PARTIAL ANALYSIS OF

HECKSCHER-OHLIN MODEL
ON
INDIAN AGRICULTURAL SECTOR

SUBMITTED TO
DEPARTMENT OF BUSINESS ECONOMICS
SRI GURU TEGH BAHADUR KHALSA COLLEGE
UNIVERSITY OF DELHI
NEW DELHI -110007
DATE – NOVEMBER 10 , 2017

1
 Certificate

We, do hereby certify that the Project Report/Dissertation (BBE) entitled “Partial Analysis of
Hickscher-Ohlin model on Indian Agriculture sector” is done by us and is an authentic work
carried out by us. The matter embodied in this project work has not been submitted earlier
for the award of any degree or diploma to the best of our knowledge and belief.

__________________________

__________________________

__________________________

__________________________

__________________________

Signature of the Student

Certified that the Project Report/Dissertation (BBE) entitled “Partial Analysis of Hickscher-
Ohlin model on Indian Agriculture sector” done by these students is completed under my
guidance.

2
 DECLARATION

We hereby declare that this project report entitled “PARTIAL ANALYSIS OF

HECKSCHER-OHLIN MODEL “ is written by us and is our own effort and that no part
has been plagiarized without citations.

The matter embodied in this project work has not been submitted earlier for the award of any
degree or diploma to the best of our knowledge and belief.

Mayank Khanna(2015BBE1051)

Anjali Kumari (2015BBE1056)

Garima Rawat(2015BBE1043)

Ayushi Prajapati (2015BBE1057)

Shubham (2015BBE1054)

Nitesh Singh (2015BBE1046)

Date: NOVEMBER 15.2017

3
 ACKNOWLEDGEMENT

We would like to extend thanks to all the people, who so generously contributed to the work
presented in this project.

Special mention goes to our enthusiastic mentor, philosopher and guide Ms. Karman Kaur.
Her dedication and keen interest, above all her overwhelming attitude to help us, was our
constant source of motivation. Her timely advice, meticulous scrutiny and scientific approach
have helped us to a very great extent to accomplish this project.

Similar profound gratitude goes to Ms. Sanchita Dhingra, Ms. Upasana Dhanda and Mr.
Neeraj Jain, who gave us unconditional support, time to time. We are grateful for their help
and co-operation throughout the study period.

It is our privilege to thank Dr.Bibhu Prasad Sahoo, Head, Department of Business

Economics, Sri Guru Tegh Bahadur Khalsa college, University of Delhi for his tremendous
academic support and for providing the opportunity and means to complete the research.

4
 CONTENTS

Serial TITLE Page

Number Number
1. Introduction 6-9
2. Literature Review 10-11
3. Objective 12
4. Economic Theory
4.1 Heckscher – Ohlin model 13
4.2 Ricardo’s Theorem 14
4.3 Difference between Ricardo’s Theorem and Heckscher 15-16
Ohlin model
5 Methodology 17
5.1 Regression Analysis 18
5.2 Cobb – Douglas Production function 19
5.3 Cluster analysis 20
5.4 Classical Linear Regression Analysis 21-22
6 Data and Interpretations
6.1 Descriptive statistics for rice 23-28
6.2 Production Function of Rice 29-30
6.3 Descriptive statistics for Wheat 31-34
6.4 Production Function of Wheat 35-36
6.5 Cluster Analysis 37-45
7 Conclusion 46
8 Limitations 47
9 References 48-49

5
 INTRODUCTION

Indian Economy has always been dependent on the Agricultural sector for its development
and growth. It has been increasingly observed that the economic system of the country is
directly influenced by the direction of trade taken by the agriculture products. Even today
agriculture continues to dominate the economic scene of India, accounting for about one-third
of gross domestic product (GDP) and one-fifth of foreign exchange. This sector provides
employment to more than 70 per cent of the total labour force in the country.

A well renowned theory of International Economics pertaining to the exports and imports of
goods is the centre of discussion in this research paper. One of the prominent models being
the Heckscher Ohlin (HO) model. HO model is a general equilibrium model of international
trade, developed by Eli Heckscher and Bertil Ohlin at Stockholm School of Economics.

The model suggested by Heckscher-Ohlin assumes that there are two factors of production,
namely, labour and capital. One country has comparative advantage over the other because of
the differences in relative amounts of each factor. The model suggests that countries should
produce and export goods using the resources that they have in abundance. Similarly, the
countries should import goods that require resources that they have in short supply.

The Indian Economy has been a labour -intensive economy since its independence
and thus it becomes important to check the application of this model on the
agriculture sector, which is the backbone of Indian economy. In this research
paper we try to analyse the extent of applicability of HO model on the
agriculture sector of the Indian Economy. The partial analysis of HO model aims
to determine that whether a country like India over the different time periods can
be categorised as labour intensive or capital intensive. Following the results, we
will test the variables like labour, capital, debt, modes of transportation,
electrification for major exporting crops to determine if the major exporting
crops that represent a significantportion of Indian Agriculture output are capital
intensive or labour intensive. Thereby, checking for the satisfaction HO model in
relation to the agriculture output of India.

6
ASSUMPTIONS OF HECKSCHER OHLIN MODEL:

1. Both countries have identical production technology: -

This assumption means that producing the same output of either commodity could be
done with the same level of capital and labour in either country. Actually, it would be
inefficient to use the same balance in either country (because of the relative availability
of either input factor) but, in principle this would be possible. Another way of saying this
is that the per-capita productivity is the same in both countries in the same technology
with identical amounts of capital.
Countries have natural advantages in the production of various commodities in relation
to one another, so this is an "unrealistic" simplification designed to highlight the effect of
variable factors. This meant that the original H–O model produced an alternative
explanation for free trade to Ricardo's, rather than a complementary one; in reality, both
effects may occur due to differences in technology and factor abundances.
In addition to natural advantages in the production of one sort of output over another
(wine vs. rice, say) the infrastructure, education, culture, and "know-how" of countries
differ so dramatically that the idea of identical technologies is a theoretical notion. Ohlin
said that the H–O model was a long-run model, and that the conditions of industrial
production are "everywhere the same" in the long run.[2]

2. Production output is assumed to exhibit constant returns to scale: -

In a simple model, both countries produce two commodities. Each commodity in turn is
made using two factors of production. The production of each commodity requires input
from both factors of production—capital (K) and labour (L). The technologies of each
commodity is assumed to exhibit constant returns to scale (CRS). CRS technologies
implies that when inputs of both capital and labour is multiplied by a factor of k, the
output also multiplies by a factor of k. For example, if both capital and labour inputs are
doubled, output of the commodities is doubled. In other terms the production function of
both commodities is "homogeneous of degree 1".

7
 The assumption of constant returns to scale CRS is useful because it exhibits a
diminishing returns in a factor. Under constant returns to scale, doubling both capital and
labour leads to a doubling of the output. Since outputs are increasing in both factors of
production, doubling capital while holding labour constant leads to less than doubling of
an output. Diminishing returns to capital and diminishing returns to labour are crucial to
the Stolper–Samuelson theorem.

3. The technologies used to produce the two commodities: -

The CRS production functions must differ to make trade worthwhile in this model. For
instance if the functions are Cobb–Douglas technologies the parameters applied to the
inputs must vary. An example would be:
Arable industry, Fishing industry
Where A is the output in arable production, F is the output in fish production,
and K, L are capital and labour in both cases.
In this example, the marginal return to an extra unit of capital is higher in the fishing
industry, assuming units of fish (F) and arable output (A) have equal value. The more
capital-abundant country may gain by developing its fishing fleet at the expense of its
arable farms. Conversely, the workers available in the relatively labour-abundant country
can be employed relatively more efficiently in arable farming.

4. Factor mobility within countries: -

Within countries, capital and labour can be reinvested and reemployed to produce
different outputs. Similar to Ricardo's comparative advantage argument, this is assumed
to happen without cost. If the two production technologies are the arable industry and the
fishing industry it is assumed that farmers can shift to work as fishermen with no cost
and vice versa.
It is further assumed that capital can shift easily into either technology, so that the
industrial mix can change without adjustment costs between the two types of production.
For instance, if the two industries are farming and fishing it is assumed that farms can be
sold to pay for the construction of fishing boats with no transaction costs.

8
It should be noted that this model differs from the comparative advantage theory that
focuses on the efficiency of the production process. Because the country produces goods
based on the resources that they have in abundance, it will be cheapest to produce these
goods. Very broadly, countries that have more capital will specialize in capital-intensive
goods and countries that countries with more labour will specialize in labour-intensive
goods. These countries will trade these goods with each other.

There has been significant amount of research work performed on the HO model in
different parts of the world and each research gave some valuable inputs in terms of
enhancing the trade mobility, economic growth of that area of research. Hence, we also
aim to draw valuable insights from the partial analysis of HO model on the agrarian
economy which can help in better understanding of the International Economy, policy
formulation and a reality check on where we stand after so many years of Independence.

9
 LITERATUREREVIEW

Heckscher and Bertil Ohlin came up with a general equilibrium mathematical model of
international trade in 1933. This model has been tested by various economists to determine
the impact on various sectors of different Economies at different time periods.

The original H-O Model contained two countries, two factors for production and two
commodities that could be produced. It has been extended since 1930s by many economists.
Notable contributions came from Paul Samuelson, Ronald Jones and Jaroslav Vanek, so that
variations of the model are sometimes called the Heckscher-Ohlin Samuelson model or the
Heckscher- Online Vanek model in neo classical economics.

Wassily Leontief was the first to test H-O Model empirically in 1953. He analysed the U.S.
data for the year 1947 using input/output tables of the U.S. exports and import substitutes. He
concluded that America was actually exporting labour intensive goods and not capital
intensive goods. Since his results were opposite to the model itself, people started calling his
findings the Leontief Paradox.

Tatemoto and Jehimura (1959) on testing found out that Japan's overall trade was not in line
with the H-O Model. Even though Japan was a labour abundant country in the 1950s, it
exported capital intensive goods and imported labour - intensive goods. When they tested just
Japan and the U.S. trade, they found that it was in line with the H-O Model.

Bharawaj (1962) studied India's trade overall pattern with the world. He found India’s
exports were labour intensive which was consistent with the H-O theory. When he applied his
test with just India and the U.S. trade, he found that India was exporting capital intensive
goods and importing labour intensive goods which was not in line with the H-O Model.

Richard Brecher (1993) tested trade between the U.S. and Canada by performing tests on
different variations of the H-O Model. He concluded that the data supports the H-O Model
after you take into account the differences in factor prices. The model that had best results
was the one that took factor price differentials between industries as consequences of
imperfect mobility.

10
In a working paper of International Monetary Fund (I.M.F.), Trevor Alleyne and Arvind
Subramanian (2001) analysed South Africa's pattern of trade. Their findings appear to
contradict the Heckscher Ohlin- Samuelson theorem. Despite South Africa being more well-
endowed with labour in physical terms (relative to capital), when compared with its high-
income trading partners, this physical labour abundance has not translated into a lower price
of labour (relative to capital) and thus not resulted in a comparative advantage in the
production and export of labour intensive goods.

Amith Jay Shetty (2014) in his paper explored the Heckscher- Ohlin- Theorem and checked if
it applied to India's exports 2009 data. Using the trade data between India and the U.S., he
tried to find if India's exports trade patterns were consistent with the theory taken from
Heckscher- Ohlin Theorem. The idea was to see if labour abundant India exported labour
intensive goods when trading with capital abundant United States. The results from the paper
found that there was no statistically significant relationship between India's export/import
ration and its capital/labour ratio using 2009 trade data with the U.S. The second test
performed using the 2005 trade data (a time before the 2008 Financial Crisis) and the third
test performed using the 2009 trade data but only limiting it to commodities in the
manufacturing sector also yielded the same results. On conducting tests to see if it was
normally distributed, it was found that it was not. Thus, it was not a good model to use as a
predictive model.

All the researches done over the years vary in their results. Also, every new paper sheds light
on some new aspect of the model. The model has been criticized for its basic assumptions
like no technological difference between countries. However, it still remains a relevant area
of study in economics and has scope for further research and empirical analysis.

11
12
 OBJECTIVE

The purpose of present study is to contribute towards understanding the Analysis of

Heckscher Ohlin Model in the Agriculture Sector of Indian Economy.

There are various objectives in line with the aim to determine the Partial Analysis Heckscher
Ohlin Model: -

1. To quantify the variables needed to determine the capital and labour proportions in
the agriculture sector. Thereby, empirically calculating the capital labour ratio for the
Indian Economy at different periods.
2. To derive production function for the agricultural output using variables such as debt,
capital, labour, electrification, temperature etc.
3. To test empirically if the major exported crops are capital intensive or labour
intensive
4. To make analysis of whether the exported crops are following partial analysis of the
theory of Heckscher Ohlin Model and define the causes for the same.
5. To determine variables significant in regression analysis of the two major exporting
crops that is wheat and rice.

These objectives are contributing towards the emerging questions about the viability of
Heckscher Ohlin Model in the Agriculture Sector of India.

13
 ECONOMIC THEORY

In the early 1900s, a theory of trade was developed by two Swedish economists,Eli
Heckscher and Bertil Ohlin, which subsequently became known as theHeckscher-Ohlin
Model.The H-O model is based on the theory that the pattern of trade is determined by
differences in factor endowments.

The model essentially states that international trade occurs because countries differ in their
relative factor endowments and commodities differ in their relative factor intensities. Relative
endowments of factors of production (land, labour and capital) determine a country’s
comparative advantage. Countries have comparative advantages in those goods for which the
required factors of production are relatively abundant and cheap locally. This is because the
prices of goods are ultimately determined by the prices of their inputs. Goods that require
inputs are locally abundant will be cheaper to produce than those goods than require inputs
that are locally scarce.

The Heckscher-Ohlin Model makes the following core assumptions,

1. Labour and Capital flow freely between the sectors.

2. The amount of labour and capital in two places differ from each other.
3. The Technology is similar among the two places.
4. The tastes are same.

It builds on David Ricardo’s theory of comparative advantage by predicting patterns of

commerce and production based on the factor endowments of a trading region. The Ricardian
model of comparative advantage has trade ultimately motivated by differences in labour
productivity using different technologies. Heckscher and Ohlin did not require production
technology to vary between countries, so the H-O model has identical production technology
everywhere.

The classical theory of international trade is popularly known as the Theory of Comparative
Costs or Advantage. It was formulated by David Ricardo in 1815. The classical approach, in
terms of comparative cost advantage, as presented by Ricardo, basically seeks to explain how
and why countries gain by trading.

14
The idea of comparative costs advantage is drawn in view of deficiencies observed by
Ricardo in Adam Smith’s principles of absolute cost advantage in explaining territorial
specialisation as a basis for international trade.

Being dissatisfied with the application of classical labour theory of value in the case of
foreign trade, Ricardo developed a theory of comparative cost advantage to explain the basis
of international trade as under:

Ricardo’s Theorem:
Ricardo stated a theorem that, other things being equal, a country tends to specialise in and
export those commodities in the production of which it has maximum comparative cost
advantage or minimum comparative disadvantage. Similarly, the country’s imports will be of
goods having relatively less comparative cost advantage or greater disadvantage.

The Ricardian Model:

To explain his theory of comparative cost advantage, Ricardo constructed a two-country, two
commodity, but one-factor model with the following assumptions:
1. Labour is the only productive factor.
2. Costs of production are measured in terms of the labour units involved.
3. Labour is perfectly mobile within a country but immobile internationally.
4. Labour is homogeneous.
5. There is unrestricted or free trade.
6. There are constant returns to scale.
7. There is full employment equilibrium.
8. There is perfect competition.

Under these assumptions, let us assume that there are two countries A and В and two goods X
and Y to be produced. In short, “each country can consume more by trading than in isolation
with a given amount of resources. Indeed, the relative gains of the two countries will be
conditioned by the terms of trade and one is likely to gain proportionately more than the other
but it is definite that both will gain.

In fact, the principle of comparative costs shows that it is possible for both the countries to
gain from trade, even if one of them is more efficient than the other in all lines of production.

15
The theory implies that comparative costs are different in different countries because the
abundance of factors which may be necessary for the production of each commodity does not
bear the same relation to the demand for each commodity in different countries.

Thus, specialisation based on comparative cost advantage clearly represents a gain to the
trading countries in so far as it enables more of each variety of goods to be produced cheaply
by utilising the abundant factors fully in the country concerned and to obtain relatively
cheaper goods through mutual international exchange.

Ricardo’s theory pleads the case for free trade. He stresses that free-trade is the pre-requisite
of gains and improvement of world’s welfare. Free trade “by increasing the general mass of
production diffuses general benefit and binds together by one common tie of interest and
intercourse, the universal society of nations throughout the civilised world.”

To sum up, what goods will be exchanged in international trade is the main question solved
by Ricardo’s theory of comparative costs.

1. More variables: -
As against Ricardian Theory which is based on two countries, two commodities and
one factor, Ohlin's Modern theory incorporates two countries two commodities and
two factors.

2. Comparative cost theory: -

According to the classical theory, the principle of comparative costs is a special
feature of international trade.According to Ohlin, the principle of comparative cost is
applicable to all trade; whether internal or international.

3. Cost difference is expressed in terms of money: -

Superiority of H.O. theory lies in the fact that it calculates the cost differences in
terms of money. Ricardo explains the difference in terms of labour theory of value of
which is full of defects and impracticable. Trade which involves exchange of goods
and services has been carried in terms of money. H.O. theory renders the cause of the
trade easy to understand.

16
4. Supply of factors of production considered: -
The main cause of the international trade is the difference in factor supplies between
the countries. Each country differs in factor endowments i.e. in their abundance or
scarcity. Difference in supply, given the demand, brings the difference in cost of
factor and finally, the difference in commodity prices. In Ricardian theory, difference
in factor (labour) efficiency is recognized but difference in factor supply is ignored.
H.O. theory, therefore provides a better explanation of price difference of factors
through the difference is their supplies.

5. Contribution to positive economies: -

Ricardian theory which no doubt explains the reason for internal trade is more
concerned with the benefits of trade. For classical economists, the welfare aspect of
trade is more important. Ohlin has adapted a positive approach in explaining the cause
of international trade. His main concern was to find out the cause of trade and not so
much of its welfare aspect. Therefore, it is pointed out that Ohlin's analysis has
contributed to positive economic analysis

17
 METHODOLOGY

To find whether India is a capital abundant or labour abundant country we have used capital
to labour ratio.

Capital labour ratio measures the ratio of capital employed to labour employed. If capital
labour ratio is more than 1 it means it is capital abundant and if it is less than 1 then it is
labour abundant.

Capital to labour ratio: -

Capital (K)/ Labour (L)

To check whether India produces labour intensive good or capital intensive good we have
collected the capital (K) and labour (L) data from the World Bank site for the period of 12
years beginning from 2000.

We took debt and electricity consumption as capital and total workforce employed in
production as labour. Then, we have calculated the capital to labour ratio for 12 years. We
can see from the data that for each year the capital to labour ratio was less than 1. So we can
say that India is a labour abundant country.

We have used Ms-excel for the calculation of capital to labour ratio.

In the next part of data analysis, we will be making a Cobb- Douglas production function for
the two major exporting crops i.e. Wheat and Rice.

Cobb- Douglas production function is used to represents the relationship between its input
namely physical capital and labour – and the amount of output produced. It’s a means for
calculating the impact of changes in the inputs, the relevant efficiencies, and a yield of
production activity.

Cobb-Douglas production function:

Q (L,K)=A*Lβ*Kα

After making the production function, we’ll be doing cluster analysis.

18
Cluster analysis is a statistical technique in which data is subdivided into groups (clusters)
such that the items in a cluster are very similar (but not identical) to one another and very
different from the items in other clusters. It is a discovery tool that reveals associations,
patterns, relationships, and structures in masses of data.

We will be using K-Means cluster analysis technique to know whether the crops are labour
intensive or capital intensive.

K-means cluster analysis: -

K-means cluster analysis is a tool designed to assign cases to a fixed number of groups
(clusters) whose characteristics are not yet known but are based on a set of specified
variables.

 REGRESSION ANALYSIS
Regression is a statistical measure used in finance, investing and other disciplines that
attempts to determine the strength of the relationship between one dependent variable
(usually denoted by Y) and a series of other changing variables (known as
independent variables). Regression helps investment and financial managers to value
assets and understand the relationships between variables, such as commodity prices
and the stocks of businesses dealing in those commodities.
Regression can help finance and investment professionals as well as professionals in
other businesses. Regression can help predict sales for a company based on weather,
previous sales, GDP growth or other conditions. The capital asset pricing model
(CAPM) is an often-used regression model in finance for pricing assets and
discovering costs of capital. The general form of each type of regression is:

Linear Regression: Y = a + bX + u

Multiple Regression: Y = a + b1X1 + b2X2 + b3X3 + ... + btXt + u

Where;
Y = the variable that you are trying to predict (dependent variable)
X = the variable that you are using to predict Y (independent variable)
a = the intercept
b = the slope

19
 u = the regression residual

 COBB DOUGLAS PRODUCTION FUNCTION

In economics, a production function is an equation that describes the relationship

between input and output, or what goes into making a certain product, and a Cobb-
Douglas production function is a specific standard equation that is applied to describe
how much output two or more inputs into a production process make, with capital and
labour being the typical inputs described.

Developed by economist Paul Douglas and mathematician Charles Cobb, Cobb-

Douglas production functions are commonly used in both macroeconomics and
microeconomics models because they have a number of convenient and realistic
properties.

The equation for the Cobb-Douglas production formula, wherein K represents capital,
L represents labour input and a, b, and c represent non-negative constants, is as
follows:

F (K,L) = bKaLc

If a+c=1 this production function has constant returns to scale, and it would thus be
considered linearly homogeneous. As this is a standard case, one often writes (1-a) in
place of c. It's also important to note that technically a Cobb-Douglas production
function could have more than two inputs, and the functional form, in this case, is
analogous to what is shown above.

THE ELEMENTS OF COBB-DOUGLAS: CAPITAL AND LABOUR

When Douglas and Cobb were conducting research on mathematics and economies
from 1927 to 1947, they observed sparse statistical data sets from that time period and
came to a conclusion about economies in developed countries around the world: there

20
 was a direct correlation between capital and labour and the real value of all goods
produced within a timeframe.

It's important to understand how capital and labour are defined in these terms, as the
assumption by Douglas and Cobb make sense in the context of economic theory and
rhetoric. Here, capital indicates the real value of all machinery, parts, equipment,
facilities, and buildings while labour accounts for the total number of hours worked
within a timeframe by employees.

Basically, this theory then posits that the value of the machinery and the number of
person-hours worked directly relate to the gross output of production. Although this
concept is reasonably sound on the surface, there were a number of criticisms Cobb-
Douglas production functions received when first published in 1947.

 CLUSTER ANALYSIS

Cluster analysis is a statistical classification technique in which a set of objects or

points with similar characteristics are grouped together in clusters. It encompasses a
number of different algorithms and methods that are all used for grouping objects of
similar kinds into respective categories. The aim of cluster analysis is to organize
observed data into meaningful structures in order to gain further insight from them.

Its primary objective is to group activities (variables, entities, etc.) that possess similar
characteristics.

 K-Means clustering

K-means (MacQueen, 1967) is one of the simplest unsupervised learning algorithms

that solve the well-known clustering problem. The procedure follows a simple and
easy way to classify a given data set through a certain number of clusters (assume k
clusters) fixed a priori. The main idea is to define k centroids, one for each cluster.
These centroids should be placed in a cunning way because of different location

21
 causes different result. So, the better choice is to place them as much as possible far
away from each other. The next step is to take each point belonging to a given data set
and associate it to the nearest centroid. When no point is pending, the first step is
completed and an early groupage is done. At this point we need to re-calculate k new
centroids as barycentre’s of the clusters resulting from the previous step. After we
have these k new centroids, a new binding has to be done between the same data set
points and the nearest new centroid. A loop has been generated. As a result of this
loop we may notice that the k centroids change their location step by step until no
more changes are done. In other words centroids do not move any more.
Finally, this algorithm aims at minimizing an objective function, in this case a squared
error function. The objective function

Where, is a chosen distance measure between a data point and the

cluster centre , is an indicator of the distance of the n data points from their
respective cluster centres.

 CLASSIC LINEAR REGRESSION MODEL

It is a statistical-tool used in predicting future values of a target (dependent) variable

on the basis of the behavior of a set of explanatory factors (independent variables). A
type of regression analysis model, it assumes the target variable is predictable, not
chaotic or random.

Assumptions: -
1. The regression model is linear in the coefficients, correctly specified, and has an
additive error term.
2. The error term has zero population mean: E ( εi) = 0.
3. All independent variables are uncorrelated with the error term: Cov (Xi ,εi) = 0 for
each independent variable Xi .

22
4. Errors are uncorrelated across observations: Cov ( εi ,εj) = 0 for two observations i
and j (no serial correlation).

5. The error term has constant variance: Var ( εi) = σ 2 for every i (no
heteroscedasticity).

6. No independent variable is a perfect linear function of any other independent variable

(no perfect multi-collinearity).

7. The error terms are normally distributed. We’ll consider all the others, and see what
we get. Then, we’ll add this one.

23
 DATA AND INTERPRETATIONS

India is an Agrarian economy and has contributed to the world production of food crops at a
large scale. So as a first step we need to check whether the economy has more of labour
contribution to its production of crops or more of agriculture crop to its use. This phenomena
can be determined by calculating the capital – Labour Ratio of the economy.

Table 1 - Capital Invested in agriculture sector from (2000-2012)

Year Debt Electricity consumed Electricity Rate Total Capital

( In lakh) (in kw/h) (rupee/ unite) ( In Crore)
2000-2001 19,513 771.551 0.2 196.673102
2001-2002 21,536 811.143 0.4 218.604572
2002-2003 23,974 876.887 1.9 256.400853
2003-2004 32,004 912.056 1.6 334.632896
2004-2005 49,247 967.15 0.75 499.723625
2005-2006 75,136 1048.67 0.79 759.644493
2006-2007 90,945 1107.82 0.8 918.31256
2007-2008 73,265 1160.41 0.95 743.673895
2008-2009 91,447 1000.44 1.01 924.574444
2009-2010 1,07,858 800.76 1.23 1088.429348
2010-2011 1,32,741 795.78 1.53 1339.585434
2011-2012 1,14,871 611.276 2.03 1161.118903

Explanation: -

In the above table we have calculated the total amount of Capital Invested in the Agriculture
sector in India during 2000-2012. The capital refers to primarily the amount of debt invested
and electricity consumption in India. The last column represents the total capital employed in
the agriculture sector.

24
 Table 2 - Labour Investment in agriculture Sector

Year Total Labour in agriculture Wages per day Total Labour force

2000-2001 125549333.9 139.4 1750.157714

2001-2002 118120453.8 139.79 1651.205824
2002-2003 117900277.5 197.35 2326.761977
2003-2004 116036241.5 193.4 2244.140911
2004-2005 112560472.4 203.6 2291.731218
2005-2006 107180448.7 204.75 2194.519688
2006-2007 102883507 202.48 2083.185249
2007-2008 102118843.3 202.94 2072.399806
2008-2009 97578823.87 195.48 1907.470849
2009-2010 120632792.4 204.94 2472.248447
2010-2011 120681379.6 212.62 2565.927493
2011-2012 112236886.9 216.3 2427.683865

Explanation: -

In the above table we derived the data for the labour investment in agriculture sector .
Basically we aim to determine the value of amount of labour employed in agriculture sector
in terms of money spent on labour.

25
 Table 3- Determining Capital – Labour Ratio

Year Total Capital Total Labour Capital /Labour Ratio

(In Crores) (In Crores)
2000-2001 196.673102 1750.157714 0.112374502
2001-2002 218.604572 1651.205824 0.132390868
2002-2003 256.400853 2326.761977 0.110196426
2003-2004 334.632896 2244.140911 0.14911403
2004-2005 499.723625 2291.731218 0.218055076
2005-2006 759.644493 2194.519688 0.346155242
2006-2007 918.31256 2083.185249 0.440821363
2007-2008 743.673895 2072.399806 0.358846731
2008-2009 924.574444 1907.470849 0.484712227
2009-2010 1088.429348 2472.248447 0.440258886
2010-2011 1339.585434 2565.927493 0.522066753
2011-2012 1161.118903 2427.683865 0.47828258

Explanation: -

In the above table we have calculated the Capital – Labour ratio for the year 2000 – 2012.
The variables are in same units, that is, (rupees Crore). The ratio is calculated by dividing the
amount of capital by the amount of labour. All results show a value less than 1 which implies
labour is more significant in contribution of yield than capital invested.

26
In this section we are going to discuss the variables used in defining the yield of rice and
wheat.

Table4- Rice Production for 2012-13 in Indian States

States Yield Contribution of labour Area under cultivationRainfall Capital investment

(kg/hectare) ( in lack rupees) (million hectare)
Andhra Pradesh 3246 26244.49022 3.44 2760.9 1105381.53
Assam 1614 11921.76395 2.5 2321.3 169745.5
Bihar 1599 11796.03681 3.21 853.6 863983.12
Gujarat 1744 13017.10544 0.68 652 231932.16
Haryana 2726 21564.27001 1.21 665.3 821558.43
Karnataka 2511 19657.2037 1.49 884.1 496988.84
Kerela 2519 19727.84392 0.23 2187.5 72281.36
Madhya Pradesh 927 6333.220955 1.45 1054.6 287448
Maharashtra 1501 10978.14413 1.47 664.2 537693.5
Orrissa 1529 11211.19252 4.37 1430.2 802139.49
Punjab 4022 33380.72804 2.8 338.9 1876152.96
Tamil Nadu 2683 21181.45555 1.85 529.1 590566.55
Uttar Pradesh 2171 16679.66962 5.19 718.2 2219691.56
West Bengal 2533 19851.52539 5.63 1258.3 1932406.24
Chattisgarh 1176 8314.171975 3.67 1366.8 800130

The above table is a compilation of data for fifteen major rice producing states in India for
the year 2012-13. The variables considered are as follows: -

1. Yield (Y)- This dependent variable refers to the production of rice per hectare.

2. Contribution of Labour (X1)- This independent variable refers to the contribution of

Labour force in the total output produced of rice for a given year (2012-13). It
iscalculated as a product of the contribution of small farm holders to the total output
produced and the value of output in the concerned states. It is calculated in (Rs lacks).

3. Area under cultivation(X2)- This dependent variable refers to the land area under
cultivation for every rice producing state in a given year (2012-13). It is calculated in
(Million Hectares).

4. Rainfall (X3)- This dependent variable refers to the annual precipitation or rainfall in
rice producing states of India in a given year (2012-13). It is measured in (MM).

27
 5. Capital Investment(X4)- This dependent variable refers to the amount of investment in
capital in major rice producing states. It is calculated by multiplying the contribution
of fertilizers in rice production to the value of output in rice. It is measured in (Rs
lacks).

Table 5 – Ln of variables used in production of wheat

States Ln of yeild Ln of contributionLn
of of
labour
area under cultivation Ln of rainfall Ln of fertilisers

Andhra Pradesh 8.08517875 10.17521135 1.235471471 7.923311993 13.91570111

Assam 7.38647085 9.386120912 0.916290732 7.749882653 12.04205554
Bihar 7.37713371 9.37551889 1.166270937 6.7494627 13.66930851
Gujarat 7.4639366 9.474019574 -0.385662481 6.480044562 12.35420019
Haryana 7.91059061 9.978793058 0.19062036 6.500238067 13.61895834
Karnataka 7.82843636 9.886199151 0.39877612 6.784570178 13.11632285
Kerela 7.83161728 9.889786314 -1.46967597 7.690514618 11.18832156
Madhya Pradesh 6.83195357 8.753564225 0.371563556 6.960916827 12.56879725
Maharashtra 7.31388683 9.303661678 0.385262401 6.498583309 13.19504397
Orrissa 7.33236921 9.32466789 1.474763009 7.265569574 13.5950378
Punjab 8.29953457 10.41573401 1.029619417 5.825705079 14.44473394
Tamil Nadu 7.89469085 9.96088134 0.615185639 6.27117745 13.28883761
Uttar Pradesh 7.68294317 9.721945869 1.646733697 6.576748082 14.61287881
West Bengal 7.83715965 9.896036129 1.728109442 7.137516883 14.47427654
Chattisgarh 7.06987413 9.025716805 1.300191662 7.22022752 13.59252949

Herein to avoid the complexities of units in computation, we have standardised the data so as
to avoid difference in units for different factors by taking Ln function or log of independent
variables.

28
 OUTPUT

Table 6.1- Regression Statistics

Regression Statistics
Multiple R 0.999994
R Square 0.999989
Adjusted R Square 0.999985
Standard Error 0.001563
Observations 15

Table 6.2- ANOVA Analysis

Significance
Df SS MS F F
Regression 4 2.213762 0.553441 226532.1 9.82E-25
Residual 10 2.44E-05 2.44E-06
Total 14 2.213787

Table 6.3- Summary Statistics

Coefficients Standard Error t Stat P-value
Intercept -0.9039 0.017439881 -51.8296 1.73E-13
Ln of contribution of
labour 0.882616 0.001219868 723.534 6.26E-25
Ln of area under
cultivation -0.00046 0.001281759 -0.35856 0.727378
Ln of rainfall -8.8E-05 0.000969904 -0.09109 0.929219
Ln of fertilisers 0.000603 0.001281222 0.470857 0.647846

Regression Statistics gives information about the Coefficient of Determination, R2. Since the
value of Coefficient of determination is very high (0.999989), we can say that the
independent variables taken together explain major variation in the dependent variable. Thus,
Ln (Land area under cultivation, labour contribution, capital investment, rainfall) together
determine major variation in crop production per hectare that is, ln(Yield). We can infer from

29
the regression statistics table that Adjusted R2 is also very high, (0.999985). This re-affirms
our idea of Goodness of Fit for our variables taken in the regression process.

ANOVA analysis table shows the regression of independent variables and the residual term.

The F statistic of Independent variables is higher than the significance level which implies F-
Statistics of the regression function lies in the rejection region. Thus, we reject the null
hypothesis that “Theimpact of Independent variable is zero in other words “Coefficient of
Xi’s equals zero”.

The summary statistics table includes all the variables and their coefficients followed by their
standard error, t stats and p value. The p value of each variable is considered and compared
with the level of significance taken as 5 percent or α=0.05. If the P value is less than the level
of significance (p value <α), then we reject the null hypothesis otherwise we fail to reject null
hypothesis.

Null Hypothesis (H0): Coefficients of Independent Variables are statistically insignificant to

determine the yield of rice, that is, βi = 0.

Alternate Hypothesis (H1): Coefficients of Independent Variables are statistically significant

to determine the yield of rice,that is, βi ≠ 0.

Significance Level (α):the probability of rejecting the null hypothesis in a statistical test
when it is true. Here it is taken as 5 percent.

P- value:In statistical science, the p-value is the probability of obtaining a result at least as
extreme as the one that was actually observed in the biological or clinical experiment or
epidemiological study, given that the null hypothesis is true.

Coefficient of Determination (R2):It is interpreted as the proportion of the variance in the

dependent variable that is predictable from the independent variable.

F-Statistics: A value you get when you run an ANOVA test or a regression analysis to find
out if the means between two populations are significantly different

T-Statistics: The ratio of the departure of the estimated value of a parameter from its
hypothesized value to its standard error.

30
 Rejection of null hypothesis
Case 1 P value < α

Fail to reject null hypothesis

Case 2 P value ≥ α

In the Summary Statistics Table, the only p value less than alpha is in case of ln (contribution
of labour) otherwise all variables have p value higher than alpha. Thus, only in case of
contribution of labour we reject the null hypothesis that there is no significant impact of
Contribution of Labour Force on the yield of rice.

31
 Production Function of Rice

The following Cobb Douglas production function of Rice has been developed based on the
coefficients of regression analysis:

Yi = b0 . Aib1. Rib2. Kib3. Lib4

Taking Ln both the sides

Ln (Yi)= Ln bo + b1Ln Ai + b2 Ln Ri + b3 Ln Ki+ b4Ln Li

In the above developed Cobb Douglas Production Function we have the Independent
variables. Each variable represents a factor of production. Herein Ai= Land Area Under
Cultivation, Ri= Average Annual Rainfall, Ki = Total Capital Employed and Li=Labour
employed.

By taking Logarithms on both sides the function will be converted to linear form that can be
that can be evaluated by multiple linear regression analysis.

The parameters and related statistical test results of Independent variables obtained from
regression analysis are given in table 1.3.3. The signs of coefficients of variables of the model
are as expected.

The positive sign of the coefficients shows positive relation between the factor and its impact
on the yield produced. The coefficient of contribution of labour is positive it shows that a one
percent in labour force leads to 0.882616 percent increase in yield produced. The coefficient
of Area under cultivation is negative which shows inverse relation between land and yield.
As land area under cultivation increases by one percent the yield decreases by 0.00046. This
may happen because Rice is a seasonal crop and producing it on large scale might lead to loss
of fertility of land and take higher time to replenish the lost nutrients. The coefficient of
rainfall is negatively related as when rainfall increases by one percent the yield decreases by
8.8E-05 percent. The coefficient of fertilizer is positively related as when fertilizers increase
the yield of rice increases. Only Labour Force makes significant contribution to the yield of
rice produced.

32
 Table 7 – Wheat Production in India for 2012-13

Area Yield Rainfall Insurance Capital Labour Contribution

States
Area (lakh (Kg/Ha)
per hac) (In mm) (Rs. Lakh) ( in rs ) ( In rs)
Uttar Pradesh 96.37 3113 718.2 1488286 64.74 329748660
Madhya Pradesh 43.41 1757 1054.6 2184003 28.02 158427932
Punjab 35.1 4693 338.9 284457.5 28.68 58758505
Haryana 25.15 4624 665.3 41729.5 20.37 16529490
Rajasthan 24.792 2910 467 810154.5 19.785 73097000
Bihar 21.035 1948 853.6 595201 20.295 345073050
Maharashtra 13.07 1761 664.2 1027207 51.825 270761400
Gujarat 12.74 3155 652 1776367 29.085 108384745
Uttarakhand 3.729 2316 450 42053 2.3535 9169500
Himachal Pardesh 3.572 1530 500 23801.5 0.82695 2680577
West Bengal 3.168 2760 1258.3 557794.5 0.0156 179202870
Jammu & Kashmir 2.907 1535 558.8 4522 0.0526 5114850
Karnataka 2.55 1094 884.1 7911362 31.65 39334788
Chattisgarh 1.108 1144 1366.8 511078.5 8.433 58930040

The above table is a compilation of data for fifteen major Wheat producing states in India for
the year 2012-13. The variables considered are as follows: -

1. Yield (Y)- This dependent variable refers to the production of Wheat per hectare.
2. Contribution of Labour (X1)- This independent variable refers to the contribution of
Labour force in the total output produced of wheat for a given year (2012-13). It is
calculated as a product of the contribution of small farm holders to the total output
produced and the value of output in the concerned states. It is calculated in (Rs lacks).
3. Area under cultivation(X2)- This dependent variable refers to the land area under
cultivation for every wheat producing state in a given year (2012-13). It is calculated in
(Million Hectares).
4. Rainfall (X3)- This dependent variable refers to the annual precipitation or rainfall in rice
producing states of India in a given year (2012-13). It is measured in (MM).
5. Capital Investment(X4)- This dependent variable refers to the amount of investment in
capital in major wheat producing states. It is calculated by multiplying the contribution of
fertilizers in rice production to the value of output in wheat. It is measured in (Rs lacks).

33
 Table 8 – Ln of wheat production variables
States Ln of yeild Ln of contributionLn
of of
labour
area under cultivation Ln of rainfall Ln of fertilisers

Andhra Pradesh 8.08517875 10.17521135 1.235471471 7.923311993 13.91570111

Assam 7.38647085 9.386120912 0.916290732 7.749882653 12.04205554
Bihar 7.37713371 9.37551889 1.166270937 6.7494627 13.66930851
Gujarat 7.4639366 9.474019574 -0.385662481 6.480044562 12.35420019
Haryana 7.91059061 9.978793058 0.19062036 6.500238067 13.61895834
Karnataka 7.82843636 9.886199151 0.39877612 6.784570178 13.11632285
Kerela 7.83161728 9.889786314 -1.46967597 7.690514618 11.18832156
Madhya Pradesh 6.83195357 8.753564225 0.371563556 6.960916827 12.56879725
Maharashtra 7.31388683 9.303661678 0.385262401 6.498583309 13.19504397
Orrissa 7.33236921 9.32466789 1.474763009 7.265569574 13.5950378
Punjab 8.29953457 10.41573401 1.029619417 5.825705079 14.44473394
Tamil Nadu 7.89469085 9.96088134 0.615185639 6.27117745 13.28883761
Uttar Pradesh 7.68294317 9.721945869 1.646733697 6.576748082 14.61287881
West Bengal 7.83715965 9.896036129 1.728109442 7.137516883 14.47427654
Chattisgarh 7.06987413 9.025716805 1.300191662 7.22022752 13.59252949

Explanation: -

Herein to avoid the complexities of units in computation, we have standardised the data so as
to avoid difference in units for different factors by taking Ln function or log of independent
variables.

34
 OUTPUT

Table 9.1 – Regression Statistics

Regression Statistics
Multiple R 0.855357
R Square 0.731636
Adjusted R Square 0.582544
Standard Error 0.293259
Observations 15

Table 9.2- ANOVA Table

Df SS MS F Significance F
Regression 5 2.110164 0.422033 4.907301 0.019188344
Residual 9 0.774009 0.086001
Total 14 2.884173

Table 9.3 - Summary Statistics

Standard
Coefficients Error t Stat P-value
Intercept 10.88492158 1.657424355 6.567372 6.567372
Ln area 0.137494653 0.088387409 1.555591 0.154229
-
Ln of rainfall 0.626151237 0.232434087 -2.69389 0.024639
Ln of labour 0.094078609 0.097823069 0.961722 0.361314
-
Ln (Debt) 0.068267288 0.076120746 -0.89683 0.393167
-
Ln (Capital) 0.071036402 0.050020828 -1.42014 0.189268

35
Regression Statistics gives information about the Coefficient of Determination, R2. Since the
value of Coefficient of determination is high (0.731636) we can say that the independent
variables taken together explain major variation in the dependent variable. Thus, Ln (Land
area under cultivation, labour contribution, capital investment, rainfall and debt) together
determine major variation in crop production per hectare that is, ln(Yield). We can infer from
the regression statistics table that Adjusted R2 is also average (0.58244). This can be
interpreted as a suitable measure of the idea of Goodness of Fit for our variables in regression
process.

ANOVA analysis table shows the regression of independent variables and the residual term.
The F statistic of Independent variables is higher than the significance level which implies F-
Statistics of the regression function lies in the rejection region. Thus, we reject the null
hypothesis that “The impact of Independent variable is zero in other words “Coefficient of
Xi’s equals zero”.

The summary statistics table includes all the variables and their coefficients followed by their
standard error, t stats and p value. The p value of each variable is considered and compared
with the level of significance taken as 5 percent or α=0.05. If the P value is less than the level
of significance (p value <α), then we reject the null hypothesis otherwise we fail to reject null
hypothesis.

Null Hypothesis (H0): Coefficients of Independent Variables are statistically insignificant to

determine the yield of wheat, that is, βi = 0.

Alternate Hypothesis (H1): Coefficients of Independent Variables are statistically significant

to determine the yield of wheat, that is, βi ≠ 0.

Significance Level (α): The probability of rejecting the null hypothesis in a statistical test
when it is true. Here it is taken as 5 percent.

P- value: In statistical science, the p-value is the probability of obtaining a result at least as
extreme as the one that was actually observed in the biological or clinical experiment or
epidemiological study, given that the null hypothesis is true

Coefficient of Determination (R2): It is interpreted as the proportion of the variance in the

dependent variable that is predictable from the independent variable

36
Student’s T test statistics: The ratio of the departure of the estimated value of a parameter
from its hypothesized value to its standard error.

Rejection of null hypothesis

Case 1 P value < α

Fail to reject null hypothesis

Case 2 P value ≥ α

In the Summary Statistics Table, the only p value less than alpha is in case of ln (contribution
of labour) otherwise all variables have p value higher than alpha.Based on above discussed
results, we compare the p value of each variable in summary statistics table (Table 2.3.3) and
observe that only the p value of Ln (rainfall) is less than α. This means that only the ln
(rainfall) has statistically significant contribution to the yield and all other factors have
insignificant contribution to yield of wheat. 27

37
Production Function of Wheat

The following Cobb Douglas production function of Rice has been developed based on the
coefficients of regression analysis

Yi = b0 . Aib1. Rib2. Kib3. Lib4

Taking Ln both the sides

Ln (Yi)= Ln bo + b1Ln Ai + b2 Ln Ri + b3 Ln Ki+ b4Ln Li

In the above developed Cobb Douglas Production Function we have the Independent
variables. Each variable represents a factor of production. Herein, Ai= Land Area under
Cultivation, Ri= Average Annual Rainfall, Ki = Total Capital Employed and Li=Labour
employed.

By taking Logarithms on both sides the function will be converted to linear form that can be
that can be evaluated by multiple linear regression analysis. The parameters and related
statistical test results of Independent variables obtained from regression analysis are given in
table 2.3.3. The signs of coefficients of variables of the model are as expected.

The Coefficient of Rainfall has a negative sign which implies that there exists an inverse
relation between rainfall and yield of wheat. Whenever rainfall increases by one percent the
yield of rice decreases by 0.6265 percent. This may happen because rice requires less water
so high moisture content may reduce its growth. The coefficient of debt is also negatively
correlated as when debt increases by one percent yield decreases by 0.0682672 percent. This
may be because debt on wheat increases the financial burden and thereby reducing
productivity. The coefficient of capital investment is also negatively correlated as when

38
capital investment increases by one percent yield decreases by 0.07103 percent. This may
happen because of ineffective utilisation of capital resource.

The coefficient of Land Area under cultivation for wheat has a positive coefficient which
implies there exist direct relation between rainfall and yield of wheat. As the land area under
cultivation increases by one percent the yield increases by 0.13749 percent. The coefficient of
rainfall has a positive relation with yield of wheat. The results may be because wheat is a
staple food grain for North Indian states. So they have high production and labour force
involved.

CLUSTER - ANALYSIS

Cluster analysis is a statistical classification technique in which a set of objects or points with
similar characteristics are grouped together in clusters. It encompasses a number of different
algorithms and methods that are all used for grouping objects of similar kinds into respective
categories. The aim of cluster analysis is to organize observed data into meaningful structures
in order to gain further insight from them. Its primary objective is to group activities
(variables, entities, etc.) that possess similar characteristics. Herein, we have used K means as
a method to form clusters based on their labour and capital inputs.

The procedure follows a simple and easy way to classify a given data set through a certain
number of clusters (assume k clusters) fixed a priori. The main idea is to define k centroids,
one for each cluster and then divide the data into these clusters.

Table 10 .1 – Input of Rice production Function : Data View

39
 States Land_Use Labour Contribution Rainfall Capital Series
Andhra Pradesh 0.21 10.18 7.92 13.92 1
Assam -0.09 9.39 7.75 12.04 2
Bihar 0.15 9.38 6.75 13.67 1
Gujarat 9.47 6.48 12.35 2
Haryana -1.66 9.98 6.5 13.62 1
Karnataka -0.92 9.89 6.78 13.12 1
Kerela 9.89 7.69 11.19 2
Madhya Pradesh -0.99 8.75 6.96 12.57 2
Maharashtra -0.95 9.3 6.5 13.2 1
Orrissa 0.39 9.32 7.27 13.6 1
Punjab 0.03 10.42 5.83 14.44 1
Tamil Nadu -0.49 9.96 6.27 13.29 1
Uttar Pradesh 0.5 9.72 6.58 14.61 1
West Bengal 0.55 9.9 7.14 14.47 1
Chattisgarh 0.26 9.03 7.22 13.59 1

Table 10.2 – Rice Production function: Variable view

Name type D label width Align Measure Role
States Numeric 2 Name of states 8 Right Unknown Input
Land_use Numeric 2 None 8 Right Scale Input
labour_contribution Numeric 2 None 8 Right Scale Input
Rainfall Numeric 2 None 8 Right Scale Input
capital_contribution Numeric 2 None 22 Right Scale Input
QCL_1 Numeric 0 Cluster Number of None
Case 10 Right Nominal Input

In this model we undertake to perform cluster analysis by classifying data into two clusters
broadly: Capital Contribution and labour Contribution.

40
All the variables like Area under Cultivation, Labour Contribution, Capital Contribution and
Rainfall have been converted into log function so as to avoid the computational errors due to
difference in units.

In table 10.1, we calculate the values of log function and the last column refers to the series
of allotment where 1 refers to Capital Intensive and 2 refers to labour intensive. Thus, these
series of classification help to determine that major states of India which produce Rice are
capital Intensive in 2012-13.

In table 10.2, we determine the input values of variables. The decimal place decides the
number of decimal places, name is the name of the column to be considered, type is the
classification of variable, role refers to the input range.

These classifications are the input ranges and helps to determine the value of variables to be
used to make clusters.

OUTPUT TABLES

Table 11.1 Initial Cluster Centers

Cluster
1 2

labour_contribution 9.72 9.89

capital_contribution 14.61 11.19

41
 Table 11.2 Final Cluster Centres
Cluster
1 2

labour_contribution 9.73 9.38

capital_contribution 13.77 12.04

Table 11.3 -ANOVA (Analysis Of Variance)

Cluster Error
Mean
Square df Mean Square Df F Sig.

Labour contribution .374 1 .190 13 1.975 .183

Capital contribution 8.846 1 .298 13 29.708 .000

42
The F tests should be used only for descriptive purposes because the clusters have been
chosen to maximize the differences among cases in different clusters. The observed
significance levels are not corrected for this and thus cannot be interpreted as tests of the
hypothesis that the cluster means are equal.

Table 11.4 -Number of Cases in each Cluster

Cluster 1 11.000
2 4.000
Valid 15.000
Missing .000

In the above cluster analysis we have tables and each table aims to classify the given averages
for fifteen major rice producing states in India. In table 1.5.4 we have two clusters and cluster
1 contains eleven observations and cluster 2 contains four observations. In table 1.5.1 which
the initial cluster formation we have taken average values of each set of observations and
presented in the table. Since the line of methodology follows K- Means cluster analysis we
have divided our data in two clusters: Capital Contribution and Labour Contribution.. As we
observe that in table 1.5.2 results are similar in nature as in case of table 1.5.1. This is
because of less number of iterations involved in the process. It becomes evident that in both
clusters formed the contribution of capital dominates the contribution of Labour. In this line
of thought we can refer to the graphical representation (Figure 1.1). Herein, the contribution
of capital shown by green has higher values compared to contribution of labour shown by
blue.

In table 1.5.3, we have ANOVA (Analysis of Variance) table presented each variable and
their p-value. These values help us determine the contribution of each variable in cluster
analysis and their contribution on yield.

WHEAT PRODUCTION VARAIBLES

43
 Table 12.1 - Wheat Production variables
States land_Use Rainfall Labour Capital Debt Series
Uttar Pradesh 4.57 6.58 19.61 4.17 14.21 1
Madhya Pradesh 3.77 6.96 18.88 3.33 14.6 1
Punjab 3.56 5.83 17.89 3.36 12.56 1
Haryana 3.22 6.5 16.62 3.01 10.64 2
Rajasthan 3.21 6.15 18.11 2.98 13.6 1
Bihar 3.05 6.75 19.66 3.01 13.3 1
Maharashtra 2.57 6.5 19.42 3.95 13.84 1
Gujarat 2.54 6.48 18.5 3.37 14.39 1
Uttarakhand 1.32 6.11 16.03 0.86 10.65 2
Himachal Pardesh 1.27 6.21 14.8 -0.19 10.08 2
West Bengal 1.15 7.14 19 -4.16 13.23 3
Jammu & Kashmir 1.07 6.33 15.45 -2.95 8.42 2
Karnataka 0.94 6.78 17.49 3.45 15.88 1
Chattisgarh 0.1 7.22 17.89 2.13 13.14 1
Jharkhand -0.04 7 17.74 0.64 12 2

Table 12.2 – Ln of wheat production variables

Name type D label width Align Measure Role
States Numeric 2 Name of states 8 Right Unknown Input
Land_use Numeric 2 None 8 Right Scale Input
labour_contribution Numeric 2 None 8 Right Scale Input
Rainfall Numeric 2 None 8 Right Scale Input
capital_contribution Numeric 2 None 22 Right Scale Input
Debt Numeric 0 None 10 Right Nominal Input
QCL_1 Numeric 0 Type of cluster none Right Nominal Input

In this model we undertake to perform cluster analysis by classifying data into two clusters
broadly: Capital Contribution and labour Contribution.

All the variables like Area under Cultivation, Labour Contribution, Capital Contribution and
Rainfall have been converted into log function so as to avoid the computational errors due to
difference in units.

In table 12.1, we calculate the values of log function and the last column refers to the series
of allotment where 1 refers to Capital Intensive and 2 refers to labour intensive. Thus, these

44
series of classification help to determine that major states of India which produce Rice are
capital Intensive in 2012-13.

In table 12.2, we determine the input values of variables. The decimal place decides the
number of decimal places, name is the name of the column to be considered, type is the
classification of variable, role refers to the input range.

OUTPUT

Table 13.1 -Initial Cluster Centers

Cluster
1 2 3
Labour 19.61 14.80 19.00
Capital 4.17 -.19 -4.16
Debt 14.21 10.08 13.23

45
 Table 13.2 - Final Cluster Centers
Cluster
1 2 3
Labour 18.61 16.13 19.00
Capital 3.31 .27 -4.16
Debt 13.95 10.36 13.23

Table 13.3 – ANOVA (Analysis Of Variance)

Cluster Error
Mean Square Df Mean Square Df F Sig.
Labour 10.631 2 .872 12 12.192 .001
Capital 33.798 2 1.776 12 19.030 .000
Debt 20.882 2 1.189 12 17.556 .000
The F tests should be used only for descriptive purposes because the clusters have been
chosen to maximize the differences among cases in different clusters. The observed
significance levels are not corrected for this and thus cannot be interpreted as tests of
the hypothesis that the cluster means are equal.

46
 Table 13.4 -Number of Cases in each Cluster
Cluster 1 9.000
2 5.000
3 1.000
Valid 15.000
Missing .000

In the above cluster analysis, we have tables and each table aims to classify the given
averages for fifteen major wheat producing states in India. In table 2.5.4 we have three
clusters and cluster 1 contains nine observations and cluster 2 contains five observations and
cluster three contains one observation. In table 2.5.1 which the initial cluster formation we
have taken average values of each set of observations and presented in the table. Since the
line of methodology follows K- Means cluster analysis we have divided our data in three
clusters: Capital Contribution, Labour Contribution and Debt. As we observe that in table
1.5.2 results are similar in nature as in case of table 1.5.1. This is because of less number of
iterations involved in the process. We can observe that in cluster one, the contribution of
labour was highest followed by positive contribution of debt and minimal contribution of
capital . In cluster 2 again it was labour dominating the contribution pattern followed by debt
and insignificant contribution of capital. In cluster 3, it was labour force with maximum
contribution followed by debt and a negative contribution from capital.

In table 1.5.3, we have ANOVA (Analysis of Variance) table presented each variable and
their p- value. These values help us determine the contribution of each variable in cluster
analysis and their contribution on yield.

In the above two cluster analysis we had formed the data and divided in groups based on the
variables dominating in the cluster. Now we see that in Analysis Of Variance technique that
majorly capital is dominating in case of labour is dominating in case of capital.

The null hypothesis in this case is” There is no significant contribution of variables in
determining the yield of rice”. While the alternate hypothesis states that” there exists a
significant contribution of variables in determining the yield of rice “. The p value which
shows the probability of obtaining an extreme value. The level of Significance is the
probability of rejecting null hypothesis when it is true.

47
Whenever p-value is less than alpha we reject null hypothesis, which means there is
significant contribution of these variables in determining the yield of rice. Since capital has
highest contribution so it is significant in rice production

Similarly in case of Wheat, the null hypothesis is “There is no significant contribution of

variables in determining the yield of wheat”. While the alternate hypothesis states that “There
exists a significant contribution of variables in determining yield of wheat”. The p value
which shows the probability of obtaining an extreme value. The level of Significance is the
probability of rejecting null hypothesis when it is true.

Whenever p-value is less than alpha we reject null hypothesis, which means there is
significant contribution of these variables in determining the yield of rice. Since labour has
highest contribution so it is significant in rice production.

CONCLUSION

This paper’s research tried to perform partial analysis of Heckscher –Ohlin model for trade in
Indian context. The idea was to see if labour abundant India exported labour intensive goods.
On calculating capital-labour ratio for Indian agricultural sector for the period of 12 years
(2000-2012), the sector was found to be labour abundant.

48
In the next section of the paper, we considered India’s two major exporting crops i.e. wheat
and rice. We collected state wise data of the major factors affecting their production in India
(2012).We did regression of Independent variables i.e. land area, labour, rainfall, debt etc. on
dependent variable yield for both the crops. From the ANOVA tables we got to know that
labour is statistically significant for rice production and rainfall is statistically significant for
yield of wheat in India (significance level 5%).

Lastly, we performed cluster analysis of the major factors for wheat and rice separately. Rice
was found to be capital intensive whereas wheat was found to be labour intensive.

Initially, when we found out that Indian agricultural sector is labour abundant, we expected
that major exporting crops will be labour intensive. Even though the results are different from
our expectations. The H-O model is still relevant in today’s economic discussion and will still
be used in future research of international trade.

LIMITATIONS

1. Categorizing the different variables into labour or capital was subjective. There might be
more and perhaps better variables that could be considered as labour or capital. It is one
of the major limitations that we cannot take all the possible variables into consideration.

49
2. While considering the data for rainfall, we took annual rainfall for the regression.
However, the crops (wheat and rice) have a sowing season and they require rain majorly
in some particular period(s) in the year. Thus, it becomes ambiguous whether the entire
annual rainfall data is relevant or not.
3. The cluster analysis and regression were performed on state-wise data for the year 2012
only. The results might not good true or might slightly differ for different years.
4. Like Ricardian Theory, the H-O theory model is also static in nature. The theory is based
on a given state of economy and with a given production function and does not accept any
change.
5. Ohlin forgot an important fact that commodity prices are also influenced by consumer
demands.
6. Leontief tested H-O theory under US conditions, and found out that USA exports labour
intensive goods and imports capital intensive goods, but USA being a capital abundant
country must export capital intensive goods than to produce them at home.

REFERENCES

Bharawaj, R. (1962). “Factor proportions and the structure of Indo-US trade”, Indian
Economic Journal.

50
Brecher, R. (1993) “Some empirical support for the Heckscher -Ohlin model of production”.
The Canadian Journal of Economics.

Leontief, W. (1953) “Domestic Production and Foreign Trade; the American capital position
re-examined”. Proceedings of The American Philosophical Society .

Tatemoto, M, and Ichimura ,S.(1959) “Factor Proportions and Foreign Trade: The case of
Japan”. The Review of Economics and Statistics .

Jay Shetty , Amith (2014), “A test of the Heckscher -Ohlin theorem on Indian Exports “.
Faculty of California State Polytechnic University, Pomona.

Alleyne Trevore and Subramanian Arvind (2001), “What does South Africa’s pattern of trade
say about its labour market?” IMF working paper , African Department.

Narayan, A. (2010),”US- India Partnerships: It matters”

Tharoor , S.(2009),”Resilient India “

Hayashi, F. (2009). “Econometrics “Princeton University Press.

Martin, M. , and Kronstandt , A. (2007) “India- US economic and trade relations.”

Congressional Research Service

Choi, E. K. (2007) “Leontief paradox”

Gunay, K., 1992. Bitkisel Uretimde Besin Urun Dengesi (Turkish).

T.C. Merkez Bankasi, Ankara.

Heady, E.O., Dillon, J.L., 1961. Agricultural Production Functions, Iowa State Univ. Press,
Ames.

Henderson, J.M., Quandt, R.E., 1971. Microeconomic Theory.McGraw-Hill, New York.

Johnston, J., 1972. Economettic Methods 2nd edn. McGraw-Hill,New York.

Kazgan, G., 1983. Tarim ve Gelisme (Turkish). Der Yayinlari,Istanbul.

Kilicbay, A., 1986. Ekonometrinin Temelleri (Turkish). IktisatFakultesi, Istanbul Univ., Pub
No. 3330/512.

Kocberber, E., Altin, S., 1990. GAP Bi:ilgesinde 1980 Sonrasin-daki Gelisme1er (Turkish).
Devlet Istatistik Enstitusu, Ankara,

51
Toruner, M., Karakaya, 0., 1975. Bitkisel Uretimde Faktorlerin

Dagilimi Verimlilik ve Uretim Fonksiyonlari (Turkish). Devlet

Planlama Teskilati, Ankara: Pub. No. DPT: 1388-SPD: 270.

Ustunel, B., 1978. Ekonominin Temelleri (Turkish). DoganY ayinevi, Ankara.

Walpole, R.E., 1982. Introduction to Statistics, 3rd edn. Macmillan, New York.

Walpole, R.E., Myers, R.H., 1990. Probability and Statistics for Engineers and Scientists, 4th
edn. Macmillan, New York.

Walters, A.A., 1963. Production and cost functions: an econometric survey. Econometrica 31,
1-66.

Wonnacott, R.J., Wonnacott, T.H., 1970. Econometrics, Wiley,New York.

Yurt Ansiklopedisi, 1984. Ana Yayincilik A.S., Vol. 11. Turkish,

52