Ind. Jn. of Agri. Econ.

Vol.69, No.3, July-Sept. 2014

SUBJECT II
LABOUR SCARCITY IN AGRICULTURE AND MECHANISATION

Farm Mechanisation, MGNREGS and Labour Supply Nexus:

A State-Wise Panel Data Analysis on Paddy and Wheat Crop
A. Narayanamoorthy*, M. Bhattarai**, R. Suresh* and P. Alli†
ABSTRACT

An attempt has been made in this study to find out the relationship among the farm mechanisation,
Mahatma Gandhi National Rural Employment Guarantee Scheme (MGNREGS), labour supply and other
factors mainly using state-wise data pertaining to paddy and wheat crops covering the period from 2000-
01 to 2010-11. To measure the regression of various growth factors including MGNREGS on the use of
farm machineries, regressions is computed using panel data with fixed effects models. The descriptive
analysis of the study shows that the machine labour cost in real value (which is used as a proxy variable to
reflect the level of farm mechanisation) incurred for cultivating both paddy and wheat has increased
considerably during post MGNREGS period in almost all the states selected for the analysis. The machine
labour cost incurred for cultivating paddy has increased substantially in states like Andhra Pradesh, Tamil
Nadu, Karnataka and Madhya Pradesh after the implementation of MGNREGS, while the same increase
was found very high in Madhya Pradesh, Himachal Pradesh, Uttar Pradesh and Punjab in wheat
cultivation. In most states where the machine labour cost has increased substantially, the use of human
labour in man-hours has declined sharply in both paddy and wheat, confirming the fact that farm
machineries are used to substitute the human labour especially after implementing MGNREGS. The
regression results computed using panel data suggest that the factors determining the use of farm
machineries is not the same between the two major crops selected for the study. Besides MGNREGS
dummy, the factors such as coverage of irrigation, yield enhancing inputs cost, land-labour ratio and
human labour use in man-hours have significantly influenced the use of machine labour in paddy
cultivation. But, in the case of wheat crop, irrigation coverage and land-labour ratio has not significantly
influenced the use of machineries. The MGNREGS dummy used to capture its impact on farm
mechanisation has turned out to be positive and significant in both paddy and wheat cultivation suggesting
that the level of farm mechanisation has increased after its implementation of national rural employment
guarantee scheme.
Keywords: Farm mechanisation, Rural employment, Farm wages.
JEL: J3, E24, Q16

INTRODUCTION

The major objective of this study is to find out whether any nexus exists among
the farm mechanisation, Mahatma Gandhi National Rural Employment Guarantee

*NABARD Chair, Professor and Head and Ph.D., Fellow respectively, Department of Economics and Rural
Development, Alagappa University, Karaikudi (Tamil Nadu), **Principal Scientists (Economics), Markets,
Institutions and Policies, International Crops Research Institute for Semi-Arid Tropics (ICRISAT), Patancheru-502
324, Hyderabad (Andhra Pradesh) and †Assistant Professor, Economics Division, Vellore Institute Technology,
Vellore-632 014 (Tamil Nadu), respectively.
 FARM MECHANISATION, MGNREGS AND LABOUR SUPPLY NEXUS 321

Scheme (MGNREGS), labour supply and other factors across the major states in
India. More specifically, it tries to find out the relationship between farm
mechanisation and MGNREGS, which has reportedly increased the farm wage rate
by creating artificial demand for labour in rural areas (Gulati et al., 2013,
Narayanamoorthy and Bhattarai, 2013). It is well known that the advent of new
agricultural technology during the mid-sixties has brought many changes in Indian
agriculture. Besides changing the land use pattern, it has increased the adoption of
modern inputs (HYV seeds, fertilisers, pesticides, etc.), changed the cropping pattern
from low value to high value crops and also considerably improved the cropping
intensity. As the modern agriculture is intensive that warrants the field operations to
be completed in time, the use of machineries became necessary for farming. This is
also confirmed by the use of tractors (an important constituent in farm machineries)
that have increased from just three per 1000 hectares in 1962-65 to 167 per 1000
hectares in 2005-08 at the all India level (Bhalla and Singh, 2012). But, since India is
a labour abundant country, it is often argued that the increased use of machineries in
farm operations is not justifiable as it will reduce the employment opportunities in
agriculture (for discussion on this see, Sidhu and Grewal, 1990). Farm machineries
are used not necessarily to reduce the use of human labour but mostly to produce
more output per unit of land, to complement other inputs use, achieve higher
productivity and to reduce the post-harvest losses in agriculture. It is often
misunderstood that the farm machineries generally reduce the employment in
agriculture, which is not correct (Sidhu and Grewal, 1990). In fact, farm machineries
are different in nature; some are labour augmenting, while others are labour
displacing (see Binswanger, 1978).
Researchers across time and space have critically analysed the possible outcome
of farm mechanisation using aggregated and disaggregated data. While mechanisation
has been well received all over the world as one of the important elements of
modernisation of agriculture, there exist certain varied compelling factors that
accelerates the pace of farm mechanisation in the Indian context. It has been
identified that cost-efficiency consideration, faster ploughing operations during the
crop seasons, the increased adoption of high-yielding varieties, relative scarcity of
labour during peak agricultural operations, sharp rise in wages, ready availability of
cheap institutional credit and availability of off and on farm custom hire possibilities
as the contributing factors for the increased farm mechanisation (Rao, 1972; Grewal
and Kahlon, 1972; Binswanger, 1978, 1986; Jose, 1984; Bhalla, 1995; FAO, 2013).
With respect to the impact of farm mechanisation on farm operations, it is the labour
displacement effect of farm mechanisation that has come up vividly in most of the
studies (Rudra, 1971; Rao, 1972, 1974; 1975; Parthasarathy and Abraham, 1975;
Government of India, 1975; Mcinernery and Donaldson, 1975; Kahlon, 1976;
Dasgupta, 1977; Binswanger, 1978; Vaidyanathan, 1978; Roy and Blasé, 1978;
NCAER, 1980; Agarwal, 1981; Zarkovic, 1987; Bhalla,1991; Narayanamoorthy,
1992). Studies have also corroborated the positive effects of farm mechanisation on
322 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

income generation, cropping intensity, productivity of crops, land augmentation and

total factor productivity (Brian, 1972; Government of India, 2006; Reid, 2011).
However, the validity of the positive findings of mechanisation have been questioned
by some scholars (Binswanger, 1978, Agarwal, 1983; Alagh, 2004). Quite a few
studies have also found a close nexus between farm size and mechanisation in
agriculture (Agarwal, 1981; Sidhu and Grewal, 1990).
In the recent years it is being argued that the implementation of MGNREGS has
increased the labour shortage and farm wage rate in the major foodgrains producing
states which have forced the farmers to resort to mechanized farming (Reddy, 2012).
To our knowledge no systematic study is available that could cross-check the validity
of the claim that MGNREGS induces farm mechanisation. Has the farm
mechanisation accelerated at a greater pace after implementing MGNREGS? Farm
mechanisation is determined by a host of factors like labour supply, wage rate, labour
use by man-hours, bullock labour use, productivity and profitability of crops, source
and coverage of irrigation, cropping intensity, land labour ratio, output price and
pattern of cultivation. One must consider all these factors along with NREGS variable
to find out the important determinants of farm mechanisation. Moreover, the existing
studies have mostly used availability of tractors or other machineries to cropped area
as the variable to capture the influence of farm mechanisation (see, Bhalla and Singh,
2012). But the availability of machineries does not always reflect its actual use in
crops cultivation. The real farm mechanisation can be studied either by taking the
cost incurred on machine labour or hours of use of machineries in crops cultivation.
In spite of proliferation of studies on farm mechanisation, not many detailed studies
are available in recent years covering all these issues raised above. Keeping this in
view, an attempt is made in this study to find out the relationship among the farm
mechanisation, MGNREGS as well as other factors using state-wise data relating to
paddy and wheat crops with the following objectives: (1) To find out whether
machine labour cost (used as a proxy variable to reflect the level of farm
mechanisation) increased in the case of paddy and wheat crops across the major states
after the implementation of MGNREGS, (2) To find out the relationship between
rural labour supply and the farm mechanisation in paddy and wheat crops across the
major states and (3) To measure the influence of various growth variables including
MGNREGS on the use of farm machineries in paddy and wheat crops using
regression analysis with panel data.

II

DATA AND METHODOLOGY

The study has utilised state-wise data of two major crops, namely, paddy and
wheat covering the period from 2000-01 to 2010-11. With 11 important states for
paddy crop and eight states for wheat crop, a panel data of 121 observations (11 years
x 11 states) for paddy and 88 observations (11 years x 8 states) for wheat have been
 FARM MECHANISATION, MGNREGS AND LABOUR SUPPLY NEXUS 323

considered altogether for the analysis. As the study aims more specifically to find out
the nexus among the farm mechanisation, NREGS and other factors including labour
supply, a total of 14 variables that one way or the other determine the use of farm
machineries have been considered for the analysis. The variables considered for the
analysis including its sources are presented in Table 1. Of the total variables
considered, variables such as BLC, HLC, HLMH, MLC (which is used as proxy
variable to farm mechanisation), output price realised, profit and YEIC have been
compiled from the cost of cultivation survey data published by CACP, while other
variables namely CI, GIA/GCA, LLR and WR have been collected mainly from
different publications of the Ministry of Agriculture, Government of India. Machine
labour cost is used as a proxy variable to reflect the use of farm mechanisation in this
study. As one of the objectives of the study is to find out the real change in the use of
machine labour cost after the implementation of MGNREGS, all costs and income
related data have been converted into real value with CPIAL base year of 1986-87. In
order to capture the effect of MGNREGS on farm mechanisation in the two selected
crops, the data period has been divided into two as pre-MGNREGS (2000-01 to
2005-06) and post-MGNREGS period (2006-07 to 2010-11). Besides descriptive
analysis, correlation and regression analysis have been carried out to trace the
determinants of farm mechanisation including the effect of NREGS.

TABLE 1. DEFINITION OF VARIABLES USED IN THE STUDY AND THEIR

EXPECTED RELATIONSHIP WITH MACHINE LABOUR COST

Expected
relationship
Variables Description of the variables Unit Source of data with MLC
(1) (2) (3) (4) (5)
HLC Human labour cost (Rs./ha) CACP -
BLC Bullock labour cost (Rs./ha) CACP -
YEIC Yield enhancing input cost (Rs./ha) CACP +
Yield Productivity (Qtl/ha) CACP
HLMH Human labour use (Man-hours) CACP -
PR Price realised (Rs./Qtl) CACP +
Profitt-1 Profit with one year lag (Rs./ha) CACP +
Profit Profit in current year (Rs./ha) CACP +
WR Male wage rate (Rs/.day) www.dacnet.nic.in +
CI Cropping intensity (Per cent) www.dacnet.nic.in +
GIA/GCA Irrigated area to cropped area (Per cent) www.dacnet.nic.in +
LLR Land labour ratio (ha) www.dacnet.nic.in +
NREGS Dummy variable Before = 0; After = 1 ---- +
Notes: CACP – Commission for Agricultural Costs and Prices; MLC – Machine labour cost.

For regression analysis, we have used fixed effects form of panel model to
quantify the marginal impact of factor-inputs on spatial and temporal variation of
machine labour cost in paddy and wheat. The fixed effects form of regression
modeling allows separating intercept of each cross-section unit (state) of India by
controlling the state-specific institutional and structural constraints affecting the use
of farm machineries. More particularly, we have used fixed effects form of panel
324 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

model with robust form of error correction by using STATA software, with VCE
option of fixed effect regression model. The VCE option provides the robust
estimators and efficient parameter estimates by correcting the heterogeneity, scale,
and size effects across the cross-section units (see, STATA version 12 manual;
Greene, 2011). In the context of machine labour use, some variables inherently are
time invariant in nature such as state specific agricultural machinery markets, state
governments agricultural and agro-machinery financing policies, development of
rural financing, etc. In such context, fixed effects form of panel model is preferred
(for details on the use of fixed effects model see, Greene, 2011). The regression
model estimated is shown below in equation (1):
MLCit = ait + b1HLCit + b2BLCit, + B3YEICit,+ b4HLMHit, + b5PRit, + b6Profitit-1, +
b7CIit, + b8GIA/GCAit, + b9LLRit, + b10NREGSDit + eit ….(1)
where; i = 1,…n states of India
t = year in number from 1 to 11(121 observations for paddy; 88 observations for
wheat)
MLCit = Machine labour cost in paddy/wheat in Rs./ha
HLCit = Human labour cost in Rs./ha
BLCit = Bullock labour cost in Rs./ha
YEICit = Yield enhancing inputs cost in Rs./ha
HLMHit = Human labour in man-hours/ha
PRit = Output price realised in Rs./qtl
Profitit-1 = Profit with one year lag in Rs./ha
CIit = Cropping intensity in percentage
GIA/GCAit = Irrigated area to cropped area in percentage
LLRit = Land labour ratio in ha
NREGSDit = Dummy variable (before implementation = 0; after implementation
= 1)
III

RESULTS AND DISCUSSION

Farm Mechanisation Before and After NREGS

It is argued vehemently that after the implementation of the national rural

employment scheme, the use of machineries in various crops cultivation has
increased to tackle the increased wage rate and labour scarcity artificially created by
this scheme (for details see, Narayanamoorthy and Bhattarai, 2013). Since there is a
complete absence of information from the literature on this issue, we have made an
attempt in this section to find out how far this argument is correct. As mentioned
earlier, the actual cost incurred on machine labour is used as the proxy variable to
reflect the intensity in the use of mechanisation in paddy and wheat cultivation.
 FARM MECHANISATION, MGNREGS AND LABOUR SUPPLY NEXUS 325

Besides studying the level of increase in the real cost (at 1986-87 prices) of machine
labour and its growth rate, we have tried to capture the changes in the share of
machine labour cost in the gross cost of cultivation (Cost A2) during pre- and post-
MGNREGS period.
Table 2 presents the details on machine labour cost by states before and after
implementing NREGS for both paddy and wheat. Let us first study the machine
labour cost incurred for paddy cultivation. It is clear that the real MLC in paddy
cultivation has increased considerably after the introduction of MGNREGS in 10 out
of 11 states considered for the analysis. In Punjab, which is all along a forefront state
in using farm machineries, the real MLC has declined marginally from Rs. 929/ha to
Rs. 892/ha between the two periods. As expected, the pattern of increase in MLC is
not the same across the states. Except for Andhra Pradesh and Tamil Nadu, the
absolute increase in real MLC between pre and post-MGNREGS was only in the
range of Rs. 32-124/ha in all other states. But, the increase was substantial in Andhra
Pradesh (Rs. 341/ha) and Tamil Nadu (Rs. 322/ha). This large increase in these two
states could be due to increased labour scarcity after the introduction of MGNREGS
which is also corroborated by earlier studies (see Gulati et al., 2013). Although the
MLC in real value has increased in almost all the states, its growth rate is not
appreciable in most of the states after MGNREGS as compared to its previous period.
However, the share of machine labour cost in the gross cost of cultivation (cost A2)
has increased in all the states after the introduction of MGNREGS. All these seem to
suggest an increase in the adoption of farm mechanisation in paddy cultivation after
implementing MGNREGS.
Similar to paddy, the real machine labour cost in wheat crop too has increased in
most states after implementing MGNREGS. Except for Bihar where the MLC has
declined marginally, it has increased in the range of Rs. 63-175/ha in different states.
As expected, the extent of increase in MLC is not the same among the eight states
considered for the analysis in wheat crop as well. MLC has increased by Rs. 175/ha
in Madhya Pradesh and Rs. 152/ha Himachal Pradesh, whereas it increased only in
the range of Rs. 63-82/ha in Gujarat and Rajasthan. Punjab and Uttar Pradesh have
also incurred more cost on machine labour in wheat cultivation after implementing
MGNREGS, but the same was not true in the case of paddy crop. This implies that
the level of mechanisation varies from crop to crop as reported earlier. The increased
MLC during MGNREGS period is also reflected through its share in cost A2 which
expanded considerably in all the states except for Bihar. Although the machine labour
cost incurred for the cultivation of paddy and wheat has increased in almost all the
states after implementing MGNREGS, one cannot say candidly that it is an
MGNREGS induced farm mechanisation. The adoption of machineries in crop
cultivation is determined by a number of factors where MGNREGS is one of among
the variables. Therefore, one must carry out a comprehensive analysis by taking into
account all the relevant variables to find out the influence of MGNREGS on farm
mechanisation, which is done in the following section.
326 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

TABLE 2. STATE-WISE MACHINE LABOUR COST INCURRED FOR CULTIVATING PADDY AND
WHEAT DURING PRE AND POST-NREGS PERIOD
(values in Rs./ha at 1986-87 prices)
Paddy Wheat
Pre- Post- Pre- Post-
NREGS NREGS All NREGS NREGS All
States Particulars period period period period period period
(1) (2) (3) (4) (5) (6) (7) (8)
Andhra MLC (Rs./ha) 550 891 705 ---- ---- ----
Pradesh Per cent MLC to A2 Cost 11.56 17.06 14.19 ---- ---- ----
CGR of MLC 4.24 7.50 6.46 ---- ---- ----
Assam MLC (Rs./ha) 55 123 86 ---- ---- ----
Per cent MLC to A2 Cost 3.00 6.51 4.63 ---- ---- ----
CGR of MLC -2.49 33.30 10.96 ---- ---- ----
Bihar MLC (Rs./ha) 291 329 308 633 629 631
Per cent MLC to A2 Cost 12.88 15.09 13.87 23.83 21.91 22.92
CGR of MLC 11.05 -0.74 2.62 1.57 -0.50 -0.46
Gujarat MLC (Rs./ha) ---- ---- ---- 552 615 581
Per cent MLC to A2 Cost ---- ---- ---- 16.49 20.10 18.05
CGR of MLC ---- ---- ---- 2.76 -0.86 2.98
Haryana MLC (Rs./ha) 697 732 713 1012 1084 1045
Per cent MLC to A2 Cost 15.35 17.46 16.27 28.12 31.79 29.74
CGR of MLC 8.12 -6.92 0.82 6.38 -1.46 1.59
Himachal MLC (Rs./ha) ---- ---- ---- 448 600 517
Pradesh Per cent MLC to A2 Cost ---- ---- ---- 28.65 33.70 31.11
CGR of MLC ---- ---- ---- 20.51 -7.68 7.18
Karnataka MLC (Rs./ha) 642 755 694 ---- ---- ----
Per cent MLC to A2 Cost 11.48 16.72 13.58 ---- ---- ----
CGR of MLC 1.79 8.00 3.72 ---- ---- ----
Madhya MLC (Rs./ha) 118 242 174 426 601 506
Pradesh Per cent MLC to A2 Cost 6.18 12.38 9.04 18.98 25.87 22.17
CGR of MLC 24.67 2.68 20.30 7.99 4.63 6.26
Orissa MLC (Rs./ha) 121 153 136 ---- ---- ----
Per cent MLC to A2 Cost 4.37 5.57 4.92 ---- ---- ----
CGR of MLC 12.42 0.44 6.70 ---- ---- ----
Punjab MLC (Rs./ha) 929 892 912 1064 1169 1112
Per cent MLC to A2 Cost 19.35 19.46 19.40 26.95 30.81 28.67
CGR of MLC 0.87 0.32 0.13 3.67 -0.60 1.73
Rajasthan MLC (Rs./ha) ---- ---- ---- 606 688 643
Per cent MLC to A2 Cost ---- ---- ---- 20.05 23.75 21.69
CGR of MLC ---- ---- ---- 2.90 -4.55 0.20
Tamil Nadu MLC (Rs./ha) 872 1194 1018 ---- ---- ----
Per cent MLC to A2 Cost 14.95 21.21 17.74 ---- ---- ----
CGR of MLC 8.07 1.69 5.96 ---- ---- ----
Uttar MLC (Rs./ha) 404 456 427 795 916 850
Pradesh Per cent MLC to A2 Cost 14.45 15.95 15.14 25.17 26.80 25.94
CGR of MLC 4.53 6.76 3.63 4.44 -2.10 2.56
West Bengal MLC (Rs./ha) 195 250 220 ---- ---- ----
Per cent MLC to A2 Cost 5.28 6.60 5.89 ---- ---- ----
CGR of MLC 4.16 9.57 7.26 ---- ---- ----
Source: Computed using data from CACP (various years).
Notes: MLC-Machine labour cost; CGR-Compound growth rate in percent per annum.
 FARM MECHANISATION, MGNREGS AND LABOUR SUPPLY NEXUS 327

Human Labour Use Before and After MGNREGS

Farm mechanisation and use of human labour have a close relationship. Increased
use of farm machineries in agricultural operations tends to reduce the use of human
labour in any crop, be it foodgrain crops or others. We have also seen above that the
costs incurred on machine labour for cultivating paddy and wheat have increased in
almost all the states after the introduction of national rural employment guarantee
scheme. Going by this, the use of human labour for both paddy and wheat must have
reduced in all the states considered for the analysis. In order to study this issue, we
have computed human labour use by man hours before and after the introduction of
MGNREGS for both paddy and wheat (see Table 3).

TABLE 3. STATE-WISE HUMAN LABOUR USE FOR PADDY AND WHEAT CULTIVATION,
PRE- AND POST-MGNREGS PERIOD
(man-hours/ha)
Paddy Wheat
Pre- Post- Pre- Post-
NREGS NREGS Per cent NREGS NREGS Per cent
States period period change period period change
(1) (2) (3) (4) (5) (6) (7)
Andhra Pradesh 975 809 -17.01 snca snca snca
Assam 732 696 -4.90 snca snca snca
Bihar 855 768 -10.12 474 416 -12.26
Gujarat snca snca snca 610 430 -29.53
Haryana 614 612 -0.36 307 300 -2.39
Himachal Pradesh snca snca snca 289 253 -12.44
Karnataka 1258 1087 -13.64 snca snca snca
Madhya Pradesh 626 556 -11.20 356 317 -10.95
Orissa 1085 1053 -2.87 snca snca snca
Punjab 456 417 -8.46 218 188 -13.85
Rajasthan snca snca snca 526 476 -15.33
Tamil Nadu 939 809 -13.86 snca snca snca
Uttar Pradesh 861 835 -3.07 468 473 +1.08
West Bengal 1199 1230 +2.61 snca snca snca
Sources: Computed using data from CACP (various years).
Note: snca – state not considered for analysis.

As expected, the use of human labour has declined considerably in all other states
after implementing MGNREGS except for one state each in paddy and wheat.
Reduction in human labour use is found to be very high in those states where the cost
incurred for machine labour is higher. For instance, the farmers from Andhra Pradesh
and Tamil Nadu have incurred higher MLC for cultivating paddy crop after
implementing MGNREGS, which has directly impacted on the use of human labour
in these two states. Human labour use in man-hours has declined by about 17 per
cent/ha in Andhra Pradesh (from 975 to 809 man-hours/ha), whereas the same has
declined by about 14 per cent/ha in Tamil Nadu (939 to 809 man-hours/ha).
Similarly, the use of human labour for cultivating wheat crop has declined
considerably during post-MGNREGS period in states like Himachal Pradesh,
Madhya Pradesh, Punjab and Rajasthan. All these states have also incurred much
328 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

higher cost on machine labour after implementing MGNREGS. The results on the
whole suggest that the increased farm mechanisation has declined the human labour
use in almost all the states cultivating paddy and wheat after the implementation of
MGNREGS.

Determinants of Farm Mechanisation

The descriptive analysis carried out above shows an increased use of farm
mechanisation after implementing MGNREGS. However, one cannot say that this
increase in farm mechanisation is only due to MGNREGS as many supply and
demand factors relating to agricultural sector play crucial role in determining it.
Therefore, as reported in methodology section, in order to find out the influence of
various factors including MGNREGS, we have computed correlation and regression
(panel data with fixed effects model) by taking into account the important variables
that are expected to have relationship with the adoption of farm mechanisation in
paddy and wheat cultivation. As the correlation value explains the one to one
relationship between MLC and other associated variables, let us first study
correlation before getting into the analysis of regression results.
The correlation value between MLC and 13 other variables presented in Table 4
shows that the intensity of use of farm machineries vary considerably from one crop
to another. Except two variables in both paddy and wheat, all other variables have
shown significant relationship with MLC. Factors such as bullock labour cost, yield
enhancing inputs cost, productivity, profit, male wage rate and irrigation coverage to
cropped area are highly correlated with MLC in paddy crop, whereas in the case of
wheat crop, the variables such as yield, cropping intensity and irrigation coverage are

TABLE 4. CORRELATION VALUES: MACHINE LABOUR COST WITH OTHER SELECTED VARIABLES

Paddy Wheat
(No of (No. of
Variables Description of the variables Unit observations 121) observations 88)
(1) (2) (3) (40 (5)
HLC Human labour cost (Rs./ha) 0.399a 0.142d
BLC Bullock labour cost (Rs./ha) -0.675a -0.340a
YEIC Yield enhancing inputs cost (Rs./ha) 0.816a 0.431a
Yield Productivity (qtl./ha) 0.760a 0.785a
HLMH Human labour man hours (hrs./ha) -0.242a -0.406a
PR Price realised (Rs./qtl.) 0.434a -0.203b
Profit t-1 Profit with one year lag (Rs./ha) 0.658a 0.491a
Profit Profit in current year (Rs./ha) 0.647a 0.489a
WR Male wage rate (Rs./day) 0.523a 0.057ns
CI Cropping intensity (Per cent) -0.023ns 0.650a
GIA/GCA Irrigated area to cropped area (Per cent) 0.546a 0.881a
LLR Land labour ratio (ha) 0.225ns 0.008ns
NREGS Dummy Before = 0; After = 1 ---- 0.155c 0.196c
Sources: Computed using data from CACP (various years) and www.dacnet.nic.in.
Notes: a, b, c and d are significant at 1 per cent, 5 per cent, 10 per cent, 20 per cent level, respectively; ns-not
significant.
 FARM MECHANISATION, MGNREGS AND LABOUR SUPPLY NEXUS 329

closely associated with MLC. Cropping intensity (CI) was expected to have positive
correlation with MLC in paddy crop, but it turned out to be negative with
insignificant value. Similarly, for wheat crop, wage rate of male which is used to
capture the overall wage rate of the state has showed insignificant correlation with
MLC, which is not expected. Surprisingly, HLC is positively associated with MLC in
paddy and wheat, instead of a negative relationship. With this understanding of
correlation between MLC and other variables, let us now focus on the regression
results.
As mentioned in the methodology section, we have estimated panel data
regression with fixed-effects model to quantify the marginal impact of various factors
on spatial and temporal variation of machine labour cost of paddy and wheat across
the selected states. The fixed-effects form of regression modelling allows us for a
separate intercept for each cross-section unit (state) by controlling the state-specific
institutional and structural constraints affecting farm mechanisation, which is not
possible through simple OLS (ordinary least squares) regression model. As
demonstrated earlier, the factors that control the adoption of mechanisation in paddy
cultivation may not be necessarily the same with wheat cultivation in any given
region. In view of this, we have estimated regression separately for paddy and wheat.
One of our objectives is to capture the influence of human labour use on MLC and
therefore, two models of regression have been estimated: one with HLC and another
with HLMH.
The regression results estimated using panel data with fixed effects model on the
determinants of machine labour cost in paddy cultivation are presented in Table 5.
The higher adjusted R2 estimated for paddy crop indicates that the machine labour
use in paddy is better explained by the independent variables used in the model. Of
the nine variables used in the regression model (1), variables such as YEIC,
GIA/GCA, LLR and MGNREGS dummy have positively and significantly
influenced the machine labour cost in paddy cultivation, while CI has negatively and
significantly influenced MLC. Increased cost on human labour was generally
expected to reduce the cost of MLC, but HLC has negatively and significantly
influence it. One of the major objectives of the study is to find out the influence of
labour supply and MGNREGS on MLC. The land-labour ratio which is used to
capture the impact of labour supply has positively and significantly influenced,
suggesting that wherever the supply of labour to cropped area is less the farmers tend
to adopt more machineries in paddy cultivation. The positive and significant
coefficient of MGNREGS dummy shows that the use of machineries in paddy
cultivation has increased significantly after its introduction. Among all the variables
used in model (1), the percentage of irrigation to cropped area seems to be
influencing the MLC more than any other variable which is evident from its higher
elasticity (1.38) value. This is plausible because of the fact that irrigated areas
generally practice intensive agriculture, where increased use of farm machineries
cannot be avoided to tackle the labour scarcity.
330 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

TABLE 5. DETERMINANTS OF MACHINE LABOUR COST IN PADDY CULTIVATION –

PANEL DATA REGRESSION RESULTS

Model (1) Model (2)

Variable Coefficient Elasticity Coefficient Elasticity
(1) (2) (3) (4) (5)
Human labour cost (HLC) (Rs/ha) 0.06 0.25
(0.80)ns (0.80)ns --- ---
Bullock labour cost (BLC) (Rs./ha) -0.16 -0.11 -0.07 -0.05
(-0.95)ns (-0.95)ns (-0.73) ns
(-0.73)ns
Yield enhancing inputs cost (YEIC) (Rs./ha) 0.14 0.41 0.19 0.53
(2.19)c (2.19)c (3.50)a (3.50)a
Human labour use (HLMH) (man-hours/ha) -0.81 -1.40
--- --- (-4.55)a (-4.55)a
Output price (Rs./qtl) 0.05 0.02 0.03 0.01
(0.51)ns (0.51)ns (0.75)ns (0.75)ns
Profitt-1 (Rs./ha) 0.01 0.07 0.012 0.08
(1.27)ns (1.27)ns (1.30)ns (1.30)ns
Cropping intensity (CI) (per cent) -2.91 -0.87 -1.29 -0.39
(-1.71)d (-1.71)d (-1.09)ns (-1.09)ns
GIA to GCA (per cent) 13.21 1.38 7.51 0.79
(1.95)c (1.95)c (1.59)d (1.59)d
Land labour ratio (LLR) (ha) 177.88 0.75 122.30 0.51
(1.66)d (1.66)d (1.62)d (1.62)d
NREGS dummy (before = 0; After = 1) 144.09 0.13 102.32 0.09
(1.66)d (1.66)d (1.44)d (1.44)d
Constant -508.19 397
(-1.31)ns (2-79)b
Adjusted R2 (overall model) 0.49 0.41
F statistics 42a 44a
Number of observations 121 121
Number of groups (states) 11 11
Rho 0.96 0.97
Sources: Computed using data from CACP (various years) and www.dacnet.nic.in
Notes: (1) Figures in parentheses are t-values; (2) a, b, c and d are significant at 1 per cent, 5 per cent, 10 per
cent and 15 per cent level, respectively; ns-not significant; (3) The elasticity values are estimated at sample mean
level of observation. (4) Rho = 0.96 in Model 1 means the fraction of the variance of intercept that is explained by ui
(state specific error term). The higher the value of Rho the more individual state effect (fixed effect) is important in
explaining variation of the dependent variable.

In model (2) on paddy crop, HLMH is used (instead of HLC) along with other
variables to capture the influence of human labour variable on machine labour cost.
As expected, HLMH turned out to be negative and significant in influencing the use
of machineries in paddy cultivation. It also turned out to be the most dominant
variable in negatively influencing the farm mechanisation; its elasticity is estimated
to be at -1.40. This negative coefficient of labour use in man-hours was expected
because farm machineries are generally used as substitutes for human labour in crop
cultivation. Similar to the results generated through model (1), YEIC, GIA/GCA,
LLR and NREGS dummy have all positively and significantly impacted on the use of
farm machineries. The regression results also provide answer to the issue of whether
the introduction of MGNREGS has anything to do with farm mechanisation? The
results of both models clearly suggest that the MGNREGS has significantly and
positively impacted on the use of farm machineries in paddy cultivation.
 FARM MECHANISATION, MGNREGS AND LABOUR SUPPLY NEXUS 331

The use of farm machineries generally varies from one crop to another because of
certain intrinsic reasons as reported earlier. Therefore, as followed for paddy crop, an
attempt has been made to study the determinants of farm mechanisation in wheat,
which is an important foodgrain crop cultivated predominantly during rabi season.
Two regression models (1) and (2) have been estimated for wheat using the same
variables that were employed for paddy. The adjusted R2 estimated through two
models presented in Table 6 clearly shows that the variables included in the models
are appropriate as they explain 76 to 82 per cent of variation in the adoption of farm
machineries in wheat. One can observe that the variations in the results estimated
between model (1) and (2) are not very substantial, except the level of significant
values. The coefficients of YEIC, CI and NREGS dummy have all positively and
significantly influenced MLC in both the models, while human labour variable (HLC
or HLMH) has negatively and significantly influenced it. Among all the variables, CI
turned out to be the strongest variable (elasticity is 1.70-1.74 in both models) in
impacting the use of farm machineries in wheat crop, which is obvious as intensive
agriculture requires increased use of farm machineries.

TABLE 6. DETERMINANTS OF MACHINE LABOUR COST IN WHEAT CULTIVATION –

PANEL DATA REGRESSION RESULTS

Model (1) Model (2)

Variable Coefficient Elasticity Coefficient Elasticity
(1) (2) (3) (4) (5)
Human labour cost (HLC) (Rs./ha) -0.07 -0.09
(-3.48)a (-3.48)a --- ---
Bullock labour cost (BLC) (Rs./ha) -0.04 -0.01 -0.03 -0.01
(-1.26)ns (-1.26)ns (-1.04)ns (-1.04)ns
Yield enhancing inputs cost (YEIC) (Rs./ha) 0.26 0.54 0.28 0.56
(3.91)a (3.91)a (4.21)a (4.21)a
Human labour use (HLMH) (man-hours) -0.24 -0.12
--- --- (-2.18)b (-2.18)b
Output price (Rs./qtl) 0.25 0.07 0.29 0.08
(1.08) ns (1.08) ns (1.29)ns (1.29)ns
Profitt-1 (Rs./ha) 0.005 0.02 0.005 0.023
(1.15)ns (1.15) ns (0.98)ns (0.98)ns
Cropping intensity (CI) (per cent) 8.31 1.70 8.46 1.74
(3.28)a (3.28)a (3.25)a (3.25)a
GIA to GCA (per cent) 1.79 0.13 0.54 0.04
(0.47)ns (0.47) ns (0.12) ns (0.12)ns
Land labour ratio (LLR) (ha) 0.96 0.005 -0.16 -0.008
(0.09)ns (0.09)ns (-0.01)ns (-0.01)ns
NREGS dummy (before=0; after=1) 79.99 0.05 70.50 0.043
(1.87)c (1.87)c (1.68)c (1.68)c
Constant -1050.54 -1000.41
(-3.03)b (-2.79)b
Adjusted R2 (overall) 0.82 0.76
Total number of observations 88 88
Total number of groups (states) 8 8
Rho 0.60 0.69
Sources and Notes: Same as in Table 5.
332 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

On the influence of human labour variable on machine labour cost, it appears

from the regression coefficients that human labour man-hours is more important than
the human labour cost not only in the case of wheat but also in paddy cultivation.
This is possibly because any increase in human man-hours would automatically
reduce the adoption of farm mechanisation, which is also observed in most states that
have been considered for the analysis. One important difference noted between the
results of paddy and wheat is the irrigation coefficient that should be explained here.
Irrigation has positively influenced the use of farm mechanisation in paddy
cultivation, which is not true for wheat cultivation. The insignificant influence of
irrigation variable does not mean that the increased coverage of irrigation is not
important (see, Narayanamoorthy and Deshpande, 2003). The irrigation variable has
not turned out to be significant because some of the states (Himachal Pradesh,
Madhya Pradesh, Uttar Pradesh) considered for the analysis have incurred higher cost
on machine labour but they have relatively less irrigation coverage. However,
NREGS dummy has come out with positive and significant value in wheat crop as
well, suggesting that the use of machineries has increased appreciably after the
introduction of the rural employment scheme.

IV

CONCLUSION AND RECOMMENDATIONS

An attempt has been made in this study to find out the relationship among the
farm mechanisation, MGNREGS, labour supply and other factors using state-wise
data pertaining to paddy and wheat crops covering period from 2000-01 to 2010-11.
To measure the influence of various growth factors including MGNREGS on the use
of farm machineries, regression is computed using panel data with fixed effects
model. The descriptive analysis of the study shows that the real (at 1986-87 prices)
machine labour cost (which is used as a proxy variable to reflect the farm
mechanisation) incurred for cultivating both paddy and wheat has increased
considerably during post-MGNREGS period in almost all the states considered for
the analysis. The machine labour cost incurred for cultivating paddy has increased
substantially in states like Andhra Pradesh, Tamil Nadu, Karnataka and Madhya
Pradesh after implementing MGNREGS, while the same increase was found very
high in Madhya Pradesh, Himachal Pradesh, Uttar Pradesh and Punjab in wheat
cultivation. In most states where the machine labour cost has increased substantially,
the use of human labour in man-hours has declined sharply in both paddy and wheat,
confirming the fact that farm machineries are used to substitute the human labour
especially after implementing MGNREGS. The regression results computed using
panel data suggest that the factors determining the use of farm machineries is not the
same between the two major crops selected for the study. Besides MGNREGS
dummy, the factors such as coverage of irrigation, yield enhancing inputs cost, land-
labour ratio and human labour use in man-hours have significantly influenced the use
 FARM MECHANISATION, MGNREGS AND LABOUR SUPPLY NEXUS 333

of machine labour in paddy cultivation. But, in the case of wheat crop, irrigation
coverage and land-labour ratio has not significantly influenced the use of
machineries. The MGNREGS dummy used to capture its impact on farm
mechanisation has turned out to be positive and significant in both paddy and wheat
cultivation, suggesting that the level of farm mechanisation has increased after the
implementation of national rural employment guarantee scheme.
Although the study suggests that the use of machineries in agricultural operations
appears to have increased after implementing MGNREGS, more studies covering
crops such as pulses, oilseeds, sugarcane, cotton, etc., need to be carried out using
more disaggregated data to validate the results of this study. Some argue that the
increased farm mechanisation is a desperate attempt of farmers to tackle the labour
scarcity that occurred after implementing MGNREGS. One must find out whether the
increased farm mechanisation is a desperate attempt or is it a normal development
dictated by the exogenous and endogenous factors. Adoption of modern technology
in agriculture also takes place partly due to profit reason which has also been
corroborated by plethora of studies. The Indian farmers are more enterprising now
than ever before due to market related reasons, which is also corroborated by the fast
change of cropping pattern from low value to high value crops (see, Bhalla and
Singh, 2012). Therefore, more detailed studies with farm level data need to be carried
out to find out whether farmers use machineries increasingly due to profit motive or
due to MGNREGS induced labour scarcity. In any case, the present study shows that
the increased use of machine labour has reduced the human labour (farm
employment) in crop cultivation which will have various social ramifications.
Therefore, as demanded by the farmers from different states, efforts may also be
taken to link the MGNREGS work with agriculture to have win-win effect for both
farmers and agricultural labourers.

REFERENCES

Alagh, Y.K. (2004), State of the Indian Farmer – A Millenium Study: An Overview, Vol. 1, Ministry of
Agriculture, Government of India, New Delhi, India.
Agarwal, B. (1981), “Agricultural Mechanisation and Labour Use – A Disaggregated Approach”,
International Labour Review, Vol. 120, No. 1, pp. 115-127.
Agarwal, B. (1983), Mechanisation in Indian Agriculture – An Analytical Study Based on the Punjab,
Allied Publishers, New Delhi, India.
Bhalla, G.S. (1995), “Agricultural Growth and Industrial Development in Punjab”, in J.W. Mellor (Ed.)
(1995), Agriculture on the Road to Industrialisation, International Food Policy Research Institute,
The John Hopkins University Press, Washington, U.S.A., pp. 67-112.
Bhalla, Sheila (1991), “Speeding up Agricultural Growth: Implications for Labour Absorptions”, Paper
presented at the National Seminar on Rural Development, April 26-27, Lucknow.
Bhalla, G.S. and G. Singh (2012), Economic Liberalisation and Indian Agriculture: A District-level
Study, Sage Publications India Private Limited, New Delhi.
Binswanger, H.P. (1978), The Economics of Tractors in South Asia – An Analytical Review, Agricultural
Development Council, New York, U.S.A. and International Crop Research Institute for Semi-Arid
Tropics, Hyderabad, India.
334 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

Binswanger, H.P. (1986), “Agriculture Mechanisation – A Comparative Historical Perspective”, The

World Bank Research Observer, Vol. 1, No.1, pp. 27-56.
Brian, L. (1972), “Patterns of Investment in Farm Machinery and Equipment”, Economic and Political
Weekly, Vol. 7, No. 40, September 30, pp. A113-A124.
Government of India (various years), Report of the Commission for Agricultural Costs and Prices,
Commission for Agricultural Costs and Prices, Ministry of Agriculture, New Delhi.
Dasgupta, B. (1977), The New Agrarian Technology and India, Macmillan Publishers, New Delhi, India.
Desai, D.K. and C. Gopinath (1975), Impact of Farm Tractorization on Productivity and Employment
(Gujarat State), Center for Management in Agriculture, Indian Institute of Management,
Ahmedabad.
Food and Agriculture Organization of the United States (2013), “Agricultural Mechanisation in India”,
in J. Kienzle et al., (Eds.) (2013), Mechanisation for Rural Development – A Review of Patterns
and Progress from Around the World, Rome, Italy, pp. 99-119.
Government of India (1975), Report on the Expert Committee on Combine Harvesters, Planning
Commission, New Delhi.
Government of India (2006), Report of the Study Relating to Formulating Long-Term Mechanisation
Strategies for each Agro-Climatic Zones/States, Ministry of Agriculture, New Delhi, India.
Grewal, S.S. and A.S. Kahlon (1972), “Impact of Mechanisation on Farm Employment in the Punjab”,
Indian Journal of Agricultural Economics, Vol. 27, No. 4, October-December, pp. 214-219.
Greene, W.H. (2011), Econometric Analysis, 7th Edition, Prentice Hall, New Jersey, U.S.A.
Gulati, A.; Jain, S. and N. Satija (2013), Rising Farm Wages in India the ‘Pull’ and ‘Push’ Factors,
Discussion Paper No.5, Commission for Agricultural Costs and Prices, Department of Agriculture
and Cooperation, Ministry of Agriculture, Government of India, New Delhi.
Jose, A.V. (1984), “Farm Mechanisation in Asian Countries – Some Perspectives”, Economic and
Political Weekly, Vol.49, No. 26, June 30, pp. A-97-A-103.
Kahlon, A.S. (1976), “Impact of Mechanissation on Punjab Agricultural with Special Reference to
Tractorisation”, Indian Journal of Agricultural Economics, Vol. 31, No. 4, October-December, pp.
54-70.
Mcinerney, John P. and Graham F. Donaldson (1975), The Consequences of Farm Tractors in Pakistan,
World Bank Staff Working Paper No. 210, Washington D.C., U.S.A.
Narayanamoorthy, A. (1992), “Impact of Tractors and Weedicides on Yield and Employment in
Sugarcane Cultivation”, Journal of Rural Development, Vol. 11, No. 3, May, pp. 351-362.
Narayanamoorthy, A. and R.S. Deshpande, (2003), “Irrigation Development and Agricultural Wages:
An Analysis Across States”, Economic and Political Weekly, Vol. 38 No. 35, August 30, pp. 3716-
3722.
Narayanamoorthy, A. (2013), “Profitability in Crops Cultivation in India: Some Evidence from Cost of
Cultivation Survey Data”, Indian Journal of Agricultural Economics, Vol.68, No.1, January-March,
pp. 104-121.
Narayanamoorthy, A. and M. Bhattarai (2013), “Rural Employment Scheme and Agricultural Wage
Rate Nexus: An Analysis Across States”, Agricultural Economics Research Review, Vol. 26, pp.
149-163.
National Council of Applied Economic Research (NCAER) (1980), Implication of Tractorisation for
Farm Employment, Productivity and Income, New Delhi.
Parthasarathy, G. and V. Abraham (1975), “Impact of Tractorisation on Agriculture”, ICSSR Research
Abstract Quarterly, Vol. 4, No. 1 & 2, January- June, pp. 60-86.
Rao, C.H.H. (1972), “Farm Mechanisation in a Labour-Abundant Economy”, Economic and Political
Weekly, Vol. 7, No.5-6-7, February 5, pp. 393-400.
Rao, C.H.H. (1974), “Employment Implications of the Green Revolution and Mechanisation: A Case
Study of the Punjab”, in N. Islam (Ed.) (1974), Agricultural Policy in Developing Countries,
McMillan Publishers, London, U.K..
Rao, C.H.H. (1975), Technological Change and Distribution of Gains in Indian Agriculture, Macmillan
Publishers, New Delhi, India.
 FARM MECHANISATION, MGNREGS AND LABOUR SUPPLY NEXUS 335

Reddy, D.N. (2012), “MGNREGS and Indian Agriculture – Opportunities and Challenges”, in A.K.
Pankaj (Ed.) (2012), Right to Work and Rural India – Working of the Mahatma Gandhi National
Rural Employment Guarantee Scheme, Sage Publications India Pvt. Ltd., New Delhi.
Reid, J.F. (2011), “The Impact of Mechanisation on Agriculture”, The Bridge, Vol.41, No. 3, pp.22-29.
Roy, S and M.G. Blasé (1978), “Farm Tractorisation, Productivity and Labour Employment – A Case
Study of Indian Punjab”, Journal of Development Studies, Vol. 14, No. 2, pp. 193-209.
Rudra, A. (1971), "Employment Patterns in Large Farms of Punjab", Economic and Political Weekly,
Vol. 6, No.26, June 26, pp. A89-A94.
Sarkar, K.K. and M. Prahaladachar (1965), “Mechanisation as a Technological Change”, Indian Journal
of Agricultural Economics, Vol. 21, No. 1, January-March, pp. 171-182.
Sidhu, R.S. and S.S. Grewal (1990), “Factors Affecting Demand for Human Labour in Punjab
Agriculture: An Econometric Analysis”, Indian Journal of Agricultural Economics, Vol. 40, No. 2,
April- June, pp.125-133.
Vaidyanathan, A. (1978), “Labour Use in Indian Agriculture: An Analysis Based on Farm Management
Survey Data”, in Bardhan et al., (Eds.) (1978), Labour Absorption in Indian Agriculture: Some
Exploratory Investigations, Asian Employment Programme, ARTEP, International Labour
Organisation, Bangkok.
Zarkovic, M. (1987), “The Effects of Economic Growth and Technological Innovation on the
Agricultural Labour Force in India”, Studies in Comparative International Development, Vol. 22,
No. 1, pp. 103-120.