.
This is not to say that the one
feature we have been discussing is
the only relevant one. A m e r i c a n cotton i n Punjab i s relevant, i n d i v i s i b i l i t y of bullocks is i m p o r t a n t , and
so are a lot of other factors that are
found i n I n d i a n agriculture. But
what can be claimed is that even
if these other special features were
not there,
the broad qualitative
observations made
earlier can be
expected, given the mode of production of I n d i a n agriculture and its
variation w i t h the size of the farms.
This is? of course, not to deny that
the other special features must be
brought in to e x p l a i n the exact quantitative observations. No agricultural study, however, can go very f a r
i n I n d i a , i f i t does not focus attent i o n on the systems of p r o d u c t i o n
u n d e r l y i n g I n d i a n agriculture.
C

IMPLICATIONS

A n obvious i m p l i c a t i o n o f a l l this
is that the fact that smaller f a r m s
2

It should be noted that while the

amount of family labour applied per
acre is much greater for smaller
farms, the amount of wage labour
applied per acre is not, in many
cases. The increase in the application of other inputs per acre partly
reflects complementarity between labour and other inputs. The greater
value of investment per acre for
smaller farms can he expected partly
for this reason, and partly for the
reason that much of the investment
for smaller farms is done through
the direct application of family labour, where a pattern similar to
the use of recurring labour can be
expected. Naturally this one phenomenon does not explain everything,
and a lot of other considerations
must be brought in to explain the
input pattern satisfactorily.

A N N U A L NUMBER FEBRUARY 1962

have
greater "losses" ( o r smaller
" p r o f i t s " ) in the sense discussed, is
no sign of inefficiency of small-scale
f a n n i n g . As an indicator of efficiency, the surplus of o u t p u t over
all "cost", i n c l u d i n g the i m p u t e d
wage of f a m i l y labour, is w o r t h very
little.
In view of the fact that smaller
farms have higher o u t p u t per acre,
can we argue that the reverse is
true, viz. that small-scale f a r m i n g is
more productive ? But that too does
not seem' to follow necessarily. If
the explanation chosen here is correct, the factor that makes the crucial difference is not size as such,
w h i c h is incidental, hut the system
of f a r m i n g . viz. whether it is wagebased or f a m i l y based.
F o r example, if a large cooperative f a r m
operates on a non-wage f a m i l y labour
basis, there is n o t h i n g in our observations to indicate that it w i l l have
a lower o u t p u t per acre. In fact in
I n d i a n private-enterprise agriculture,
a whole lot of subsidiary operations
are wage-based, e g: b u i l d i n g canals,

THE ECONOMIC WEEKLY

dams, etc. A cooperative f a r m m i g h t
be able to include some of these
activities
within
the
non-wage
sphere, at least upto a pointy w i t h a
corresponding gain in efficiency.
On a s i m i l a r trend of thought, the
growth
of
wage-based
capitalist
f a r m i n g , as is noted in some parts
of India, should not be necessarily
viewed as a progressive trend, because in an economy w i t h structural
unemployment.
non-wage f a m i l y based f a r m i n g ' h a s some efficiency
advantages
that capitalist f a r m i n g
does not have, as we have seen above.

Barren Economic Concepts

Finally- a remark on the implications for
the concepts used in the
study of I n d i a n agriculture. W h a t
is the point of i m p u t i n g to f a m i l y
labour a wage equal to the market
one? What does the wage rate represent in this context? It does not
represent the m a r g i n a l social opport u n i t y cost of labour, for there may
be no alternative employment opportunities at the m a r g i n . Nor can it

246

be j u s t i f i e d in terms of extra con'

s u m p t i o n i n v o l v e d that is relevant
for long-run g r o w t h (cf, D o b b ,
Galenson and L e i b e n s t e i n ) , because
no extra consumption is induced
merely by the a p p l i c a t i o n of a
marginal
u n i t o f extra f a m i l y
labour,
except perhaps in
the
shape
of extra
requirement of
calories f o r extra
effort, w h i c h
in any case is not measured by
the r u l i n g wage rate in the labour market. W i t h positive propensities to save of the peasants, a
policy that maximises income also
maximises savings. So the use of
the wage rate in this context is not
at all easy to understand. As economic concepts, "profits"' and "loss'
es" over total costs i n c l u d i n g the
cost of f a m i l y labour valued at the
market wage rate, seem to be q u i t e
barren. This is j u s t an example of
how ideas relevant to one economic
system when
i n d i s c r i m i n a t e l y app l i e d to another, produce quite
absurd results.