ABSTRACT

REDISTRIBUTING INCOME THROUGH TECHNOLOGICAL
CHANGE IN AGRICULTURE: IMPLICATIONS FOR
SRI LANKA'S PADDY SUBSECTOR

BY

S. Nicholas Samuel

This study sought to identify policies to increase
paddy output in ways contributary to both growth and equity
objectives of Sri Lanka‘s Five Year Plan. Technological
and institutional variables were examined as to interacting
effects for increasing paddy productivity among small
farmers and significantly expanding rural employment oppor-
tunities.

Agricultural technology constituted a link between
growth and income distribution. Modern technology which is
the chief source of agricultural growth is not neutral with
respect to either personal or functional income distribu-
tion; it implicitly influences both. Hence explicit
cognizance of equity factors in formulating policies to
foster growth reduces the need to rely upon fiscal and
welfare policies for income redistribution.

A conceptual framework was developed to explain
the interaction of the primary variables of technology

and institutions upon output and the resultant income

8. Nicholas Samuel

distribution. iResource endowments were shown to influence
the interaction of technology with institutions but only
to the extent that the political process chooses to dis-
cover shadow price signals, and once discovered, chooses
to be guided by them.

If equity considerations are to be explicitly
incorporated into policies aimed at increaSing growth,
such policies need to acquire a focus by being directed
at preidentified groups. The program objectives were
defined as increasing the productivity of paddy holdings
below 4 acres in size, and effectuating institutional
modifications that will concurrentlyincrease employment
opportunities.

A regression of five technical and five institu-
tional variables on interdistrict paddy yields revealed
that for policy purposes the seed-fertilizer package was
(as expected) highly significant for yields. Other
variables that were significant were row transplanting,
rotary weeding and handweeding, the risk factor in pro-
duction, and farm size. Chemical weeding and the renting
of holdings had a positive and important though sta—
tistically insignificant effect on yields. The statis-
tical analysis indicated that tractOr use and production
credit were not important for explaining differences in
inter-district paddy productivity. An R"2 of .76 indi-
cated that 24 percent of paddy yield was not explained by

the included variables.

S. Nicholas Samuels

The variables were then assessed as to applica-
bility to small farms (size neutrality) and capacity for
generating employment. The relation between farm size and
output indicated more sc0pe for double cropping among
small farmers than is generally supposed. Tractorization,
found not significant for yields, was antagonistic to
equity. But there were a need to ensure technical flexi-
bility in the use of tractors. Despite scale neutrality
in the physical seed-fertilizer package, a positive rela-
tionship exists between farm size and adOption, mainly
attributed to small farmer lags in adoption. Since techni-
cal information flowed unevenly through the farm size
structure, small farmers postpone innovation until per-
ceived risks are reduced through adequate information from
the demonstrations of early adoptors, the development of
informal local channels, and formal extension contact.
Hence time (adoption lags) was a proxy variable for
information costs to the farmer.

Possible factors inhibiting small farmer access
to technical information were: indivisibility in the
time needed to learn how to use new inputs, extension
biases against small farmers, and non-peer group control
over some multipurpose c00peratives.

Input costs were higher for small farmers due to

high fixed costs in input acquisition.

S. Nicholas Samuels

Production credit was discounted as a factor in the
perceived positive relationship between farm size and adop-
tion of biological technology, although this may not con-
tinue. The analysis suggested that credit was mainly
facilitative and that less reliance should be placed on
credit as a primary means of inducing technological change.

Contrary to theoretical presumptions, tenants have
higher adoption rates than owner-operator counterparts.’
But this factor justified extension rather than tenancy
because it underlines the importance of information and
reveals limited productivity potentials in concomitant
crOp-sharing arrangements.

The risk factor was statistically significant for
inter-district paddy yields and affected small farmers
much more than large farmers for several possible reasons;
the limited credit alloted for pesticide use, lack of
technical knowledge, and externalities in pesticide use.

Considerable congruence exists between growth and
equity for row transplanting and handweeding, but their
relatively greater revenue benefits vis-a-vis chemical
weeding seem offset by higher relative costs.

. Policy suggestions sought to de—emphasize vari-.
ables antagonistic to the objectives of the Five Year
Plan, and to emphasize those contributary to these objec-

tives. Limitations of the study were expressly recognized.

S. Nicholas Samuels

The tentative recommendations'suggeSt alternatives for
further exploration and empirical study, not a blueprint
for action.

The analysis suggested consideration of the follow-
ing: (1) Changes in the foreign exchange rates, interest
rates, and tax allowances for machinery and (change in)
labor costs applicable to agriculture. (2) Elimination
of quantity controls over machinery and spare parts and
the use of both public and private distribution channels.
(3) Reactivation of the credit scheme but assigning it a
neutral role in development strategy. (4) Greatly increas-
ing the emphasis on extension, reducing costs of input
acquisition to small farmers, reducing search and organi-
zation costs of labor utilization, and socialization of
the costs of pest and disease control. (5) Conducting
research into the development of simple water control
systems in the wet zone, task specific machines, and
defining relative costs of labor intensive versus capital

intensive techniques.

REDISTRIBUTING INCOME THROUGH TECHNOLOGICAL
CHANGE IN AGRICULTURE: IMPLICATIONS FOR

SRI LANKA'S PADDY SUBSECTOR
by

S. Nicholas Samuel

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Agricultural Economics

1973

To My Parents

ii

ACKNOWLEDGMENTS

During the course of my research I have been
fortunate to receive the generous and able assistance of
many persons. I am pleased to acknowledge here my
indebtedness to them and to recognize those that provided
special help.

Sincere gratitude is acknowledged to Dr. Lawrence
W. Witt, my major professor and chairman of my thesis
committee, for his immeasurable aid, unfailing courtesy,
and scholarly guidance in directing my thesis and making
my work both pleasant and rewarding. Thanks are also
extended to Dr. Robert D. Stevens and Dr. Lester
Manderscheid for their constructive counsel, expert gui-
dance, insightful suggestions, and enthusiastic assistance.

General acknowledgement is due to the Agricultural
DevelOpment Council for its financial support, and to the
Central Bank of Ceylon for the leave of absence granted.

Finally, I am pleased to acknowledge my gratitude
to my wife, Suramya, for perseverence in decoding my
hieroglyphics and indefatigability in typing numerous
drafts. My thanks to my son, Migara, for bolstering my
spirits at times when the need for such assistance was

dire.

iii

 

TABLE OF CONTENTS

Page
LIST OF TABLES . . . . . . . . . . . . . vii

LIST OF FIGURES o o o o o o o o o o o o o Viii

LIST OF APPENDICES . . . . . . . . . . . . ix
Chapter

I. INTRODUCTION . . . . . ._ . . . . . 1

Problem Setting . . . . . . . . . _ l

The Problem . . . . . . . . . . . 3

Importance of the Problem . . . . . . 6

Objectives . . . . . . . . . . . 10

Methodology . . . . . . . . . 12
Agricultural Production Goals in the Plan

and Their Significance . . . . . . 13

Organization of Thesis . . . . . . . 17

II. AN EXAMINATION OF THEORETICAL AND CONCEPTUAL
MATTERS RELATING TO INCOME DISTRIBUTION AND

GROWTH AND THEIR INTERRELATIONSHIPS . . . 19
Introduction . . . . . . 19
A Theory of Technological Change and

Agricultural Growth . . . . . . . 20
Theories of Income Distribution . . . . 24

Relationships Between Growth and Equity
Factors . . . . . . . . . . . 27
Introduction . k. . . . . '. . .~ . 27
The Separation Issue . . . . . . . 29
The Conflict Issue . . . . . . . . 40
Summary and Conclusions . . . . . . . 54
III. CONCEPTUAL FRAMEWORK FOR RESEARCH . . . . 58
Technology . . . . . . . . . . . 59
Institutions . . . . . . . . . . 62
Resource Endowments . . . . . . . . 63
The Distribution of Power . . . . . . 66
Constraints . . . . . . . . . . . 68
Focus of Variables . . . . . . . . 69
Conclusion . . . . . . . . . . . 71

iv

Chapter Page

IV. ANthENTIFICATION OF EQUITY NORMS AND
TARGET GROUPS 0 o o o o o o o o o o 7 3

Introduction . . . . . . . . . 73
Income Distribution Objectives in Sri
Lanka . . . . . . . . . . . 75
Some Relevant Measures of Income
Distribution . . . . . . . . . 78
Output Objectives . . . . . . . . . 82
Equity Objectives . . . . . . . . . 84
Program Focus . . . . . . . . . 84
Size Groupings . . . . . . . . . 86
Target Groups . . . . '. . . . . 89
Summary . . . . . . . . . . . . 105

V. AN ASSESSMENT OF VARIABLES AFFECTING

INTERDISTRICT YIELD . . . . . . . . . 108
Introduction . . . . . . .. . . . 108
The Econometric Model . . . . . . . 109
The Regression Equation . . . . . . 109

Explanation of the Structure of the
Model . . . . . . . . . . . 110
Results . . . . . . . .. . . . 125
Unincluded Institutional Variables . . . 128
Objective Institutions . . . . . . 129
Subjective Institutions . . . . . . 135
Summary . . . . . . . . . . . . 137

VI. POTENTIAL OF VARIOUS PRODUCTION PRACTICES
INFLUENCING INCOME DISTRIBUTION . . . . . 139

Introduction . . . . . . . . . . 139
Objectives . . . . . . . . . 139
Sources and Methods . . . . . . . 140

Identification of Equity Criteria . . . 141
Scale Neutrality . . . . . . . . 142
Employment Potential . . . . . . . 143

Congruence . . . . . . . . . 143
Assessment of Revenue Variables . . . . 144
Yields . '.. . . . . . . 145
Reasons for Negative Relationship . . . 148
Cropping Intensity . . . . . . . . 149
Results of Model C . . . . . . . . 152
Assessment of Cost Variables . . . . . 153
Mechanical Technology . . . . . . . 155

Biological Technology . . . . . . 165
Other Institutional and Economic Varia-

' bles That Affect the Use of Both

Mechanical and Biological Innovations . 181'

V

Chapter Page

The.Significance.of.Techniques . . . . 197
Summary . . . . . . . . . . . . 204

VII. ALTERNATIVE POLICY SUGGESTIONS FOR SRI .
WKA . O C C C O C I C O C O O 210

Introduction . . . . . . . . . . 210
Policy Alternatives . . . . . . . . 212
Fiscal Changes . . . . . . . . . 212
Elimination of Market Control on

Machinery Supply . . . . . . . . 214

Production Credit . . . . . . . . 215

Extension . . . . . . . . . . . 217

Input (Acquisition) Costs . . . . . 221

Labor Utilization Costs . . . . . . 223

Pest and Disease Control . . . . . . 224

Research Needs . . . . . . . . . 225

VIII. SUMMARY . . . . . . . . . u . . . 228

BIBLIOGMPHY . O O O O O O O O O O C . C 2 54

vi

Table

4.1.

4.3.

LIST OF TABLES

Distribution of Households and Total Income
by Income Groups in the Rural Sector . .

Household Income Received Annually from
Paddy Cultivation . . . . . . . .

Output Target for Paddy . . . . . . .

Households Classified by Size and Number of
Income Receivers in the Rural Sector . .

Area under Paddy in Sri Lanka by Size of
Holding . . . . . . . . . . .

Average Cost of Production Per ACre and
Per Bushel--A11 Island . . . . . .

Structural Data in Income Terms . . . .

Categorization of Farms in Terms of Their
Income Potential . . . . . . . .

Land Tenure Structure of Holdings under
50 AcreS-—1962 o o o o o a o o 0

Amount of Cr0p Failure Classified by Cause
in Sri Lanka . . . . . . . . . .

Maximum Sums Granted for Meeting Input
Costs under 1967 Credit Scheme . '. . .

Double Cropping by Size of Paddy Holding .

Coefficient of Relationship between
Independent Variables and Farm Size . .

Reasons for Credit Default . . . . . .

Amount of Loan Returned Classified by Size
of Area Cultivated . . . . . . . .

vii

Page
79

83
84

89
92

93

95
98
117,
120

124,

150

154

186

186

LIST OF FIGURES

Figure Page
1. Conceptual Framework Showing Short-Run Primary
Relationships Between Categories of
Variables Affecting Output and Income
Distribution 0 O O O O O O O O O O 60

2. Lorenz Curve of Distribution of Rural
Income and Paddy Holdings . . . .. . . . 77

3. Income of Households in the Rural Sector . . 77

4. Effect of Information and Tenancy on the ,
Diffusion Process . . . . . .. . . . 172

5. Effect of CrOp Sharing on Input Usage
(AdOption) . 1. . . . . . . . . . . 172

viii

LIST OF APPEND ICE S

Appendix Page

A. Sources of Prior Beliefs for the Evaluation
of Statistical Results . . . . . . . 241

B. Results of Regression Regarding Interdistrict
Paddy Yields--Model A . . . . . . . . 244

C. Results of Regression Regarding Interdistrict
Paddy Yields--M0del B o o o o o o o o 246

D. Results of Regression Regarding Cropping
Index--MOdel C o o o o o o o o o o 2 4 8

E. Correlation Matrix . . . . . . . . . 250

F. Tabulated Data . . . . . . . . . . . 252

ix

CHAPTER I

INTRODUCTION

Problem Setting

 

There is widespread evidence that technological
changes in agriculture, through the use of modern scien-
tific mechanical and biological inputs, have increased
and have demonstrated a potential for further increasing
agricultural output in less developed countries.1 In
three South Asian countries the growth in agricultural
output from 1966 (the year before the first major distri-

bution of new high-yielding varieties) to 1971 was fairly

substantial.
Sri Lanka 14 percent
India 25 percent
Pakistan 17 percent

While on a per capita basis the increase is much less,
due to fairly high rates of population growth of over 2.5

percent, the change is nevertheless positive varying

 

1For instance, Lester Brown regards the new seeds
as being "likely to be a greater force for change than
any technology ever introduced into the poor countries."
The Social Impact of the Green Revolution (International
Conciliation, 1971), p. 44.

2Food and Agriculture Organization, Production
Yearbook 1971 (Rome: FAQ). All quantitative data in
this chapter will be from this same source unless other-
wise stated.

 

 

 

between 5 and 12 percent between 1965 and 1971, depending
upon the base chosen. Productivity increase in wheat
and rice yields have been substantial, due to the esti-
mated acreage planted with new high-yield varieties
increasing from 200 acres in 1964 to 30 million in 1971,
and the adoption of modern methods. For instance, in
Sri Lanka rice yield increased from 38.8 bushels per
acre in 1963-64 to 51.1 bushels per acre in 1969-71,1
while total production increased from 50 million bushels
to 77 million bushels in the same period.2

However, the adoption of modern methods has not
been widespread on an overall basis. At the time the
productivity increaSes were being hailed as constituting
a "Green Revolution," some writers cautioned against
3

exaggerating the magnitude or scale of the changes.

Some writers considered it merely an "enclave" development

 

1Department of Census and Statistics, Statistical
Pocket Book of Ceylon, 1971 .

2While the innovations were mainly in respect to
inputs, it was also a product innovation to the extent
that the new wheat and rice products differed from the
traditional products in taste and cooking characteristics.

3See Welter Falcon, "The Green Revolution: Gen-
erations of Problems" (paper presented at AAEA meeting,
Columbia, Missouri, 1970). See also U.S., Department of
Agriculture, Rice Situation, USDA ERS RS-lS (March,
1970); and C. R. Wharton, "The Green Revolution: Cornu-
COpia or Pandora's Box?" Foreign Affairs, April, 1969.

 

 

Pa
III

.
on,

0V.

0..“

 

 

 

leading to "dualism."1 By 1971 only about 12 percent of
rice land and 25 percent of wheat land in Asia was in
improved varieties.2 Hence there is considerable scope

for further changes.

The Problem

 

The very process of technology diffusion has the
inherent characteristic that it affects income dispari-
ties, and hence the distribution of income.3 More
knowledge about this relationship between growth and
equity will make it possible to identify policies for
Sri Lanka that will increase output, and at the same
time contribute towards an acceptable distribution of
income.

The process of diffusion affects income dispari-
ties in both relative and absolute terms. Future increases

in income will occur as those who have already adopted

 

1K. N. Raj, "Some Questions Concerning Growth,
Transformation and Planning of Agriculture in Developing
Countries," Journal of Development PlanniES; No. l (1969).

2Although according to FAO statistics the trend
rates of growth in production as well as productivity
before and after the "Green Revolution" do not differ
materially except for wheat, these results cannot be
relied upon too much for information on South Asia because
it probably subsumes regional trends. FAQ, The State of
Food and Agriculture 1970 (Rome: FAO), p. 131.

3The concept of income distribution is wider than
the concept of income disparities. A redistribution of
income could either increase or decrease income dispari-
ties depending on the relative income levels of those who
pay and those who benefit.

 

 

 

modern methods move towards a more complete (optimal)
ad0ption of a given technology, and as those who have not
yet adOpted move towards experimental adoption. Cur-
rently, the. degrees of adoption are infinitely many,
varying from complete adoption of a given technology to
no adOption of that technology. This implies that the
process of diffusion inevitably involves differences in
the rates of increase of output on different farms,
depending on the degree of innovation. Such differences
therefore have a direct bearing on relative income dis-
parities. Two cases in this "relative disparity" category
need to be distinguished: differences among adoptors of
technology and differences between adoptors and non-
adoptors, given the farm size group. Absolute income
diSparities can increase due to structural differences in
(irrigated) farm size. Thus even if the rate of increase
in output is the same for different farms, absolute
income differences could increase due to difference in
the size of (irrigated) land holdings.

The fact that the very process of technological
change influences income disparities suggests a need to
understand more about the nature of this relationship,
both for national-level policy formulation and for local-
level implementation. A better knowledge of these
relationships likely will provide improved insights into

the policies needed in Sri Lanka that will contribute

toward a desired distribution of income, while at the
same time increasing output. It seems preferable that
the "Green Revolution" stimulate an agricultural trans—
formation rather than an agricultural dualism. More
specifically, more knowledge of these relationships may_
indicate the type of projects needed, the form such
projects should assume, their direction and orientation,
their focus, and the manner in which they could dovetail
into integrated programs at the national or regional
level, in order that the "Green Revolution" may be made
more pervasive. Suchfiknowledge also likely‘will suggest
the type of complementary institutional modifications
needed to enable the programs to achieve their desired
goals.

Several research studies, as listed in Appendix
A, have dealt with technological change and its effect on
income distribution either centrally or incidentally.
These studies have been of a micro nature and related to
Specific locations in India, Pakistan, and Bangladesh.
They could be expected to contribute useful information
for policy formulation and implementation at the local
level. Less attention however appears to have been paid
to the assembly of such information into the broad inte-
grated picture necessary for macro-level, national,
decision making. To the extent that these issues have

been considered at this national level, the decisions have

 

 

 

'9

tended to be based on conceptual formulation and intuitive
judgments. Some of these micro-level studies may have
made a modest input into these judgments, or will in the
future (Since many are only recently completed). But
there does not appear to be an attempt to integrate these
studies and develop analyses of national policy options
based on such firm empirical foundations. Similarly
there does not appear to be an adequate reaching out by
national policy formulators for empirical evidence to
enable them to better assess the potential performance

of relevent alternative programs and institutional modi-

fications.

Importance of the Problem

 

Although fairly substantial rates of growth of
agricultural output have occurred, there are indications
that the poverty rate, as measured by the percentage of
the persons below a predetermined poverty line, has
increased. A recent study in India shows that 40 to
50 percent of the total population has a per capita
income below the official poverty line. Moreover, the
per capita .income of this group has declined over the
last two decades, while the average per capita income

increased.1 In Sri Lanka, despite an increase in per

 

lM. Haq, "Employment in the 19705: A New Per-
spective," in Inernational Development Review, December
1971 (Ottawa: Society for InternatiOnal Development),
pp. 9-10.

 

capita agricultural output of around 12 percent in the
period 1966 to 1971, a growth of real GNP of 14 percent

in this same period, and an average increase in per

capita real incomes of roughly 2.1 percent per year during
the period 1959 to 1971,1 unemployment has increased from

3.5 percent2 of the labor force in 1959, to 15 percent3

of the labor force in 1971.4 Also, there are indications
that the extent of poverty increased during the period
1960-68. A recent study based on the data gathered by a
1969 Consumer Survey in Sri Lanka revealed that the extent
of poverty (measured by the number of persons below the
$100 per annum income level) increased from 49 percent

to 60 percent during the period 1960 to 1968.5 There is
also evidence that real wages in agriculture have drifted

downward and, if the mid-1950's are compared with the

early 19703, they have fallen about 4 percent.6

 

. 1Central Bank of Ceylon, Annual Report 1971
(Colombo).

2National Planning Council, The Ten Year Plan
(Colombo: Planning Secretariat, 1959).

3Departmentof Census and Statistics, Socio-
Economic Survey 1969-70, First Round (Colombo: 1971).

 

 

 

4Different methods of measuring unemployment in
the two years probably were used, thus limiting the reli-
ance that can be placed in comparing these data.

5Ranier Schickele, "Improving Income Distribution
as a Development Goal," Staff Paper P72-17 (Ag. Econ.
Department, University of Minnesota, 1972).

6Central Bank of Ceylon, Bulletin (Colombo:
August 1971), Table 37.

 

That poverty is widespread is clear. In 1969-70,
44 percent of the non-estate (i.e., traditional) rural
households received less than Rs. 200 ($33) per month
per person in an (average) household of five persons.1
Poverty also gives every indication of growing. One
forecast expects the number of peOple economically
active in agriculture to grow by 50 percent from 2.4
million today to 5.3 million in 1985.2

Growing evidence of this phenomenon has stimula-
ted increasing concern for equity considerations and
contributed to the demise of the ideology of economic
growth. Interest has shifted from the narrower considera-
tions of eConomic growth to the broader issues of economic
development. Development is increasingly viewed in multi-
dimensional terms, and defined to include issues of
poverty, unemployment,and the distribution of wealth,
besides growth. Concern has thus tended to move from
increasing productivity to concern for whose productivity
is to be increased. The United Nations, by addressing
itself to the problems of poverty, has in the Second

Development Decade moved away from its reliance on the

rate of growth of GNP as a primary indicator of economic

 

1Department of Census and Statistics, Socio-
Economic Survey (Colombo: 1972).

2International Labor Organization, "Matching
Employment Opportunities and Expectations," Report
(Geneva: ILO, 1971), p. 86.

 

development.l WOrld Bank president Robert McNamara has
commented as follows:

The state of development throughout most of
the developing world is unacceptable. It is unac-
ceptable because hundreds of millions of people are
living at levels of deprivation that simply cannot
be reconciled with any rational definition of human
decency. . . . Current develOpment programs are
seriously inadequate because they are not signifi-
cantly reducing the poverty which shapes and limits
these lives. And though the matter is complex,
basically we know why. . . . the developing coun-
tries are not moving decisively enough to reduce the
severe social and economic inequities among their
own pe0ples. . . . The broad statistical evidence is
clear that there is dangerously skewed distribution
of income both within developing nations. . . .
Development simply cannot succeed unless that mas-
sively distorted distribution of income Es brought
into a more just and reasonable balance.

It is thus increasingly being recognized that
growth does not necessarily imply develoPment,3 and that
development is a multi-dimensional concept that encom-
passes growth, employment, and equity factors.4

Development has been equated with "the fulfillment of

 

1United Nations, Report on the Second Deve10pment
Decade, UN Sales No. E 71, ll, A.2.

2Robert S. MacNamara, "The Environmental Dilemma,"
in war on Hunger (Washington, D.C.: U.S., Department of
State, December, 1971).

 

 

3See for example the experience of Liberia: R. W.
Clower et a1., Growth Without Development (Evanston, Ill.:
1966). Also, W. J. Barber—IA Critique of Aggregate
Accounting Concepts in UnderdevelOped Areas," Bulletin of
the Oxford University_Institute of Economics and Statis-
tics (November, 1963); and I. Adelman and M. Taft, Societ ,
POlltiCS and Economic Development (Baltimore: '1967).

4D. Byerlee and C. K. Eicher, Rural Employment,
jMiqration and Economic Development: _Theoretical Issues
and Empirical Evidence from Africa (East Lansing: Depart-
ment of Agricultural Economics, Michigan State University,
1972).

 

 

 

 

 

10

human potential,"1

the "upward movement of the entire
social system,"2 a reduction of poverty,3 and reduced
unemployment.4

Sri Lanka's economic objectives, as enunciated in

The Five Year Plan, are consistent with the foregoing con-

 

ceptions of development. The Plan states:

The trend in Gross National Product does not
indicate the growing crisis in unemployment, in
income disparities, and in the balance of payments.
Statistics of Gross National Product at per capita
income are insensitive both to the distribution of
income and to the virtual stagnation of certain
vital sectors of the economy. Nor do these sta-
tistics indicate'the maturing crisis within
society.

Objectives

 

The broad objective of this study is to examine
how technological changes in paddy cultivation could be
made to contribute to the growth and equity objectives
of Sri Lanka's Five Year Plan. This will require an

identification of the factors that presently inhibit

 

1Dudley Seers, "The Meaning of Development,"
Communications Series Br. 44 (Sussex: Institute of
Development, 1970). '

 

2G. Myrdal, Asian Drama (New York: 1968),
p. 1869.

 

3Ranier Schickele, Agrarian Revolution and Econ-
omic Progress (New York: Praeger, 1968).

4International Labor Organization, Towards Full
Employment; A Programme for Columbia (Geneva: ILO, 1970).

 

 

 

 

5Ministry of Planning and Employment, The Five
Year Plan (Colombo: 1971). This document will hereafter
be referred to as the Plan.

 

11

increases in productivity among small farmers, and pre-
vent increased employment opportunities from developing
in the rural sector. Conversely, the intention is to
identify variables that have a potential for increasing
paddy productivity and improving the distribution of
income, with a View of activating or changing them for
desired performance.

The specific objectives of the study are of a
sequential nature that cumulatively facilitate the ultimate
objective of identifying the policies that will permit
the paddy subsector to deliver expected performance with
respect to both output and equity. These specific objec—
tives are as follows:

1. To examine the theoretical and conceptual
relationships between growth and equity factors from a
policy standpoint, and to utilize the insights gained
to develop a conceptual framework for the subsequent
analysis of the relationships between technological
change and equity.

2. To translate policy norms from a social welfare
function in Sri Lanka into concrete objectives for policy,
by identifying socio-economic groups upon whom programs
should be focused.

3. To identify the variables that are significant
for paddy output and income distribution in Sri Lanka by
means of statistical techniques and a review of other

primary data based micro studies.

12

4. To use the insights gained from all of the
above to identify policy changes, at primarily a macro
level, that will facilitate the type of institutional
change and agricultural programs that will yield desired

results with respect to output and equity.

Methodology

 

Linear multiple regression analysis will be
utilized to conduct a statistical analysis of macro data
pertaining to Sri Lanka's paddy subsector. The data
utilized will be primarily from published secondary
sources.

The analysis will be essentially Bayesian in that
the statistical results will be assessed in the light
of prior beliefs. In formal terms this involves adjust-
ing a prior distribution to take account of observed
data, with the weight attached to the observed data
depending on the probability of the posterior distribu-
tion.

The dominant source of this prior belief will
be the several primary data based micro studies listed
in Appendix A, which examine various aspects of the rela-
tionship between technological change and income
distribution, in India, Pakistan, and Bangladesh. In
designing agricultural modernization policies it must be
acknowledged that no universal answers exist. The

experience of different localities with broadly the same

13

ethnic, social, and cultural backgrounds, however, could
be expected to provide insight and material that may be
considered as possibilities for application in Sri Lanka.

The cross section of information in the micro
studies, when assembled in the light of the statistical
results, is expected to provide a picture of the patterns
of interrelationships between growth and income distribu-
tion, and hence a reliable basis for the policy
prescriptions ultimately to be made.

Agricultural Production Goals ip the
Plan and Their Significancel‘

 

 

The Plan demonstrates an appreciation of the con-
tribution that the agricultural sector could make to
economic development.2 Since agriculture occupies a
dominant position in the Sri Lanka economy, an increase
in agricultural output and an equitable distribution of
such output is likely to have an overall effect on
economic development. In 1970 the agricultural sector
contributed 35 percent of the Gross National Product and
employed 55 percent of the work force. The dominance of

the agricultural sector thus makes the interplay between

 

1The data in this section is from the Five Year
Plan, unless otherwise indicated.

2It has to be recognized that: "Neither growth
or equity problems can be solved by . . . the agricul-
tural sector. The employment problem in particular is
total-economy in character. . . ." Falcon, Op. cit.

 

 

14

development means and development objectives important
in that sector.
The Five Year Plan states:

. . . the agricultural sector presents the greatest
sc0pe for the expansion of output and employment in
relation to investment outlay. In agriculture
there are substantial growth possibilities which
are not constrained by a shortage of foreign
exchange. This is due to the scope for more inten-
sive use of land and labor, for increasing output
with high yielding strains in a variety of crops,
and for better use of available water resources.
The possibilities of import substitution are also
greatest in agriculture. . . . The gap between
population and resources has to be met by an expan-
sion and diversification of small-scale agriculture
on a wider and more systematic basis. . . . Since

a large prOportion of the unemployed live in rural
areas this task is urgent."

In terms of strategy, the Plan shifts the emphasis
for agricultural development from the three major export-
oriented plantation crops of tea, rubber, and coconut to
the mainly traditional crops2 of rice, pulses, chillies,
and onions. A significant nontraditional crop to be
expanded is sugar cane. Trends over the past fifteen
years suggest a very limited potential for cumulative
development in the plantation crops due to low income and
price elasticity of demand factors,3 and the oligopolistic

structure in production and marketing which, particularly

 

1The Plan, p. 33.

2A traditional crop would be one which has his-
torically been cultivated under traditional as opposed to
modern methods.

3Terms of trade declined by 7.6 percent in 1971.

 

15

for tea, is controlled mainly by foreign interests.l
During the period since 1956 tea production increased by
approximately 35 percent, rubber production by approxir
mately 40 percent, and coconut by some 15 percent. Yet
even these modest gains in output were offset by declining
prices.2

The shift to traditional cr0ps represents an
attempt to earn foreign exchange through import displace-
ment rather than through export expansion. The Central
Bank Report for 1971 states:

As in previous years, the commodity(terms of
trade moved unfavourably and had a more severe
effect in absolute terms than in any previous
year. . . . It would appear that the major obsta-
cle to economic development in Ceylon continues
to be the shortage of foreign exchange. In fact,
in 1971, due to the constraints of the external
payments position, restrictions had to be applied
on a wide range of goods, including intermediate
and investment goods, which probably contributed'
to the lower rate of growth of the economy. . . .
Given the generally unfavourable prospects for
the prices of the three major exports, it is
inevitable that development strategy must, in the
Short run, lean morg heavily on a policy of import
substitution. . . .

The substantial need for foreign exchange can be inferred
from the fact that the value of investment goods imported

was cut back by 23 percent in 1971, at a time when the

 

1Food and Agriculture Organization, Tea Trends
and PrOSpects, Commodity Bulletin Series (Rome: FAO, 1960).

2Central Bank of Ceylon, Annual Reports 1956-71
(Colombo).

3

 

 

 

Central Bank, Annual Report 1971, p. 233.

 

16

value of rice, pulses, onions and sugar imports increased
by over 4 percent.1

Considerable development potential exists in the
development of the paddy subsector alone. Although
government programs have been directed at the traditional
subsector producing paddy in the pasteight years, and
an increase in yield of about 6 percent per annum occurred
during the period 1966 to 1971, yet 40 percent of Sri
Lanka's rice requirements were still being imported in
1971. Rice accounted for 46 of the import bill in 1971,
which was 8 percent of the GNP.2 Hence, the import dis-
placement that is likely to occur with an increase in
domestic rice production could, if complete, release
nearly half the total import bill for the import of
investment goods. In fact,this is substantially an
objective of the Plag which envisages a 45 percent increase
in rice output in five years. If achieved, this would
meet 97 percent of domestic rice requirements by the end
of the Plan period in 1976.

The Flag places primary emphasis on intensive, as
Opposed to extensive, paddy production. Although the
program for irrigation and land development is expected
to bring into production about 100,000 new acres, the
major portion of the increase in output (80 percent) would

take place through increasing productivity on land already

 

l 2

Ibid. Ibid.

17

under paddy cultivation. The newly prepared lands are
expected to account for only 9 percent of total rice
output in l976-—up from_l percent in 1972. This is con-
sistent with the urgency of the need to increase output,
as stimulating output is usually easier than initiating
production; the former requiring only a complement of
resources to combine with existing land, labor, and

capital.

Organization of Thesis

 

Chapter Erexamines the theoretical and conceptual

 

relationships between growth and equity factors with a
view to developing a conceptual framework for subsequent
analysis, obtaining clues with regard to the subsequent
direction of the study, and for generating relevant
hypotheses for later verification.

Chapter III develops a short-run conceptual frame-

 

work which is expected to be helpful for both analyzing
the micro studies cited in Appendix A and for tidying up
the analysis of relationships between variables in the
rest of the study.

Chapter IV translates political values in the

 

social welfare function into specific objectives for
equity policy by identifying the socio-economic groups

upon whom policies Should be focused.

18

Chapter V identifies the variables that are sta-

 

tistically and economically significant for explaining
interdistrict paddy productivity.

Chapter VI assesses the variables already identi-

 

fied, from the vieWpoint of equity; their applicability
to small farms and their employment generation potential.

Chapter VII uses the concepts and facts previously

 

developed to identify policy steps that should be taken
that will contribute towards the growth and equity objec-
tives of the Five Year Plan.

Chapter VIII summarizes and synthesizes the

 

thesis.

CHAPTER II

ANIQQMDWBHONCE‘THEHETEHElAND(INKEPEEL
NETHflEBRHJflINSUDIDKDMEIHSTKEMTKIJ

AMDGEOWHIANDCHEERIRHERNEATElEmIHS

Introduction

 

This chapter first examines the theories of
growth and income distribution, and then proceeds to
examine their interrelationships. Two aspects of equity
factors will be explored: (l) The question of whether it
is desirable or feasible to leave matters relating to ‘
equity to be dealt with independently of policies designed
to increaSe agricultural output (the separation issue),
and (2) the basis for, and validity of, the belief that
the two goals of growth and equity are in conflict (the
conflict issue). The discussion of these interrelation—
ships is directly relevant for and related to the
development of a conceptual framework for analysis pre-
sented in the following chapter. It will also be shown
that these interrelationships are central to the objective
of this study, that the conclusions with respect to them
are useful for suggesting a direction for the study to
take, and for generating relevant hypotheses for later

verification.

19

20

A Theory of Technological Change and
Agricultural Growth

 

 

Increasing agricultural productivity is more than
a problem of investment, as investment in traditional
factors involves a movement along an existing production
function yielding only low returns at the margin. T. W.
Schultzl argues that comparatively few significant ineffi-
ciencies exist in the allocation of the factors of H”
production in traditional agriculture. To the extent that
this is true, no great increase in agricultural production
is likely to occur either by reallogating the factors
already at the disposal to farmers, or by applying more
units of traditional inputs to the production process.

His argument stresses concentration on new tech-
nology, which clearly brings much greater and more rapid
increases in agricultural output. Such technology pushes
out the production function through the use of inputs
embodying modern scientific knowledge, and provides both
the incentive and the opportunity for investment in agri-

culture through potentially higher returns.2 Since most

 

1T. W. Schultz, Transforming Traditional Agricul-
ture (New Haven: Yale University Press, 1964).

2This model fits into the transitional stage II of
the Perkins-Witt, Johnston-Mellor and Hill-Mosher models of
the agricultural development process that are based on his-
torical analogies. M. Perkins and L. Witt, "Capital
Formation: Past and Present," Journal of Farm Economics,
Vol. 43, No. 2 (May, 1961); B. F. Johnston and J. W. Mellor,
"The Role of Agriculture in Economic Development," American

 

 

21

such inputs are purchased, one index of modernization
would be the ratio of purchased to total inputs in a farm.
This ratio would be useful for determining the degree of
modernization and for deciding if it is significant.1
The extent of the increase in output resulting from
the application of modern inputs depends on the availarh,
bility of complementary inputs that together constitute
a package, and the institutional context in which such
technology is applied. Thus, if the input package is
incomplete, or if the institutional context is inappro-
priate, the output effect could be less than optimal.
Agricultural technology is important not only
because it is the primary source of agricultural growth
but also because it has direct implications for the func—
tional and personal distribution of income. Technology
constitutes a primary element in the conceptual framework
to be developed. 'Hence a theoretical classification of

technological changes appears relevant.

 

Economic Review, Vol. 51, No. 4 (September, 1961);

E. F. Hill and A. T. Mosher, "Organizing for Agricultural
Deve10pment" (New York: Agricultural Development Council,
1963; mimeographed).

1This index of modernization could, however, be
an unreliable indicator of the degree of integration of
farms in the product market, as although the degree of
modernization may be significant, an inadequate land
base may prevent the generation of a marketable surplus
above subsistence needs.

 

22

The following analysis is based on Hicks' theory1

as applied to agriculture by Heady,2 and is postulated in
a Cobb-Doublas world which assumes constant returns to
scale, constant product prices, equilibrium in the factor
market, and two factors of labor and capital. Under
these assumptions a change in technology which affects
the relative shares imputed to the two factors in the.
total product may be classified as labor saving, neutral,
or capital saving, depending on whether the change r
increases the marginal productivity of capital by rela-
tively more, in the same proportion, or relatively less.
These changes may be relative or absolute. A
labor saving technical change may increase the marginal
product of both factors, but the marginal product of
capital is increased by relatively more. Thus it becomes
profitable to apply more of both resources to the produc-
tion process, but proportionately more units of capital
will be applied. Also, technical changes may increase
the marginal productivity of capital, while at the same
time decreasing the marginal productivity cf labor. In
this instance, fewer units of labor will be employed.

Thus, while in both cases the relative share imputed to

 

'1J. R. Hicks, "Distribution and Economic Progress:
A Revised Version," The Review of Economic Studies, Vol.
IV (1). pp. 1-12.

2E. L. Heady, "Changes in Income Distribution in
Agriculture with Special Reference to Technological
Change," Journal of Farm Economics, XXVI (3), pp. 435-47.

 

 

 

23

labor is changed in the same direction, the absolute share
moves in an Opposite direction.

The distinction is made in the literature between
mechanical and biological technOlOgy, the former being
regarded as labor displacing while the latter is regarded
as being land replacing.1 This distinction, although
conceptually useful, is not mutually exclusive. For
instance, tractors that break a timeliness bottleneck can
be land rather than labor replacing over the Spectrum of
one year. This difficulty of making\a distinct classifi-
cation between labor-saving and land—saving technologies
could be traced to the fact that when new factors are
applied the whole production function may change and
hence induce changes in the use Of other factors not
directly applied.

The foregoing analysis assumes that product
prices remain unchanged. However, it is possible for
prices to fall within limits without reducing the
profitability of (or incentive for) investment in agricul-
ture. Such profitability is determined by the difference
between per unit price and per unit costs. There is a
theoretical presumption that the unit costs from package
cultivation are usually much less than from traditional

methods. While the application of modern inputs increases

 

1This distinction can be traced to Earl O. Heady's
Economics of Agricultural Production and Resource Use
(New York:' Prentice Hall, Inc., 1952), pp. 818-19.

24

total costs Of production, the productivity of these
inputs can greatly increase output, and thereby reduce
average costs considerably. Thus, innovators who increase
net output faster than the product price falls could still

gain.

Theories of Income Distribution

 

The theories Of distribution in the literature
can be classified into two categories: (a) those dealing
with functional distribution and (b),those relating to
personal distribution.l‘ The former attempts to explain
the distribution of income among factors of production
while the latter relates to the distribution of income to
people (households) by size brackets Of income.

The theory of functional distribution could be
traced to the Ricardian theory which seeks to explain the
distribution of income to the various factors of produc-
tion in terms Of their marginal productivity. It is in
fact the theory of factor prices--an extension of price
theory. Relative shares in the distribution of income
depend on (a) relative factor prices that have been

determined by the interaction of supply and demand and

 

1Some writers distinguish a third category called
the Theory of Distributive Shares, which is an aggrega-
tion of functional distribution at the macro level. This
is of limited use for this study. See, for instance, Jan
Pen, Income Distribution (London: Allen Lane Penguin
Press, 1971), and J. Marchal and Ducros, eds., The Distri-
bution Of National Income (New York: Macmillan, 1968).

 

 

 

25

(b), the elasticity of substitution which is a technical
factor, having its basis in the production function.
Market forces will generate a set of factors to clear the
market and make their individual price equal to the value
of their marginal product.

Generally, however, the neoclassical theory of
income distribution per se, which is £22 theory of
income distribution, is of limited usefulness for the
purpose of this study. It has been labeled a theory of

1 This functional distribution

”pseudo distribution."
accepts the ownership pattern of incOme generating
aSsets (wealth) as given and therefore directly reflects
the extant distribution Of wealth. It is also based on
the assumptions of perfect competition in both product
and factor markets. Besides these shortcomings, its
relevance to this study exists only to the extent that
it coincides with personal income distribution. This
implies that its relevance depends upon the payment Of
labor being the sole source of low income, while capital
income accrues to high income recipients. This require-
ment therefore limits the usefulness of the neoclassical
theory in a subsistence and small farmer situation where

Often no division exists between the ownership of capital

and labor, and where returns to both can be low. This

 

lTerm attributed by Jan Pen, Op. cit., to .
E. Cannon, "Division Of Income," Quarterly Journal of
Economics (1905).

 

 

26

latter point implies that even if levels of income can
be associated with factor shares, such factor Shares do
not usually correspond to distinct socio-economic groups.
Of greater relevance in a welfare and policy
sense is the size distribution of income to people. The
distribution of personal income provides the link between
technological change and income disparities. For reasons
of convenience and practicability, such personal distri-
bution will be disaggregated into economic groupings.
Hence this study is concerned with the incOme that
accrues to economic groups, rather than whether the
income consists of wages, profits, rent, etc. A quite
different concept underlies personal income distribution.
Changes in governmental transfer payments to nonproducers,
for example, will affect Size distribution through the
change in personal income, but leave functional distri-
bution unchanged. Conversely, labor intensive technology
could change the personal distribution of income by
changing the functional distribution.
A search Of the literature reveals that no general-
theory concerning personal income distribution exists.
The redistribution of personal income involves interper-
sonal comparisons of utility and non-Pareto better changes.
Fundamentally, questions relating to income redistribution
deal with matters of relative property rights, where

property rights are defined to be "the set of rights and

27

Obligations established by law, custom and covenant which
define relations among members of a community in respect

their control over resources."1

Questions addressed to
the bundle Of prOperty rights are prior to those asked by
classical economists, who accepted the institutional

variable as datum and subsumed it in ceteris paribus.

 

Ultimately the distribution of personal income must be
determined by the political process and reflect the
distribution of political power. Economic theory there-
fore cannot be expected tO provide all of the guidelines
concerning the redistribution of income. Such guidance
must derive directly from some exogenous social welfare
function. 7

Relationships Between Growth
and Equity Factors

 

Introduction

 

Technical changes that greatly increase output do
not necessarily promote development. As previously stated,
economic growth is a necessary but inadequate condition
for develOpment, which has been defined in multi-.
dimensional terms to contain both growth and equity
components. Both in the theoretical literature and at
the level of practical planning, the two goals of

"economic growth" and "equity" (i.e., greater equality

 

1J. D. Shaffer, Property, Market Structure and
Efficiengy, Agricultural Economics Miscellaneous Report
1966-69 (East Lansing: Department of Agricultural Econ-

omics, Michigan State University).

 

 

28

in incomes) often have been considered as amenable to
separate treatment and also Often regarded as being in
conflict.

The question of whether policies directed at
equity Objectives are amenable to separate treatment, and
can be dealt with independently of policies designed to
increase agricultural output, relates to planning and
programming strategy. The second aspect of whether the
Objectives of equity and growth are in conflict, deals

with the conseggences Of growth and income redistribution

 

policies. Therefore, although the policies,may be
implemented separately, a trade-Off could still result
in terms Of effects. To the extent that growth policies
can be dealt with separately, the income redistributional
aspects could be left for some subsequent (second) stage,
through the device of fiscal policy, after growth has
generated increased income flows, and after such income
flows have accrued to the owners of the various factors
of productiOn. To the extent that the objectives are in
conflict, a trade-Off between growth and equity is
implied, and growth could be expected to cause, and be
caused by, the skewness in the distribution of income.
This section first explores the extent to which
it is desirable and feasible from an Operational stand-
point to deal independently with equity policy--the

separation issue. Later it examines the theoretical

29

reasons for believing that growth and equity may be in

conflict with each other-—the conflict issue.

The Separation Issue

 

Theories of separation.--Several writers have

 

advocated a separation between growth and equity Objec-
tives. Two schools of thought may be identified:

(a) those who believed in "contrived dualism" and (b),
those who regarded growth as an intermediate goal with
equity being an ultimate one.

The contrived dualism model advocated by Owen1

and modified by Thiesenhusen2 evinces a concern for both
growth and employment. These Objectives are to be
achieved concurrently through the creation of two subsec-
tors, one emphasizing growth of output (a marketable
surplus), and the other emphasizing equity (the creation
of employment Opportunities). However, contrived dualism
in effect argues for disparities in income between house-
holds.

The other school of thought considers growth as

a means to an end, the end being equity. Hence policy

should concentrate on maximizing rates Of growth in the

 

lwyn F. Owen, "The Double Development Squeeze on
Agriculture," American Economic Review, Vol. LVI, NO. 1
(March, 1966), p. 64.

 

2William C. Thiesenhusen, "Latin America's
Employment Problem," Science, Vol. 171 (March 5, 1971),
p. 871.

30

first instance, and on dealing with redistributional
equity aspects at some subsequent (second) stage. The
following quotation seems representative of this view-
point:
we should hypothesize that economic efficiency
is a less ultimate goal than an appropriate dis-
tribution of income. As a consequence, economic
efficiency becomes a means to an end rather than
competitive to those ends. . . . Economic growth
appears (here) clearly in a suppOrtive role to
higher Objective of men. It is not likely, how-
ever, that such ultimate goals can be achieved
when the production of economic wealth in a
society is at a low level . . . improved national
income is viewed as an input and individual ful-
fillment is considered to be the product.
W. Arthur Lewis also supports this notion when he says,
"The advantage of economic growth is not that wealth
increases happiness, but that it increases the range of
human choice."2 To advocate the foregoing supports the
following positions by implication: (a) that income
redistribution refers to the redistribution of income
flows through government taxing and Spending and (b),
that the redistribution is to take place ex post, after

income vests-~i.e., after such income accrues to the

owners of the various factors of production.

 

. 1Emery Castle and Russel Youmans, "Economics in
Regional water Research and Policy," American Journal
Of Agricultural Economics, Vol. 50 (December, 1968),

p. 1662.

 

 

2W. A. Lewis, "Is Economic Growth Desirable?" in
Theory of Economic Growth (Homewood, Illinois: Richard
D. Irwin, Inc.)

 

31

Three reasons can be identified against the
advisability Of leaving income distribution matters to
be dealt with at a later stage:

(a) To regard income redistribution as the
transfer of income flows through fiscal policy raises the
issue of the possibility of a trade-Off between capital
formation and equity, and thus between higher incomes and
employment in the future versus higher income and employ-
ment in the present.1

(b). TO concentrate on growth in the first
instance may give rise to factors that render difficult
or wasteful the movement toward the ultimate equity
objective of increasing employment and reducing poverty.
There is evidence that growth can be rapid if concen-
trated in a small number of larger farms utilizing large-
scale capital-intensive techniques.2 But this gives rise
to fixed assets and costs in the form of capital equip-
ment, as Well as institutional and organizational
infrastructure geared to the needs of large farms, which
would be the major market participants. It also gives

rise to vested interests, which are the political

 

l O I 0 I t C I
This lssue is examined later in thls subsection.

2W. Falcon, "The Green Revolution: Generations Of
Problems," American Journal of Agricultural Economics _
(December, 1970). Also, ILO, Towards Full Employment, 02:
cit.; Carl K. Eicher, et a1., Employment Generation in
African Agriculture, Institute Of International Agricul-
ture, Research Report NO. 9 (East Lansing: Department of
Agricultural Economics, Michigan State University, 1970).

 

 

32

manifestations of fixed assets. Once these fixed assets
get established they are by definition difficult to
change, and so limit the feasibility of moving towards the
ultimate equity Objective. Such difficulty could for all
practical purposes make the intermediate growth Objective
an end in itself, at least in the "short run."1
(c) The third Objection to separating equity
from growth policy is because of administrative insuffi-
ciencies in the rural areas of developing countries, which
limit the efficacy of government taxation and spending.
Such machinery tends'to have an inelastic supply in the
short run, and cannot be expanded except at high financial

and social costs, due to the high Opportunity cost of

scarce talents.2

Redistribution Of income sources.--The aforemen-

 

tioned difficulties can be reduced by a redistribution of
the sources or assets from which such income flows are
generated. ‘This accomplishes the same Objective in a
rather more fundamental way, and provides a direct link
between income redistribution and growth pOlicy. It
involves changing the pre-tax distribution and thus tends

to reduce the need for subsequent redistribution through

 

l"Short run" is defined as the period during
which the MVP of an asset is greater than its salvage
value.

2This point is probably Of limited applicability
tO Sri Lanka which has an abundance of high school and
university graduates.

33

fiscal measures. It refers to redistributing the capacity
for income, rather than income per se, which in tradi-
tional agriculture involves the provision to low income
farmers of what T. W. Schultz calls "low priced sources

of permanent income streams."1 These include modern
scientific inputs like high yielding seed, fertilizer,

water control, weedicides, pesticides, inter alia, as

 

well as education and knowledge, which act to broaden
horizons, increase farmer receptivity to innovative
techniques,~and contribute toward their capacity for
making economically rational choices. This.notion is
therefore broader than the human capital notion, which
was originally stated by Alfred Marshall when he wrote
that "capital consists in great part of knowledge and

. . 2
organlzatlon."

Redistributional impact Of public intervention.--

 

When income sources are to be redistributed through
governmental agencies, the intervention by the state
designed to provide farmers with low priced sources of
income streams in order that they may increase their 93;:
ppp, has implications for income distribution.

1. TO the extent that these low priced sources

of income streams have to be publicly provided, they

 

lSchultz, op. cit., p. 130.

2Alfred Marshall, Principles of Economics (8th ed.;
New York: Macmillan, 1961), p. 139.

 

34

must be publicly funded, and to the extent that they are
funded out of progressive taxation, these policies
implicitly have an impact on income distribution.

2. The subsequent expenditure of these funds
for the provision of low priced sources of permanent
income streams and human capital investment has distri-
butional effects, directly through its effect on “H
productivity and employment, and indirectly through
its effect on the institutional wage.

The-final effect on income disparities Of the
foregoing will depend on the relative income levels of
those groups who pay and those who benefit. A recent
study supports the conclusion that the larger the
government's share in total investment, the smaller the
share of the wealthiest 5.tO 10 percent.1

The case for public provision of low priced
sources Of permanent income streams is partly stated by
T. W. Schultz. These "sources" have to be produced by
research, because "there are very few reproducible agri-
cultural factors in technically advanced cOuntries that

2 Public provi-

are ready-made for most poor countries."
sion is necessitated by externalities and indivisibilities.

The output Of research is in effect a public good with

 

lI. Adelman and I. Morris, "An Anatomy of Income
Distribution Patterns," Development Digest,_Vol. Ix
(October, 1971).

2

 

Schultz, op. cit., p. 147.

35

public good characteristics.l As the output of research
is of limited appropriability, underinvestment by pri-
vate investors in research is a certainty, because when
marginal private gains have fallen to zero, marginal
social gain will still be positiVe. Another reason why
public production of technology is required is because Of

known indivisibilities, primarily in the methods

and staff of scientists required when a firm under-

takes the production Of modern factors suited to

the agriculture of particular poor communities

starting with the known ssientific and technical

(agricultural) knowledge.
Moreover, such inputs have also to be distributed through
publicly created marketing channels for the Same reasons;
namely, externalities and indivisibilities. The distri-
bution of modern inputs could set in motion a cumulative
spiral of growth the (external) benefits of which could
only be captured (internalized) by the comprehensiveness
of the state. The low rate of capital formation in
traditional agriculture would probably preclude private
pricing policies that will cover costs.

In general it may be stated that because techno-

logical change is non-marginal change, market prices

cannot be expected to provide ex ante signals indicating

 

. 1F. Bator, The Question of Government Spending
(Collier, 1960), Chapter 6. The two components of public
goods are: (1) MC = 0, which means that as the cost of
additional use is zero or near zero, consumption by one
individual does not deprive others; (2) Potential users
cannot be excluded if the product exists at all.

2

 

Schultz, Op. cit., p. 150.

36

the direction and quantum Of change. This is because
larger price changes result from change when the change
is non-marginal. This factor increases the uncertain-
ties Of investments. Uncertainty would tend to discourage
private investments, probably requiring that techno-
logical changes be publicly induced in less developed
countries, rather than be left to the "non-dictates" of
underdeveloped markets. This position appears to be
supported by implication by Galbraith's description of
the various~"anti-market" institutions that firms in the
U.S. must adopt in order that they may cope.with the
uncertainties that inevitably accompany non-marginal
technological changes.1 Institutional devices like
vertical coordination, integration, contracting, and
advertising, enable U.S. firms to "supersede" and "sus-
pend" the market. However, they are of limited
availability and applicability in the less developed
economies Of less developed countries, due to traditional
institutiOns like familism, lack of capital and capacity
for risks, absence of standard weights and measures and

grading systems, an inadequate legal code, inter alia.

 

Public intervention to induce technological
changes in less developed countries has a dual effect

on income distribution. First, when public funding is

 

1John K. Galbraith, The New Industrial State
(Boston: Houghton Mifflin, 1967)"Chapters 2 and 3,

 

37

out of progressive taxation, income redistribution
occurs. Whether income disparities will become reduced
will depend on how these funds are subsequently Spent, and
the income levels Of the beneficiaries whose productivity
is increased by their acquisition of sources of income
streams. Only when such "sources" are provided to

groups whose income levels are relatively low, arer
income disparities likely to be reduced. Second,
appropriate technological changes could, if combined

with suitable institutional variables, increase employ-
ment opportunities and thereby reduce income.disparities.
Technical changes could shift the production function,
increase the marginal value product of the abundant
factor, which is labor, and make it worthwhile to apply
more units of labor to the production process.1 Also,
quick maturing new varieties make multiple crOpping
possible--a factor likely to have a most significant
effect on employment and thus income. For instance,
Japanese farmers using new mechanical and biological
technology employ three to four times as much labor per

acre of land than Indian farmers.2

 

lTwo points may be noted: Whether employment
effects are negative or positive will depend on the type
of technology used and the institutional context in which
such technology is applied; second, even though employment
per acre Of land may increase, the labor component in each
unit of output is likely to decline.

2See, for instance, John Mellor, The Economics of
Agricultural Development (Ithaca, N.Y.: Cornell, 1970),
p. 229. Also, the official estimate for 1964 Shows that

 

 

 

38

To the extent that technical innovations reduce
unemployment in agriculture they influence income distri-
bution. Where underemployment and unemployment are high,
to achieve a higher level Of employment results in a
reduction in income disparities. Dudley Seers says:

Unemployment is itself a big element in the

explanation of the shape Of the income distribu-

tion. Not only does it deprive large numbers of

people of income (including many who do not seek

work because they think the search is hopeless and

therefore do not appear in unemployment statistics),

it also depresses wage levels and causes thousands

of people to stay working on tiny farms or in small

family businesses.
The distribution of income therefore reflects the existence
of unemployment. If employment is provided for those per-
sons who are disguisedly unemplOyed with zero or near zero
marginal value productivity, the income of higher income
groups remaining the same, income disparities are reduced.
To the extent that the incomes of higher income groups are
reduced, income disparities are reduced still further.

A reduction in disguised unemployment has an

indirect effect on income distribution through its positive

 

rice cultivation consumed about 180 man-days per hectare
per crop. In Taiwan the National University has estimated
the labor requirement at 150-160 per hectare. These
figures compare with a range of 60 to 120 man-days per
hectare for crops of traditional varieties in most of
South and Southeast Asia. Robert A. Shaw, "Jobs and
Agricultural Development," Development Digest, Vol. Ix,
No. 1 (January, 1971), pp. 88-97.

1Dudley Seers, "Income Distribution and Employ-
ment: Some Issues Raised by the Colombia Report,"
Bulletin Of the Institute of Development Studies (Brighton,
England), Vol. 2, NO. 4 (1970).

 

 

39

effect on the institutional wage. Disguised unemployment
has its basis in the institution of familism and exists
when labor receives a wage that is higher than its MVP.
In a situation where the institution of familism facili-
tates disguised unemployment, the average product per
family is the institutional wage. As the number Of
persons who are subsidized by the family decrease, there
could be an increase in the institutional wage, because
then the total product will be shared among fewer persons.
In the alternative, the product will be larger as the
newly employed person's earnings or part Of.them are
added to the family income. An increase in the institu-
tional wage will reduce income disparities if we assume
that the incomes of higher income groups do not increase.
To conclude, although the actual effect of
particular technologies on income disparities is a matter
for empirical investigation, the foregoing analysis seems
to demonstrate that the same factors that cause growth
also influence income distribution. Policies designed
to influence growth are usually not neutral with respect
to income distribution. In fact such policies, when

implemented, influence income distribution ipso facto.

 

Ikance there are fairly strong reasons for believing that
inacome distribution considerations are beSt explicitly
taken into cognizance at the time that policies are

fornmlated to increase the rate of growth of agricultural

40

output. Furthermore, there are fairly good practical
reasons (like the establishment of fixed assets, and
administrative difficulties) for not leaving income
distributional questions to be dealt with separately at
some subsequent stage. The above two factors reinforce
each other in suggesting the advisability of incorporating

equity Objectives into growth policies.

The Conflict Issue

 

As was mentioned, there is a widespread belief
that the Objectives of growth and equity are in conflict
with each other, and thus competitive rather than comple-
mentary. This implies that trade-offs are involved in
pursuing both policies, whether concurrently or sepa-
rately. A primary objective of this study is to examine
the empirical evidence on this point. This subsection
examines the a priori reasons for the possible existence
Of a trade-Off with a view to determining its theoretical
foundations. The degree of acceptability of the concep-
tual framework presented in the following chapter depends
on whether a conflict between growth and equity exists
and, if so, whether it can be regarded as Significant.
A.theoretical background also contributes to understand-
ing the interrelationships between growth and equity,
‘when micro studies are analyzed. To this extent it will
facilitate an identification of the institutional vari-

ables that do or do not Cause a conflict situation to

41

arise. An understanding of the conflict issue could also
be useful in identifying socio-economic groups who are
potential beneficiaries of agricultural development
programs. Hypothetically, the degree of conflict between
growth and equity resulting from the provision of sources
of income streams would vary with the characteristics of

different socio-economic groups.

Theoretical and conceptual relationships.--The
relationship between agricultural growth and income dis-
tribution has been conceptually develOped and historically
documented for developed countries. Kuznets has demon-
strated that over long periods as GNP increases income
disparities are reduced because of the structural trans-
formation in the national economy that has accompanied
economic growth.1 Some evidence exists that, even in
Asian countries, develOpment tends eventually to reduce
inequality for Similar reasons.2

Within this argument, the possibility Of conflict

between the two objectives arises at a national level in

 

1Simon Kuznets, Modern Economic Growth (New Haven:
Yale University Press, 1966); and Simon Kuznets, Six Lec-
tures on Economic Growth (New York: The Free Press, 1959).

Recent U.S. evidence, however, seems to indicate
that income disparities, both before and after taxes,
appear to have increased moderately.

2Harry T. Oshima, "The International Comparison
Of Size Distribution of Family Incomes with Special Refer-
ence to Asia," Review Of Economics and Statistics
(November, 1962), pp. 444 ff.

 

 

42

develOping countries from structural rigidities which stem
from state intervention, traditionalism, underdeveloped
markets, and high rates of population growth (which
reduces the coefficient of differential growth).

Besides these structural reasons, other (mainly
monetary) reasons for a possible conflict between growth
and equity Objectives may be of greater relevance in
respect to the agricultural sector. These reasons have
their basis in the Keynesian premise that the rate Of
savings is a function of the level of disposable income.
For this reason, high income groups have a higher marginal
propensity to save than lower income groups. Henceit
is assumed a redistribution of income tends to reduce
the savings rate, capital formation, and growth. There-
fore, a trade-Off could arise between higher incomes in
the present and higher incomes in the future. This is a
basic ingredient of the HarrOd-Domar growth model which
is a prototype Of post—Keynesian growth theory. Some
empirical support for this theory is found in a study of
Brazilian macro data, which concluded that the redistri-
bution Of incremental income in favor of the middle
stratum Of income caused the gross domestic savings ratio

. . . . 1
to decrease, even though per capita incomes were rlslng.

 

1N. H. Leff, "Marginal Savings Rates in the Devel-
opment Process: The Brazilian Experience," Economic
Journal (September, 1968).

 

43

Keynes, however, did not regard the propensity to
consume as being competitive with investment. Assuming
as Keynes does that the aggregate supply function is given,
the thesis Of the "General Theory" is that employment is
determined by aggregate demand, with the propensity to
consume and the amount of investment at a given time as
complementary components. Hence, the transfer of hoards
ings, or funds earmarked for imported "luxury" goods, to
the consumption of those goods that may be produced
domestically, have multiplier effects. These multiplier
effects would increase effective demand and stimulate the
agricultural economy into achieving higher levels of
income and employment, given the relatively high income
elasticity of demand for agricultural goods in developing
countries. If so, then growth and equity would be com-
plementary rather than competitive objectives.

But Keynes' theory was meant to apply to a
depressed, developed, national economy, and not an under-
developed, traditional agricultural sector. Even with
regard to a developed economy, some doubt exists as to
whether an income-leveling policy will significantly
help raise consumption and increase employment.1 Con-
sumption is less likely to increase in a stagnant

traditional economy. If workers are unemployed or

 

lH. Staehle, "Short Period Variations in the
Distribution Of Incomes," Review Of Economic Statistics
(1937), pp. 133-43.

 

 

 

44

underemployed when idle productive capacity exists, the
under-consumption theories probably have a message. But
this would hardly apply to a traditional agriculture where
output must increase primarily through technological
change. Even if idle capacity in productive equipment
(like large tractors) exist, the under-consumption
theories would apply only if the idle capacity is due to
a lack Of effective demand and not to other constraints
like the lack of foreign exchange for Spare parts or

raw materials, or the absence Of skilled manpower. The
Operation of the multiplier in a (traditional) economy
where the supply is inelastic is likely to generate
inflation rather than growth.

Even the notion that income redistribution reduces
investment and hence future growth seems Of doubtful
validity, especially in a developing economy. There are
four broad considerations, one at the level Of pure theory,
which question the notion Of a trade-Off, and which when
taken together, may weaken it:

1. The theories Of Duesenberry, Friedman, and
others do not establish a clear connection between the
level Of income and the rate Of savings.

2. When income redistribution is regarded as a
redistribution of the sources Of income streams, resources
are not transferred from investment to consumption but

investments by high income groups, both productive and

45

nonproductive, are replaced by investments by low income
groups.

3. Even if the funds are transferred from the
savings Of high income groups, this redistribution of
personal income is likely to have a positive effect on
growth through its functional redistributive effect and
the release Of high Opportunity cost foreign exchange
resources.

4. From a negative standpoint, the skewness in
income distribution which the Keynesian hypothesis seems
to advocate for high rates Of growth, may not yield a
high savings rate if high income recipients engage in
conspicuous consumption instead of saving.

The validity of this fourth factor rests upon a
definition of savings that may be of limited practical
value. Each of these factors will be examined in turn.

Taking the first Objection; other hypotheses on
the relationship between the level of personal income and
the savings rate do not support the Keynesian notion that
income redistribution necessarily reduces aggregate
savings. With Milton Friedman's permanent income
hypothesis,l redistribution of income would not affect
savings so long as all families believed the new incomes

to be permanent. In Duesenberry's relative income

 

1Milton Friedman, A Theory of the Consumption
Function (New York: National Bureau of Economic Research,
Inc., 1955).

 

46

hypothesis,1 the effect of income redistribution on
savings and thus future growth cannot be stated a priori.
This is so because the average propensity to save depends
on the level of income relative to the national average
income, and the effect of income redistribution on

savings depends on the specific nature of the relationship
between actual and average income.

In other theories, savings depend on factors
other than the level of income, such as the rate of
interest, age Of saver, expected income, etc.

The supposedly adverse effect of income redis-
tribution On the savings rate, the second Objection,
assumes that the redistribution consists of primary income
flows already vested, and that this redistribution will
be accomplished through fiscal policy. In this case,
differences in marginal propensities to save between
high and low income groups could cause the redistribution
to substitute consumption expenditure for investment
expenditure. However, when the redistribution refers to
the provision of sources of income streams, investment

by high income persons is replaced by investment by low

 

income groups.
Besides, small farms make up such a large pro-

pxution of the rural acreage that an increase in

 

1J. S. Duesenberry, Income, Saving and the Theory

(If Consumer Behaviour (Cambridge, Mass.: Harvard Uni-
versity Press, 1949).

 

47

productivity on these farms is likely to have a signifi-
cant effect on output.

Moreover, the redistribution of sources of income
streams also minimizes the possible adverse effect of
income redistribution on incentives Since Whatever adverse
effect such transfers may have on the incentives of high
income groups are Offset by the Opportunities created for
low income groups.

Furthermore, the growth in agricultural output
per se can have equity effects because it increases the
supply of goods for which the income elasticity of demand
is relatively high, from 0.6 to 0.9 percent.1 The bene-
fits would be negated, however, if the shift in the 7
supply schedule results in a more than proportionate fall
in the price (due to the inelastic nature of the demand
curve) and thus smaller incomes to farmers. But this is
unlikely because of the interdependence between supply
and demand factors in the underdeveloped rural economy.
The level of per capita income in the rural economy is
largely determined by the level of agricultural output;
per capita income cannot increase without an increase in
agricultural output. Consequently, a supply shift in
agricultural goods will, through its effect on per capita
income and high income elasticity factors, actuate an

outward demand shift which dampens the price-reducing

 

lMellor, Agricultural Development, Op. cit.,
pp. 73-77.

 

48

effectof the initial supply shift. Even if prices do
fall, real incomes do not decrease if unit costs are
reduced proportionately more.
Trade-offs could arise, however, if within a
given time span relatively high return private investment
is replaced by relatively low return public investments
in rural infrastructure. Since such investments tend to
yield lower returns over a longer period, their discounted
present value would be lower, given a discount rate that
reflects the scarcity Of capital and the urgency of the
need to develop. Also, public investment in regions where
infrastructure is weak would, by being deprived of
"Marshallian" externalities that derive from physical
propinquity like pools of skilled labor, common services,
infrastructure, etc., yield a lower return per unit of
capital investment. This is what Gunner Myrdal refers
to as the "backwash effect," in terms of which externali-
ties Of infrastructure cause capital to pour into the
more advanced "development poles."l
Even if the funds transferred to low income
farmers yield a rate of return per unit of capital invest-
ment that is equivalent to that which might have accrued
to high income investors, the differences in marginal

,pIOpensities to consume is likely to result in the lower

 

1G. Myrdal, "The Inequality Issue," in The

(Shallenge of WOrld Poverty (New York: 1970), pp. 49-77.

 

49

income persons spending on consumption a larger portion
Of the incremental income generated by their newly
acquired sources Of income streams.

Third, if progressive taxes transfer funds from
high income groups to low income groups for consumption,
this redistribution of personal income could, though not
necessarily, have favorable effects on growth through
its effect on the functional distribution of income, and
the reallocation Of high Opportunity cost foreign exchange
resources from consumption to investment. If progressive
taxes are utilized to redistribute income to,lower income
groups, this would transfer control over real resources
from persons with a relatively high marginal prOpensity
to consume income elastic (luxury) goods to those with a
relatively high marginal propensity to consume the more
income inelastic (basic) goods. If luxury goods have a
high import content, and basic goods a low or nil import
content, such a transfer would shift outward the demand
curve for domestically produced income inelastic products
like food. To the extent that the foregoing is valid,
such a transfer has a potential for reducing unemploy-
ment, and thereby income disparities, in the following
‘ways.. First, it achieves a functional redistribution
(i.e., an increase in the share of labor in the total
‘product) by increasing the demand for goods that have a

Iiigher domestic labor content. An increase in the demand

50

for basic goods implies that these goods can be produced
locally and by relatively more labor-intensive methods of
production. The increase in the demand for goods with a
higher labor content contributes to reducing unemploy-
ment, and therefore reduces income disparities (on the
assumption that the income of higher income groups do

not increase). Second, the shift away from goods that
have a high import content releases foreign exchange
resources for the impOrt Of capital goods, and therefore
increases employment potentials in the economy. These
two favorable effects on unemployment actuate a third
indirect effect on income disparities by exerting an
upward pressure on the institutional wage, which is

the average product per household in the rural sector.
Thus reduced unemployment not only reduces income dispari-
. ties directly by providing persons with low MVPs with

employment and incomes, but also ceteris paribus, reduces

 

income disparities by increasing the institutional wage.
A qualification to the foregoing argument exists.
All movements toward a more equal income distribution
will not necessarily increase total.employment unless
relatively labor intensive final goods are also labor
intensive with regard to intermediate goods in comparison

With capital intensive final goods. Also, there must

51

be no significant degree of involuntary short-time work-
ing in the labor intensivesector.1

Mention might be made of the "Economy of High
Wages" factor in terms of which higher consumption levels
could lead to better nourishment and hence higher pro-
ductivity levels. In this case a rise in real wages
becomes another form of investment.2

Fourth, the Keynesian hypothesis with regard to
savings and income levels seems to advocate a skewed
distribution of income for a high rate of capital forma-
tion and growth. However, there seems to be a growing
consensus that the level Of income does not affect the
savings rate in developing economies. The proposition
that savings is not a function of the level of income has
been strongly supported by several writers, including
the authors Of the ILO studies on Columbia and Ceylon.3

These writers have taken the position that in developing

countries income constitutes a means Of purchasing

 

1John Weeks, Employment and the Distribution of
Income (unpublished; Brighton: Ms. IDS, 1972).

2Myrdal seems to attach considerable, and prob-
ably exaggerated, importance to this factor. Myrdal,
Asian Drama, 9p: cit., Vol. II, p. 745 ff.

 

 

3See Dudley Seers, "New Approaches Suggested by
the Columbia Employment Programme," International Labor
Review (October, 1970), as well as V. K. R. V. Rao,.
”Re-distribution Of Income and Economic Growth in Under-
developed Countries," in Income and Wealth, Series X
(1964). This same point has also been made by Kaldor
and Myrdal in UN Publication E/CN/S/AtlB/R3/Addl.

 

 

52

status and hence gets disassociated from production.
Thus despite the skewed distribution, the savings rate
remains low and the rate Of growth is consequently slow.
However, these theories tend to become merely
academic, and lose their practical usefulness for policy
if "savings" is defined to include both actual and poten-
tial savings--a distinction which corresponds to the
dichotomy between voluntary and forced savings. The
foregoing theories attempt to explain actual and volun-
tary savings. Potential savings refers to funds that
would have been spent on "nonessential" consumption and
"nonproductive" investments like land speculation, and are
potentially available for conversion into public savings
through taxation. In this case the savings potentially
available will depend on (a) the private expenditure
that is normatively, though not necessarily arbitrarily,
deemed by the state to be nonessential or socially
unproductive and (b), the prevalent level of income.
Thus if the voluntary savings rate is low, and public
investments are needed for agricultural deVelOpment,
savings could be "generated" through taxes. To regard
savings as the surplus above "necessary" consumption
appears to be realistic if, as in the case of Sri Lanka,
the distribution of wealth fails to match the distribu-
tion of real political power. However, public savings
:mnst in the long run depend on an efficient fiscal system

and a widening tax base.

53

Some writers also discount the possibility of a
conflict between growth and equity Objectives on the
basis that redistributional policies promote social and
political stability and thereby create a climate suitable
for investment and growth.1 This expectation seems valid
only if we assume a more or less fixed standard of
living with an upward moving level of living that pro-
gressively closes the gap between attainment and desires.
But this belief does not seem to have empirical validity,
as distributional policies could widen the gap between
levels and standards of living by activating latent
needs. This could bring into existence social discontent
which did not previOusly exist, not because there was no
cause, but because there was no hope.

One summary of the evidence on the relationship
between growth and equity was presented at AID's 1970
Spring Review of Land Reform where E. Lang noted:

The most important finding of the review is

that the social and political goals Of wider dis-
tribution of Opportunity, power, and employment
among farm people are not in conflict with

increassd agricultural productivity and effi-
ciency.

 

1See, for instance, ILO, Towards Full Employment,
0 . cit., and G. Myrdal, "Differences in Initial Condi-
tions," Asian Drama, Op. cit., Vol. I, Chapter 14.

 

 

2Quoted in report on subsession by Peter Dorner,
"Discussions on Equity and Productivity," in the American
JOurnal Of Agricultural Economics (December, 1970),
P0 7170

54

Probably, the degree Of conflict between the goals
will depend on whether the groups at whom agricultural
development programs are directed are willing and able
to utilize such technology to increase their productivity
and levels of income. Some groups at the very low level
of the income spectrum or having paddy lands Of very
small size probably would be incapable of utilizing the
sources of income streams for the purpose of generating
greater income flows. This may be due to low levels of
ability, inadequate lipdw.or other factors. If so, the
extent to which technology can promote both growth and
income distribution Objectives will depend on the recep-
tivity to and responsiveness of different groups to
agricultural modernization. Conflict is likely if the
groups are not responsive. Therefore the degree Of
success in achieving both growth and equity Objectives,
and the degree to which they are complementary, appear to
depend On the judiciousness with which policies are

focused.

Summary and Conclusions
The foregoing analysis has demonstrated that techno-
logical change constitutes a link between growth and income
distribution. The main source of agricultural growth is
net reduced productivity slack (i.e, more efficient

allocation, or more investment in traditional inputs),

_———- * * *"——__—

 

55

but new technology-—modern, Off-farm, purchased inputs,
that embody scientific knowledge.

Technology links growth to both functional and
personal income distribution. Technology directly

affects the functional distribution of income, depending

 

on whether it is labor-saving, neutral, or capital-saving.
The precise effect of technology on functional distribu-
tion depends on the nature of the shift in the production
funciton that such technology actuates. For this analy-
sis, however, the functional distribution of income is
important, not for its own sake, but only insofar as it
influences personal income distribution.

The link between agricultural technology and
personal income distribution arises from the following
factors:

(a) Most agricultural technology has to be
produced Off-farm (because it can be produced more
cheaply under large-scale specialized conditions) and
marketing channels therefore need to be developed to bring
such technology to traditional farmers. Externalities,
indivisibilities, and uncertainties with regard to
technological change make public production and distri-
bution necessary. Public funding to provide inputs and
complementary infrastructure through progressive taxation

has redistributive effects.

56

(b) The acquisition of technology by farmers has
income effects, directly through the impact of such tech-
nology on productivity and employment, and indirectly
through its influence on the institutional wage.

Technology, the primary source of agricultural
growth, therefore is not usually neutral with respect to
income distribution. In fact, implementing policies for
promoting agricultural growth implicitly affects income
distribution by increasing or decreasing income dispari-
ties, either relatively or absolutely. The same .
instruments that promote growth can be utilized to move
toward a more desired distribution of income. Conse-
quently, it seems advisable to explicitly take income
redistribution objectives into cognizance at the time
policies are formulated to increase Output. This con-
clusion seems reinforced by the practical difficulties
of leaving income distribution questions to be dealt
with at some later stage through fiscal policies. These
difficulties include the possibility Of conflict between
the policy goals Of growth and equity, the establishment
of fixed assets, and administrative insufficiencies
affecting the efficacy Of fiscal policies. As a corollary
conclusion, income distribution should not be regarded
as a matter exclusively for fiscal policies, to be
carried out independently of policies designed to

increase output.

57

To the extent that the foregoing is valid, a
major task of planners is to identify groups, on the
basis of some homogeneous characteristics, who are to
be the beneficiaries Of policies, and that agricultural
development programs should be focused on these specific
socio-economic groups. This conclusion would be sup-
ported by the position, if valid, that the degree of
conflict between growth and equity will vary with
different socio-economic groups. Therefore, the degree
of success in achieving growth and equity Objectives,
and the degree to which they are cOmplementary or com-
petitive, partially depend on the judiciousness with

which such policies are focused.

CHAPTER III

CONCEPTUAL FRAMEWORK FOR RESEARCH

The conceptual framework relies on the belief
that technological changes cannot be isolated from socio-
political milieu. Gotsch1 has proposed that the effect
of technological change on income distribution be ana-
lyzed within the conceptual framework of four broad
interacting variables: (1) the characteristics of
"abstract" technology (efficiency and factor intensity);
(2) the absolute magnitude and distribution of productive
assets (land tenure factor); (3) the type and distribu-
tion of institutional services; and (4) local social
Customs and traditions. This conceptual framework (a)
shows how the characteristics of technology, local insti-
tutions, and the rural social structures are related to
each other at a point in time, and (b) how these rela-
tionships can be expected to evolve in a dynamic rural
system.

The conceptual framework prOposed here utilizes

Gotsch's variables but adapts their arrangement and flow

 

1D. H. Gotsch, "Technical Change and the Distri-
bution of Income in Rural Areas," Economic Development
.Re ort, NO. 205 (Cambridge, Mass.: Centre for Inter-
natlonal Affairs, Harvard University, 1971).

58

59

directions to suit the purpose in hand, which is a
review and analysis of micro level studies. The frame-
work is Of a hierarchical nature, with two broad sets of
(primary) variables, which interact at the local level
to determine the distribution of personal income and
agricultural growth. These two sets Of variables have
their basis in other (secondary and tertiary) variables
which determine their efficacy, efficiency, viability,
and character. These secondary and tertiary variables
could also determine the focus of the primary variables.
The framework is valid primarily.in.a shOrt,run (fivew
year) situatiOn.. The conceptual framework is diagrammed
in Figure l.

The two sets of primary variables that interact
to determine personal income distribution and growth in
agriculture are: (1) technology, and (2) institutional-

organizational factors.l

Technology

 

The effect of a technological change in agricul-
ture on inter-farm income distribution would depend on

the nature of the technology itself. Thus if the

‘

1As with Gotsch, ibid., and Schultz, O . cit.,
institutions are treated as variables. Subscr1ption is
made to the belief that small farmer development has to
be studied in the context of surrounding economic and
Political processes. See also Walter E. Coward and
Wayne Schutjer, "The Green Revolution: Initiating and
SuStaining Change," in Civilizations, Vol. 20, NO. 4
(1970), pp. 473-83.

6()

 

 

 

 

 

 

Ann—«002 EH03 H1303

g .3HEOA no aha—Pagan

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Auoaoduom.
ouauuonum
goon: 05.3.55.
upououm noduuusoaoHnEH A
.aoa»«ao~ nuance ofluawuoaom
Manda—":0 annum
23335.»: 2.3 gunman
H
Hm mewmsow P. memawrwfl H“
3.225. 23:35 33331 30332233
0: on a o» no u Bonuses: annexe—nae
u an m «a who mhmwmuwW mazux=oczm momsomme soroHOactm sauntaaca resort
EOHEEHEWZH lzgmo HOSOZZUNF mg...—
I I

 

 

 

 

 

 

 

  
   

 

 

e

was mogoam oma<>m 80¢.

 

Bamboo AtmDPADUHMU<

.lll'l'l'

MtoozH A<=omflbml no onSmHmemHD

 

 

 

nodusnauunwo @5005 «En unease wcauoouuc meanness, uo nuauOOOumo
cows—yon amaze—noduaaofl ans—33m :37 uuonm 9.1.0:...“ xugoauum Hmsumuocoo I... 0.5.5.:

61

technology is capital intensive it will be biased against
the capital-poor farmers or, if it requires water control,
it will favor farmers with irrigated farm land.

_Technology has two main asPects to it: (a) the
level and change in the efficiency factor, and (b) the
input intensity factor. The efficiency factor refers to
the physical productivity of inputs. It is the purely
technical engineering notion of the ratio of physical
output to physical input. Gotsch states with regard to
this:

. . . a critical.ingredient to.any broad-based
rural development effort is a significant
improvement in the value-added by agriculture

. . . the increase to resources owned needed to
overcome the reticence of traditional farmers
to innovite must be on the order of 30 to 50
percent.

Changes in efficiency may be due to the broader
use or Optimal use of existing technology, or due to
innovations that are new to the system.

The second aspect Of technology relates to
factor intensities. This has its origin in the Hicksian
theory2 as applied to agriculture by Heady,3 and postula-

ted in a Cobb-Douglas world which assumes constant returns

to scale, constant product prices, equilibrium in the

 

lGotsch, Op. cit. This is consistent with the
findings Of Dobbs who considered a return of 25 percent
necessary if new practices are to be adopted. (J. L.
Dobbs, NO. 1, Appendix A.)

2Hicks, loc. cit. 3Heady, loc. cit.

 

 

62

factor markets, and two factors of labor and capital.
Then, a change in technology which affects the relative
shares imputed to the two factors in the total product
may be classified as labor saving, neutral, or capital
saving, depending on whether the change increases the
marginal productivity of capital by relatively more, in
the same proportion, or relatively less.

These two aspects determine the divisibility Of
technology (i.e., the scale factor) and their potential
for generating externalities (i.e., creating a divergence
between private and social returns). Technology in this
framework is determined by the state Of scientific knowl-
edge, and by policies such as the extent and type of
research funded, interest rates for capital investment,
the level at which the rate of exchange is fixed, and a
wide spectrum of fiscal policies, such as investment

credits, tariffs, and depreciation allowances.

Institutions

 

The second set of variables is the institutional
variables (including organizational factors) which interact
in the conceptual framework with technology to determine.
income distribution and increases in output. These
institutional variables are of two types. First are the
subjective institutional factors like social values,
farmer goals, rural perceptions, etc., that stem from

culture and religion. These factors affect receptivity

63

to innovation and the desire to change. The other compo-
nent is the subset of objective variables that create the
Opportunities for farmers to be gplg_to change. These
include tenure conditions, credit availability, price
supports, and the whole gamut of factors that can be
manipulated through programs for agricultural develOpment.
The framework recognizestnu>aspects to programs--the
design Of programs, and the efficacy with which they are
implemented. Thus, well designed programs may not yield
desired performance due to unsatisfactory implementation.
Organizational variables are an appendage of institutional
variables because the viability of organizations are to

a large extent determined by the rules under which they
operate; i.e., institutions. Thus, organizational
performance to a considerable extent is determined by

organizational rules.

Resource Endowments

 

Resource endowments have been included in the con-
ceptual framework, and some resources are shown to be
capable of influencing and be influenced by both primary
variables of technology and institutions.

Two categories Of resources can be distinguished.
One category consists Of immutable factors, the other of
changeable factors. Some resources like the number of
monsoons or amounts Of sunlight may be given and unchange-

«able either because change is technically infeasible or

64

because it is economically unprofitable, even though tech-
nically feasible. Hence a conceptual differentiation
separates those factor endowments that can be changed
(e.g., through irrigation, fertilization, etc.) and

those that cannot. The flow directions in Figure 1 indi-
cate that resource endowments send relative price signals
on the one hand, and directly contribute to the final
resource mix available on the other.

Recent writings, based on empirical research,
propose that factor endowments, technology and human
capital, account for 95 percent of labor productivity
growth and labor productivity differences.1 The concep-
tual framework recognizes that resource endowments could
influence both technical and institutional changes.

Such recognition infuses into the framework the fundamental
notion in the Hayami-Ruttan induced development model,

that institutions and technology can be endogenously
determined by relative prices. To say that these primary
variables are endogenously determined proceeds beyond

T. W. Schultz's high pay-Off input model in which they are
exogenous and unresponsive to economic forces. In

the Hayami-Ruttan induced development model, factor

endowments could generate relative prices that provide

‘

lYujiro Hayami and Vernon W. Ruttan, "Sources of
A9ricu1tura1 Productivity Differences Among Countries,"
in Agricultural Development: An International Perspective
(Baltimore: John Hopkins Press, 1971), pp. 86-107.

 

 

65

appropriate signals to public. and private sector sup-
pliers Of knowledge and technology, and stimulate the
responses of institutions to new economic Opportunities.
For instance, the desired physical properties Of plants
can be predetermined by policy makers, and the research
biologists can then be asked to design such a plant.1

However, while relative prices gppl§_send such
signals to innovate, it does not follow that they actually
do so successfully. Evidence exists that economic
signals have not succeeded in influencing the design and
focus Of programs, policies, and laws that determine the
nature and type Of institutions and technology. In
fact, numerous instances can be cited where capital
intensive technologies have been accepted in labor
abundant situations.2

Hence the conceptual framework would be defi-
cient for analyzing real world situations if it failed
to take cognizance Of the possible lack of efficacy of
relative price signals in actually influencing the
policies of public and private agencies. Part Of the
discrepancy between what is and what should be may be

attributable to the fact that the induced development

 

1The process referred tO as "biological engi-
neering" by the Rockefeller Foundation.

2See the instances cited by Eicher, Gotsch, and
Falcon, loc. cit. '

 

66

model is built upon the foundation of shadow rather than
market prices.1

Shadow prices are invisible prices that have to
be painstakingly discovered through careful analysis
and computation. Whether social shadow prices convey
appropriate signals to public and private agencies.
depends on whether such agencies choose to discover them
in the first place and whether, having done so, choose

to use them for policy guidance.

The Distribution of Power

The conceptual framework proposed here seeks to
explain the factors that actually influence the type and
nature Of technology and institutions as they exist in
the real world. This it does by tracing back through
policies and laws to their source, which is the distri-
bution of real political power. As the lines of
interaction indicating flow directions suggest, whether
the (shadow) price signals do influence technology and
institutions, and the extent to which they are successful

in doing so, depends on whether they are permitted to

 

do so by vested interests and the political process.

Since both technology and institutions determine relative

 

lTime lags may also partially account for a lack
of response in the short run. '

67

property rights,1 those who hold power can be expected to
perceive policies, laws, the distribution of new flows

of productivity, etc., as possible threats or reinforcing
to their existing position. Dorner recognizes this when
he says:

The Significant distributional questions are

not those defined by marginal productivity theory,
but those concerned with economic and political

' power and the patterns Of resourge ownership and
control to which they give rise.

A recent publication by Nulty arrives at essen-
tially the same conclusion with regard to technical changes
in west Pakistan. Nulty3 says:

Technical innovations in and of themselves

would appear to be insufficient to the demands of
development; they inevitably force some kind of
adjustment in prevailing institutional relation-
ships and the outcome of this adjustment is
likely to be determined by the distribution Of
political power and the relative political
awareness of the affected groups.

Thus the conceptual framework recognizes that,
since the relative level of technology among users, and
the relative rate at which users and potential users of
improved technology adopt it, determine relative rights,
the question of who should gain and who should lose is

ultimately a political decision.

 

. 1Warren Samuels, Welfare Economics, Power and
Property, Department of Economics, Michigan State Uni-
ver31ty. (Mimeographed.)

 

2Dorner, Op, cit., p. 718.

3Leslie Nulty, The Green Revolution in west
Pakistan:- Implications of Technological Change (New York:
Praeger, 1972).

 

68

Similarly, with regard to institutions the frame-
work demonstrates that the distribution Of income and
ownership Of productive assets must ultimately reflect
(and later interact with) the distribution of real
political power. The nature, design, and responsiveness
of institutions cannot adequately be explained solely in
terms of relative prices. The nature, design, innova-
tiveness, focus, and radicalism of policies and programs
that modify institutions, and the degree of effectiveness
with which such policies and programs are implemented, can
(must) ultimately be explained in terms of the actual
and expected distribution Of real political power.

Given resource endowments, the primary variables
of technology and insitutions interact in the conceptual
framework to influence the quantity and mix Of resources
finally available for increasing output and the produc-

tivity Of low income groups.

Constraints

 

The exogenous constraints identified are probably
valid only in the short run (five years). They are:
(a) Cultural factors, including religion, which are the
subjective institutions which influence the farmers'
receptivity to new technology. This variable recognizes
the importance Of the farmers' subjective evaluation of

the technology, and that the success of programs depends

69

on the response of people at whom they are directed.1

The flow directions indicate that the urgency associated
with increasing paddy productivity does not permit the
modernization of farming communities through a change of
existing culture and values to modern practices.

(b) The state Of scientific knowledge limits the type,
quality, and variety of the technology that can be pro-
duced. (c) The level Of skills of experts and
administrative personnel will have a pervasive limiting

influence on all of the relevant variables.

Focus of Variables

 

The framework develOped here can be used to ana-
lyze the reasons for both growth and income distribution
in agriculture. This stems from the assumptions that
income distribution in developing countries should be
dealt with by increasing the productivity of low income
groups, rather than through transfer payments using the
device of fiscal and welfare policies. While increasing
output depends on the quality of policies and programs
and their implementation, their consideration for income
distribution hopefully only requires in addition that

legislation, policies with regard to factor intensities,

 

1C. R. Wharton has discussed the importance of
this factor in "Risk, Uncertainty and the Subsistence
, Farmer" (paper presented at the Joint Session of the
American Economic Association and Association for Compara-
tive Economics, Chicago, December, 1968; mimeographed).

70

and the design of programs be given a focus--i.e., be
directed at preidentified economic groups, and reflect a
concern for gppsg productivity should be increased.
This is a matter for subsequent verification. To the
extent that this is valid, the framework illustrates
that the same instruments (variables) that have a
potential for increasing growth may, if appropriately
focused, be made to contribute toward a desired distri-
bution of income. This it does through changes in the
relative control over resources exercised by the pre-\
identified groups.

Whether programs and policies gill_be focused
is a political decision. But whether programs and
policies ggp be focused will depend on:

(a) the divisibility of rice technology (scale
factor),

(b) the magnitude Of externalities involved
(divergencies between private and social benefits/
costs),

(c) the extent to which institutional and organi-
zational variables can be designed to capture economies
of scale and internalize (externalize) enough of the
external benefits (costs), given the small size and
large number of producing units (households), and

(d) the extent to which (at least the) early

benefits can be concentrated in the target groups.

 

71

Conclusion

 

The two primary variables of technology and insti-
tutions interact to determine both agricultural growth and
the distribution Of net benefits. These variables are
endogenously determined, through the interaction of the
resource endowment situation with the power distribution.
The distribution Of power and the attitudes (values) of
those who wield it, have a dominant influence on the
extent to which relative prices influence the design and
focus of agricultural develOpment programs and laWs, as
well as the manner in which they are implemented. Thus
the conceptual framework recognizes that the extent to
which relative prices could influence the various variables
depends ultimately on the pattern of distribution and use
of real political power. The effectiveness with which
political power can bring about desired results is
limited by the exogenous constraints of culture, the
state of scientific knowledge, and the skills or lack of
them on the part of experts and administrators.

The conceptual framework presented has attempted
to compromise between the need for consistency with the
intricacies of the real world, and the need for a rela-
tively simple tOOl for analytic application. The number
of permutations have been kept to a minimum through a
hierarchical arrangement, a distinction between primary

and nonprimary variables, and a short run perspective.

72

The short run perspective makes it possible for the
primary flow lines to be the only ones shown. In the

long run other lines of interaction would be needed.

CHAPTER IV

AN IDENTIFICATION OF EQUITY NORMS AND

PROGRAM OBJECTIVES

Introduction

 

In order to incorporate equity goals into pro-
grams for agricultural development, these goals must be
stated in concrete Operational terms, with primary refer-
ence to Sri Lanka. This indicates the direction in which
to proceed, suggests the relevant measures of inequality,
and helps identify the groups for whom the programs
should be focused.

Program objectives serve as a link between
political values and program specifications. The char-
acteristics of programs cannot be determined directly
from a social welfare function.1 If program specifica-
tiOns are regarded as inputs, and program Objectives the
output, these inputs would have to be translated into

outputs so that they can be evaluated with reference to

 

1Charles L. Schultze, The Politics and Economics
of Public Spending (Washington, D.C.: The Brookings
Institution, 1968), Chapter 4.

R. N. McKean states with regard to the social
welfare function: "There is no mysterious, unique set of
'correct' group of goals that can be identified by think-
ing deeply Of reading the right books, or turning over
the right rock." Public Spending (New York: Mc Graw Hill,
1968): P— 4.

 

 

 

73

74

political values. If we work downward, therefore, we go
from political values (norms) to program Objectives
(social outputs) to program specifications (inputs).
Questions relating to income distribution his-
torically have posed difficulties for economists. The
classicists subsumed the institutional variables in

ceteris paribus. Income redistribution involved inter-

 

personal comparisons Of utility which did not permit
objective or reliable measurement. This difficulty led
positivists like Lionel Robbins to regard economists as
being "social" engineers or technicians, who are neutral
between ends, which they must regard as given.1 If so,
however, "an economist would have nothing to say about the
merits of policies, and this would have virtually ruled
out his education and policy role."'2
In dealing with such interpersonal comparisons,
a rough guide can be found in the Kaldor-Hicks compensa-
tion principle as refined by Scitovsky.3 According to

this criterion, policy is socially desirable if the

gainers can potentially bribe the losers into accepting

 

the change and if, in turn, the losers cannot bribe the

 

lLionel Robbins, An Essay on the Nature and Sig-
nificance of Economic Science (2nd ed., 1952; London:
Macmillan, 1935), p. 30.

2Luther Tweeten, Foundations of Farm Policy
(Lincoln: University Of Nebraska Press, 1970), p. 503.

3Tibor Scitovsky, "The State of welfare Econom-
ics," American Economic Review, 41, pp. 303-15.

 

 

 

75

gainers into not making the change. In effect this con-
denses to recommending policies if the social benefit-
cost ratio.is greater than one.

The process of ensuring that the Scitovsky cri-
terion is satisfied does not assume significant prOportions
in this study for several reasons. For one thing, income
redistribution is conceived in terms Of redistributing
sources Of permanent income streams. If appropriate
groups utilize these sources to generate as much income
flows as previously, the resultant increases in produc-
‘tivity (and possibly employment) could stimulate a
cumulative spiral of develOpment, the gains from which
are likely to be so great and widespread that the losers
would be unable to bribe the gainers in terms of the
Scitovsky criterion. Secondly, this conclusion is rein-
forced if it can be assumed that the marginal utility
attached to-a monetary unit is greater for low income
groups than it is for high income groups--the rationale
used for the initial redistribution Of income streams.
In any case, the social welfare function, as manifested
in the Objectives of the Five Year Plan (quoted below),
palpably indicates that income disparities should be
reduced.

Income Distribution Objectives
in Sri Lanka

 

 

The equity goals of the Five Year Plan have been

clearly stated:

76

While the immediate social objective of the
Plan is to provide employment, it also aims to
bridge the disparities in income and living
standards by raising the income and living con-
ditions of the low-income households.
This goal of reduced income disparities implies in nega-
tive terms that income disparities should not be permitted
to increase. This statistically translates into making
the Lorenz curve in Figure 2 an outer bound for income
inequality, thus endowing it with significance for policy.
While stating positively that income disparities
should be reduced, the Flag fails to expreSsly indicate
a norm. Nevertheless it can be inferred that it is pgp_
a policy goal to reduce income disparities to the point
that it disappears altogether. That absolute equality
is not a norm can be inferred from the many references in
the Plan to the need to "encourage . . . the cultivator,"

2 These state-

and to provide "incentives" to "work hard."
ments, which emphasize incentives and rewards for added
effort (including risk bearing), tend to be incompatible
with an absolute equality norm. ‘Absolute equality in
incomes is antithetical to incentives because Of its ten-
dency to generate externalities that encourage free
riding.

The conclusion that emerges from a statistical

standpoint is that the line of equal distribution (Figure

2) fails to assume normative significance for policy

 

1The Plan, p. 2. 2The Plan, p. 36.

77

Figure 2. Lorenzo Curve of Distribution
of nurel Inna-e end Paddy Holdings

p

Percent of Total Ineoee/Icrpege

 

umtmmmmhlnp

Source: Seeio-Roonenie Qurvey 1972.

Figure 3.1noome of floueehelde 1n the Rural

sector
Ho¥§3§§1de 9+...

6d

1n

 

 

 

 

 

 

 

o ' Hw—qiifiwm
““3305???"

Source: Soelo-Boonoeio Survey
1972

78

formulation. Even the principle of equal Opportunity
(or equal access to sources of increased productivity)
is not necessarily compatible with the goal of reducing
income disparities. Given the initial inequality, the
application of the equal Opportunity principle could
merely reinforce or increase the extant disparities in
income. This suggests that the goal of reducing income
disparities requires that disadvantaged groups receive
privileged treatment. This gives added credence to the
hypothesis that agricultural development programs should
be focused upon specific preidentified socio-economic
groups.

Some Relevant Measures Of
Income Distribution

 

 

This subsection will examine complementary
measures of income distribution that seem to be relevant
to the Objectives of the study, and useful in defining
program Objectives. These include the frequency distri-
bution, Lorenz curve, and poverty rate. The discussion
will be in the context of the distribution of household
income in the rural sector of Sri Lanka.. Table 4.1 indi-
cates the distribution of income in the rural sector by
size brackets of income Of households.

The data Show that 46 percent Of households in
the rural sector receive incomes of less than Rs. 200

($34) per mensem, while over 80 percent receive incomes

TABLE 4.1.--Distribution of Households and Total Income by
Income Groups in the Rural Sector.

 

tlmxmm.

Ibtal

 

 

Groups No. of 8 of Income % of
(Rs.) Households Households (Rs.) ”Income
Below 100 142,500 9.4 10,671 2.7
100-199 554,800 36.7 82,138 20.6
200-399 558,600 37.0 154,270 38.6
400-599 I 167,200 11.1 79,958 20.0
600-799 57,000 3.8 37,208 9.8
'800-999 .17,100 1.1 15,632 3.9
1000 and over 14,280 0.9 17,529 4.4
Total 1,511,450 100.0 399,406 100.0

 

Source: Extracted and synthesized from Socio-Economic
Report, Department of Census and Statistics,
Colombo, 1971.

of less than Rs. 400 ($68). The data are graphed in
Figures 2 and 3.

Whether the achievement of a normal distribution of
income should be an objective of policy will probably
depend on the validity of the "untestable" hypothesis that
innate abilities are normally distributed, at least within
homogeneous ethnic groups. If so, a skewed distribution
results from socio-economic, institutional factors.1
Asymmetry could be attributed to a multiplicity of reasons,
some relevant to traditional agriculture, including

attitudes of different persons toward uncertainty-bearing

 

1This theory was propounded by A. L. Bowley, "The
National Income in the U.K.," Economica (1933).

 

80

(Friedman)l wealth inheritance, mainly land (Fisher),2
distribution Of educational Opportunities (Becker),3 and
access to, and distribution of, institutional Services
(Gotsch).4

As stated, the Lorenz curve in Figure 2 provides
an outer bound for income disparities, but the line of
equal distribution fails to provide a limiting case.
Hence the desired pattern of income distribution must
lie somewhere in the concentration area. The exact sta-
tistical location Of the norm will depend on a definition
of a poverty line and poverty rate, which together deter-
mine the poverty gap.

. The concept of the poverty rate is supplementary
to the measures just discussed, and is tied to the notion
of a poverty line. Once a poverty line is defined, the
proportion of the total pOpulation (households) below

this line Could constitute the poverty rate, while the

 

1Milton Friedman, "Choice, Chance and Personal
Distribution Of Income," Journal of Political Economy
(August, 1953). -

 

2Irvin Fisher, Elementary Principles Of Economics

 

(New York: Macmillan, 1912), where he attributes a
skewed distribution to "inheritance, constantly modified
by thrift, ability, industry, luck and fraud."

3Gary S. Becker, Human Capital (New York:
Columbia University Press, 1964), pp. 61-66; also Gary
S. Becker and Barry R. Chiswick, "Education and the Dis-
tribution Of Earnings," American Economic Review (May,
1960), pp. 365-68.

4

 

 

Gotsch, loc. cit.

 

81

poverty gap would refer to the income it would take to

reduce the poverty rate tozero.1 According to Schickele

the poverty line is at the levelof income at which a

‘family suffers.from lack of certain minimum needs, and

which he says the economist should regard as given. He

says:

Clearly, the concept of the 'poverty line' is
not a subjective or arbitrary judgement of some
individual, but represents a consensus of society,
based on studies of consumer budgets, on what peo-
ple concur to consider necessities of life in
their country and at the present period of time.
This socio-economic consensus is evidenced by 7
general public acceptance and by legislative or
administrative action with reference to,the pov-
erty line and renders it 'Objective' in the same
sense as many other institutional facts such as

.standards Of individual and group behaviour estab-

lished by faw or custom are Objective facts to an
economist.

If we accept the findings of the ILO Repprt on Sri

Lanka, we can fix the poverty line at Rs. 200 per mensem,

or Rs. 2400 ($600) per annum per family. The Report states:

. . . In Ceylonese conditions the income level of
Rs. 200 per month per household may be taken as
above the poverty line, at least amongst those with
medium or small families. The minimum estimated to
be required for adequate nutrition of a family of
two children was Rs. 178. The poverty line falls
generally somewhere in the Rs. 100 or Rs. 200 group,
depending on size (and composition) 3f family and
on living costs in particular areas.

 

53.

lSchickele, Agrarian Revolution, Op. cit., pp. 49-

 

2Schickele, "Improving Income Distribution," Op.

cit., p. 15.

BILO, Report, 9p. cit., p. 35.

 

82

This conclusion is supported by an earlier study1

which arrived at the same poverty line, but for an average
family of six persons.

The Objectives of the Flag, however, evidence a
concern for a broader class in the income spectrum;
namely, those receiving less than Rs. 4800 per annum
($800), who constitute 80 percent Of households. Hence
while Rs. 2400 may be a povery line, Rs. 4800 may be
regarded as an adequate level of living line.

The arithmetic mean for rural household incomes
from paddy cultivation is Rs. 157 ($26) per.mensem or
Rs. 1880 ($313) per annum, as Shown in Table 4.2. The
measures of income distribution discussed and data on
incomes will be utilized for the purpose of identifying

program Objectives in the following sub-section.

Output Objectives

 

As stated, program Objectives constitute a link
between the social welfare functiqn as manifested in the
Objectives Of the Flag, and the program characteristics
that have yet to be Specified. These program Objectives
could be expected to have two components; a growth com-

ponent, concerned with efficiency, and an equity component

designed to give the program a focus. A base will need

 

1T. Jogaratnam and Rainer Schickele, Practical
Guidelines to Agricultural Deve10pment Policies in Ceylon
(Peradeniya: Agricultural Economics Research Unit, Uni-
versity of Sri Lanka, 1970).

 

83

TABLE 4.2.--Household Income Received Annually from
Paddy Cultivation.

 

Imputed Value

 

NO. Of Total Average
Income Money Income Of Paddy Income Income
Receivers consumed
(Rs. '000), (Rs. '000) (Rs.'000) (Rs.)
551,000 68,734 17,476 86,210 156.73

 

Source: Extracted from Socio-Economic Survey data,
Department of Census and Statistics, Colombo,
1971.

 

to be identified in respect to each Of these components
in order that the desired change, as embodied in the progam
Objectives, may be measured.

The Objectives with regard to the output Of
paddy are stated in the Agriculture Sector Program, as in
Table 4.3.

Given the time horizon of the Plgp (five years),
over 80 percent Of the total increase in output is
expected to come from land already under paddy. Paddy
output is expected to increase by roughly 40 percent,
over present levels in traditional areas, and by around
10 percent on newly develOped lands in the Dry Zone. As
mentioned, scarce water and low population density in the
Dry Zone require that output expansion in that region be
brought about through medium and long term capital
investments in irrigation and colonization schemes. In

absolute terms, the required increase in paddy output

 

84

TABLE 4.3.--Output Target for Paddy..

 

1972 1973 1974 1975 1976

 

On land now aswedhmmised

(million bushels) 77.0 90.0, 95.0 101.0 107.0

 

 

On land newly aswedhumised
(million bushels) 1.0 4.0 5.0 6.0 9.0
'Ibtal 78.0 94.0 100.0 107.0 116.0
Less 10% for seed, etc. 70.0 85.0 90.0 96.0 104.0

 

Source: Agriculture Sector Programme, The Pigs Year Plan,
Ministry of Planning and Employment,COlombo, 1971.

 

 

from presently cultivated paddy land is about 30 million

bushels (80 percent) by 1976.

Equity Objectives

 

Program Focus

 

When income distribution is conceived of in
terms of redistributing low priced sources of permanent
income streams, achieving equity objectives could reduce
essentially to a matter of focus. Groups will need to
be identified on the basis of some homogeneous economic
characteristic(s) toward whom programs could be directed.
Identifying such groups is also important because Of the
presumption that the degree of congruence between the
twin goals of growth and equity will depend partly on the
success with which such groups are chosen. This implies
that some groups at a very low level of the income spec-

trum, or having paddy holdings of very small size, may

 

85

be incapable, for technical or institutional reasons, of
utilizing the sources of income streams for the purpose
of generating sufficiently increased income flows.
Technical reasons may stem from limits to divisibility in
technology, for instance, while other reasons may be.
such factors as low levels of intelligence, poor health, kfi‘
extreme traditionalism, etc. Therefore, as was concluded

in Chapter II, the degree of complementarity will depend 5

 

on the receptivity to and responsiveness of different WI;
groupings to agricultural innovation. Hence the success

of agricultural develOpment programs will depend nOt

only on the Specifications of programs and their manner

of implementation, but also on the judiciousness with

which such programs are focused.

When programs are focused it implies that projects
that constitute such programs will be directed at pre-
identified groups. The focusing of programs for promoting
rice technology could vary in stringency from concentra-
tion of administrative attention on preidentified groups,
to absolute exclusion of other groups. The degree of
focus that should be applied would probably need to be
decided at the project level, when the Objectives of
projects are Specified. The basis for deciding the

degree of focus is examined below.

 

86

Size Groupings

 

The groups will be distinguished in this study
on the basis of size of paddy holdings, and within each
group a distinction will be made between owner operator,
tenant cultivators, agricultural laborers, and Off-farm
workers. The benefits of new technologies are expected
to be different for each group.

Groups are classified therefore on a categoric
as Opposed to geographic basis. The geographic approach
may be appropriate in a situation of socio-economic
homogeneity within a specific region or district. On
the other hand, a categoric approach would be more suit-
able when the farm households that a specific program is
supposed to reach are scattered in the midst Of others
that are outside the scope of that Specific program.

The basis of classification or grouping is a
compromise--a compromise between generality needed for
ease Of implementation, and the detail really needed to
identify the target households or farms. The simpler
the classification the easier and cheaper it would be to
administer. At the same time it is recognized that

simplicity may reduce overall program accomplishment.

Some households, particularly at the end points of a class

may more appropriately be in a different group than in

the one assigned. As the foregoing implies, the classifi-

cation inevitably contains an element Of arbitrariness.

 

 

87

Economic size classification.--A more useful

 

grouping of farms from the viewpoint Of income distribu-
tion is on the basis of economic rather than physical size
Of farms. Such a classification would help minimize
some of the difficulties caused by the overlap Of the
small farm category with the agricultural laborer and
Off-farm worker category, and include the effects on
family income of differences in resource endowments and
family size. A classification on the basis of economic
farm size would also help in overcoming the difficulty
caused by the possibility that the physical size distri-
bution of farm units may be egalitarian, but the
distribution of land ownership may be very unequal. (In
such a situation yields and technology may vary less with
farm size than with the form Of tenure.)

The economic size refers to the income generation
potential Of a farm in relation to the poverty line.
Thus whether a farm size is economically small, medium,
or large will depend on whether its annual income is
below the poverty line (Rs. 2400 p.a.), between the
poverty line and the adequate level of living line
(Rs. 2400-4800 p.a.), or above the adequate level Of
living line (above Rs. 4800 p.a.), respectively. Hence
increasing the economic size Of farms could be synonymous
with improving the distribution of income (if we assume

that the incomes of "high" income groups dorxfizincrease).

88

The variables that influence the economic size
of farm are thus crucial to this analysis. They are
in fact substantially the same variables of technology
and institutions which were said to be the primary
determinants of income distribution and growth in the
conceptual framework in Chapter III. An identifica-
tion of these variables and EH1 assessment Of their
applicability to Sri Lanka is a matter for subsequent
chapters.

One institutional variable which has a direct
bearing on the economic size of farm is the institutional.
wage or the average value product per household. It
depends on the level of income of the household divided by
the number of persons in that household. The level of
income is the aggregate Of income earned by the income
receivers in the household. A decrease in the ratio of
income receivers to non-income receivers (including

unemployed members) would ceteris paribus increase the

 

level Of the institutional wage and hence the economic
size of farm.

The relationship between the size of household
and the institutional wage in Sri Lanka is evident in
Table 4.4. The data indicate the household variable
assumes significance in determining the economic size of
farm in Sri Lanka. A decrease in the size of household

or an increase in the number of income receivers ceteris

89

TABLE 4.4.--HOusehOlds Classified by Size and Number of
Income Receivers in the Rural Sector.a

 

Average No. of

Household Size . .
Income Receivers

 

2 1.2
3 1.3
4 1.4
5 1.5
6 1.6
7 1.6
8' 1.8
9 1.9
10 2.0
11 2.6

 

Source: Socio-Economic’Survey data, Department Of Census
and Statistics, Colombo, 1972. .

aThe arithmetic mean size of rural households is
5.8 persons and a median 5.0 persons, while the average
number Of income receivers per household is 1.6 persons.
The discrepancy'between the mean and the median size

suggests that the rural household size distribution is
somewhat positively skewed.

paribus increases the level of the institutional wage, and
therefore the economic size of farms. The positive cor-
relation between the household size and the number of income
receivers nevertheless does indicate that the decline in

the institutional wage is less rapid than if the number

of income receivers had been fixed.

Target Groups

 

Introduction.--It was stated earlier that the

 

degree Of congruence between growth and equity Objectives

probably depends on the judiciousness with which paddy

90

develOpment programs are focused. Groups have to be
distinguished on the basis of their income generation
potential, given the extant structural distribution of
holdings.

It is assumed for the purpose of group identi-
fication that (a) the attitudes of farmers toward
technical changes, (b) the entrepreneurial ability of
farmers, (c) costs per unit Of output, and (d) the quality
of farm resource endowments, are not correlated with farm
size.

Given these aSSumptions, the holdings Of some
groups may be identified as being too small to enable them
to utilize rice technology to rise above the poverty line.
On the other hand, some groups may be found to be not in
a poverty condition, and would hence need to be treated
somewhat differently from those who are. A third middle
group of farmers may be distinguished who are presently
below the poverty line, but whose structural circumstances
will permit them to utilize rice technology to rise abOve

it.

Specification Of Criteria for identifying mean

 

groups.--Two criteria can be applied to identify the
(third) middle group of farms, One criterion is to define
the poverty gap for the average farm in each size group,
and then determine whether an optimum or near-Optimum

application of technology, combined with appropriate

 

1.1:;

91

institutional modifications, will increase output suffi-
ciently to close the poverty gap, given the time horizon
of five years. The criterion is whether this is realistic;
that is, technically feasible. The second uses the output
target in the Plgp to see whether, if the average farm in
the size group succeeds in increasing its output by a

target percentage, the mean farm will succeed in rising

xV—J —_ J a":
e
I.
I' .1 ’.

above the poverty line.

The former criterion approaches the issue from "

i

 

the side Of family income with a residual output effect,
while the latter approaches it from the size Of output,
with a residual income effect. The second criterion is
more practical, but only concerns itself with relative
increases in income. This is because even though the
mean farms in each size group increase their output by
the same percentage, the differences in farm sizes could
cause absolute income differentials to increase.

Both approaches will be used in this study to
enable an informed judgment to be made on the selection

of groups that are to be the focus of different programs.

Transformation Of structural data.--In order that

 

this judgment be made, structural information on paddy
holdings will need to be transformed into income and
output information. This will be done by identifying the
mean farm by Size in each of the structural groupings,

and then applying average income data to determine

 

92

approximate income levels, subject to the assumptions
enumerated at the beginning of this subsection. This
transformation, although approximate, is nevertheless
expected to be useful.

The structural information on paddy holdings in
Sri Lanka is provided in Table 4.5. The data indicate
that nearly 74 percent of paddy holdings are below 5
acres in Size. A Lorenz curve of the data in Table 4.5
indicates that there is greater inequality in the dis-
tribution of paddy holdings, than there is in the
distribution Of income (Figure 2).

As stated, it would be possible to utilize this

structural information to determine the increase in output

TABLE 4.5.--Area under Paddy in Sri Lanka by Size of

 

 

 

Holding.
N0.0f
% Of Extent Of % of
Size of Holding Holdings .
with Paddy Holdings Paddy Iand Paddy Land
Less than 1 acre 67,630 12 25,150 2
l - 2 '1/2 acres 179,230 32 147,780 13
2 1/2 - 5 acres 173,131 30 299,901 27
5 - 10 acres 111,761 20 378,232 33
Greater'tha“ 10 35,901 6 284,125 25
acres
Total 567,653 100 1,135,125 100

 

Source: SEANZA Lectures compiled by the Central Bank Of
Ceylon, Colombo, 1968.

93

per acre needed to reduce the poverty rate, if the
distribution of returns to land is transformed into the
distribution of returns to households-~persona1 income
distribution being the focus of interest.

This in turn will require that ngt_returns to
farmers be calculated by deducting costs of production
from the gross value of output. Such cost of production
data is provided in Table 4.6.

Too much reliance cannot be placed in the fol—
lowing data bOth because of inflation and changes in the
cost structure and perhaps the method of charging for

labor since 1969. Latest reports indicate that the cost

TABLE 4.6.--Average Cost of Production per Acre and per
Bushe1—-A11 Island.

 

 

 

 

 

Cost per Acre .Cost per Bushel
Inputs
Rs. Percentage Rs. Percentage
1. Labor 211.15 56.57 6.32 56.58
2. Fertilizer 17.57 4.70 0.52 4.66
3. Seed 30.56 8.17 0.91 8.15
4. Rent
(a) Land 26.19 7.08 “0.79 7.07
(b) Machinery 48.37 12.94 1.45 12.98
(c) Buffaloes 11.97 3.20 0.36 3.22
5. Maintenance cost ’
(a) Machinery 5.94 1.59 0.18 1.61
(b) Buffaloes 9.76 2.61 0.29 2.60
6. Weedicides and .
Pesticides 4.34 ”1.16 0.13 1.16
72 Other 7.39 1.98 0.22 1.97
A11 Island 373.90 100.00 11.17 100.00

 

Source: Central Bank of Ceylon, Cost of Production Survey,
1969.

 

 

 

94

per acre for the average farm is around Rs. 650 per crop.
At an average yield of 50 bushels per acre per crOp, this
amounts to a cost of Rs. 13.00 per bushel, without
including the imputed cost of family labor. If, as the
study indicates, about 15 percent of labor cost is for
imputed family labor, then the net return per acre for the
average farm would be Rs. 700 ($117) per annum. The

method of computation is as follows:

Average output per acre per annum (bushels)1 2 100
Value of Output @ Rs. 18 per bushel under GPS Rs. 1800
Cost of Production @ Rs. 13.00 Rs. 1300
Annual return per acre 3 Rs. 500
Add imputed cost of family labor \ Rs. 200
Total net return per acre per year Rs. 700

This information could now be related to the
structural data in Table 4.7 to determine the poverty gap
under the assumptions previously enumerated.

As can be inferred from the data in Table 4.7, the
size of farm needed for a net average return at the

poverty line is about 3.4 acres, given present levels of

 

., .1The choice of a spectrum of one year seems advis-
able not only because it enables a more realistic
assessment of the average monthly income and the compari-
son of income levels with the poverty and adequate level
lines, but also because it permits an assessment of the
income effects of multiple cropping.

2The GPS refers to the Guaranteed Price Scheme,
under which the Paddy Marketing Board is the sole legal
outlet for paddy, at the assured price of Rs. 18.00 per
bushel. ,

_ 3According to the Central Bank of Ceylon, Survey

on Cost of Production of Paddy (Colombo: Economic Research
Department, Central Bank, 1969), 15 percent of total cost
was imputed to family labor.

 

 

 

 

 

95

.mom muuo>om may mmoao ou mam>ma usomonm Bonn cocoon mufl>fluodpoum on» as mmcmao

 

m

.mwm muum>om

map mmoao ou meoocfi usoaoflmmdm mumumcom ou cocoon umwm mom muom mom maonmon mo Hogan: 6380

.ocfla muuo>om mnu pom mEoocH Hudson Hmouom mo Hm>oa
ocu cmmzumn mocmnmmmwo on» ma scan: .mmm muum>om on» omoHo ou poowos mEOOCfi GA omnmuoqfl oxen

.wuflamuusos oamom cam .msoum oNHm Aomo CH coausnfluumfio Hafiuoc m mcHEdmmm pom .msoum
mNHm map Ga Show mo oNHm mmmum>m man an ouom Mom suduou nos momnm>m mcflmamwpasa an nousmaoom

 

 

In II II In oomv Mo>o h um>o
I: II I: Has oomvuoomm n u v
wmv swap mmmq mv cwau mmmq mva cmnu mmmq HHGIOOOH commuoova v I N
mvlmva mwlmva mvalmvm OOOHIOOhH oovHIOOP m I H
mva Ho>o mva Hm>o mom Hm>o coma um>o oon swap mmoq H coca mmmq
.Ammmn Nemav aumom mom onom Anomw Mom muom Ammomsmv Ammomomv Amouodv
.mwu auno>om omoau Mom maonmsmv mom mamsmsmv mosfloaom mod 0
on oopmoz wuw>flu mow mafl>fluoopoum omdo unmuoo ammo wuum>om ommum>¢ ha om» mo.MMHM

Iosooum ca mmcmnu w o

Imuocmw meoocH

.mEHmH. 05006:” CH MU.MQ HMH5#05H#WII.N.V mqmdfi.

96

adoption of presently available technology, and the
extant institutional context. This is tantamount to
stating that farms below 3.43 acres are economically
small. This, however, does not mean they are Egg_small.
Whether this is so will depend on their output poten-
tial--the potential to increase paddy output sufficiently
as to enable the farm household to rise above the poverty

line.

Application of criteria for identifying middle

 

group.--In terms of the first criterion proposed, whether
a group will qualify as the middle group will depend on
its ability to meet the output gap. we find from Table

4.7 that the average paddy holding in the below 2 acre

class would have to increase its output by over 143 percent
in order to qualify, while the holdings between 2 and 4
acres in size would need to increase their output by

less than 43 percent. If a 43 percent increase is regarded
as being reasonable (given the time horizon of five years
and the technology available), while an increase of 143
percent is regarded as not, then "viability" will com-
mence somewhere in the upper region of the l to 2 acre
class.1 Since the specific cut-off point cannot be uni-
versally determined, it is considered appropriate in this

instance to accept the supportive opinion of the ILO Report

 

lA viable paddy farm is one which can utilize
paddy technology to rise above the poverty line.

97

that for all Ceylon an effective floor for an adequate
level of living would be 2 acres of paddy.l In terms of
the analysis so far, such a floor would seem to be somewhat
too high (since viability would seem to commence in the
upper ranges of the 1 to 2 acre class) but will be accepted
because it seems reasonably consistent with the data in I
Table 4.8.

The second criterion will now be applied to approach
the same question (determine the size of a viable unit for
paddy) from the side of income. In terms of this criterion
the groups are distinguished on the basis of the likely
increase in annual income for the average farm in the size
category, if the Plan target in respect of paddy output-
is to be met. Since the increase in output expected from
paddy land already under cultivation is 40 percent or
30 million bushels in five years, it is assumed that
paddy output will increase by at least 40 percent in five
years on the average farm in each size group. The
resultant income information is tabulated in Table 4.8.

The data in Table 4.8 indicate that of the size
groups, only the 2 to 4 acre category of holdings will
be able to change its income level from below the poverty
line to above it if it were to meet its share of the

output target. This conclusion substantiates the

 

lILO, loc. cit. How they arrived at this con-
clusion is not clear. But this same conclusion was
reached in a 1970 study by Jogaratnam and Schickele, op.
cit. ‘

 

 

 

98

.huwamuusmc mamom mcflasmmm .muomumu usmuso

nova ou usuoumm ow commouocfl ma Show omouo>m on» no nomuoo ma oaooaw :H ommonosfl Umuommxmn

.oovm .mm
mo mafia huumbom on» new Hm>ma fisooafl ucwmoum cmw3uon wosmuomwwo msu mafixmu an pounmaoom

 

 

ommm Hm>o Ha: ommm um>o ooa ma ooo.mo h Hm>o
comm AH: ommm am on ooo.mHH 5 moon: I v
ovmm oom ooam Hm mm ooo.ooH v moon: I N
cued omma omoa mm Hm ooo.omH m Moon: I H
omv omom 0mm NH NH ooo.mo H Hopes
dmmomnmv Amoomsmu AuMma mom Hosmsn “monocy
Hum mommsmv Ao#mawxoummmv
waoosH Gd sham ommum>¢ om sound m . om: mas o
nommouocH ammo muum>om mo mam>mq m>Mu seam oomucmouom cams. on o.wwam
nouommxm maoocH ucmmmum .uma U m 02 m .m

 

.mmmmuosH mEoooH Hmwusmuom on wNHm Emma mo QHSmGOHumaomII.w.v mqm¢a

99

conclusion reached in terms of the first criterion;
namely, that 2 acres constitutes a safe and effective
floor, for purposes of determining viability.

The target groups can therefore be identified as

being the below 4 acre groups. These groups qualify as

 

targets of primary focus of programs designed to promote
a more widespread and optimal utilization of rice tech-
nology. They are also to be in the vanguard of the
effort to achieve self-sufficiency in rice and improve
income distribution. This group consists of 408,000 or
72 percent of all households in the paddy subsector,

and occupies 32 percent of the total paddy acreage. A

40 percent increase in output from this acreage will meet
roughly 48 percent (14.2 million) of the targeted

increase in paddy output of 30 million bushels.

Target group characteristics.--Three target

 

groups are distinguishable on the basis of the income
data: (1) the below 1 acre category, (2) the 1 to 2
acre category, and (3) the 2 to 4 acre category. The
first two groups (below 2 acres) are similar in that
they are made up of largely nonviable holdings, while
they are different in degree in that the income problems
of the smaller group are likely to be more acute. Using
the definition of economic farm size made earlier, the
below 2 acres category would be "too small," while the

above 2 acre group is of adequate size. This adequacy

 

 

100

of farm size with respect to the 2 to 4 acre group
depends upon the ability to utilize rice technology to
rise above the poverty line in the time horizon contem-
plated.

On the other hand, the data in Table 4.7 suggest
that the chances of being able to render viable, through
the application of existing rice technology, the class of
small farms that are below 2 acres in size, seems exceed-
ingly small. The 2 acres refer only to the upper limit
of this class of farms. Moreover, the size distribution
of holdings within the below 2 acre class is not normal,
as the mean size in the below 2 acre class is 0.8 acres.1
This structural factor would require that such (average)
farms have a level of productivity per acre that is
almost one and one-half times the national average in
order for them to be able to rise above the poverty line.
Even this, moreover, is based on assumption that the farm
household is not larger than average (5.8 persons) in
size.

The degree of specialization in paddy increases
with farm size.2 Hence the smaller farms are more diversi-
fied than large farms. These farms have supplementary

sources of income from off-farm sources and/or the

 

1Department of Census and Statistics, Census of
Agriculture, 1962 (Colombo).

 

 

2Jogaratnam and Schickele, 0p. cit.

 

101

cultivation of non-paddy crops, such as vegetables,
coconut, rubber, or tea.1 The bulk of the agricultural
labor force comes from the households on these farms,
their participation in the agricultural labor market
probably related to the amount of income accruing to them
from non-paddy agricultural products. The income from
paddy would, for those who own tiny paddy plots, be a
supplementary income source that provides insurance
against unemployment and/or the vicissitudes of the market

for non-paddy crOps.

Program effects.--While this analysis suggests

 

rice development programs be focused on the below 4 acre
groups, the relative effect of such programs could be
expected to differ depending on the structural condition
of affected farms. While all affected farm households
could be expected to experience increases in the institu4
tional wage emanating from increases in output and
employment, the income from output is likely to be greater
relative to the income from employment for the households
on the large farms when compared to small farms. Con-
versely, the potential increase in income from employment
is likely to be greater than that from output for the
households on the smaller farms. It is obvious, in fact,

that under the assumption that family size is not

 

lIbid.

 

 

 

102

correlated with farm size, the potential returns from
capital should be relatively greater for the households
on larger farms.

Surplus labor from small farms would need to
find work on the large farms for a fulfillment of such
potential, due to differences in the per farm labor input
requirement, given structural differences, and independence
between the farm size and household size variables.

Given similar production functions it is preferable that
labor be employed on a larger number of smaller farms, as
in that case the demand curve for labor is likely to have
an elasticity near to zero, and a greater number of persons
are likely to be employed.

However, the extent to which the surplus labor on
the small farms could expect to gain from employment Oppor-
tunities on the larger farms will depend on (a) the extent
and nature of the shift in the production functions of
larger farms, and (b) the employment slack already on,
those farms. The extent of the shift of the production
function is determined by the absolute impact of tech-
nology on production, while the nature of that shift
(bias) is a function of the type of technology applied.

The former changes the absolute levels of labor demanded,
while the latter changes the relative amount of labor
demanded.

If labor requirements increase on large farms

consequent upon their application of improved technology,

103'

the extent to which the surplus labor on the smaller farms
could expect to benefit will depend on the extent of dis-
guised unemployment already existing on larger farms.
Theoretically, if the marginal product of labor is less
than the market wage level on the larger farms, the
adOption of new technology which increases the marginal
product of labor would not lead to an increase in the
demand for hired labor until the marginal product rose
above the wage level. It is possible, however, that the
larger farms may withdraw some (female) members of the
family from the labor market When they experience an
increase in the level of their institutional wage.

But unlike the case of larger farms, rice devel-
opment programs will constitute only a partial solution
to the income problems of farms below 2 acres in size for
several reasons: UJ; Rice technology alone may not be
able to create a sufficient increase in demand for.labor
to absorb the some 400,000 unemployed persons on the
below 4 acre paddy farms; (2) Since there is greater
diversification on smaller farms, rice technology would
have a smaller absolute impact on incomes, unless there
is increased specialization; (3) Inadequacies in mana-
gerial and other talents may be prOportionately greater
on the smaller farms due to lower incomes and fewer
educational opportunities; (4) There may be limits to

scale neutrality.

 

 

104

However, the below 2 acre paddy groups that culti-
vate other crops such as tea, rubber, coconut, and
onions could expect to benefit from other programs that
make available improved technology relating to those
crops. The below 2 acre groups that do not cultivate
these other crops would constitute the target of other
programs pushing rural industrialization, land settle-

ment, and colonization in the dry zone, inter alia. A

 

detailed examination of these other programs falls outside

the scope of this study.

Other groups.--The groups that generally fall

 

outside the direct focus of rice development programs
are the categories above 4 acres in size, who constitute
38 percent of farm households in the rice subsector.
These households already have incomes above the poverty
rate, and as relatively major market participants may

be presumed to have preference and production functions
susceptible to demonstration effects. They are of their
own volition more capable of translating felt needs into
technological changes in response to market signals.

No absolute exclusion of these groups is possible
and desirable because private market alternatives may not
be available to them. Their eligibility for specific
access to benefits from programs may be decided on at the
project or scheme level depending on the market alterna-

tives available. For instance, all farmers must have

 

 

105

access to the Guaranteed Price Scheme because the Paddy
Marketing Boards are the sole legal outlet for paddy
output. However in the case of production credit, a
network of state-owned Peoples' Bank branches are avail—
able, besides the local level money market, as alternatives
to supervised credit provided by the state through the
quasi-government agency of multipurpose cooperatives.

The performance of the larger farms could be made
to contribute to equity objectives by making their Opera-
tions more labor absorptive. A low cost method would iden-
tify and use the critical institutional variables that need

to be manipulated to yield socially desirable results.

Summary
To be able to identify the program specifica-

tions needed in contributing to the output and equity
goals of the Five Year Plan, it would first be necessary
to determine program objectives. Program objectives con-
stitute the link between the social welfare function as
manifested in the Objectives of the Five Year Plan and
the program characteristics that need to be identified.
The.need for determining cardinal utility does
not arise because the ends or goals are regarded as
given. In any case, an effective increase in the pro—
ductive capacities of a large number of small farmers
could set in motion a cumulative process of development,

which would probably make it difficult for the losers to

 

106

bribe the gainers, in terms of the Scitovsky criterion.
These conclusions seem reinforced if it can be assumed
that marginal utilities of monetary units vary between
high and low income groups.

As income disparities are not permitted to
increase, according to the stated develOpment plan, the r
present Lorenz curve constitutes an outer bound. How-

-ever, since income is not to be equalized, the line of

 

equal distribution fails to acquire normative signifi- *lfi
cance.' Neither does the normal distribution curve provide

a practical guideline for policy. The concept of the

poverty rate, when used in conjunction with the Lorenz

curve, seems useful for identifying the objectives and

focus of programs.

Program objectives with regard to output are clear.
Since the subject of interest was distribution of income
to households rather than to land, the distribution of
output data must be transformed into distribution of
income data. Such a transformation enabled the judgment
to be made that the target groups were to be those with
paddy holdings below 4 acres in size.

The group with holdings above 4 acres usually is
above the poverty line. From a social point of view, the
activities which these larger farmers should be encouraged
to emphasize include those which increase their demand

for labor, and thus supply employment opportunities for

 

107

other rural residents whose land base is inadequate.
They, too, must increase productivity if the output

goals of the agricultural sector are to be attained.

 

 

CHAPTER V

ASSESSMENT OF VARIABLES AFFECTING

INTERDISTRICT PADDY YIELD

Introduction

 

Given the focus for agricultural develOpment
programs proposed in the previous chapter, an attempt
Will now be made to identify the technological and
institutional variables (including concomitant techniques
and infrastructure) that are important for explaining
interdistrict variation in paddy yields in Sri Lanka.

An important but not exclusive means for identi-
fying the relevant variables for Sri Lanka is through an
econometric assessment of variables. Such an analysis}
is likely to provide or deny statistical support for the
prior beliefs which exist with regard to the significance
of variables included in the model, and the unincluded
variables for which the included variableS'are proxy. It
may also provide an indication of the relative importance
or weights of the significant variables. The econometric
analysis is thus expected to improve the relevance and

focus of the study.

108

 

 

 

109

The Econometric Model

 

The Regression Equation

 

This subsection specifies the model in terms of
a linear multiple regression equation, and thereafter
discusses the rationale for (a) the exclusion of some vari-
ables from the model, and (b) the past and present role 57‘
in agricultural develOpment of the variables that are
included.

The linear regression Model A is as follows: . .u

 

X1 = (1+ 82X2 + B3X3 + B4X4 + 85x5 + 86x6 + 87X.) + 89X9

+ B10x10 + B12x12 + B13X13 + e

where:
X is the average annual paddy yield per acre

a is the intercept term determining the height of the
regression

2 is the percentage paddy acreage in district subject to
tractor plowing

X is the percentage paddy acreage in district subject to
chemical weeding

X is the percentage paddy acreage.in district subject.to
row tranSplanting

X is the percentage paddy acreage in district under
improved seed varieties

is the percentage paddy acreage in district that is
irrigated

X is the percentage paddy acreage in district that is
rented

 

110

x9 is the ten-year (1962-71) average of harvested extent
as a percentage of extent sown

Xlo is a dummy variable where X10 = 1 if in wet zone and
X10 = 0 if in dry zone

12 is the percentage of paddy holdings in district below
2 acres in size

X13 is the prOportion of government production credit
loans defaulted in the previous year, 1970-71

e is the error term, where e. are assumed to be inde-
pendent random variables with a mean of zero and
constant variance.

The data utilized for this analysis are tabulated
in Appendix F. Generally, the data are for 1971-72 unless
otherwise indicated, and are district averages for the
22 districts in Sri Lanka.

The hypothesis to be tested is the null hpothesis

H0 : 81 = 0 against the one-sided alternative HA : Bi < 0
for variables X6 and X13 and HA : 81 > 0 for the rest, at

a 10 and 5 percent level of significance.

Explanation of the Structure
of the Model

 

 

Unincluded variables.--The problem of deciding

 

which variables should be included condenses to one of

a trade-off between multicollinearity and specification
error. While multicollinearity does not cause the esti-
mated coefficients of the independent variables to be
biased, their statistical significance may not be estab-
lished because of the largeness of the standard deviation

of their coefficients. This would therefore reduce the

reliance that could be placed on the coefficients for

111

policy guidance. At the same time, the exclusion of
relevant variables will increase the stochastic component
in the error term and cause the coefficients of the
included variables to be biased, with the same adverse
result for policy decisions.

The foregoing econometric reasoning could be [7?
utilized to identify some unincluded variables and explain

the rationale behind their exclusion. Thus the area that

 

is row transplanted is included, but not the area that is ;'
seeded directly by broadcasting (on a dry or mud field), k,
or the area that is transplanted in a random((ordinary)

way. The reason for this exclusion is because of probable

high multicollinearity between the area that is not subject

to row transplanting and the area that is chemically

weeded (which is included). This presumption is based

on the belief that it would be economically irrational

to spray chemical weedicides when transplanting is in

rows, given the fact that 70 percent of farms are below

five acres in size. weeding in these circumstances would

Abe carried out by rotary methods on the larger holding or

by hand on the smaller ones. Hence the "chemically

weeded" variable (X3) that is included serves as a proxy

for the "direct seeded," "random transplanted," and

"rotary weeded" variables that have been excluded.

 

112

' As regards seedbed preparation, only the per-
centage of land area that is tractor plowed is included.
The areas that are subject to the traditional practices
of animal and hand plowing are excluded because they
are together the reciprocal of tractor plowing with
which they are collinear. Besides, they are presumed
to be closely cOllinear with the traditional practice
of hand weeding, which is excluded for the reasons
stated above.

The paddy area under new seed varieties is
included as a variable.notwithstanding its expected
collinearity with the area that is tractor plowed (X2),
chemically weeded (X3), transplanted (X4), irrigated
(X6), and possibly via the "irrigated" variable (X6)
with the "risk" variable (X9).. The importance of this
variable is considered large enough to justify its
inclusion. Conversely, the misspecification resulting
from its exclusion would be expeCted to cause the coef-

ficients of variables X2, X3, X , and X to be

4 6

"significantly" biased.

The inclusion of the percentage of land under new
seed varieties as a variable has, however, justified the
exclusion of a variable for fertilizer application. This
is due to the strong complementarity between these inputs,
and hence the high degree of collinearity that can be

expected between them.

113

Similarly, the share of output received by
the landlord as rent per district has been omitted
.because of its expected relationship with the included

variable X7 (the percentage of land rented by district).

Description of included variables.--The previous '
subsection discussed the relationship between included and
excluded variables, with a view to explaining the rationale
for the noninclusion of some variables in the model. This
subsection attempts to describe the nature and role in
Sri Lanka of the variables that are included,7

X2 (Acreage Tractor Plowed): As much as 46 percent
of paddy land is tractor plowed. Government encouragement
of tractor imports through the granting of a favorable

2

exchange rate,1 liberal allocations of exchange, and

numerous tax allowances3 resulted in the importation of
over 10,000 tractors between 1965 and 1969.4
The capital allowances allowed for tractors and

other agricultural equipment include a lump sum depreci-

ation of between 66 2/3 and 80 percent and development

 

1Sri Lanka has a two—tier exchange rate system.

2Almost all imports require import licenses.

3Department of Inland Revenue, Tax Concessions
for Agriculture (Colombo: 1966).

4

 

 

ILO, Repgrt, 0p. cit., p. 74.

114

rebates of 40 percent on the cost of the asset in addition
to the depreciation allowance.1

Many of these tractors are used in the dry (irri-
gated) zone, where ". . . very considerable use of tractors
. . . is probably without parallel in South Asia."2
Besides tilling, tractors are also used for threshing and
transport. In fact, the extent to which tractors are used
in Sri Lanka suggests that tractorization may be induced
by factors other than farm size.3 While there is evidence
that the intensity of tractor use in Sri Lanka is generally
higher in areas where relatively large-sized(holdings
predominate, the area tilled by tractors have in several
districts exceeded the area under cultivation on farms
above 5 acres in size--thereby implying that tractors
were used even in holdings below 5 acres in size. In
fact as early as 1962, the total area under paddy in
holdings of 5 acres and above fell short of the area
tilled by tractors in 10 out of 22 districts. Signifi-
cantly, almost all of the 10 districts belong in the

dry zone--which, according to Raj, accounts for 60 percent

 

1Department of Inland Revenue, op. cit.

2ILO, Report, loc. cit.

 

3The remainder of this paragraph draws heavily on
K. N. Raj's article entitled "Mechanization of Agriculture
in India and Sri Lanka," International Labor Review, Vol.
106, No. 4 (Oct. 12), pp. 315-34.

 

115

of the area sown and about 75 percent of the area tilled
by tractors.1

X3 (Acreage Chemically Weeded): The percentage of
paddy land in Sri Lanka that is chemically weeded averages
31. As previously mentioned, the need for chemical
weeding arises when the seed is either broadcast or trans-
planted randomly. About 20 percent of all paddy lands
are hand weeded, while less than 4 percent are rotary
weeded.

7 X4 (Acreage Row Planted): The percentage of

row transplanted paddy land is around 3.5, which'
corresponds to the percentage of acreage that is rotary
weeded. ~

X5 (Acreage under Improved Varities): The per- .
centage of paddy acreage under improved seeds averages
75 percent for all of Sri Lanka. Two categories of
improved seeds are used in Sri Lanka: the "old improved
varieties" (OIV--H8, H4, Ponnulot, Dickwee) which mature
in four to six months, and the "new improved varieties"

(BG 34/8, GB 34/11, IR, MI, K, etc.) which mature in

three to three and one-half months. Of about 1.4 million

 

1As Raj points out, the dry zone, which receives an
annual rainfall of less than 75", does not have the benefit
of the South-West Monsoon, and the soils are therefore apt
to be hard until the North-East Monsoon set in. Draught
animals would probably lack the intensity of energy required
to till very hard soil. Tractors substitute for both water
and animals, and even land, to the extent that land with
very low Opportunity costs is brought under cultivation.
Thus even small holders may find it worthwhile to use (hire)
tractors in the dry zone.

116

acres subject to new varieties, the "old improved vari-
eties" account for 45 percent of the acreage under both
categories of improved varieties, and the "new improved
varieties" 55 percent.1

X6 (Irrigated Acreage): The irrigated paddy area
constituted 66 percent of the total cultivated paddy in
1967-68.2 Of the total irrigated area, 20 percent is
served by major works for the two seasons, while about 46
percent is served by minor works. About 34 percent of
the annual cultivated acreage is thus rainfed. Most of
the irrigated land (70 percent) is in the dry zone, where
about 98 percent of the land cultivated in the Yala
season is irrigated. I

A major and costly input provided almost free is
I water. Although there is a flat tax of Rs. 5.00 per acre
in the colonization schemes, hardly any amount is col-
lected.3 According to the ILO Report, the state is

subsidizing paddy lands under major irrigation schemes to

the tune of at least Rs. 300 ($50) per acre, and lands

 

1Ministry of Agriculture and Lands, Implementation

 

 

Programme 1973 (Colombo: 1972).

 

2Department of Census and StatistiCs, Statistical
Abstract of Ceylon, 1969 (Colombo: 1970). All data in
this paragraph is taken from this source unless otherwise
indicated.

In view of the fairly long gestation period for
irrigation schemes, these figures are probably indicative
of the current position.

3

 

 

ILO, Report, loc. cit.

 

 

117

under minor schemes to the extent of Rs. 60 ($10) per
acre, by not collecting any irrigation rate.

X7 (Acreage Rented): According to the source of
this data,1 rent of paddy land is measured by the share the
tenant gives to the landlord, less the expenses borne by
the landlord. About 25 percent of the cultivated area had
been rented out on this basis in 1969. The percentage was
highest in the Hambantota District (57.3 percent) where the
crop share received by the landlord was also high (80.7
percent). The actual rent paid per acre, less the expenses
borne by the landlord averaged 24 percent in 1969.2

The land tenure structure is summarized in

Table 5.1.

TABLE 5.1.--Land Tenure Structure of 1962 Holdings under
50 Acres, 1962.

 

Percentage of

Ownership Pattern Total.Holdings

 

Full owners 62
Part owners 20
Full tenants 18

 

Source: T. Jogaratnam and Rainer Schickele, Practical
Guidelines to Agricultural Development Policies
in Ceylon (Peradeniya: Agricultural Economics
Research Unit, University of Sri Lanka, 1970).

 

 

 

 

1Central Bank of Celen, Cost of Production of
Paddy, op. cit., p. 36.

2151a.

 

 

 

118

Crop sharing (Andé) arrangements fall in the
part ownership class, because the landlord provides some
inputs. Andé accounts for 70 percent of tenure arrange-
ments in the part ownership class.

X9 ("Risk" Variable): The percentage of area
harvested to area sown may be regarded as a variable show-
ing the partial risks in cultivation by district. In
general,-the greater the risks, the fewer would be the
variable inputs applied, and hence the lower the yield
per acre. The shortfall in the area harvested which
averaged 5.3 percent over a 10 year period (1962-1971)
can be classified as those emanating from: (a) natural
factors (e.g., floods, drought, etc. that would be
regarded as being outside the control of the individual
farmer): (b) institutional-organizational factors (e.g.,
timeliness of input availability, price of output, etc.);
(o) technological factors (e.g., seed failure, failure to
apply fertilizer at the appropriate time, neglecting to
use complementary inputs like insecticides, weedicides,
etc.). The risk from natural factors is likely to be
higher in the wet zone where a major portion of the paddy
land area is rainfed.

1

The crop failure data for 1966-67 reveal that at

a national level the percentage difference between the

 

1Department of Census and Statistics, Statistical
Abstract, op. cit., pp. 118-19.

 

 

W

W1

 

 

 

119

area sown and the area harvested was 4.8 percent in Maha
and 6.5 percent in Yala. This information is however

not consistent with the 10 year average data which were
synthesized for use in the regression analysis as inde-
pendent variable X9. The 10 year average data show that
the difference between the area sown and the area harvested
is 6.2 percent for the Maha season, and 4.4 percent for
Yala season.1

The relative causes for crop failure is apparent
in Table 5.2. As the data in Table 5.2 indicate, the
largest single reason for crop failure (between 67 and 71
percent) could be traced to the water control factor. It
is also evident that the risk from this source is roughly
the same in both the dry and wet zones and for both
seasons. The second most important cause for the
shortfall in the acreage harvested (about 15 percent)
is attributable to institutional and other factors, while
pests account for 9.3 percent, and seed failure for 7.6
percent.

X10 (Dummy Variable): The dummy variable has
been included to make a distinction in the intercept term
between dry zone and wet zone regions in Sri Lanka. This
is likely to reduce the bias and variance in the coeffi-

cients of the other included variables.

 

1It should be recalled that 1966-67 was a non-
‘typical year in that the shortfall in the harvested area
was less for Maha.

 

tint-w. .. J'-!n--.;.1 a $1.: I
. I’A
la‘

 

 

120

.AONmH "OQEOHOUV moma
.coHNmU mo uomuumn¢ Hmowumflumum .moaumflumum can momcmo mo ucmEuummma "mousom

 

 

o.ooa hmo.oH o.ooa mom.va o.ooa mmm.NH o.ooa omm.mm Hmuoe

 

A. ouo .Hmcoflu

 

 

 

 

 

 

 

usuaumch
m.o~ mma.~ m.e ~5H.H m.~m Hom.m H.ma emm.v umzuo

m.sm a . . a a a. . ems. .oe m.m Houucoo
mm m o a o OH H mm m m ma m “mums mo gums
m.m mam m.ma mmm.~ o.e Ham m.e mam.m mummm
m.m mum m.m ems 5.4 «cm m.m mmm.m musaamu 666m

w mmuofi w mmuofl w mouofi w monod
snow one: snow can: mmsmu
ozoN umz mcoN mun

 

.smmaummma .mxcmq sum as mmsmo an nmamammmao musaflmm mono mo unsoeauu.m.m mamas

121

The basic difference between the two zones is
obviously climatic and ecological, the most important
single difference being in the limited amount of annual
rainfall (less than 75 inches) in the dry zone, which
is confined to the months from November to March. A
fairly distinct drought season between the months of May
and September causes soils to be hard and evapotranspiration
(which determines the precipitation that is effective for
plant growth) to be low.1 This in turn influences the
variables included in the model.

The dry zone condition influences the included
variables in the following ways. The hard soil makes
tractor plowing necessary and the lack of rainfall makes
it worthwhile to use certain types of weedicides and
pesticides which otherwise would be washed away by
"overhead" water. Much of the land is irrigated and
colonized under government sponsored schemes. This
factor influences the "irrigated land" variable,
"improved seed" variable (since its yield response is
partly a function of water control), and the "land rented"
variable (as much of the population is made up of colo-

nists who are outside traditional tenure patterns).

 

1George Thambyahpillay, "Climatological Research
and Agricultural Deve10pment in the Dry Zone," in Devel-
opment of Dry Zone Agriculture, ed. by O. S. Perera
(Colombo: Ceylon Association for the Advancement of
Science, 1967). '

 

 

122

These factors also influence the average size of farm,
and perhaps also the risk variable.
X12 (Paddy HOldings Size): This variable measures

the prOportion of farms below 2 acres in size by district.

 

This variable was found to be preferable to taking the
average farm size in each district, due to the possibility F“)
of a disporportionate influence on the mean of,a few 5 '1
very large farms in each district.

The average size of all holdings reporting paddy :r‘€
1

 

is about 1.6 acres. The average size of holdings is v
larger in the dry zone (2.6 acres) than in the wet zone
(0.6 acres). The wet zone contains most of the tradi-
tional areas of cultivation, and although no data are
available, the man-land ratio in this zone is known to
be high. The wet zone has about one—half the total paddy
acreage, while containing about three-fourths of the rural
population.2

X13 (Credit Default): Sri Lanka's Agricultural
Credit Scheme was introduced on a nationwide basis in
1967. The Scheme of 1967 was designed to accelerate an
incipient "Green Revolution," and its timing was expected

to coincide with the adOption of new technology by farmers.

The salient features of the scheme were as follows:

 

1A farm may cultivate other crops besides paddy.

2This conclusion is inferred from the data in
Department of Census and Statistics, Socio-Economic Survey,

op. cit.

 

 

123

(i) The ultimate source of funds was to be the
Central Bank, a quasi-autonomous body, rather than the
Department of Agrarian Services, a department of the
government.

(ii) The rate of interest was to be a maximum of
12 percent with provision for a 3 percent rebate for
prompt payment.

(iii) The maximum granted was to be Rs. 262 per
acre for non-IR-8 cultivation and Rs. 370 per acre for
IR-8 cultivation up to a maximum of 10 acres (Table 5.3.).

(iv) Government and extension workers were expec-
ted to collaborate with cooperative societies in
preparing a detailed cultivation program for the season,
showing the total area to be cultivated, the area to be
subjected to improved methods, the estimated requirements
of inputs and cash capital required.

(v) The collateral was to be farmers' ration books
of one year validity which entitled them to a pound of
"free" rice per week supplied by the government. Default
was to disqualify farmers from receiving further loans the
next year, but not from receiving "free" rice.

(vi) The loans were to be released in three stages
and were to be partly in kind.

One of the provisions of the scheme made defaulters
ineligible for further loans, Hence theoretically the

default rate in the previous year could be said to affect

124

TABLE 5.3.--The Maximum Sums Granted for Meeting Input
Costs Under the 1967 Credit Scheme.

 

 

 

IR-8 Loan Non-IR—B Loan

Purp°se- Rs. Rs.

Seed paddy 25 32
Ploughinga 60 60
Fertilizer 160 95
Transplan ing and row '
seedingg 25 25
Hand weedinga 25 , 25
Pesticides 75 25
Total 370 262

 

Source: Report of Cooperative Commission, 1970, as quoted
in the Report of the Cooperative Commissioner for
1968-69.

 

aPayment for these is necessary in order to hire
labor for short periods for timing reasons. These payments
are made in cash.

yields negatively in the following year by imposing capital
rationing on farmers. But this would be true only if
production credit loans were utilized for their intended
purpose of increasing paddy productivity.

There is evidence that credit is not being used for
its intended purpose. In looking at the performance of
the credit scheme we observe a negative correlation between
changes in productivity and changes in repayment levels.
Although average annual per acre yields have increased

from 41 bushels in 1966-67 through 51 bushels in 1969-70,

125

to about 55 bushels in 1970-71,1 repayment levels show a
declining trend in the same period. Voluntary repay-
ments, which were 70 percent of loans granted in Maha
1967-68, declined through 49.7 percent in Maha 1969-70

to 39.0 percent in 1970-71.2

Results
Overall, a coefficient of multiple determination
or fig of .76 indicates that 24 percent of the variance in

paddy yield per acre is not explained by the variation in

the independent variables that are included.

Statistical significance of independent variables.--

 

The least squares estimates fitted to 22 observations are

as follows:3

Constant Tractor weeded Transplanted
x = - 127.55 + 0.02 X + 0.12 x + 0.60 x
l 2 3 4
(8 =0.10) (S =0.l3) (S =0.32)
82; B3: 84
(t2 =0.21) (t'3 =0.87) (t4 =l.87)*

*Significant at 5 percentwhere t = 1.80.4

 

1Department of Census and Statistics, Statistical

Pocket Book, 0p. cit., p. 61; and Ministry of Agriculture
and Lands, Implementation Programme, op. cit.

 

 

 

 

 

2Central Bank of Ceylon, Annual Report 1971, 0p.

 

cit.

3The computer printout is in Appendix B.

4The t-value at the 10 percent level of probability
is not shown because all t—values of estimates that are
significant at 10 percent are also significant at 5 percent.

126

Seed Irrigated Rented Risk
+ 0.30 X5 - 0.03 X6 + 0.17 X7 + 1.71 X9
(S =0.10) (S =0.09) (S =0.13) (S =0.57)
B5 B6 B7 89

(t5 =3.18)**(t6 =0.28) (t7 =1.28) (t9 =3.01)**

Dummy Size Default

- 0.67 Xlo - 0.28 X12 + 0.06 X13

(SB =8.89) (SB =0.12) (SB =0.07)
10 12 13

(tlo =0.08) (t12 =2.36)* (tl3 =0.88)

*Significant at 5 percent where t =
**Significant at 1 percent where t = 2.72.

S8].- is the estimated standard errors of coeffi-
cients, and ti the corresponding t-value. Since there are
11 independent variables and 22 observations, there are
11 degrees of freedom. The critical point with 11 degrees
of freedom at the 5 percent level of probability is 1.80
and at the 1 percent level of probability, 2.72. If the

null hypothesis H is true, 61 will be centered on B1 = 0

0
and there will be only a 5 or 1 percent probability of
observing a t-value exceeding 1.80 and 2.72, respec-
tively. These critical values define the rejection
region for determining the degree to which (if any), the
independent variables are significantly related to the
dependent variable (paddy yields).

The observed values that fall in the region where
the null hypothesis (HO : Bi = 0) is rejected are t4, t5,

t and t at the 5 percent level of probability and

9' 12

t5 and t9 at 1 percent.

127

Economic significance of variables.--In Bayesian

 

terms, a distinction may be drawn between economic and
statistical significance in the light of prior beliefs.
Although an observed t-value may fall in the "acceptance"
region, it does not necessarily mean that the null
hypothesis (that the estimated coefficient is zero) would
be accepted. Whether or not it is accepted depends on
the reliance that can be placed on the observed t-value.
In formal terms, the prior distribution is adjusted to
take account of observed data, with the weight attached
to the observed data depending on the probability of the
posterior distribution.1

It may be that the lack of statistical signifi-
cance is due to the sample being too small, and the
standard errors consequently large. Multicollinearity
may also have contributed to the largeness in the standard
errors, thus causing the observed t-values to be statis-
tically insignificant.

In applying Bayesian logic to the foregoing
results we may conclude that the renting of land and the
practice of chemical weeding have at least some positive
effect on yields. The coefficient of the variable
measuring the credit default rate (X12) is positive,

thus justifying a reversal of a prior belief that the

 

1See, for instance, Ronald J. Wonnacott and Thomas
H. Wonnacott, Econometrics (Toronto: John Wiley and
Sons, Inc., 1970), pp. 209-12.

 

128

disqualification of farmers from receiving paddy loans

due to default adversely affects average productivity the
following year. The result supports the hypothesis that
credit disbursement may not have contributed to produc-
tivity increases. The weakness in the t-value with
respect to the effect of irrigation on yields makes it
advisable to regard the results with respect to this
variable as being inconclusive. However, the results with
regard to tractor use fit a prior belief that tractors are
more important potentially for cropping intensity rather

than yields.l

Unincluded Institutional
Variables

 

 

In looking at the results overall, the R2 value
indicates that 24 percent of paddy yields is not explained
by the included variables. There is reason to believe
that most of the unexplained variation is attributable to
institutional-organizational and resource endowment
factors. The reason for this belief lies in the fact
that most of the variables relating to technology (inclu-

ding techniques) were included in the regression analysis,

 

1For instance, in the Pakistan Punjab tractors had
very little influence on yields. The tractor farms did
not have tractor implements for deep plowing, row crop
sowing, and fertilization. Also tractor use had not been
accompanied by a simultaneous increase in the bundle of
inputs and improved practices. See Bashir Ahmed, "Farm
Mechanization and Agricultural Development: A Case Study
of the Pakistan Punjab" (unpublished Ph.D. dissertation,
Michigan State University, 1970).

129

either explicitly or by proxy (to the extent that the
included variables were representative of other excluded
variables with which they were collinear). Hence much of
the unexplained variation is probably due to unincluded
institutional, infrastructural, and resource endowment
(including weather) factors. The institutions examined
below are those that may account for some of the unex-
plained variation in the econometric analysis. They are
categorized into objective and subjective institutions
and are examined from the standpoint of their possible

effect on interdistrict adoption and hence on yields.

Objective Institutions

 

Information.--Information to farmers flows chiefly

 

from extension channels, informal local channels, and the
demonstration effects of early adOptors (3).1 While some
studies identify extension as having the most effect on
adoption, other studies emphasize less formal local
channels. The Comilla study reports that "the training of
model farmers from the village cooperatives is a major edu-
cational input which is multiplied by the subsequent
discussion held by these individuals in their villages,
often aided by written material" (8, p. 51). Also, that
"through local communication channels, the agricultural

knowledge extended through the cooperative passes to all

 

1The number in parentheses refers to the study
listed in Appendix A.

130

members of the village. . ." (8, p. 59). Similarly, the
West Pakistan study also reports the importance of informal
channels (5, p. 85). On the other hand, in the Barant
Development Block, North India, "dealing personally with
extension agents was found to be the most important factor
after per capita income influencing the adoption behaviour
of farmers" (7, p. 25).1

It is not clear to what extent less formal local
channels substitute for deficient extension. To the
extent that this is so, the importance of less formal com-
munication channels will diminish relatively if extension
is strengthened.

There is evidence which suggests that: (a) exten-
sion is deficient in Sri Lanka, and that the distribution
of extension agents varies among districts, and (b)
informal channels are resorted to because of extension
weaknesses.

No data are available to develop a statistical
assessment of extension on yields in Sri Lanka. It is
known, however, that in Sri Lanka there are some 1,669
extension workers to serve the needs of 1.2 million farmers
with holdings below 50 acres in size and that extension

efficacy varies among districts.2 Experience in other

 

1See also Arthur Niehoff, A Casebook of Social
Change (Chicago: Aldine Publishing Co., 1966).

 

2Director of Agriculture, Administration Report
1967-8 (Colombo: Department of Agriculture, 1970).

 

131

Asian countries suggests than an extension worker rapidly
loses impact if he has to serve more than 250-300
farmers.l Yet in Sri Lanka a village level Operator has
to deal with l,500-3,000 farmers, which is about five

to ten times this rate. Not surprisingly the recent
Central Bank Survey on loan defaults reports, "Extension
work does not seem to have been conducted in most of the
areas to which these (credit) defaulters belonged. The
extension officers had not visited farmers even once in

certain areas."2

The survey also reports that ". . . input
utilization is (often) left entirely in the hands of the
farmer. He himself does not supervise its usage as he

has no understanding of modern scientific methods."3
These numerical weaknesses in extension are probably
exacerbated by quality weaknesses.

There is also evidence which suggests that
informal channels in Sri Lanka are a substitute for
extension. The Central Bank Survey states that due to
"absence of help of a scientific nature from the exten-

sion officers he (the farmer) relies on his own 'hunches'

or he is influenced by his neighbours about modern

 

lRainer Schickele, Ceylon Papers 1967-70 (New
York: A/D/C, 1972). In Japan the ratio is 1:550 accord-
ing to Naseem, #1, Appendix A.

 

2Central Bank of Ceylon, Survey of Defaults in the
Repayment of New Agricultural Loans (Colombo: Department
of Economic Research, 1972), p. 67. The word in paren-
theses is the writer's.

3

 

 

Ibid., p. 35.

132

methods of production."1

That interpersonal localite
channels are a less preferred substitute is also suggested
by the Recife area study where it is reported that farmers
who use "direct" sources of information (extension) rely
less on "indirect" (informal) sources (2, p. 175). This
factor, to the extent that it is valid, has the effect

of further emphasizing the importance of the extension

variable in explaining the unexplained variance in inter-

district yields.

Input supply.--Differences in the efficacy of

 

institutions governing the adequate and timely supply of
inputs could also account for a considerable portion of
the unexplained variation in the regression analysis.

Fertilizers in adequate quantities are often not
readily available when needed2 due to :

(a) structural defects in the organization of
some cooperatives which are exclusive outlets for inputs
including credit, and

(b) the need to satisfy cumbersome bureaucratic
requirements on the part of multipurpose cooperatives in
order to get fertilizers for farmers, and the consequent
imposition of complex procedures on farmers in order for

them to qualify for such subsidized fertilizer.

 

Ibid. Ibid.

133

1. The major structural deficiencies in coopera-
tives are as follows:1

(a) Unsatisfactory personnel in some coopera-
tives leading to poor organization and
management inefficiencies;

(b) Lack of consolidation and centralization
which makes supervision of management
difficult and inhibits economies of
scale;

(c) Lack of small farmer influence due to
large farmer control over certain c00pera-
tives.

2. Elaborate bureaucratic procedures have been
established to ensure.that subsidized fertilizer is used
for specific crops. Fertilizer subsidies vary.among
crops, while some crOps do not qualify for any subsidy.
Fertilizer for paddy cultivation receives a 50 percent
subsidy, necessitating intricate bureaucratic checks to
ensure crop specificity. Hence multipurpose cooperatives,
which have exclusive distribution rights, have problems
in ensuring the timely availability of inputs to farmers.
Once such fertilizer becomes available at the local
cooperative, farmers have to satisfy complex and time-
consuming procedures to receive fertilizer. These defects
impose costs on farmers that serve to at least partially

offset the benefit of the 50 percent subsidy. These costs

may also be said to vary between districts depending on

 

1The sources of this information are "The Royal
Commission Report of 1969," and the report on the Socio-
Economic Survey of the Elahera Colonization Project in
Ceylon Papers by Schickele, op. cit.

 

134

the number of farmers serviced by a given cooperative, and
the quality and influence of the local personnel.

In general, the price subsidy on strategic seed and
fertilizer inputs can be expected to increase paddy output,
including paddy productivity. It would, by reducing aver-
age costs of production increase the profitability of
paddy cultivation (given demand) in relation to other less
profitable crops. This would involve a more intensiVe
use of paddy land and a transfer of land, labor, and
capital resources from the cultivation of other crops
to the cultivation of paddy (since the aggregate supply
response for all agricultural products is probably very

low) .1

Input subsidies are more likely to encourage
innovation than price supports (as it enables the farmer
to discover the MVP of fertilizer), and with input subsi-
dies government expenditure is in direct proportion to
levels of input adetion.2 But as innovation becomes
widespread such subsidies could prove to be an enormous
burden on the exchequer. Moreover, peasant farmers

would be more familiar with output prices than.input

prices and hence more responsive to increases in the former.

 

1See, for instance, John Mellor, "The Functions of
Agricultural Prices in Economic Deve10pment," Indian
Journal of Agricultural Economics, Vol. XXIII, No. l
(Jan.-March, 1968).

2See, for instance, Raj Krishna, "Agricultural
Price Policy and Economic Deve10pment," in Agricultural
Development and Economic Growth, ed. by Herman M} SoutHworth
and Bruce F. Johnston (Ithaca, New York: Cornell Univer-
sity Press, 1967).

 

 

 

135

Subjective Institutions

 

Some subjective institutions may also contribute
to explaining the unexplained variation of 24 percent.

Those examined are religiops, cultural, and social factors.

Religion and culture.--This variable may’influence

 

productivity by modifying farmers' responses to economic
stimuli as predicted by received economic theory.1 It may
assist in explaining the resistance of some farmers in
Sri Lanka to technological change. The rural sector is
78.4 percent Buddhist2 and it would not be surprising if
many slow adoptors subscribe (perhaps below a perception
threshold) to the basic Buddhist belief, as stated by
Alfred Marshall,.that "real riches.consist not in the
abundance of goods, but in the paucity of wants."3

To desire less material goods in terms of the
individual's self-concept4 is still maximizing behavior,

although the goods demanded are nonmarket in character.

 

lAn assumption subscribed to by writers like
Schultz in Transforming Traditional Agriculture, op. cit.,
but disputed by writers like Kusum Nair in The Lonely
Furrow (Ann Arbor: University of Michigan Press, 1969).

2Department of Census and Statistics, Socio-
Economic Survey, op. cit.

3

 

 

 

 

Marshall, Principles, op. cit., p. 136.

 

4The individual's self-concept is a critical factor
in behavior, according to Carl Rogers:h1"Client-Centered
Therapy" as described by H. S. Hall and G. Lindzey in
Theories of Personality (New York: John Wiley and Sons,
1970), Chapter 11.

 

136

However, it could certainly influence the farmer's atti—
tude towards technological changes which promise the
means of purchasing only material goods.

Questions also arise as to whether farmer resis-
tance to more widespread use of pesticides could be
ascribed to one of the Fundamental Precepts of Buddhism
which forbids all killing, and is widely interpreted as
applying to insects, perhaps even those that compete for

Man's food.

Other social factors.-—Sociological case studies

 

of rural village life in.Sri Lanka1 suggest that a con-
siderable interdependency of preference and production
functions exists at the village level that may influence
paddy productivity.

The concept of interdependency among preference
functions has been incorporated into consumption theory in
the form of bandwaggon, snob, and veblen effects, and
between production functions into development theory as
demonstration effects.. From a sales standpoint such
interdependency is generally regarded as a low cost means

of gaining a more wideSpread acceptance of commodities.

 

1These case studies are contained in an UNRISD
study titled Cooperatives and Planned Change in Asian
Rural Communities, edited by Inayatullah (Geneva:
United Nations Research Institute for Social Deve10pment,
1970), pp. 175-341.

 

 

137

However, such interdependency seems to have
effects of an opposite nature in Sri Lanka by exercising
a restraining influence on more progressive early adop-
tors. Numerous instances of hostility toward fellow
villages who show signs of progressivity have been cited

in the studies.

Summary

This chapter has further contributed to the objec-
tive of this study, which is to determine how to use the
variables of technology and institutions to increase pro-
ductivity on (small) farms below 4 acres in size. The
chapter assessed the relative importance of variables
as they influence interdistrict paddy yields in Sri
Lanka. Since the quantitative analysis failed to explain
24 percent of interdistrict yield, other objective and
subjective institutional variables were identified that
could be expected to account for a considerable prOportion
of this unexplained variation.

It was found that the statistically significant
variables were row transplanting, improved seed, risk
factors in production, and farm size. The number of
significant variables for policy purposes was expanded
to include fertilizer use (which was unincluded but

represented by the seed variable), chemical weeding, and

the renting of holdings. In the case of tractor use and

138

and production credit, the statistical results suggested
they were not important for past increases in paddy pro-
ductivity.

The other institutional variables identified but
not quantitatively assessed were the distributiOn of
extension personnel, input supply factors, religious and
cultural factors, and social factors, mainly those
stemming from the interdependency of preference and pro-

duction functions of rural inhabitants.

CHAPTER VI

POTENTIAL OF VARIOUS PRODUCTION PRACTICES

INFLUENCING INCOME DISTRIBUTION

Introduction

 

Objectives

 

This chapter attempts to assess the relative
importance of variables already identified, from the
standpoint of income distribution ianri Lanka. The
econometric analysis just concluded, supplemented by
other information, provided some insight into the relative
importancecfiftechnology (including techniques) and insti-
tutions (including infrastructure) that explain~
interdistrict paddy productivity. But this information
is incomplete in terms of the objective of this study,
which is to see how not just to increase paddy produc-
‘tivity, but to increase paddy productivity in a manner
that will enable low income groups to gain Control over
sources of income streams, and see that employment Oppor-
tunities are adequately generated.

The variables that determine income can be divided
into those that influence farm revenue and those that

influence farm costs. Revenue to farmers from the sale of

their output is a function of both quantity and price.

139

140

Since price is guaranteed and administratively fixed in
Sri Lanka, the critical variable from the standpoint of
income is output per acre per annum. Output per acre per
annum is in turn a function of yield and cropping inten-
sity, given cropping patterns. Cost per unit is also
important as income benefits from higher yield could be
offset by higher costs.

Variables affecting both revenue and costs will
be examined with a View to determining their relative and
absolute significance for income distribution from both a.
statistical and an economic (policy) standpoint. Farm
size will be used as the central variable against which
each of the revenue and cost variables will be assessed,
from productivity and employment standpoints, so as to

relate to equity objectives.

Sources and Methods

 

The sources of information for assessing the
equity significance of variables are as follows:

1. The results of the quantitative analy-
sis. ‘

2. .0ther information with regard to Sri
Lanka.

3. The experience of other countries, 1
mainly India, Pakistan, and Bangladesh.

These three sources are complementary and equally

important. The statistical results will be assessed in

 

1The number in parentheses will refer to the
study listed in Appendix A.

141

the light of prior beliefs, which will serve to reinforce
or question the statistical findings.1

Irrespective of whether the prior beliefs confirm
or contradict the statistical findings, they serve to
strengthen the foundations on which to base the policy
prescriptions to be made in a subsequent chapter. The
aggregate information to be gained is expected to influence
the degree of certitude with which policy suggestions are

made, and generally contribute to the quality of judgments.

IdentifiCation of Equity Criteria

 

As mentioned in connection with the conceptual
framework, the interaction of pure technology with insti-
tutional variables (given resource endowments) determines
growth and income distribution. It was also suggested
that the same variables that determine growth could prob-
ably be made to influence income distribution if they could
be given focus. The targets of primary focus were deter-
mined in this study to be farms below 4 acres in Sri
Lanka. Since attention is directed at small farms, the
ability to focus programs depends on the economies of

scale in the primary variables.

 

;The methodology applied is a loose variant of

Bayesian logic, but the rationale is the same: to temper
statistical analysis by common sense.

142

Scale Neutrality

 

The objective of program focusing is to introduce
the dynamism of growth into the production system of
farms below 4 acres in size. The degree of success with
which the focusing of programs could be accomplished is
mainly through divisibility of inputs, which refers to
scale (size) neutrality in the variables significant for
paddy productivity, and the degree to which the diffusion
process could be made pervasive among the target groups.
Size neutrality may be said to exist if the overhead
costs of technological change are not significantly
different between small and large farms, at a given level
of productivity per acre.

Since the interaction of technology with insti—
tutions determines both output and income distribution,
consideration for scale neutrality in technology alone is
inadequate for the successful focusing of programs;
attention will also have to be paid to scale or size
neutrality in institutions. It appears in fact that the
primary variables can be successfully focused only if
scale neutrality in one is not permitted to be offset by
indivisibilities in the other. For instance, indivisi-
bility in physical technology can be offset by 1.
institutional-organizational factors, as is the case
when small farmers gain access to large tractors through

custom hiring or cooperative ownership. Conversely,

143

physical divisibility in technology can be offset by
indivisibilities in institutions as is the case when, for
instance, small farmers do not apply the seed-fertilizer
package because they cannot gain access to adequate
technical information due to extension biases in favor

of large farmers.

Employment Potential

 

The second criterion which will be applied for
the evaluation of variables from an equity angle, is their
labor absorption potential. To the extent that scale
neutrality in variables actuateseulappropriate outward
shift in the production functions of small farms it
would create employment Opportunities on affected farms,
depending on the type of technology used (whether labor
or capital intensive), and the employment slack or dis-
guised unemployment on larger farms (since some of the
labor supply for large farms comes from landless or

landpoor categories).

Congruence

 

If a technological or institutional variable that
is scale neutral and/or labor absorptive is also signifi-
cant from the standpoint Of productivity, there would then
be congruency between growth and equity Objectives. This
quality of congruency would be a most desirable attribute
in a variable from the standpoint of the Objectives of the

Five Year Plan.

144

There is no presumption that simply because an
alternative technology is socially more desirable, it
will be automatically adopted. Actual adoption is a
function of private profitability at the point of output
where the added felt costs are equated with added per-
ceived returns. The challenge is to identify institutional
modifications that will bring private and social benefits/

costs into reasonable alignment.

Assessment of Revenue Variables

 

The income to (small) farmers from farming Opera-
tions depends on the difference between average revenue
and average costs at a given level of output. The revenue
aspect examined here is a function of price and output.

Since the farm level price of paddy is adminis-
tratively fixed, it is unlikely that the prospective price
is discounted by a risk fraction by farmers, as would be
the case if a free market prevailed. However, since the
guaranteed price is the delivered price, some marketing
costs are internalized to the farm. These are the fixed
costs of transporting the produce to the nearest collec-
tion point, and the costs associated with uncertainties
regarding the timely availability of transportation
facilities soon after harvest. The costs of uncertainty
will depend mainly on farmer "holding power" which in
turn is a function<xfthe.urgency of the need for cash, and

amount of storage capacity available. A discussion of

145

these factors properly belongs in the subsequent subsec-
tion assessing costs.

The investigation here is directed at finding out
how farm size influences yield per acre and cropping
intensity. If yields and intensity are already high on
small farms, or if there is limited potential for yield
increases on small farms, then strategies for increasing
small farmer incomes will need to concentrate more on
the cost side. If they are not, the potential for
catching-up with large farmers could be great, and atten-
tion will need to be directed at both cost and revenue

aspects.

Yields

Evidence suggests that small farmers have a high
yield potential, probably even higher than large farmers.
For instance, a study of the Budaun District in India
shows a negative relationship between size and yields.
This study concludes that subject to operating capital
being available, and fertilizers, etc. being free to move
between farms in response to market stimuli, the predicted
net crOp return per acre on the small farms (those below
ten acres in size) will average 30 percent higher than
medium or large farms (9, p. 203-04).l A similar phe-

nomenon between farm size and yields was Observed in the

 

1The number in parentheses refers to the listing
in Appendix A.

146

Pakistan Punjab (2, p. 108). The study of West Bengal
and Andhra Pradesh reaches a general conclusion with
regard to rice growing areas that after an "experimental
period" of three seasons, the new varieties were adopted
to a greater extent by small farmers than by large
farmers. Whereas in the post-experimental period large
farmers adOpted on 20-30 percent of their acreage, small
farmers adopted on a much higher proportion (35-80 per-
cent) of their acreage. The Central Bank Survey in Sri
Lanka1 also concluded that a negative relationship
between farm size and yields existed in Sri Lanka--the
average yield on farms below 2 acres in size was 34.5
bushels, while the average yield of farms larger than

2 acres was 33.7 bushels.

The results of the statistical analysis appear
to contradict the foregoing evidence. The regression
coefficient of the farm size variable (X12) is signifi—
cantly negative at the 5 percent level of probability,
where farm size is the percentage of farms below 2 acres
in each district. This relationship was confirmed by a

regression of yields per acre on the average size of

 

1Central Bank of Ceylon, Cost of Production of
Paddy, op. cit., p. 54.

 

 

147

farm and other variables.1 The resulting coefficient was

positive but statistically not significant, probably due
to the influence of a few large farms on the mean size
of holdings in each district.2
However, there may not in fact be any contradic-
tion between the prior knowledge of a negative relationship
between farm size and yields and the posterior knowledge
of a statistically significant positive relationship between
them. This is because mixing adOptors and non-adOptors
at an aggregated national level, particularly in a situ-
ation where early innovators are large farmers (an aspect
to be examined later in this chapter), makes it possible
to observe a significant positive relationship between
yields and farm size without refuting the prior belief
held.
Two conclusions emerge from this analysis. One
is that small farmers are capable of increasing yields
considerably, even more than large farmers. Hence there

is greater potential for increasing incomes through

increases in yields than is generally believed to be the

 

1The model (B) is:
Xl=a+82X2+B3X3+B4X4+855X +866X +8.7.7X +88X
+ 89X9 + Bloxlo + e
where X is the average size of farm and the other vari-
ables are the same as in Model A. The computer printout
of results is in Appendix C.

2The critical point at 10 percent was 1.36 while
the Observed t-value for X8 was only 0.24.

148

case. Second, if small farmers tend to lag behind in the
adoption of new technOlOgy in Sri Lanka, then the sig-
nificant positive relationship observed at the national
level between farm size and productivity may be a valid
but transient one, caused by the particular location of
the Sri Lanka paddy subsector on the growth path.

Reasons for Negative Relae
tionship

 

Four main reasons may be identified for the preva—
lence Of a negative relationship between yields and farm
size among innovating farms. One is the need for small
farm households to make the most of what land is available,
and their possibly higher utility from added income.
Another stems from the inherent nature of the agricultural
production process. Agricultural production processes
cannot be routinized, and require frequent monitoring with
many on—the-spot "snap" decisions needed to betaken.
Since a farmer is able to know a smaller farm more inti-
mately than a larger farm, there would be greater Opti-

mality in smaller farms.l

A third reason is that the ratio
of permanent labor to land and Operating capital on small
farms is much higher than on larger farms; consequently

small farms require very little expensive hired labor in

the labor scarce months. In Uttar Pradesh, India, for

 

1A point made by Schickele in Agrarian Revolution,
9p. cit. But this would be so only if entrepreneurial
ability is not correlated with farm size.

 

 

149

instance, this ratio is twice that on medium sized farms
farms and five times that on large fams (8, p. 204). A
fourth reason is probably a need for innovating small

farmers to put a larger proportion of acreage under new
varieties so as to spread the overhead costs of techno-
logical changes among a greater number of output units--

i.e., lower their average costs of production.

Croppinngntensity

 

It is now hypothesized that even if a positive
relationship between farm size and productivity exists,
it may be offset overifluaperiod Of one year if there is
a strong enough negative correlation between farm size and
cropping intensity. The cropping index, which measures
the intensitycflfpaddy land use over a period of one year,
seems to be negatively correlated with farm size. Evi-
dence of this was found in the Pakistan Punjab, where the
cropping index was 125.8 percent on the below 12.5 acre
farms and 118.6 on the larger than 12.5 acre farms (2,

p. 108). A similar phenomenon is evident in Sri Lanka,
as can be perceived in Table 6.1. -

The data in Table 6.1 suggest a negative rela-
tiOnship between farm size and crOpping intensity in Sri
Lanka. For the target (below 4 acre) groups the percent-
age of land double-crOpped is about 51 percent while for
other (above 4 acre) groups it is 26 percent. However,

this data is neither sufficient nor conclusive. For one

150

TABLE 6.1.--Double Cropping by Size of Paddy Holding.

 

 

Size of Holding Percent of Land
(acres) Double-Cropped
Under 1 53
2.5 - 5 50
5 - 10 45
10 - 100 32
100 - 250 17
Over 250 . 10

 

Source: Census of Agriculture, 1962. Vol. VI as repro-
duced in Table 5, p. 7, by T. Jogaratnam and
Rainer Schickele, in Practical Guidelines to
Agricultural Deve10pment Policies in Ceylon,
Agricultural Economics Research Unit, University
of Sri Lanka, Peradeniya, September, 1970.

 

 

thing it does not show how farm size influences land use
intensity. For another, there may be other factors besides
farm size that enter into the relationship.

Improved insight into the aforementioned rela-
tionship between land use intensity and farm size may
be gained by regressing the cropping index data by dis—
trict on farm size and other relevant variables.1 Such

an analysis would:

 

1This third model (C) is as follows:
x11 = a + B2x2 + B6X6 + B7X7 + B9X9 + B10x10 + 81le2
where
Xll-is the cropping index,

X12 is the size variable Obtained by taking the percentage

of farms in each district below 2 acres in size, and

15.1

(a) Facilitate a verification of whether or
not a negative correlation coefficient between farm
size and land use intensity exists and, if so, if it
succeeds in offsetting any positive correlation between
farm size and paddy land productivity, and thereby
provides a more complete income picture over the period
of one year.

(b) It would help to identify the variables that
are important for gaining a more widespread adoption of
double crOpping, particularly among the target groups,
in view of its considerable income (including employment)
generation potential.

(c) It would, if valid, make a cogent contri-
bution to the case for a redistribution of paddy .
holdings in Sri Lanka, and hence provide a palpable
example of congruence between growth and equity objec-

tives.

 

X2, X6’ X
Model A.

7, X9 and X10 are the same as in the original

The computer printout showing results is in
Appendix D.

The crOpping index is computed from data in the
Implementation Programme, 1973, by taking the percentage
of the Maha season acreage that is harvested for the Yala
season and adding it to 100. The acreage figures are a
lO-year average. For example, since an average of 1.2
million acres were under the Maha crOp in the past 10 years,
’but only 0.6 million acres of the same land was under
Yala in the same period (which is 50 percent of the maha
season acreage), the national cropping index is 150.

152

Results of Model Cl

 

The results from regressing the cropping index on
farm size and other variables show a negative coefficient
for farm size, where farm size refers to the percentage
of farms below 2 acres in size; i.e., as the percentage
of farms under 2 acres increases, the cropping index
decreases.2 The results seriously question the validity of
the relationship in Table 6.1. In fact, a t-value of
-l.55 (the critical point is t = 1.35) is sufficiently
strong to justify a reveral of a prior belief that there
'is a negative relationship between farm size and cropping
intensity.

The policy significance of these findings is that
since potentials exist for increasing both yields 92g_
double cropping on small farms to an extent greater than
envisaged previously, there seems to be scope for
increasing income more appreciably than otherwise on the
revenue side. Now if average costs of production can also
be reduced, even greater scope will exist for increasing
the incomes of low income groups. These cost aspects are

examined next.

 

1See the computer printout in Appendix D.

2This finding was verified by regressing the
cropping index on the average size of farm in each dis-
trict, and other variables. The coefficient in this
case was also found to be significantly positive at
the 10 percent level.

153

Assessment of Cost Variables

To the extent that variables affecting the felt
costs of innovation to small farmers can be made scale
(size) neutral and divisible it will decrease average costs
for small farmers. Lower costs can be expected (given
returns) to increase the profitability of new technology,
gain its more widespread acceptance, and so increase the
productivity and incomes of affected farm households. It
could also actuate an outward shift in the demand curve
for hired labor, and thereby increase household incomes on
other farms on which labor is surplus.

The central variable with regard to scale neutral—
ity is physical farm size. The statistical measure of
the coefficient of correlation between farm size and each
of the other variables that have been quantitatively
assessed (Table 6.2) is expected to be at least mildly sug-

gestive of the degree of scale neutrality (divisibility)

 

 

 

in a specified variable.1 The variable of farm size X8 is
the average size of farm by district.
1The verbal interpretation of the correlation
coefficient in Appendix B is as follows:
Coefficient (r) Description
r = 0 Perfectly divisible
0 < r 3 .24 Highly divisible
.25 < r S .49 Fairly divisible
.50 < r S .74 Low divisibility
.75 < r S .99 Indivisible
r = l Perfectly indivisible

Since the correlation coefficient is a measure
that is inherently biased upward, it may suggest a higher
degree of indivisibility than may exist.

154

TABLE 6.2.--Coefficient of Relationship between Independent
Variables and Farm Size.

 

 

Variables Average Farm Size X8
x2 Tractor plowing 0.75
X3 Chemical weeding 0.46
X4 Transplanting -0.l4
X5 Improved seeding 0.54
X6 Irrigation 0.39
X7 Renting -0.66
x9 Risks -0.06
Xlo Dummy (Wet zone=51) ‘ -0.73
Xl3 Credit defaults -0.55

 

Source: Correlation Matrix, computer printout, Appendix E.
(Average farm size data is from the Central Bank
of Ceylon, Survey on Cost of Production of Paddy,
Economic Research Department, Colombo, 1969.

 

Due to the possibility of spurious cOrrelation,
the cOefficient of correlation will not qualify as a full-
fledged coefficient Of divisibility; to regard it as such
would be to attach exaggerated importance to it. The
utility of the correlation coefficient for the following
analysis will lie in its ability to confirm or raise
questions with regard to prior beliefs that already exist
with respect to divisibility. The source of such prior
beliefs will, as stated earlier, be information from Sri
Lanka and other countries, principally India, Pakistan,

and Bangladesh. ,The studies are listediJLAppendix A.

155

An integrated approach will be used for assessing
the equity impact of the variables. The organization
followed will discuss the following aspects in the indi-
cated sequence: (1) Scale neutrality--whether the
technology is applicable to small farms; (2) employment--
whether the technology or technique is labor intensive;

(3) congruency—-whether growth and equity objectives could
both be achieved by the same technology/technique; and

(4) general conclusion.

Mechanical Technology-

 

This subsection investigates the equity and "con-
gruity" implications of tractor use, and the use of
mechanical means of exterminating weeds. The variable
X3 which measures the area chemically weeded is a proxy
variable for the spray equipment that is necessary for
chemical weeding. Power spray equipment would be
regarded as being more "mechanical" than hand spray

equipment.

Tractor plowipg.--The r coefficient of .75 in

 

Table 6.2 confirms a prior belief that tractor use is
positively correlated with farm size and that economies
of scale place this variable in the "low divisibility"
categOry. The reason why the coefficient is not even
higher is probably because of the countervailing
influence of institutional-organizational factors such as

custom hiring and ownership of tractors by multipurpose

156

cooperatives, which introduce scale neutrality into
tractor use. This is supported obliquely by the evidence
adduced earlier, that in the dry zone even farms below
5 acres in size use tractors.1

But economies of scale associated with the use
of tractor technology in Sri Lanka could be said to have
an adverse effect on income distribution. For one thing,
scale neutrality in tractOr use does not extend to tractor
ownership. Farm size denies to small farmers the financial
capacity to own tractors and the economic ability to use-
them profitably on their own fields. Therefore fiscal
policies designed to encourage tractorization by causing
tractor prices to private buyers to diverge from social
shadow prices, confer (regressive) social subsidies on
larger farmers, and are thus fundamentally antagonistic
to the objective of reduced income disparities. More-
over, while custom hiring facilitates scale neutrality in
tractor use, it does not ensure it. The actual benefits
to small farmers from such scale neutrality will be
directly proportional to the monopoly power exercised by
the local tractor owner, as reflected in hiring rates.
Such a monopoly element is very likely to exist in a
given locality in view of the relative numerical dif-

ferences between tractor owners (large farmers) and

 

lRaj, "Mechanization of Agriculture," Op. cit.

 

157

and tractor hirers (small farmers).1 Very small holdings
would be denied the use of tractors at all due to the
minimum physical size required for tractor Operation.
Tractor use is also biased against small farmers in view
of the limited sum (Rs. 60 per acre) loaned in cash for
plowing under the 1967 credit scheme which is much less
than the actual cost of hiring a tractor (which is about
Rs. 140 per acre). To the extent that the tractor owner
extends credit to small farmers, interest rates which
reflect the risks in lending could siphon away benefits
of new technology from low income groups, and thus be
inimical to the objective Of improved income distribu-
tion. A

At the same time there is much evidence that
tractor use has adverse implications for the second com—
ponent of income distribution; namely, employment. At
least three different studies reach the conclusion that.
tractor use in Sri Lanka displaces labor as well as
draught animals. The ILO Report states that tractors
caused the man-days required for plowing to be reduced
by five to seven days per acre.2 FAO evidence from

South Asian countries, including Sri Lanka, indicates

 

1For instance in the Pakistan Punjab about 80
percent of tractors were owned by 8 percent of farmers
having holdings of 25 acres and above (1, p. 49).

2ILO,'Report, Op. cit., p. 74.

 

 

158

that while the man-hours per acre needed for primary and
secondary tillage (including plowing, harrowing, puddling,
etc.) is 632 with nonmechanical methods, the man-days
needed become reduced to 43 per acre when mechanical
means are employed.1 That tractors displace labor in Sri
Lanka is further confirmed in a study which reports that
rice farms using modern technology in 1969-70 were found
to use 13 percent less labor when tractors rather than
bullocks were used.2
Tractors also adversely affect employment indi-
rectly in Sri Lanka since tractor purchase,~IOperation,
and maintenance involve the use of foreign exchange
resources that otherwise would have been used to import
raw materials that would have enabled a fuller utilization
of existing industrial plant capacity. The foreign
exchange allocation for tractors, tractor implements, and
spare parts is about 15 million per annum, whereas the
cut-back in the foreign exchange allocations for invest-
ment goods imports was around 23 million in 1971, according

to the Central Bank.3

 

1J. C. Moomaw and H. P. Curfs, Some General and
Special Aspects of Rice Soil Tillage, AGS:MRP/7l/ll-l,
(Netherlands: Food and Agriculture Organization, July,
1971).

 

 

2William H. Bartsch, Employment Effects of Alterna-
tive Technologies and Techniques in Agriculture (unpubliShed
draft, 1972; mimeographed).

3Central Bank of Ceylon, Annual Report 1971, Op.
cit., p. 235.

 

 

159

This labor displacement effect of tractors is in
conformity with effects Observed in India and Pakistan.

By synthesizing the information in the studies listed in
Appendix A, we find that the change in labor requirements
consequent upon the application of tractors could reduce
up to 90 percent the man-hours necessary. The man—hours
per acre are reduced from 1500 to 150 when tractors are
employed. The percentage rate of reduction of labor per
unit of output is roughly the same as in the case of labor
input per acre.

Overall a judgment can be reached that when shift—
ing from traditional methods, the labor-increasing
effects of biological technology tend to be more than
offset by the labor-displacing effects of mechanical tech—
nology (when such technology is used), unless tractor use
enables the breaking of timeliness bottlenecks and so
facilitates multiple cropping that otherwise would not be
possible.

However, Model C, presented earlier, which was a
regression of cropping index on tractor use and other
variables, showed that tractor plowing has a statistically
significant negative relationship with the cropping index
_in Sri Lanka. This justified a reveral of the prior
belief that tractor use increases the cropping intensity
in Sri Lanka. However, while tractors may not be gener-

ally effective in increasing crOpping intensity in Sri

160

Lanka, specific instances may well exist where the oppo-
site is in fact the case, particularly in the dry zone.
As a west Pakistan study indicates, the ability of trac-
tors to influence the cropping intensity depends on the
availability of (irrigated) water and the existence of a
cropping pattern in which the overlap of harvesting and
planting times can be eliminated by machines (1, pp. 64-66).
It is conceivable also that as more irrigation water
becomes available in Sri Lanka in the future, tractor
mechanization could have a positive effect on the inten-
sity of cropping, provided the cropping pattern permits
(as is likely to be the case). This points to a need for
ensuring technical flexibility in tractor use.

In conclusion, tractor use does not appear to
have in the past significantly contributed to either growth
or to equity. Tractors were found to be statistically
insignificant (at the 10 percent level) and economically
weak in explaining yields.l Tractors were also found to
be unimportant from the standpoint of timeliness. It
appears that the verdict with regard to tractor use is
also unfavorable from the viewpoint Of equity. Since
tractor use is positively and "significantly" correlated
with farm size, it is inherently biased against small

farmers. It thus has the capacity fOr widening income

 

1In fact, the results for Model B show a negative
regression coefficient (-.042) for the tractor variable
X - See the computer printout in Appendix C.

2.

 

161

disparities. Further, by displacing labor, tractors con-

tribute to further skewness in the income distribution.

Spray weeding equipment.--The degree of scale neu-
trality in chemical weeding in any given situation will be
influenced by whether power sprayers or hand sprayers are
used. Information is not available as to the extent to
which power spray equipment is used. Overall, a correla-
tion coefficient of .46 suggests that chemical weeding is
a fairly divisible input, although the degree of scale
neutrality appears to have been limited.

The statistical analysis showed that chemical
weeding was statistically insignificant for yields at a
10 percent significance level, althOugh the regression

coefficient was not negligible, and its positive charac-
ter fitted a prior belief that chemical weeding did in
fact have an effect on interdistrict productivity.1

No clear conclusion can be drawn from the informa—
tion available as to the significance of spray equipment
for paddy output. Further research needs to be undertaken
on the implication of spray weeding on small farms, to

further verify their effects on yields and the comparative

net rate of return from random planting and chemical weed-

ing vis-a-vis row planting and hand weeding.

 

 

1The t value for the "chemical weeded" variable

X was 0.87, while the critical point at a 10 percent
lavel of confidence was t = 1.36.

162

Other types of mechanization.--Mention may be
made briefly of other types of mechanization in Sri
Lanka. Dryers, seeding and rotary weeding machines, and
a few water pumps are used on less than 5 percent of paddy
acreage.1 Seeding and rotary machines will be discussed
in connection with the techniques with which they are asso~
ciated, and with regard to which quantitative information
is available. (No information is available regarding the

use of the other types of mechanical equipment.)

Conclusions regarding tractors and spray equipmentaw-
The evidence from Sri Lanka and other countries makes a
good case for a substantial reduction in the emphasis on
tractorization and perhaps the use of (power) spray
equipment. However, there are circumstances that may
justify their use in exceptionally in the present and
'more commonly in the future. There is, in other words, a
need for technical flexibility necessitated by ecological
differences within and between regions, and constantly
changing points of location of different paddy growing
areas along the development path. Hence policies with
regard to agricultural mechanization have to be designed

for Sri Lanka that will contribute to equity and increase

 

1Ministry of Agriculture and Lands, Implementa-
tion Programme, Op. cit.

 

 

 

163

output, and at the same time assure maximum technical
flexibility.1
Numerous variables need to be taken into con-
sideration in deciding whether tractors should or should
not be used. The situation is so complex and possibilities
so varied, that no blanket endorsement or prohibition of
tractorization is possible. There are benefits in particu~
lar circumstances in terms of paddy output (not necessarily
yields) and employment (from multiple crOpping and employ-
ment generation in secondary and tertiary sectors),
increased labor productivity, changes in cropping patterns
(toward more intensive crops), land reclamation, increase
in livestock and dairy produce (by reducing the need for
draught animals), and reduction in search and organiza-
tional costs of labor- There are also at the same time
costs that need to be considered; costs in terms of labor
displacement, tenant evictions, widening income dispari-
ties, use of possible crop lands for fodder cultivation,
and foreign exchange efflux (deprivation of industry of
raw materials),to name the major items. It should be
noted, however, that tenant evictions as a by-product of
tractorization may not be a significant problem in Sri
Lanka because tenancy of the traditional type predominates

in the wet zone, where-the small size of farm is not

 

1
chapter.

These policies are identified in the following

164

1 Tenant evic-

conducive to significant mechanization.
tions are an important result of tractorization in the
Pakistan Punjab, however. But as the study notes, any
increase in the productivity of land can engender evic-
tion (1, p. 126).

Tractorization also requires an ongoing release
of foreign exchange for the purchase of spare parts. In

a situation of acute foreign exchange scarcity like that

which presently exists in Sri Lanka, tractor owners face

 

considerable uncertainties concerning the ready availa-
bility of spare parts. There is good reason to believe
that draught animals are maintained in reserve to safe-
guard against machinery breakdown. In the Pakistan
Punjab, for instance, the difficulties in obtaining spare
parts caused tractorized farm to reduce (by about one-
fourth) but not eliminate the number of bullocks
maintained (1, p. 96). This consequently results in
duplication and wastes, besides denying to society the
benefits from the transfer of fodder lands for crops. In
the Pakistan Punjab the area under fodder was less on the
tractor as compared to the bullock farms, except for the
Lyallapur district, where it was equal on both (1, p. 97).
From a "micro" standpoint, project analysis for

ascertaining the social benefit-cost ratio is the obvious

 

1Only about 25% of acreage in the wet zone is
tractor plowed, according to Raj in Mechanization of
Agriculture, loc. cit.

 

 

165

solution. But at the broader national level the task of

identifying specific types of mechanization that can,

in a.given situation, increase the amount of labor needed

for rice cultivation, while at the same time increasing

output, is likely to require a cadre of experts of a size

that Sri Lanka or for that matter most developing countries E

simply cannot afford.

Biological Technology,

 

This subsection evaluates the biological variables

 

and associated techniques from the standpoint of income
distribution and equity. The pivot of the analysis is

the seed-fertilizer (S-F) package labeled X which was

5!
found to be statistically significant for yields at one
percent. The relationship between farm size and adoption

will be examined using information from Sri Lanka and

other regional studies. If a positive relationship seems

 

to exist (as a priori appears to be the case), the
remainder of the subsection will be devoted to identi-
fying the (mainly) institutional variables that contribute
to this positive relationship, and to assessing their
relevance and applicability to Sri Lanka.

Such an analysis is expected to pinpoint the
variables that will permit or actuate (l) more wide-
spread, and (2) Optimal acceptance of the S-F package in
Sri Lanka, where Optimality is regarded as containing the

components of (a) dosage and (b) techniques. Hence a

166

sub-optimal application of technology would exist if

(a) any component of the biological technology package
is applied in other than prescribed dosage, and/or

(b) the technique corresponding to the technology is

not (fully) applied. Hence extension potentials would
exist if either or both of these conditions are observed

to exist on farms.

 

Extent of scale.neutrality.--The physical divisi- ;

 

bility of the S-F package notwithstanding, a positive
correlation coefficient of .54 was observed between farm.
size and area under improved seed varieties (Table 6.2).1
This suggests that despite its physical divisibility H
factors exist which inhibit the acceptance of the S-F
package by the smaller farms at a given point in time.
The information on farmer utilization of production
credit in Sri Lanka also substantiates this finding. As
was shown in the previous chapter, productivity increase
have occurred despite increasing credit default .
which rendered defaulters ineligible for more credit. The
statistical results confirmed that noneligibility of

defaulters for further credit did not affect paddy yields.

The analysis in a subsequent subsection will show that

 

1Due to technical complementarity between improved
seed and fertilizers, the "seed" variable X was proxy
for the unincluded fertilizer variable in the statistical
analysis.

167

small farmers are bigger defaulters.l Hence the conclu-
sion emerges that much of the productivity increases have
come from large farms (unless one subscribes to the
unlikely View that small farmers who default have other
sources of credit for investment).

Even stronger support for the positive relationship f
between farm size and S-F adoption was seen in the statis-

tically significant positive relationship between farm

 

size and yields discussed earlier.2 :
Added credence to the positive coefficient observed E

between farm size and the seed-fertilizer variable is given

by the experiences of other regions. In west Bengal and

Andhra Pradesh a statistically significant positive rela-

tionship between adOption and farm size at the 5 percent

level was found to exist, indicating that adoption levels

increase with farm size. It was found, however, that

differences in levels of adoption between size of farm

groups were greater in bajra than in paddy-growing areas.

A one acre increase in farm size typically led to a 2-3

percent increase in the probability of adoption in rice

growing areas (6, p. 4). This is somewhat lower than that

suggested by the coefficient r with respect to Sri

58
Lanka. Survey results from the Pakistan Punjab show that

 

1See below under Production Credit.

2This refers to the statistical results in
Appendix C.

168

farmers with holdings above 12.5 acres were both "lead-
ers” and "maximum users" (2, p. 106). In 1970-71,
90.3 percent of this group planted Mexican wheat cover-
ing 92.4 percent of the wheat area, but only 73.1
percent of farmers in the below 12.5 acre size devoted
only 66.1 percent of wheat acreage to high-yielding
varieties. Similar relationships were Observed in ferti-
lizer and pesticide use. All these results were highly
significant statistically. In the Recife area in Brazil
it is reported that sampled small farmers typically prefer
to observe some local experience with new practices
before they try them (3, p. 162).
This positive relationship between farm size and
the level of adoption of the S-F package is primarily a
manifestation of the uneven flow of technology through
the farm size structure, and is indicative of small farmer
response at a later point in time. This phenomenon,
observed in Sri Lanka and elsewhere, exists despite the
physical divisibility of the seed-fertilizer package,
which permits its application on paddy holdings of any
size. Since the positive relationship cannot be attributed
to the technology per se, it must be attributed to the
other interacting primary variable; namely, institutions.
What institutional variables may contribute
to the positive relationship observed between adoption

levels and farm size? This subsection explores

169

this point using the concepts of felt costs and perceived
benefits. These concepts recognize that subsistence farmer
receptivity to technology is determined not just by
actual money returns and costs, but by the ratio Of
subjectively perceived benefits to subjectively felt
costs.l Institutions affecting the adoption of new
technology, if biased against small farmers, increase the
real costs of innovation to small farmers by increasing
either the money or psychological component in cost
(given perceived benefits), and thereby render the S-F
package less attractive to small farmers, its physical
divisibility notwithstanding.

Two such institutional factors that limit the
scale neutrality in the S-F package are discussed here.
These are (i) the technical information factor and
(ii)tflmzinput acquisition factor. Other possible inhibit—
ing variables (namely, the production credit, land tenure,
and risk factors will be discussed separately because
they influence the adoption and application of both

mechanical and biological technology by small farmers.

(i) Information: A synthesis of this aspect in

the studies listed in Appendix A indicates that

 

1This position is similar to Wharton's in that the
farmer's subjective evaluation is the critical factor deter-
mining farmer's receptivity to new technology. C. R.
Wharton, "Risk Uncertainty and the Subsistence Farmer,"
Development Digest, Vol. VII, NO. 2 (Washington, D.C.:
Agency for International Development, U.S. Department of
State, April, 1969), p. 7.

 

 

170

difficulties (costs) in gaining access to technical
information by small farmers has contributed substantially
to the fairly significant positive relationship observed
between farm size and seed adOption in Sri Lanka (Table
6.2). It appears that the positive relationship is sub-
stantially due to adoption lags by small farmers, which F1
in turn is due to the reduced risk associated with later

adoption. Reduced risk in turn mainly stems from

"mlnnwr - v '

improved information that reaches small farmers from the

 

demonstration effects of early adOptors, the develOpment I
of informal localite channels, and formal extension con—
tact. Time (lags) therefore is primarily a proxy variable
for information costs. If so, institutions which have the
capacity to reduce the costs of technical information to
small farmers (mainly extension) would need to be empha-
sized in the policy prescriptions to be made subsequently.
AdOption lags may be regarded as a means for
reducing information costs through the internalization of
externalities. Later adOptors, it is suggested, are those
farmers who could not previously afford the'cost of
acquiring enough technical information to enable them to
reduce risks of adoption to acceptable levels. Adoption
lags permit the experience of others to confer an external

benefit on laggards in the form of reduced risk. This

externality, when internalized into the production systems

171

of laggards, assumes the form of a cost subsidy that

Operates to make adoption an attractive proposition.l
The reported experiences of small farmers in the

Pakistan Punjab indicate that Observed lags were due to

felt risks associated with the adOption of new technology

(2, p. 87). In West Bengal and Andhra Pradesh small F“

farmers were later adOptors (6, p. 5). When these

farmers eventually adopted after an "experimental period"

of three seasons, they adopted to a greater extent than

 

large farmers; while large farmers adopted on about 20- fig
30 percent of their acreage, small farmers adopted on

35-80 percent. This phenomenon was attributed to the

need for small farmers to spread the overhead costs of

technological change over a larger number of units.

Figure 4 illustrates that the pattern of diffusion

of technical change follows an elongated "S"-shaped curve

 

1This suggests that the adoption process may be
inherently sub—Optimal. If the added risks assumed by
early adOptors remain inadequately compensated by the
change in output alone, it is reasonable to presume that
less people adopt early than otherwise would be the case.
In a free market situation early adOptors receive rough
justice in the form of higher initial prices, while later
adOptors receive lower prices due to increased supply. If
the change in output alone is inadequate compensation for
early adOptors, there would exist a need to further
compensate farmers who achieve a large change in output
(in a given locality and in a given season) in some
psychic manner by social recognition. The importance of
this factor was demonstrated by the success of the model
farmer program'in Comilla (7, pp. 35, 47, and 51).

 

1372!

Figure ‘3. Effect of Inter-stint end Tenency on the Diffusion Process

 

 

 

 

AdOption ’
Level Line of Complete Adoption
(yield)
l
I
I
l
:1 tk use
(Adoption Legs)
Figure 5, Effect of Crop Sharing on Input Usage (Adoption)
Output

HFC

MVP:

 

7

b----

 

0 :1 x2 Inputs

3‘1

 

IF '2...

173

OBCD,l where the path traced out is an intersection of

2 The

short-run supply and demand curves for new inputs.
"S" shape results from interaction effects "which is the
process through which individuals in a social system who
have adOpted an innovation influence over those who have

not yet adOpted."3

As the ”S" shape suggests, the adop- Ema
tion level picks up after a lag of 0ti time due to
interaction effects (externalities), but tends to flatten

out as the optimal level is approached. II.

 

In adopting the S-F package farmers need not only a,
the physical input but also the technical information
that goes with it, indicating its improved benefits
and, more importantly, describing the method of its appli-
cation. The role of such informatiOn is to win farmer
acceptance through risk reduction. Risk reduction is
important for farmers near the margin of subsistence who
place a high premium on security because of the high

penalty associated with error. There are two types of

 

1This pattern is the same as that observed for
hybrid corn in the U.S. See zVi Griliches, "Hybrid Corn
and the Economics of Innovation," Science, VOl. 132,
NO. 3422 (July 29, 1960), pp. 275-80.

2The demand curve would be the dominant variable
if it is possible to assume that the supply curve is rela—
tively more elastic, as is likely to be the case in Sri
Lanka. '

3Eugene A. Havens and Everett M. Rogers, "Adoption
of Hybrid Corn: Profitability and Interaction Effect,"
Rural Sociology, Vol. 26, No. 4 (December, 1961), p. 411.
Such a diffusion pattern was observed in the sampled areas
of Hazara, Pakistan (4, p. 65).

 

174

risks in new varieties that have to be borne in addition
to risks from weather that are always present. One
category of risks beyond the influence of farmer informa-
tion—gathering are those which stem from the proneness of
new seed to disease and its increased sensitivity to

weather conditions. These risks must be reduced through F**

“I ._ -

biological research. The second category of risks are

I» is.“ ‘1. 13.:

those that cause farmers to adopt later rather than sooner.

Much of this type of risk can be reduced if farmers had

 

appropriate and adequate knowledge.

It is very important for farmers in Sri Lanka to
know which of 22 different improved varieties are most
suited to their particular conditions. For instance, some
varieties like H4 are robust and hardy and require much.
less water control and ancillary inputs than the higher
yielding but riskier varieties like BG 11. Farmers also
need information on how to deal with pests, recognize and
treat disease, effect water control, use the correct

techniques like row transplanting, inter alia. Hence

 

since new technology increases farmer options, information
enables the correct (risk-reducing) choices to be made,
and significantly influences farmer willingness to adopt.
The foregoing would apply strictly to potential

(later) adOptors: those farmers who in the absence of
information cost reduction in the present will bideh.

their time until the results of adoption by others reduce

 

175

risks sufficiently to make them willing to innovate.
It does not apply to those who simply will not adopt in
the foreseeable future.

Several factors seem to inhibit small farmer
access to technical information, thereby in effect
increasing for the farmer the (average) cost of informa-
tion per unit of results (adoption).

1. Time costs needed to acquire new inputs and
learn how to use them are indivisible. Regardless of
whether a farm is large or small, approximately the same
amount of time and effort would need to be expended in
learning how to apply new inputs to the production
process. A small farmer may not consider it worth his
while to expend time and exert himself mentally to gain
the necessary information for himself in view of the
limited impact in absolute terms that increased output
will have on his income. Time has an opportunity cost
at least in terms of leisure foregone.

2.. If extension incentives are tied to produc-

tivity of acres rather than persons, there would be'

extension economies in dealing with larger farmers through

whom a greater number of acres could be influenced at
almost the same cost (extension effort). Extension con—
tact is one of the more important variables affecting
adoption behavior, the other being contact with informal

agricultural leaders, although it is not known whether

 

 

176

informal channels are a less preferred substitute for
formal channels.1 If extension efforts are relatively
biased against small farmers, the cost of acquiring
information per unit of results (adoption) will be
higher for this category.

3. It is also likely that farmer receptivity Emh
to information is positively correlated with farm size,
as progressivity in outlook and literacy probably depends

on income levels, which in turn are a function of farm F

 

size. In the Recife area study, statistically signifi-
cant positive relationships were Observed between farm
size, education levels, and adoption of new technology
(3, p. 171). In this sense, given education levels,
physical farm size is a partial proxy variable for the
farmer's cost of information per unit of adoption.
Moreover, smaller farmers who tend to be more
tradition bound (partly due to the positive relationship
between progressivity in outlook and income) would also
tend to be more influenced by the inhibiting effects of
"econological" costsz-—i.e., the economic costs associ-
ated with cultural, religious and Social factors referred

to in Chapter V. For instance, in the Recife area social

 

1This aspect was examined in some detail in the
previous chapter (V).

2Term from Thomas R. McHale's "Ecological Analysis
and Differential Economic Growth Rate," Human Organization,
Vol. 21, No. 1 (Spring, 1962).

177

pressures were found to be statistically significant
in inhibiting individual innovativeness (3, p. 302).
In intuitively assessing the profitability of innova-
tion, the fears of social, cultural, and religious
sanctions may be expected to enter the calculus on the
side of felt cost. These "econological" factors would E
make the benefits and methods of new technology seem

less attractive to small farmers.1

———- .13va '- ‘n ’ \

4. Control over some cooperatives by large

 

1w

farmers in Sri Lanka (referred to in Chapter V)
probably contributes to the cost of information and
input acquisition for small farmers. Small farmers
may be relucant to patronize COOperatives to seek
information from elite (non-peer) groups to whom they
have been traditionally subservient.

Indirect proof that improved channels of informa-
tion have the capacity for accelerating the adoption
process (and thus reducing the positive relationship

between adoption and farm size) is found in the evident

 

1This notion finds support in Boulding's concept
of an Image, in terms of which man's behavior is deter-
mined by his subjective perceptions Of the world. The
impact of a message on the individual's Image depends
on: (a) the strength of the message, and (b) the con-
truence of the message with the existing Image.
Kenneth E. Boulding, The Image (Ann Arbor, Michigan:
Ann Arbor Paperback, 1956), pp. 1-49.

 

178

influence of (a) the perceptibility of benefits and
(b) tenancy,<x1the adoption process.

There is evidence that when the benefits of a
given technology are more easily perceived, the technology
tends to be more readily accepted (provided adequate
,information at low enough cost is available about methods
and techniques). For instance, the basic technology
introduced in Comilla (tubewells and water pumps) were

of such an obvious nature that even tradition-bound

‘3.

 

‘l. :...'I‘

farmers needed to incur very low costs to gain information
about its advantages.1 The importance of low cost percepti-
bility of benefits for farmer receptivity to innovation is
also apparent in the response Of smallholders to high
yielding wheat in west Pakistan (5, p. 72). It is claimed
in this study that the physical differentiability between
old and new varieties of wheat probably contributed to
the more rapid adoption of wheat rather than maize, the
new varieties of the latter not being significantly
different physically from the old varieties. Rochin
states as follows with regard to the perceptibility
factor:

. . . perhaps theénost important characteristic

(of dwarf wheats) 18 the ab111ty to see the dif-

ference between Old and new innovations. Dwarf

wheats are short-statured, with high tillering,

and dark green compared to desi varieties; the
physical contrasts between dwarf and desi wheats

 

1This is an inference from (7, pp. 12-13).

2The words in parentheses are the writer's.

179

are very noticeable. During the second round of

interviews, farmers were interviewed right next

to demonstration plots with new, high—yielding

varieties of maize and asked if they had ever

heard Of or seen any improved varieties of maize

in the area. Surprisingly, many said no, even

though a field of maize, which gave thirty per-

cent more yield, was sitting in front of the

respondents. What this illustrates is that new

varieties need something to attract attention,

whether a visible symbol or a visible character-

istic such as dwarfness (5, pp. 72-73).

Second, the impact of information on adoption at

a given point in time is probably evident in the produc-
tivity data of tenant farms. The effect of tenancy on
adOption is examined in some detail in the following sub-
section. It might be mentioned here, however, that there
is evidence that landlords, including those in Sri Lanka,
perform an information providing extension function which
causes the adoption levels of crop-sharing tenants to
be higher than that of their owner—Operator counterparts.
at a relatively early stage of the development process.
This phenomenon is also observed in the adoption levels
of tenant farmers in other counties. However, it is not

clear as to what degree of landlord sanction accompanies

the information providing process.

(11) Input Costs: It was stated in Chapter V that
seed and fertilizer inputs were distributed through the
quasi-government agency of multipurpose cooperatives with
sole distribution rights, and that seed received a 10

percent subsidy while fertilizer received a 50 percent

 

180

subsidy. It was also stated that since subsidized ferti-
lizer was only to be used for specific crOps (including
paddy), (a) the distributing agency was called upon to
satisfy complex procedures to receive the fertilizer (thus
causing timeliness problems in fertilizer supply), and
(b) the agency in turn imposed similar complex procedures E
on farmers seeking to purchase fertilizer.

While the former defect affects all farmers

served by the distribution agency (multipurpose cooperative)

 

regardless of size, the latter affects small farmers more é

v.6.
v‘ v I.‘

than large farmers, and therefore has probably contributed
to the positive relationship observed between farm size
and adoption. Since small farmers purchase smaller quanti-
ties than large farmers, but all farmers irrespective of
size have to satisfy equally complex bureaucratic require—
ments, the acquisition costs per pound of fertilizer is
higher for small farmers. Also, since small farmers tend
to be less literate than large farmers, the filling of
forms, generally knowing what procedures to satisfy and
how to best satisfy them, imposes higher costs (greater
difficulties) on small farmers than on large farmers.
Since average fertilizer acquisition costs are
higher for small farmers than for large farmers in Sri
Lanka, large farmers in effect receive a larger subsidy
than small farmers per unit Of fertilizer applied. This

may have caused lower adOption levels, and therefore

181

probably contributes totfluapositive correlation between
farm size and adoption.

Besides the acquisition costs, the subsidy itself
is antagonistic to the objective of reducing income dis-
parities. Since large farmers have higher adoption
levels, the subsidy benefits large farmers more than small 51
farmers on a per acre basis. On a total acreage basis the

transfer from taxpayers to large farmers must be consider-

able.

 

There are alternative institutional modifications g
that could be effected that would eliminate or at least
considerably reduce these differential costs. These will
be examined in the following chapter. 9
Other Institutional and Economic
Variables That Affect the Use Of

Both Mechanical and Biological
Innovations

 

 

 

 

This subsection examines other institutional and
economic variables that influence small farmer acceptance_
of both mechanical and biological technology. As has been
observed, a positive correlation exists between the
average size Of farm and tractor use (.74), chemical
weeding (.46), and new seed varieties (.54).

While the adOption of mechanical technology by
small farmers is inhibited because of limited physical
divisibility, the adOption of biological technology is

inhibited in spite of physical divisibility. Nevertheless,

182

since both mechanical and biological technology can only
influence growth and equity through their interaction
with institutional variables, the institutional variables
determine small farmer receptivity to both types of
technology.

The relevant economic and institutional variables

are to be examined from the standpoint of their applica-

.. _.- ....- I.'.- II‘IK'E!“

bility to small farms (i.e., their scale [size] neutrality)

and, where relevant, their employment generation poten-

 

tial. These variables are: E-
1. The production credit factor
2. The tenancy factor

3. The risk factor

Credit.-—This variable (X12) has been described
in some detail in Chapter V. As stated, the credit
scheme of 1967 provided large loans for the adoption of
new varieties (Rs. 370 as Opposed to Rs. 262 for tradi-
tional varieties), and the amount of credit to be received
by each farmer was to be worked out through the collabora-
tion of extension workers with c00perative societies.
Seventy percent of the loan (for the purchase of seed,
fertilizer, and pesticides) was in kind, while 30 percent
(for the hiring of labor for plowing, planting, and
handweeding) was paid in cash. It was also shown that
paddy productivity and credit default have moved in the

same direction, despite the ineligibility Of defaulters

183

to receive further credit. The statistical analysis also
showed that the default rate in the previous year had a
statistically insignificant but nevertheless appreciable
positive relationship with interdistrict productivity.
These factors justified a reversal Of a prior belief and
led to the conclusion that the ineligibility to receive F7
credit by high default districts has not appreciably
affected interdistrict paddy yields.

The concern here is to examine how the credit

 

scheme could be made to increase the productivity of E‘
small farmers. Production credit furnishes,low income
farmers with the capital capacity to purchase Off-farm
inputs which a low savings rate and high money lending

rate would otherwise preclude. The savings rate is

lower for low income farmers than high income farmers,
although it is possible that the savings rate may be

lower than necessary among small farmers due to a lack of
perceived Opportunities for investment. (To the extent
that this is so, it further underlines the importance of
the information factor as a means of increasing levels of
application of new technology.) In West Bengal and

Andhra Pradesh, a statistically significant negative
relationship was perceived between farm size and the use

Of cooperative credit (6, p. 31). The study concludes that
large farmers meet a large part of their capital expendi-

ture from their own funds.

184

The economic Objective of increasing the produc-
tivity of low income farmers tends to be antagonistic to
the financial principles of lending. Biases in credit
disbursement against small farmers are partly attributable
to scale economies in terms of administrative costs, the
reduced risk in the granting of loans to large farmers, a5
and to the distribution Of political power. .Biases in
credit disbursement hence partly depend on the endogenous
factor of the real administrative incentives in lending--

whether the rewards are in terms of money loaned, number E

 

of loans amounts recovered or changes in productivity.
Ultimately it depends on the political intent of who
should gain from credit disbursement. If such biases
exist in credit disbursement, the resultant increase in
the cost of credit to less favored groups (small farmers)
must influence the relationship between adoption and
farm size, and small farmer demand for mechanical tech-
nology.

There is good reason to believe that established
procedures for credit disbursement do not generally
inhibit small farmer access to biological technology in
Sri Lanka.1 However, large farmer control over some

cooperatives (referred to in Chapter V) makes it difficult

 

1S. N. Samuel, "An Analysis of Institutional Modi-
fications for Effective Farmer Credit in Ceylon"
(unpublished Master's thesis, Department of Agricultural
Economics, Michigan State University, 1971).

185

to rule out the possibility in all cases in view of the
loan disbursement procedure which requires that extension
workers collaborate with cooperatives to determine loan
quantums.

‘ The credit scheme inhibits small farmer access to
mechanical technology to a greater extent than it inhibits
small farmer access to biological technology. As Table
5.5 indicates, the loan quantum allotted for plowing,

weeding, and the use of pesticides are far less than that

 

‘3‘ ear—.1: '. -
1

required for the use of mechanical technology for these

purposes. For instance, the loan granted for plowing is

Rs. 60 per acre (Table 5.3), although it costs Rs. 140

per acre to hire a tractor. If there are no facilities

for hiring spray equipment (as is probably the case),

the sum of Rs. 25 per season for weeding would not be

adequate for the purchase of spray equipment and weedicide.
AlthOugh credit factors may not have influenced

small farmer access to biological technology in the past,

it is likely to do so in the future because of a high rate

of default reported among small farmers, which renders them

ineligible for further loans. The reasons for default are

identified in Table 6.3. It appears that since most small

farmers are also low income farmers, the small farmers

are probably bigger defaulters. A negative correlation

coefficient of -.55 (Table 6.2) between farm size and credit

default, and the information in Table 6.4 confirm this.

186

TABLE 6.3.--Reasons for Credit Default.

 

 

 

Cause % of Defaults
Low income 49 -~«
weakness in lending agency 17
Borrower irresponsibility 18
Other 7 16 -
I

Source:

 

Central Bank of Ceylon, Survey of Defaults in
the Repayment of New Agricultural Loans (Colombo:
Economic Research Department, Central Bank,
1972), pp. 28-30.

 

 

TABLE 6.4.--Amount of Loan Returned Classified by Size of

Area Cultivated.

 

Amount Repaid

Size of Area (with Interest)

 

cultivated as a Percentage
(Acres) of Amount Borrowed
1 - 2 39.0
Above 2 - 3 41.8
Above 3 - 4 49.4
Above 4 57.4

 

Source:

Central Bank of Ceylon, Survey of Defaults in
the Repayment of New Agricultural Loans (Colombo:
Economic Research Department, Central Bank,
1972), Table A 204.

 

 

187

The recent Central Bank Report states that "the bulk of
these defaulters are generally small farmers to whom agri-
cultural credit is another form of static credit as it
does not help them very much to raise their incomes. . . ."
Moreover, the survey found that 71 percent Of defaulting
units received less than Rs. 250 per mensem and received 57
37 percent of total income.1 Since most defaulters are E
identified to be the smaller farmers, this high default 1

rate in the past can be expected to contribute to a

 

positive coefficient between farm size and adOption in k‘
the future by disqualifying past defaulters,(mainly small
farmers) from further access to government credit.
A further concusion that can be drawn is that the
mere disbursement of credit does not ensure greater pro-
ductivity and incomes. Since 49 percent Of the loans were
defaulted for reasons of low income, it appears that at
least half the funds disbursed have not been put to their

intended purpose Of increasing productivity sufficiently

to enable the loan repayments with interest to be met out

of the resultant change in income. The Central Bank Sur—
vey reveals that the average yield per defaulting cultivator
was 35 percent less than the national averge. The average
yield per defaulting cultivator was 37 bushels as against

a national average of over 50 bushels. Thus credit

 

1Central Bank of CEylon, Survey of Defaults,
Op. cit., p. 37.

 

188

0

disbursement in an essentially static income situation
appears to have contributed substantially to the high
default rate Observed. It follows that the mere disburse—
ment of credit to small farmers without regard to their
ability or willingness to innovate is not likely to
contribute to a reduction in the positive relationship F7
observed between farm size and adoption. This conclusion
1

is consistent with the findings of an earlier study.

Conversely, there is evidence that special credit

 

“:31. e». -" .

facilities are not a prerequisite to innovation if the
technology is sufficiently profitable and fits the
existing agricultural system. Farmers can find funds
when significant net benefits are perceived, and income
streams become identified (5). Hence the whole question
of whether production credit disbursement will succeed
in increasing the productivity Of small farmers traces
all the way back to the basic question of whether small
farmers perceive enough net benefits to wish to innovate
in the first place.2

It may generally be concluded that Credit costs

will contribute to a positive relationship between farm

 

l . . . . .
S. N. Samuel, "Inst1tutional Mod1f1cations,"

Op. cit.

2This conclusion is consistent with Mosher's
identification Of production credit as an accelerator
rather than as an initiator of development. Arthur T.
Mosher, Getting Agriculture Moving (New York: Fredrick
A. Praeger, 1966).

 

189

size and adoption only when small farmers who are already
willing to innovate are unable to do so because of the high
cost of credit to them. Although a lack of credit facili-
ties does not appear to have contributed significantly to

the positive relationship that exists between farm size and

‘1

adoption Of biological technology in Sri Lanka, this may
not continue to be the case. If due to previous default,

small farmers who are desirous of innovation find themselves

 

ineligible to receive credit, the credit factor could be p
expected to enter into the positive relationship between gv
farm size and adoption in the future. With regard to

mechanical technology, the position is somewhat different.

The credit factor could be said to have already contributed

to the positive relationship between farm size and the

application of mechanical technology, because the amounts

loaned in cash for the operations of plowing, weeding,

and pesticide use fall far short of the actual amounts

required for the hiring or purchase of mechanical equipment

for the performance of these Operations.

Tenancy.--As stated in Chapter V, about 40 percent
of paddy holdings are rented, Of which 70 percent are on
a crop sharing (andé) basis. Hence about 26 percent of
paddy holdings are under crop—sharing tenancy arrange-
ments. Since these traditional arrangements are largely

confined to the wet zone, it may be said to apply mainly

190

to small farms. A coefficient of correlation of -.66

between average farm size and land rented confirms this.1
Theoretical presumptions exist that tenants are

less likely to adopt than their owner-operator counter-

parts (i.e., those in the same size group) because:

(a) the lack Of security reduces incentives for farm u

investments of a fixed nature (such as drainage and con-

struction of bunds, fences to keep out stray animals,

etc.); (b) tenants often have to bear the costs of inno-

 

E‘J" ..'. .‘.-. _
1.:

vation but are required to share the benefits, thus
reducing the attractiveness (profitability) of innova-
tion. As Figure 5 illustrates, the marginal value
product facing a share cropper is less than that facing
an owner-Operator, causing maximum profitability (where
MVP = MFC) to be reached at a lower level of input
application (adOption).

However, the evidence from Sri Lanka and other
countries shows that tenants have a higher level of pro-
ductivity per acre than their owner-Operator counterparts.
The statistical analysis showed that the renting of land
has an appreciable though not statistically significant

effect on yields.2 Since most tenancy occurs in the

 

lVidé r in the correlation matrix in Appendix E.

78

2The t-value of this X variable was 1.28 whereas
the critical point at 10 percent was 1.36. However, the
regression coefficient for the tenancy unuuirented) vari-
able X was statistically significant at 10 percent in
Model 8. See the results in the computer printout in
Appendix C.

191

traditional areas of the wet zone where the farms are
small, tenancy could be said to have its effect on smaller
farms, as suggested by a correlation coefficient of -0.66
between farm size and land rented (Table 6.2).

The experiences of other regions confirms the rela-
tionship. In West Pakistan a land owner is reported to us
have been instrumental in introducing high-yielding vari- '
eties to traditional farmers (5, p. 62). This study also

quotes the results of a field study in the Philippines,1

 

according to which the pattern of spread of high-yielding

'13! I.

varieties in Gapan owed much to the fact that three major
landowners had sufficient political or economic power to
secure IR-8 seed and cause it to be planted by selected
tenants. The study of the Recife area in Northeast
Brazil also found that farmers who worked rented holdings
were more innovative than their owner-Operator counter-
parts (3, pp. 214-16).2 Peacock reports:
. . . the percentage of tenants using tractors and
insecticides seems to more closely resemble his
landlord rather than his owner-Operator counter-
parts. . . . If any generalization may be made, it

appears that landlords and tenants are more inno-
vative than owner-Operators.

 

lRochin quotes R. E. Huke and J. Duncan, "Spatial
Aspects of HYV Diffusion," Philippines, International
Rice Research Institute, 1970. (Mimeographed.)

2In Brazil the institutional relationships between
landlords and tenants may be different from those preva-
lent in South Asia in terms of degree of coercion, etc.

192

It appears, therefore, that tenancy has the effect
of reducing information costs to tenant farmers through the
practice of input provision by landlord. When landlords
provide inputs like new seed, fertilizer, weedicides, and

pesticides, they introduce tenant farmers to new tech—

 

nology and thereby perform an information cost-reducing m:
extension function. This hastens the adoption process for

tenant farmers, and probably accounts for the appreciable

size of the regression coefficient that exists for the

tenancy variable. ,,. é?

But the yield-increasing influence of tenancy seems
to be an essentially short-run phenomenon. Although the
(information) cost-reducing effect of the institution Of
input provision by landlords may initially result in
adOption levels for tenants that are higher than for owner-
Operators, over the longer run the concomitant crop-Sharing
arrangement could be expected to cause maximum profitabil-
ity to be reached at a lower level of output than for
owner-Operators. These relationships are illustrated in
Figure 4. As indicated, the diffusion curve for owner
Operators is OBCD, and for tenant Operators, OECF. At
time Oti tenant farmers have a relatively higher level
of adoption (OM2 > 0M1), which is however overtaken at

the point in time t due to the inherently yield-limiting

k
effect of crop sharing. Sri Lanka has apparently yet to

reach such a point in time.

193

Figure 5 shows Why the yield-increasing effect Of
input provision is Offsettnrthe yield-limiting effect of
crop sharing in time tk. For the owner-operator the mar-
ginal value product curve MVP makes it worthwhile to

0

apply 0x2 units of input, given marginal factor costs MFC.

rm
However, the marginal value product curve MVPt facing the

crop-sharing tenant does not make it worthwhile for him

to increase the number of units of inputs applied to

greater than x1. Lg

 

Moreover, higher-levels of adoption (and yields)
by tenants do not mean that tenants are necessarily better
Off than their owner-operator counterparts; the landlord
may not permit a sufficient proportion of the benefits of
innovation to accrue to tenants.

The importance of this "tenancy" variable
therefore lies not in its justification of tenancy but
in its ability to reinforce the case already made in
support of the crucial role that information can play in
reducing adOption lags and in accelerating the adoption

process.

Risk.--The statistical analysis revealed that
the "risk" variable X9 was highly significant (at the
1 percent level of confidence) in its effect on inter-

district paddy yields in Sri Lanka. This variable takes

the lO-year average difference between the area sown and

 

194

the area harvested. This and other aspects have already
been examined in fair detail in Chapter V.

The Objective here is to examine how the risk
factor influences equity--the receptivity to new technology
of small farmers with holdings below 4 acres in size. The
risk variable is potentially if not actually congruent.

It is significant for productivity, and its reduction
(measured here as an increase in the area harvested as a
percentage of area sown) will benefit small farmers more
than large farmers (since small farmers who are nearer the

k ‘ 0*, #2?

margin of subsistence have to pay a higher price for

 

error, they tend to be bigger risk averters). Also, there
is evidence that in Sri Lanka crop failure (risks)
affects smaller farms more than larger farms. The Central
Bank reports that, "There was a steep reduction in the
percentage area Of crop failure to cultivated area as the
size of area cultivated increased."1
The following analysis attempts to examine two Of
the components in "risk" which cause crOps on small farms
to be more prone to crop failure. These factors are
(a) the pest control factor and (b) the water control
factor.

Although on an overall basis only 10 percent of

risk from crop failure can be attributed to inadequate

 

1Central Bank Of Ceylon, Survey of Defaults,
Op. cit., p. 30.

 

195

pest control (Table 5.2), the data indicate that a reduc—
tion in cultivation risks through better pest control will
benefit small farmers more than large farmers. As Table
5.2 evidences, crop failure attributed to pests is more
than twice as high in the wet zone (18 percent) where the
average size of farm is less than 1 acre, than in the dry
zone (7 percent) where the average size of farm is over
2.5 acres. Information from the recent Central Bank
Survey is in conformity with the View that the pest factor
affects small farmers more than large Farmers. The report
states:1
Lower income groups are adversely affected by

their inability to purchase these pest killers in

adequate quantities . . . they (also) considered

it less important to use such weedicides and pesti-

cides than the large cultivators and high income

earners who had better knowledge of cultivation.
Apart from a lack of cash capital resources and technical
knowledge, there are other possible reasons for the
inadequacy of pesticide use by small farmers. One such
reason is the probable inadequacy of credit (Rs. 25)
allotted for this purpose. Yet another reason can be
traced to externalities associated with pest control
efforts. Due to the large number of small holdings within
a given area, the efforts Of one farmer to control pests

is likely to involve higher costs than if all farmers

made concurrent efforts. Conversely, a farmer using

k

1

author's.

Ibid., p. 32. The word in parentheses is the

 

196

pesticide on his farm may confer uncompensated (external)
benefits on contiguous farms. In either case, the
externalities factor causing a divergence between private
and "social" benefits/costs must, on the assumption of

economic rationality, result in an under—application of

 

pesticides. In short, it appears that pest control has ““
considerable "public good" characteristics, and should 3
therefore be treated as such. This matter will be dealt ;
with when policy alternatives are proposed for improving :
the productivity of small farms in the following chapter. 1?

Floods and drought account for as much as 70
percent of crop failure (Table 5.2). Also, since they
influence both zones equally, and there are twice as many
farmers in a.given area in the wet zone, more small
farmers than large farmers are affected. Although weather
is an exogenous factor, improvements in infrastructure could
minimize its adverse effects. For instance, water conser-
vation in tanks (an ancient means of ensuring an even
supply of water over the spectrum Of one year), irrigation,
and improved drainage can reduce the crop failure result—
ing from floods and drought. A positive correlation coef-
ficient of .39 (Table 6.2) between farm size and irrigated
acreage suggests that large farmers have some advantage

in access to irrigation water.

197

In the wet zone, where smaller farmers predominate,
"unsuitability of land for paddy cultivation (arises) from
excessive water logging and drainage facilities."1 This
points to a need for the emphasis on minor irrigation works
in traditional areas rather than the major works in new
areas. The latter have tended to be favored in the past.
As the ILO Report points out,2 a long series of reports has
questioned the economic viability of major colonization.

Minor works in traditional areas are congruent in that

 

besides favorably influencing productivity and cropping E;
intensity, such works would also benefit more households

for a given acreage, and tend to be more employment

generating. Moreover, the money costs of minor works are

3 esti-,

considerably lower than for major works. The ILO
mates that 1 acre of land is brought under irrigation under
minor works at Rs. 300 per acre, whereas major works

(including colonization) cost about ten times as much;

namely, Rs. 3,000 per acre.

The Significance Of Techniques

 

At the beginning of this subsection it was pro—

posed that two conditions need tO be fulfilled if the

 

1Central Bank of Ceylon, Survey on Default, Op. cit.
The word in parentheses is the writer's.

2

 

ILO, Report, Op. cit., p. 90.

3International Labor Organization, "Matching
Employment Opportunities and Expectations," Technical
Papers (Geneva: ILO, 1971), pp. 109-110.

 

 

198

biological input package is to be regarded as being
optimally applied: (1) The seed, fertilizers, pesticides,
water, and other components of the input package should be
applied in prescribed dosage, and (2) the apprOpriate
techniques corresponding to the technological package
should be fully applied. For instance, seed should be
planted in a manner (in rows) that will permit maximum
fertilization and weed control. If not, a sub-optimal

condition may be said to exist.

Congruence in techniques.--The evidence indicates

 

that techniques important for Optimality are highly con-
gruent. The statistical analysis indicated that row
planting was a technique that was statistically signifi-
cant (at the 10 percent level of confidence) in explaining
interdistrict paddy yields. It was also previously stated
that the variable for row transplanting (X4) represented
the concomitant techniques of hand weeding and rotary
weeding.l

Not only is row planting (and hand weeding) signifi-
cant for yields, it (they) is (are) also more suited to
small farms, and is (are) highly labor intensive. The
existence of "divisibility" in row planting is suggested

by the negative coefficient -.14 (Table 6.2) for the linear

 

1To have included additional variables for hand
weeding and rotary weeding in the econometric model would
have resulted in less reliable coefficients due to multi-
collinearity.

 

 

 

199

relationship between farm size and area row planted. This
is in conformity with a prior belief that row transplanting
and hand weeding are Operations that are more suited to
small farms.l

Besides being "divisible," row planting and hand
weeding are also highly employment generating. This
employment potential exists in the performance of the
Operations described below.

(a) To replace broadcasting Of seed with row
transplanting increases the demand of labor for the plant-
ing operation itself. For instance 1J1 Minipé (Sri Lanka),
the extra man-days required for transplanting the IR-8
rice variety is 17 to 25 days if complementary practices
are adopted, and an extra 15 to 18 man-days otherwise.2

(b) Row (trans)planting facilitates access to
weeds with a minimum Of soil and plant disturbance,
and hence enables hand weeding to take place. The average
increase in the demand for labor for hand weeding is
reported to be as high as 20 man—days per acre for all
varieties together.3

(c) TO the extent that row transplanting increases

yields (it has been found to be statistically significant

 

1The ILO Technical Papers make a cursory reference
to this. ILO, Technical Papers, op. cit., p. 101.

2

 

 

 

Ibid., p. 202.

31bid.

 

 

200

in this connection), it also increases the man-days needed
for harvesting and post-harvest Operations. Transplanting
reportedly increases yields from 15 to 20 percent, and
this increase requires an extra 20 to 25 man—days per acre
for harvesting.1 This information is supported by the

Central Bank's findings in 1969,2 according to which the

 

'?“3
average yield for transplanting was 65 percent higher than i 5
for broadcasting. The national average man-days required 3
for transplanting was 79.6 per acre, but only 34.4 for i
broadcasting.3 ;; 7

It may be concluded generally that a considerable
employment potential exists in this area. The dimensions
Of this potential can be gauged from the fact that only
about 20 percent of paddy land in Sri Lanka is hand weeded,
while only 55 percent of all paddy land is subject to
apy sort of weeding.4

The employment benefits from row (trans)planting
can be partially negated, however, if the rotary weeder

is used. The use of this machine reduces the employment

 

llbid.

2Central Bank of Ceylon, Cost of Production of
Paddy, Op. cit., pp. 21—27.

3It needs to be emphasized, however, that each
of these Operations will be adopted only if the added
perceived returns justify the added felt costs. NO informa—
tion is available with regard to this crucial aspect.

 

 

4Ministry of Agriculture and Lands, Implementation
Programme, loc. cit.

 

 

201

created from hand weeding by almost 5 man-days per acre
and even more (25 man-days) if a rotary weeder is used

in conjunction with a seeding machine.

Farmer recgptivity to techniques.--It seems

important to distinguish between technology and technique

because farmer response differs between the two; farmers #7
tend to be relatively more receptive to technology than to 3
techniques. In Sri Lanka only 3.5 percent of paddy E
acreage is subject to row transplanting despite its sta- iw
.3“

 

tistical significance for yields, while only 20 percent
of the acreage is subject to hand weeding. On the other
hand, the S-F package has been applied to 75 percent of
paddy landsl though not necessarily in optimal dosages.
The ILO Report states: "For a country with severe unem-
ployment it is surprising to what extent tried and tested
labor-intensive techniques in agriculture are . . .
ignored."2

Evidence of increased farmer resistance to tech-
niques is also observed in the experiences of other areas
researched. For instance, evidence from Western Uttar
Pradesh in North India indicates that farmers growing
wheat were more prepared to change the physical input than

the methods that were associated with them (9, pp. 51-52).

 

llbid..

2

 

ILO, Repopt, Op. cit., p. 91.

202

There were very few changes in regular crop Operation in
the survey farms, and even the basic irrigation technique
was unchanged in spite of the use Of purchased water.
Given the definition of Optimum adoption made
earlier in the subsection, the effect of farmer resistance
to techniques on the diffusion process is illustrated in 7“
Figure 4. Such resistance has the effect of flattening
the diffusion curve, so that it moves downward from OBCD

to OBCG.

 

Do row transplanting and hand weeding generate ifi
significantly greater returns thrOugh increased yields
than chemical weeding? If so, are these benefits recog-
nized by farmers? What factors on the cost side of farmer
calculus offset the benefits in returns from row planting
(hand weeding) and thereby cause such techniques to be
rejected in favor of chemical spraying? This would appear
to be an important area for research and possible exten—
sion emphasis.

The information already available suggests that
although the private (and social) returns from row planting
and hand weeding may be higher than from chemical weeding,
cost factors tend to offset these benefits. The "row
transplanted" variable (which represented unincluded vari—
ables for rotary weeding and hand weeding) was found to be

statistically significant at a 10 percent level of proba-

bility, while the "chemical weeded" was not. While the

203

coefficient 83 for chemical weeding was .12, the
coefficient 84 for row transplanting was .60. The cost
factors that probably Operate to render chemical weed—
ing preferable to row planting are as follows.

(a) The tax allowances specified for Spray
equipment such as the 80 percent development rebates, FT
together with low interest rates and over-valued for-
eign exchange rates, must influence the cost of random 3
planting and chemical weeding relative to that of row

planting and hand weeding. J

 

(b) Changes involving organization and routine
must have high felt costs, especially for small farmers
who are more tradition bound. For instance, in Sri
Lanka it is a strong tradition that consanguineous
farmers work each others' fields. Hence labor appli-
cation may be limited by the size of the extended
family. As the West Pakistan study indicates, so long
as new technology fits into the existing cultural
routine and agricultural calendar, they will be more
readily accepted (5, p. 72).

(c) The farmer appears to have to incur high search
and organization costs in labor utilization. The ILO

Reportl recognizes a need to "improve the accessibility

 

lILO, Report, Op. cit., p. 21.

 

204

of information concerning seasonal and casual employment."
The Report states: "The dry zone is still short of labor
at certain seasons when workers have nothing to do in
other parts of the country."1

(d) The credit scheme fails to distinguish between
traditional and new varieties in its allocation of cash
for row planting and hand weeding. The sum allocated is
only Rs. 50 (Table 5.5) whereas about 40 extra man-days
are required for these Operations, according to the ILO.2

(e) Techniques make demands on the entrepreneurial
(organizational) abilities<1ffarmers, which abilities may
be latent or inadequate.

(f) There may be extension biases in favor of the

(statistically) more visible changes in output than in the

(statistically) less visible changes in employment.

Summary
The chapter sought to assess the variables
from the standpoint of equity. Two criteria were
suggested for making this assessment: ~(a) scale
neutrality, and (b) employment generation. In evalu-
ating scale neutrality, the correlation coefficient

between farm size and each of the other independent

 

5:31;, p. 127 .

2ILO, Technical Papers, Op. cit., p. 202.

 

 

205

variables was regarded as a weak but nevertheless useful
indicator. It was used to either confirm or raise ques—
tions regarding the prior information available with
regard to the degree of scale neutrality in a variable.
Major sources of this prior information were the regional
and Sri Lanka primary—data-based micro studies listed in
Appendix A. If variables that were employment generating
and/or divisible were found to be significant from the 5

standpoint of paddy productivity, the variable was then

 

said to possess the desirable quality of congruence between :5
equity and output. ”" ‘”“”“ .
The variables that were important from the stand—

point Of paddy revenue in Sri Lanka were yields and cropping

 

intensity. It appeared prima facie that there was a
contradiction between the prior knowledge Of a negative
relationship between yields’and farm size and the posterior
knowledge of a statistically significant positive rela-
tionship between them. It was shown, however, that this
conflict may be more apparent than real, because mixing
adOptors and non—adOptors at an aggregated national level,
particularly in a situation where early innovators are
large farmers, makes it possible to Observe a significant
positive relationship between yields and farm size, with—
out refuting the prior belief held. The results with
regard to the relationship between cropping intensity and

farm size indicated that there may be more income-increasing

206

potential on small farms from increasing crOpping intensity
than is generally presumed to be the case.

It appears there is a considerable degree Of
consistency and congruence between output and equity
variables. It transpired that tractorization, which was
antagonistic to equity because of its employment-limiting
or labor—displacing effects, was also unimportant from
the standpoint of both yields and crOpping intensity,
although the latter may not continue to be the case. The
practice of chemical weeding through the use Of spray
equipment, which was not statistically significant for
yields (although there was evidence that it did have some
appreciable positive effect), was also found to be unim-
portant for employment. In fact, chemical weeding was
found to be associated with techniques that were compara~
tively weak with respect to both output and equity. On
the other hand, row planting, which was statistically
significant for yields, also has considerable employment
generation potential.

The general conclusion that emerged from the
analysis of agricultural mechanization was that although
application of mechanical technology resulted in a clear
discrepancy between private and social benefits, no
general verdict on agricultural mechanization was possible.
The variables determining the desirability Of specific

machines in specific situations are so numerous that

 

207

policy measures taken at a macro level need to leave room
for technical flexibility.

Notwithstanding the scale neutrality in the physical
seed-fertilizer package, its applicability to small farms
was found to be limited. The positive relationship between
farm size and adoption was partly attributed to small Fix
farmer lags. Small farmers were found to be later adOptors,
and the only discernible reason why small farmers adopted

later rather than sooner (ceteris paribus) was due to the

 

 

reduced risk from later adoption. The reduced risk in $3
turn stems mainly from improved information that reaches

small farmers from the demonstration effects Of early

adOptors, the develOpment of informal local channels, and

formal extension contact. Since technical information

flowed unevenly through the farm size structure, time

(adoption lags) is a proxy variable for information costs

to the farmer.

Informal local channels may be a substitute for
deficient extension. Early adOptors seem to confer an
external benefit on laggards which tends to make the
diffusion process inherently sub-optimal.

Possible factors inhibiting small farmer access
to technical information (thus in effect increasing the
average cost of adoption for small farmers) were indi-
visibility in the time needed to learn how to use new

inputs, extension biases against small farmers, the lower

208

receptivity to new information on the part of small
farmers, and control over some multipurpose cooperatives
by non-peer groups.

Indirect proof that improved channels of informa-
tion have the capacity for accelerating the adoption

process was found in the influence of the perceptibility

$23.3
of benefits and tenancy on adOption and productivity. i g
Acquisition costs of inputs were also found to g
be higher for small farmers. i
Production credit was discounted as a factor :5

 

entering into the perceived positive relationship between
farm size and adoption of biological technology. This
conclusion did not, however, extend to the effect of
credit on the positive relationship between farm size
and mechanical technology. It is conceivable that unless
the credit disbursement rules are changed, small farmers
who were bigger defaulters in the past may be prevented
from adopting despite a desire to do so in the future.
Contrary to theoretical presumptions, tenants seemed
to have a higher level of adoption than their owner-Operator
counterparts. But this factor justified extension rather
than tenancy because it served to underline the importance
Of the information factor, and reveal limited productivity
potentials in concomitant crop-sharing arrangements.
The risk factor was found to be statistically

significant for interdistrict paddy yields, and there was

209

evidence that small farmers were affected more by crop
failure than large farmers. Crop failure due to pests
and plant disease affected small farmers more than large
farmers due to the limited sum allotted for pesticide use
under the credit scheme, the lack of savings and
technical knowledge, and possible externalities in pesti-
cide use.

There is no evidence that small farmers are particu-
larly disadvantaged in their access to irrigated water.

However, minor irrigation works seem to be more congruent

 

than the major works that tend to be presently emphasized.
There is considerable congruency in the techniques
Of row transplanting and hand weeding, but their rela-

tively greater revenue benefits vis-a—vis chemical weeding

 

seem to be offset by probable higher relative costs. This
was identified as an important area for research and

extension emphasis.

CHAPTER VII

ALTERNATIVE POLICY SUGGESTIONS

FOR SRI LANKA

Introduction

 

Policy Options cannot be separated from policy
goals. The broad goals of growth and equity were con—

densed to the more specifh: Objective of increasing the j

 

*1

productivity of paddy farmers with holdingskbelow 4 acres
in size in a manner that will adequately increase employ-
ment for surplus agricultural labor. This Objective

was narrowed still further to focus on specific variables
(that were seen to be important for increasing paddy pro-
ductivity on small farms and increasing employment
Opportunities.

The analysis indicated that the use of tractors,
and perhaps the use of spray equipment for weeding, should
be de-emphasized (although there is a clear need for
technical flexibility), and that less reliance should be
placed on the disbursement Of credit as a major means Of
increasing small farmer productivity. At the same time
it is necessary that factors that encourage a more Optimal
and widespread application of the seed—fertilizer package

and congruent techniques (chiefly row transplanting/

210

211

'hand weeding)kmeemphasized. The constricting institu-
tional variables here were the high cost of information to
small farmers (mainly due to evident extension weaknesses),
the high real cost of acquiring fertilizers and other
inputs, and the technical risks in production from lack
of pest control and water control.

The objective of this chapter is to suggest policy
Options that will cause those variables that are unfavor-
able to growth and equity to be de-emphasized, and those
variables that are congruent or potentially congruent to
be emphasized.

It needs to be expressly stated, however, that

 

actual acceptance of the biological package and associ—

 

ated techniques depends not on the characteristic of

 

congruency, but on profitability--whether the added

 

perceived benefits to the farmer are worth the added felt

 

costs. Therefore no specific extent of adoption of
congruent techniquescn:technologies can be determined

in the absence of information with regard to relative
prices and costs at the community and farm level. Hence
no blueprint for action will be suggested.

Nevertheless, a number Of alternative suggestions
for economic, institutional, and organizational modifica—
tions came into View as the analysis proceeded. Since
these suggestions are tentative in nature, they need to

be examined and defined more specifically within relevant

 

212

agencies. They should be implemented so as to reduce
certain existing limitations to production expansion
and income improvements on small farms and to create
programs and emphases which stimulate expansion in
production and employment on small farms, and thereby
improve income distribution. Such changes would be
consistent with the economic movement and performances

desired in the Five Year Plan.

Policy Alternatives

 

Fiscal Changes

 

Fiscal changes have the ability to change rela-
tive prices so as to bring social and private benefits/
costs into closer alignment. Appropriate changes in
relative prices can discourage the use of those technolo-
gies and techniques considered less desirable for the
productivity of small farmers and employment generation

(income disparities), vis-a-Vis those considered more

 

desirable for these purposes.

The quantitative effect theoretically depends
on relative prices and cross elasticity of demand
factors. The strategy proposed will in effect rearrange
the manner in which the chips fall so that the resultant
performance contributes more to the growth and equity
objectives of the Five Year Plan.

It may be advisable to apply the higher of the

two exchange rates to imports of tractors and (power)

 

“e...

213

spraying machines, so that the c.i.f. value of imports will
ireflect the prevalent acute foreign exchange scarcity.
Consideration may be given to reducing or abolishing some
or all of the capital allowances (tax concessions) with
regard to these types of agricultural machinery. This
recommendation would specifically apply to the lump sum
depreciation of 80 percent of the cost on "short-lived"
equipment and 66 2/3 percent on "normal” machinery,

the exemption from business turnover tax given agricul-
tural machinery, and the develOpment rebates of 40 percent
of the cost of the assets that are allowed in addition to
lump sum depreciation allowances. In the interests of
improving the income distribution, it may also be advisable
to eliminate the tax holidays to large farmers who engage
in food production on jungle land.

Consideration may be given to increasing the com-
mercial borrowing and lending interest rates to reflect
both the scarcity of capital and the rate of inflation,
and to replace the tax concessions for machinery with
tax concessions for labor costs (or change in labor
costs) incurred in agricultural production.

These changes will increase the price Of agricul-
tural machinery to reflect their social valuation and

require that the present value of added future output from

 

214

'the use of machines exceed the discounted cost of invest-
ment valued shadow input prices, before their use is
considered worthwhile. As the foregoing implies, these
changes permit the chips to fall where they may with
regard to mechanical technology thus facilitating techni-
cal flexibility (although not ensuring it--see below). E73

Elimination of Market Controls
on MachinerypSupply

 

 

 

Concurrently with fiscal changes, steps may be

taken to make machinery and spare parts freely available

 

at the new (higher) price in order to ensure maximum
technical flexibility in their use. This involves abol-
ishing or reducing or changing physical import and exchange
controls on machinery and spare parts imports, which
presently Operate in a cumbersome manner due to inevitable
bureaucratic rigidities. There is evidence that over half
the tractors in Sri Lanka are not used at any one time due
to the nonavailability of spare parts,1 thus keeping
farmers from attaining the very flexibility that is the
most desirable attribute of tractors.

The ready availability of mechanical equipment
and spare parts is also likely to eliminate wastes in
the use of energy by reducing the uncertainties associ-
ated with tractor use. Since many farmers retain tradi-

tional sources of power (draught.animals) in reserve for

 

lILO, Report, Op. cit., p. 15.

215

use in case of mechanical failure, the free availability

of spare parts would reduce duplication and wastes, besides

making available to society the social benefits from the
release of fodder land.

The existing marketing structure can be retained.

The private sector has well developed distribution channels

that may be used. In addition, quasi public channels
which have been recently established in the Tractor
Corporation could be expanded to deal with all types of
agricultural machinery importation and distribution in
competition with private sector channels. The payment

of commissions on sales to quasi public employees, and
the use of public pricing policies to ensure reasonable
private pricing, are expected to contribute to efficiency

in the marketing of machinery and spare parts.

Production Credit

 

Production credit, it was shown, plays a largely
facilitative role. There was also evidence that small
farmers were bigger defaulters and that the availability
of credit to them per se has not induced them to innovate
to the same extent as large farmers.

Provision may be made to reactiviate the credit
scheme without writing Off loans outstanding (and thereby
encouraging even greater default in the future). The
recovery of outstanding loans to date may be spread over

five to six seasons, or even ten, in order to make past

 

:4!

216

defaulters re-eligible to receive credit. Recovery should
be a less important problem now that the government-owned
Paddy Marketing Board is the sole legal outlet for paddy.
However the repayment rate would cease to be an index of

productivity increases.

 

Loans may be confined to small farmers with hold— {“5.
ings below 4 acres in size. If subsidized inputs are 5
also confined to these target groups then the credit may i
take the form of physical input provision. When cash f .2
payments are called for, they could be made in the form éj

of vouchers which would be encashable only by the recipient
at a local bank or coop. For instance, if labor is paid
in the form of vouchers, encashment at the local coopera-
tive or bank may be confined to the person who Offered
his services. This procedure may reduce the possibility
that credit would be put to nonproductive uses, although
not ensuring it. This suggestion could, however, lead
to results that are counterproductive. It is possible
that labor will tend to show a preference for the more
affluent who pay in cash and this could result in higher
wages (either in terms of the money wage rate or only
being able to attract less efficient workers) for the
smaller farmers who pay in the form Of vouchers.

It seems necessary to considerably increase the
credit allotment for row planting and hand weeding (and

perhaps pesticide use).

217

It may be advisable to tie the production credit
scheme to a supplementary scheme providing credit for the
settlement of accumulated past loans owing to private money
lenders. These accumulated debts may be settled by the
state, and amortized by farmers in small installments from
the sale Of proceeds of future paddy output. Specific
loans for ceremonial needs would need to be provided to
prevent future indebtedness to private money lenders.

Improved extension discussed below should help
considerably in increasing the efficacy of the credit

scheme by increasing the farmer's ability to repay.

Extension

 

The wider application of the congruent variables
of the S-F package, row transplanting, hand weeding,
water control (through minor irrigation works and improved
drainage), and pest control can contribute much to the
Specified goals, given the planning horizon Of five years.
To do so two broad institutional—organizational vari-
ables need to be made operationally more effective.
Input supply channels need to ensure timely and adequate
supplies, and the extension service should be able to
provide the conconitant technical information that will
reduce information costs sufficiently to accelerate the
adoption process. The expanded use Of the techniques of

row transplanting, hand weeding, minor irrigation, and

.
1 angu-

 

218

dranage also require an intensive extension effort,
besides the other measures recommended.

The information with regard to the extension
service in Chapter V indicated that the service was defi-
cient, while Chapter VI drew attention to its crucial role
for the adOption of new technology by small farmers who
tend to be laggards.

Steps are already underway to use more stream-

lined methods Of extension program planning and evaluation.1

This is to be accomplished through the provision of a cadre
of "Extension Specialists" (who will act as liaison between
research agencies and extension staff), in-service training
in extension work for university graduates, and an increase
in the number of demonstration plots and the use of the
radio medium.

Other measures need to be considered for speedily
increasing the presently deficient extension ratio. One
means for doing this may be by a large-scale recruitment
of unemployed or underemployed science graduates with
farm backgrounds and providing them with a crash course
in extension and subsequent employment in the service.
Another alternative that may be explored is to concentrate
a majority of the present extension cadre in each of
the administrative divisions for a period of one (or

two) years, so as to improve the extension ratio in the

 

1Director of Agriculture, Op. cit., pp. 191-95.

 

 

219

short run. Particular attention may be paid to the tar-
get groups.

The success Of this strategy to intensify exten-
sion depends greatly on the enthusiasm and efficacy of

the extension personnel. The successful mobilization of

 

extension workers appears to be a prerequisite to the 373%
mobilization Of farmers. Extension workers may need to ~ i
be adequately compensated for the inconvenience costs of i
temporary relocation. This compensation may be composed Arr
of a fixed component and a variable component, the latter “J

being tied to output increases in a given area. Output
information would be readily available at the local
Offices of the Paddy Marketing Board, which is the sole
legal outlet for paddy.

An alternative strategy that is more experimental
and flexible may be adopted. Within one division it may
be possible to intensify extension contact to a ratio of
perhaps 1:250; in another, the existing structure and
ratios may be retained, but have extension personnel
concentrate on small farmers with holdings less than 4
acres; in a third division, it may be possible to
increase the number of personnel through additional
recruitment, and confinetfluaefforts of the new staff to
target group farmers. This strategy would be less disrup-
tive to individuals in their living conditions, and thereby

be conducive to stimulating innovative extension projects.

220

Studies would be needed to determine the best
extension strategy. For instance, reluctant farmers may
be mobilized by putting a small fraction Of their land
under new varieties, and assuring them of at least the
income foregone from traditional varieties. Those who
have already adOpted may be shown how to move toward
Optimum adoption. Risk reductions may be possible through
a diversification of (new) varieties applied, so as to
balance at the margin the utility of added returns against

the disutility of added uncertainty.l

 

lSuch diversification is very similar to the port-
folio strategy that investors in developed countries
adopt with regard to their investments. The concept Of a
"portfolio" is taken from liquidity preference theory in
macroeconomics and qualifies for transfer to a traditional
context because traditional farmers are risk averters,
and there is similarity in the direction of the relation-
ship between uncertainty and prospective yields. The
various types of new seed varieties in a traditional
economy have the identical attribute of an inverse
relationship between yield and uncertainty as do securi-
ties in a developed money economy. Thus a new improved
variety (NIV) like BGll may have high expected return but
a relatively low probability that this return will be
realized; an Old improved variety (OIV) like H4 or IR-8
may have a comparatively lower yield potential than the
NIVs, but because they are more robust have less yield
variability; the expected return from traditional vari-
eties (TV) may be even lower, but there is greater cer—
tainty about its realizability. Hence it would be
possible to arrange the "portfolios" of traditional
farmers in such a way as to balance at the margin the
utility Of additional return against the disutility of
additional uncertainty” This concept can be traced to
J. Tobin's "Liquidity Preference as Behaviour Towards
Risk," Review of Economic Studies, 25(2):65-86, 1958.

 

 

221

Input (Acquisition) Costs

 

It is necessary to reduce input acquisition costs
which are higher for small farmers. As was indicated in
Chapter V, the input supply problem has two dimensions to
it: the lack Of timeliness in supplies affects all

farmers in a given location while the supply procedures

 

impose greater burdens on small farmers. In either case, ' i
the costs of acquisition must be higher than otherwise a

would be the case and must consequently have an effect on V
productivity and incomes. ijvw

As argued in Chapter V, input subsidies are more
likely to encourage innovations than price supports, and
they tie government expenditures to the degree Of innova—
tion, unlike price supports. As was stated in Chapter VI,
input subsidies constitute a large regressive transfer
from taxpayers to large farmers because large farmers use
more subsidized inputs than small farmers. Hence measures
are needed to ensure that input subsidies continue to
have their expected favorable effect on innovation, without
being regressive and thus inimical to the distribution of
income.

An alternative to be considered is to make unsub—
sidized fertilizer and other inputs easily available
through private outlets, while confining subsidized seed
and fertilizer to farmers with holdings below 4 acres,

irrespective Of crop. Such an innovation in fertilizer

222

marketing is expected to yield three major benefits.

It would firstly reduce the regressive element in the
present scheme; secondly, reduce the acquisition costs
for small farmers; and thirdly, contribute substantially
to input marketing efficiency.

If the subsidized fertilizer is made available
purely on the basis of farm size regardless of the type
of crop, it would remove much of the time-consuming
procedures presently in Operation to ensure that ferti-
lizer utilization is crop-specific. These reforms do
not, however, safeguard against the resale of subsidized
fertilizer by eligible farmers. This danger may be
reduced by limiting a small farmer's eligibility to
receive subsidized fertilizer to a specified number of
seasons (two or perhaps four). Such limitation would have
the further benefit of compensating early adOptors for
the greater risks of early adoption.

Another alternative would be to provide subsidized
fertilizer to all innovators irrespective Of size for a
specified number of years.

Yet another alternative is to reduce the subsidy
considerably (e.g., from 50 percent to 10 or 20 percent)
but make fertilizer easily available through both private
and public sources.

As a supplementary measure, cooperative employees
may be paid a commission on subsidized or unsubsidized

input sales.

 

223

The upshot of these measures would be to increase,
the efficiency and efficacy of input market channels.
The development of private channels would introduce a
competitive element into input supply channels that would
Operate to the advantage of farmers by increasing their

bargaining power vis-a-vis input suppliers. It would also

 

strengthen the local market system in terms of timeliness
and adequacy of supplies, and hence be conducive to

optimum levels of adoption (where the MVP of an input

 

equals its price).

The financial savings from confining the ferti-
lizer subsidy to small farmers (amounting to about Rs. 18
million) could be diverted to agricultural research,
minor irrigation works, and improving extension. An idea
of the potential in this area can be gauged from the fact
that if only half the present fertilizer subsidy savings
(Rs. 9 million) is applied to minor irrigation works,
25,000 additional acres can be brought under improved

irrigation and drainage per annum.

Labor Utilization Costs

 

There is a need to reduce search costs of labor
not so much to reduce costs of labor hiring for the large

farmers,knn:rather to facilitate an expansion in their

 

1The cost of minor irrigation averages Rs. 300
per acre, while annual maintenance costs average around
Rs. 60 per acre, according to the ILO, Techndcal Papers,
Op. cit., p. 110.

 

224

use Of hired labor, thus stimulating employment. The
present local—level network of multipurpose cooperatives
may establish an employment exchange division, supple-
mented by a notice board, to bring together the suppliers
and demanders of labor.

An alternative proposal suggested by the ILO
Report1 is to create a network of labor c00peratives at
the district level rather than at the local level that could
match up the job seekers and jobs, and also pass on
requests which could not be met to neighboring (district)
COOperatives.

NO matter which alternative is finally adopted,
it would be necessary to provide incentives to Officials
who perform these functions that will cause them to Operate
in the general social interest. It may be necessary to
make the head Of the employment division an elected official
so as to ensure his responsiveness to the needs Of (local)

job seekers and employers.

Pest and Disease Control

 

It was seen earlier that crop failure attributable
to plant disease and pests accounted for between 12 to
15 percent Of "risk" and that this factor affected small

farmers more than large farmers.

 

lILO, Report, Op. cit.

' ’F.‘ .
r-um

 

“he.“ ‘

225

A solution may be to increase the credit allotted
for pesticide use. However, this does not solve the
problem of externalities associated with pesticide appli—
cation.

Since pesticide use has public goods character-
istics, it may be advisable to leave pest and disease
control under extension aegis where the costs would be
socialized. But the success of program implementation
measured in terms of adequacy and timeliness in pesticide
application will depend on the extent to which extension
interests coincide with farmer interests. If extension
incentives are tied to changes in output (as suggested
earlier) success is more likely to be achieved than
otherwise.

If the pest and disease control program is effec-
tively administered, the credit allotment for pesticide
use may be withdrawn. But it is suggested that the credit
allotment be maintained for use by farmers in the event of
extension deficiencies in program implementation.

The alternative of leaving pest and disease con-
trol to cooperatives may be examined, although the
existing cooperative structure would need to be reformed

before success can be expected.

Research Needs

 

There is an important and pervasive need for

continuous, relevant and adaptive research.

 

MI: I. 7

226

In the area of mechanical technology there is a
clear need for task specific machines and a need to
develop simple water control system in the wet zone. The
lack of task—specificity in machines causes fixed costs
to be higher than otherwise, and this in turn results in
more labor being displaced than otherwise would be the
case. In Sri Lanka tractors make possible the plowing of
hard dry soil in the dry zone for which draught animals
would lack the power intensity. But the purchase of
tractors needed for the specific task of breaking up hard
soil introduced into the production function a fixed cost
that necessitates the use of tractors for other tasks like
puddling, threshing, hauling, etc., for which the use of
tractor power would otherwise not be warranted.

Alternatives to tractor use for plowing may also
be explored. Instead of using tractors for this purpose,
it may be worthwhile for instance to simply divert for a
few days the power from the engines of ordinary diesel
trucks used for transport to the task of plowing, by
develOping appropriate attachments such as pulleys and
cables that permit the power source itself (truck) to
be positioned outside the field.1 Water also is a sub-
stitute for tractor power in the sense that it may be

possible to use water from the existing network of tanks

 

1That this may be technically feasible was commu-
nicated by an agricultural engineer who further said that
similar devices had been used in the U.S. at one time.

 

 

227

to soften the ground sufficiently to permit plowing by
draught animals. The benefit-cost ratio of using tubewells
for this purpose might also be explored.

With regard to biological technology, it is neces-
sary to know which of the 22 varieties are Suited to which
ecological locality, how pests and diseases could be EB%L
better controlled, and how weed control by water can be

replaced by weed control by correct methods of planting

and mechanical and hand weeding. .1

 

This study has also delineated the major areas
for economic research. More information is required on
the relative costs and elasticity of substitution between
the more labor-intensive and capital-intensive technologies
and techniques, particularly in planting and weeding, and
on the benefit-cost ratio of simple water control systems
utilizing pumps, lift irrigation, and tubewells as against
the capital-intensive large-scale irrigation projects
that are now emphasized.

In order that these and other alternatives may
be researched, adequate funding and apprOpriate directives
need to be given to research agencies. The funds may be
raised from foreign aid, and expansion of the regional
research agency in Central India may be considered due
to the considerable economies of scale and externalities

involved. This admittedly is a long—range solution.

CHAPTER VIII

SUMMARY

This study sought to identify policies that would rafig
contribute to increasing paddy output in a manner con-
tributory to the growth and equity objectives of Sri

Lanka's Five Year Plan. This required the identification 3'“

 

and assessment of technological and institutional vari-
ables, and an investigation of how they could be made to
interact to increase paddy productivity among small
farmers and significantly generate employment opportuni—
ties in the rural sector.

Technology constituted a link between growth and
income distribution. The primary source of growth is not
through the reallocation of traditional factors or greater
application of such factors; rather, it is through the
application of new off-farm purchased inputs embodying
modern scientific knowledge. But the actual nature of
the impact of technological change on income distribution
is determined by the socio-political milieu in which it
occurs. Technology is not neutral with respect to its
effect on personal and functional income distribution; it

implicitly influences both.

228

229

Since technology has an inevitable impact on
income distribution, policies designed to achieve equity
should be explicitly incorporated into policies designed
to increase productivity. If the redistributive impact
of technology can be harnessed and directed toward desired
equity goals, it reduces the reliance that needs to be
placed on fiscal and welfare policies for income redis-
tribution.

Redistributing income sources rather than income
streams has several advantages over conventional methods.
First, it minimizes the effect of a probable trade-off
between growth and equity ensuing from the transfer of
(potential) savings for consumption. It involves instead
a substitution of investment by low income groups for
investment or consumption by high income groups. Since
the transfer is of investment goods, it also contributes
toward a socially desirable pattern of production.
Second, it reduces the possible adverse effect that the
transfer of income flows may have on incentives of poten-
tial investors, since it provides capital to persons who
presumably haveEigreater marginal utility for the incre-
mental income subsequently generated. Third, since a
redistribution of income sources changes the pre-tax
distribution, it preempts possible obstacles to subsequent
redistribution by fiscal means that may arise from the

existence of fixed assets and the resistance of their

 

230

political manifestation (vested interests). Fourth, the
transfer of income sources (new agricultural technology)
increases the supply of goods (food) for which income
elasticity of demand is high, and at the same time arranges
for its sale at fairly stable prices by increasing rural
incomes and actuating an outward shift in the demand
curve. Fifth, since in a traditional economy signifi-
cant output increases must occur along a new (higher)
production function, the activation of a consumption
multiplier through a redistribution of income streams
rather than income sources (by the transfer of income
from persons with a relatively high marginal propensity
to save to those with a relatively low marginal propensity
to save) is likely, due to supply inelasticity, to stimu-
late inflation rather than generate output and employment.
The conceptual framework developed demonstrated
that the interaction of the primary variables of tech-
nology and institutions determined output and the manner
in which the resultant income was distributed. Resource
endowments influence the interaction of technology with
institutions but only to the extent that the political
process chooses to discover shadow price signals and,
once discovered, chooses to be guided by them. The
distribution of political power ultimately determines who
should gain or lose by the interaction of the primary

variables.

 

'_ w.
'W-I’H'I}

231

Program (policy) benefits (direction) in Sri
Lanka cannot be determined without reference to the social
welfare function as embodied in the Objectives of Sri
Lanka's Five Year Plan. Moreover, only when these program
benefits (outputs) are determined can the variables
(inputs) needed to achieve them be identified. An
inference was drawn that the Five Year Plan sought to
reduce disparities but not to the point that they disap-
pear altogether. Such an inference made the concentration
area of the Lorenz curve relevant because it deprived the
line of equal distribution of normative significance,
and made the Lorenz curve a normative outer bound. The
concept of the poverty rate, in conjunction with the notion
of the economic size of farm,vuu5used to identify target
groups (those with holdings below 4 acres in size) which
were to be the focus of programs. Those farm households
with incomes above the poverty line were to be outside
the general scope of agricultural development programs
providing access to subsidized income sources (like ferti-
lizer and seed), although they were to have access to
the benefits of specific projects or schemes to which
market alternatives were not available (like the guaran-
teed price scheme).

With the focus established, it becomes possible
to determine program objectives, and therefore to identify

the variables relevant for achieving them. These objectives

 

 

232

called for the identification and assessment of variables
from the standpoint of both growth and equity. Accord-
ingly, variables that were significant for paddy produc-
tivity and the cropping index in Sri Lanka were identified,
and thereafter their applicability to small farms and

their employment generation potential were evaluated.

A regression of five technical and five institu-
tional variables on interdistrict paddy yields revealed
that for policy purposes, the seed-fertilizer package
was (as expected) highly significant for yields. Other
variables that were significant were row transplanting,
rotary weeding and hand weeding, the risk factor in pro-
duction, and farm size. It was concluded that chemical
weeding and the renting of holdings (the tenancy factor)
had a positive and important, though statistically
insignificant, effect on yields. For tractor use and
production credit, the statistical analysis indicated that
they were not important for interdistrict differences in
paddy productivity. An R2 of .76 indicated that 24 per-
cent of paddy yield was not explained by the included
variables (and the excluded variables for which they were
proxy). Since most of the variables relating to tech-
nology (and techniques) were included in the regression
analysis (either expressly or by proxy), much of the
unexplained variation was attributed to omitted institu-

tional, infrastructural,anuiresource endowment factors.

 

 

233

The major excluded objective institutions were identified
as technical information (mainly extension) and input
supply differences between districts. The subjective
institutions were religious, cultural, and social factors.

The relation between farm size and output indicated
more scope for yield increases and double cropping among
small farmers than is generally supposed. Tractorization,
which was found to be not significant for yields, was
also found to be antagonistic to equity. Despite the
general conclusion that emerged that the application of
mechanical technology resulted in a clear discrepancy
between private and social benefits in Sri Lanka, no
blanket condemnation of tractorization and the use of
power spray equipment was possible because of a need to
ensure technical flexibility with regard to the use of
such equipment.

Notwithstanding the scale neutrality in the
physical seed-fertilizer package, its applicability to
smal farms was found to be limited. The positive rela-
tionship between farm size and adoption was partly
attributed to small farmer lags--small farmers were later
adoptors. The only discernible reason why small farmers

adopted later rather than sooner (ceterisgparibus) was the

 

reduced risk from later adoption. Since technical informa-
tion flowed unevenly through the farm size structure,

small farmers postpone innovation until perceived risks are

 

.‘. _ -;"|
“mv~

 

234

reduced through adequate information from the demonstra-
tion effects of early adoptors, the development of informal
local channels, and formal extension contact. Hence time
(adoption lags) is a proxy variable for information costs
to the farmer.

Possible factors inhibiting small farmer access
to technical information (thus in effect increasing the
average cost of adoption for small farmers) were indivisi-
bility in the time to learn how to use new inputs, extension
biases against small farmers, the lower receptivity to new
information on the part of small farmers, and control over
some multipurpose c00peratives by non-peer groups.

Indirect proof that improved channels of informa-
tion have the capacity for accelerating the adoption
process was found in the influence of the perceptibility
of benefits of technological change and tenancy on adop-
tion and productivity. ’

Perceived input costs were found to be higher for
small farmers due to high fixed costs in input acquisition.
Production credit was discounted as a factor

entering into the perceived positive relationship between
farm size and adOption of biological technology. The
existence of this positive relationship despite the availa-
bility of credit for small farmers indicated that less
reliance should be placed on credit as a means of inducing

technoloqical change in agriculture. It is likely,

 

235

however, that the credit factor did enter into the posi-
tive relationship between farm size and mechanical
technology due to inadequacy of loan allotments for the
utilization of mechanical equipment. Conceivably,
unless credit disbursement rules are changed, small

farmers who are ineligible to receive further credit

 

because they were bigger defaulters in the past, may be
prevented from adopting despite a desire to do so in the

future.

 

Contrary to theoretical presumptions, tenants
seemed to have a higher level of adoption than their owner—
operator counterparts. But this factor justified extension
rather than tenancy because it served to underline the
importance of the information factor, and revealed limited
productivity potential in concomitant crop-sharing
arrangements.

The risk factor was found to be statistically
significant for interdistrict paddy yields, and there was
evidence that small farmers were affected more by crop
failure than large farmers for several possible reasons:
the limited sum allotted for pesticide use for those
participating in the credit scheme, the inadequate amount
of operating capital, the limited technical knowledge, and
externalities in pesticide use.

There is no evidence that small farmers are par-

ticularly disadvantaged in their access to irrigated water.

 

236

However, minor irrigation works seem to be more congruent

than the major works that tend to be presently emphasized.
There is considerable congruency in the techniques

of row transplanting and hand weeding, but their relatively

greater revenue benefits vis-a-vis chemical weeding seem

 

to be offset by probable higher relative costs. This was

 

identified as an important area for research and extension L ;
emphasis.
The policy suggestions made sought to de-emphasize
variables that were antagonistic to the objectives of the 45
L

Five Year Plan with regard to paddy, and to emphasize
those that were contributory to those objectives. The
limitations of the study were expressly recognized. The
recommendations were tentative in nature; they sought in
the main to indicate general directions, and suggest
alternatives for further exploration and empirical study
--not a blueprint for action. No specific extent of
adOption of congruent techniques or technologies can be
determined in the absence of information with regard to
relative prices and costs at the community and farm level.
Fiscal changes were proposed that would change
relative prices in a manner that will bring social and
private benefits/costs into closer alignment. It was
suggested that foreign exchange and interest rates
applicable to agricultural machinery imports and purchases

have market prices that reflect their true cost to the

237

economy. Tax concessions applicable to machinery should

be reduced or abolished, and replaced with tax concessions

for (changes in) labor utilization. Concurrently, physical

market controls over machinery and spare parts utilization
may be eliminated so as to ensure maximum technical flexi-

bility in their use. The use of both public and private

channels for machinery distribution is likely to contribute

to marketing efficiency.

Provision may be made to reactivate the production
credit scheme by staggering recovery of outstanding loans,
confining loans to small farmers, increasing the credit
allotment for row planting, hand weeding (and perhaps
pesticide use), amortizing past loans to private lenders,
and providing loans for specific ceremonial needs (to
safeguard against future indebtedness to private sources).

The analysis revealed that extension is a key
variable that needs to receive greater emphasis. It may
need to be applied regionally (intensively) in order that
the externalities and economies in the adoption process
may be captured, and the time span needed for a given
level of adoption may be telescoped. Such levels of
intensity may be achieved over a longer period by recruit-
ment of trained extension personnel into the service.
Reforms are needed that will tie extension incentives to

desired performance.

Fm-

 

 

238

It is imperative that the high (fixed) costs of
input acquisition be reduced for small farmers by con-
fining subsidized inputs to small farmers, perhaps for a
specified number of seasons (while making unsubsidized
fertilizer freely available through private channels), and
eliminating the crop-specific character of the fertilizer
subsidy that presently necessitates intricate bureaucratic
checks. Another possible alternative is to provide inputs
at reduced levels of input subsidy to all farmers, through
both public and private channels. These measures would
reduce input acquisition costs for small farmers, reduce
the regressive element in the input subsidy, and con-
tribute to marketing efficiency by increasing outlet
Options for farmers.

Search and organization costs of labor utilization
may be reduced by setting up employment exchanges at the
local or district level. Costs of pest and disease con-
trol need to be socialized due to externalities in
pesticide application. But success in such a program
(measured in terms of timeliness and adequacy of pesti-
cide application) can be expected only if incentive
mechanisms are devised that tie the motivations of program
administrators to desired performance.

An important and pervasive need exists for con-

tinuous, relevant and adoptive research, particularly for

 

239

developing simple water control systems in the wet zone,
task-specific machines, and examining the relative costs

of labor-intensive as against capital-intensive techniques.

 

 

 

APP END ICES

240

 

 

APPENDIX A ' fig

SOURCES OF PRIOR BELIEF FOR THE EVALU-

ATION OF STATISTICAL RESULTS

241

 

 

 

APPENDIX A

SOURCES OF PRIOR BELIEF FOR THE EVALU-

ATION OF STATISTICAL RESULTS

Regional Studies .?

 

Ahmad, Bashir. "Farm Mechanization and Agricultural
Development: A Case Study of the Pakistan Punjab."
Ph.D. dissertation, Michigan State University, 1970.

 

Naseem, M.l "Small Farms and Agricultural Trans- E;
formation in Pakistan Punjab." Ph.D. dissertation, ~
University of California, 1972.

Peacock, D. L. "The Adoption of New Agricultural
Practices in Northeast Brazil: An Examination of
Farmer Decision~Making." Ph.D. dissertation,
Michigan State University, 1972.

Rebello, N. S. "Agricultural-Program Planning to
Reflect Potentials of Small Farms, with Application
to Mysore State, India." Ph.D. dissertation, Uni—
versity of Tennessee, 1972.

Rochin, R. I. "Microeconomic Analysis of Smallholder
Response to High-Yielding Varieties of Wheat in West
Pakistan." Ph.D. dissertation, Michigan State Uni-
versity, 1972.

Schluter, M., and Mellor, John W. "New Seed Varieties
and the Small Farm/'Cornell International Agricul—
tural Development Reprint 52. Department of
Agricultural Economics, Cornell University, Ithaca,
New York, June, 1972.

 

 

Singh, R., and Leagans, J. P. Communication and
Socio-Economic Variables Related to Fertilizer Use
by Indian Farmers. Department of Education, New
York State College of Agriculture, Cornell, 1972.

 

 

 

Stevens, Robert D. "Rural Deve10pment Proqrams for
Adaption from Comilla, Bangladesh, Agricultural Econ—
omics Report No. 215. Michigan State University,
June, 1972.

 

 

242

10.

ll.

12.

l3.

14.

15.

l6.

l7.

l8.

19.

243

Willis, I. R. "The Implications of the Green Revolu-
tion for Future Production Income and Employment in
Agriculture in W. Uttar Pradesh, India." Ph.D. dis—
sertation, University of Illinois, 1970.

Sri Lanka Studies

 

Bartsch, William H. Employment Effects of Alternative
Technologies and Techniques in Agriculture. Unpub—
lished draft copy, 1972. (Mimeographed.) .n,?

 

 

Central Bank of Ceylon, Surveyion the Cost of Produc-
tion of Paddy. Department of Economic Research,
Colombo, 1969.

 

 

Central Bank of Ceylon. Survey on Defaults in the
Repayment of New Agricultural Loans. Department of 5r
Economic Research, Colombo, 1962. SJ

 

 

 

Department of Census and Statistics. Census of
Agriculture. Colombo, 1962.

 

 

International Labor Organization. "Matching Employ-

ment Opportunities and Expectations," Report.
Geneva, 1971.
International Labor Organization. "Matching Employ-

ment Opportunities and Expectations," Technical
Papers. Geneva, 1971.

 

Ministry of Agriculture and Lands, Implementation
Programme 1973. Colombo, 1972.

 

 

Moomaw, J. C., and Curfs, H. P. Some General and
Special Aspects of Rice Soil Tillage. AGS:MRP/7l/ll-l,
FAO, Netherlands, July, 1971.

 

 

Raj, K. N. "Mechanization of Agriculture in India and
Sri Lanka." International Labor Review, Vol. 106,
No. 4 (Oct. 12), pp. 315-34.

 

Samuel S. N. "An Analysis of Institutional Modifica-
tions for Effective Farmer Credit in Ceylon."
Unpublished M.S. thesis, Department of Agricultural
Economics, Michigan State University, 1971.

 

 

APPENDIX B

RESULTS OF REGRESSION REGARDING INTER-

DISTRICT PADDY YIELDS--MODEL A

244

 

 

‘U‘

245

 

 

oanno. :n~n~. omn- «905.
”waao. . mnmka.ah one. noqm.m
owsko. asuwo.l «3a. smog.
mnwkn. «make. Nan. mano.o
gnome. mace». 0-o m«:o.«
Jaxno. mennooo mouc dose.
naaON. unavo. woo. Okro.o«
n~mno. ~Nfiw4. can. :oo4on
yahoo. ~mrmw. “3:. omen.
nmnno. o~nmoo 5M0. ~::o.
wwwauo muuou ddoQ can on
~¢ Acubucm
auh~nounoo
u—«tuumu no romdu Oxaozcum
m
«on. ouch-h
on“ .L
nsx-xoo coo.~

A4 a “01¢ o~.mn.o«
1‘.

(

a-ao
«<0 8

ocusunos.o:
Hoomn~uu.o«n
u¢<nam 21m:

oauup

ouna<4u

wrap

oases.
@«oma.
mafiln.
mum’s.
a33ma.
-ksd.
«:nms.
comes.
«~Nud.
can-.

m<hun no
madm .Okm

:«os.
~ mac 1

muwou axon uaaubsaz

«N
«a

an

nonna-
ummxmnl
~Mb~a.l
adamc.
0cm:~.
«ammo.l
on~n:.
Nannw.
Nona“.
m3~ac.

muzoumx
(hwm

:mno.
u

zooumxm no uuo

onmmmmomd 44¢¢u>o don >01

omxo
~¢

aquIIUJaduddt hzwo2uauo

nh\o°:\N«G zo ombdmdo

 

m x~azmmm<

0n~yanoo.
:3J~:~««o
monvmawo.o
03nuacon.
akmosqmao
smvnqmno.
n-n\:mo.
-oyaodno
oyymm‘ma.
obmmnooao

mhzuuunuuuou no

wxnmma .OFM

muruoao~.momn
Oswavano.os:
nn®90o~uoutun

muddaam an ::m

.m~<> nu 92¢ umm«u -

 

Nomnmkma.
:aaukouw.u
~umnm-:.n
s:.rn~nn.«
annuaa-.
WN’nJ’NJoI
.«xawuaun.
humminnn.
oH~«::H~.
unaaflfl~a.

Ngmmmnmm.-uc

w»:m~o«u¥unu

ZuHmmuNomx

«tam: hfiOOa. ddbo»

muwan

n“ hinudmo
~4 MNHm

a“ yrrno

a YMHm

u amazur

a nenafimwu
m ovum

: 1.an«~h
n whom“:

~ wohnnwe
a h2¢hmzou
we)

mnvrm

.z(mt hfioodu zenmmamnmu

.m~m>4<zq zo~merou~

Mom Xurhqt aura mqmnoo
tagwbtam

mm! .na.~«.oa.o.uo..~.nna

a;

.m~m»4<z< zenmmmdowm Ingmatcm

APPENDIX C

RESULTS OF REGRESSION REGARDING INTER-

DISTRICT PADDY YIELDS--MODEL B

246

 

247

cancn.. ammo«.o 43n0nnma.oa m~»o~¢-.cu o“ prune
emn~«. :Nmon. o~r3mrno. wranm~:o.« v vam
«Mk_~. mssmo. :MKJchm.~ muuungam. c m-m
«Esau. cuoiw. noamfllyfl. r.mkm9~«. s (murmo
codmm. qstfl. ommmwnox. myrmrmmu. 0 reached"
omen“. mamn:. -~:o«~a. :mcm«Mm~. m nmwn
kmnma. mcnqw. amaqm:~n. o~m-r~m. : Chum;uaw
:n~o~. osmaw. :oc«m«4“. nkmm.cgu. n «ovum:
oncun. on:««.n Nnkcro««. mmmoq~:u.u ~ an»uuw»
mr4~:m«~.m~«o a prxhwzou
mqpuo no mpxouwx Mpzuuouuuuou no upszuHuuuou 84>
wymu .c»m «pun waooqw .ohm onwaVoum
coco. coon. coco. -
~ «an m a Na .
wuuou «you wan-43x mumau

u~ coeommos.ocmn .zqut »:o¢<. a<~o~

~u Jos~:«o:.noo comaw

v m:~o:con.ooo~ .zqmr bacon. roummmaoma

roommau uo uwo

muacacm no tam

zeawmmmoum sugamso «or so.

assoc. .aczz.- moo. -iz. ems...
wast. nom3m. “Jo. :Nnm.m onr~.~
«scam. «wqbo. :«r. rsmd. palm.
:«mcuo :Jnum. New. .a:mn.« ~m0«.«
annum. bone“. ¢Mm. nan:- Juno.
onman. onwom. 3mm. oumn.m ommn.N
~o«-.. mmoQJ. oyu. :mkm.~ oo:m.u
Manon. oqa¢~. nan. oqnn.« mmmo.u
omsuo. oNnuH.c own. Now". :omn.o
muhmdwo V Muwou adou OHM mu . a»
Na 4<H»Uan
o~onmJJmo~ code.
whdtupmu mo Yommw omcordhm cam w
m
Noosanmo.om
non. caso.m neonnmwoo-n
saw a ma.:am 2.“:
cums»
n~\ms\ma .o.¢.~ ohmassm

x b . Jan;

Jn.nu.w~ urn»

.«_xuuwgr¢~¢<> pzuOZwamo

nsxmot\mao zo Owhqwao .maq> ca

0 KHszmnm<

02¢ mmmao - mom xHa»<:
.m~w>4¢z< znummwaouo

Mun

oWuw>J<z< ZOHWnHIoux

uwma uJDJoo
ItJuDraw

.oaooowvana
ma

IIJJDIGW

APPENDIX D

RESULTS OF REGRESSION REGARDING

CROPPING INDEX--MODEL C

248

 

249

u~omoI
n:~oo.
”gnaw.
nmwom.
mwtnm.
mambo.

wuwwamo
at

 

asen.~
Nana.
-mo.
«can.o
mono.»
«moo.

at

cannon

sum.»

amounol n:«.
osmo«.u ms:.
ceam~. as”.
msoam. n~o.
annum.l nuc-
nqmawoo urn.
uuuou xuou cum
J¢Hhu¢a
s~nuaumnoba
whatnbwu to woman O¢¢Dz<hn
m
noo-
cum
nsxuu\oo
u uu¢a

Naooo.ou

 

umcm.uo nmoouo :ooonoo ¢::««4an. onsmcoo:.a Na u~Hm
vans-o ~3amn. -~m~.u owno-mn.ou omuuuma:.nau on prino
~n~w. om-uo unomu. mmommonn.« :mm»mm-.« w xwuw
«owm.~ o:oa~. c~s4mo oooooaon. noJmNNos. h owhzma
o:ao.~0 n¢~«~. m«oom.0 sn:~nooa. oQNmJNmn.o a o»uumamu
Ouoo.l omsono Nnmun.l nonooanuo moowno-.c N vapuua»
soc:wanm.~c o .rrcpmzou
a» «chum no uptown: mwzuuunumuoo no mbzunu~uuuou ¢<>
«Maw ochm «rum Mdcxau .ohw zoummwaoua
one». . «ska. ammo. ~owo. . -
«co a N «40 a a we
muuou «goo uamupaaz . wuwau
«N h~h--hoooocu arauz bacon. adwor
cocaooooouan ma eunucnnn.o«uc «ocuu
«cocoonsouosu o shounona.omaa« azawx haom<u onmmuuuma
Hudson zqu: :oouuam so one mucqaam to tan
IOHMmUKuuu 44¢¢w>o com >01
xazntouo .au.xuuuao¢nu<> pauozmuuo
assooz\««o zo ouh¢u¢o .m¢¢> nu 02¢ mumqu - «on xmdhax omen unmaoo
.muwraqza onmmuuomu nuauaxam
II Mme .~«.o«.o.s.o.~.annu
324.; Q x325:
urn» .mnm>4¢z¢ acummuaouu unamozcm

APPENDIX E

CORRELATION MATRIX

250

 

 

 

 

251

 

.rrso ymae , SNHM OMS:ma caqsayaa shun oliflqup quash: rsiqqx_ ass”.
3 a s K o n .. . a u a

onoywo comma. _ nkumn.n mn:u:. :‘aua.u n;u¢«.u n9J:o. n«.~u.u enema.n vu.a,. “a wl‘a.iv

smuflufls m07m0- HDLNJII OMAMJ- Oua_»...I‘-O 1m..eo:.oI .DHDKUQI fiauqo’ol tum ,.J.I 4....,\..-l K“ 1H”.

mamas. Nmuflz. swo,m.- ”more. o.~im.- «Us; .- an.”.. “.35“.. is .p.- qrruq. ii .1: an.”

annoa.« nomnw. onwn~.o mflrnh. «Knxo.o nunfl:.n «nnJA. «1.53.. num.$.. ”4 I..n .A e. .1

wagon.“ women.- Nunan. monu. year“. «\NNq. ansflw.u n ,wy.- vbl... y v.u.
cocoa.“ ~:mmo.u okqna. mmmaa. n~::fi.o mhn31. manrx. :».. . n “emu
conga.“ ommmm.o «ANJm.I swam“. navam.v :Twco.u .zdu:.u n amp.ux
cacao.“ abflmn. omrmm. mlrmn. um~:. ~c.nr. r anamunwm
0000”.“ Gains. q.,mw. ”hp... ...,N. a a...
conco.u nam.m.- cymnu.u qn\nx. J n;.a.n~»
annum.“ awn.n. an.4n. a 1-.-“:
nouns.“ runnu. w yuhaawh
sauna.“ « nJmH>
.oz ~n>
_ monhesmmaou mamas”
n--.omm:a ooooo.~oooo canao.moo« «snwn.o~ navan.m: :aoa3.oo 103:1.a a“ “Jjuumo
n--.~sms osmoo.oszao anqon.omn~ ninso.- ”lsin.m: od.so.- as..s.m« Na “Nam
--~.ooo;. ..oqnsa.-zo.m oofiuo.n~.n «samo.o~ Nozfia.mmd ca,s..q.a assas.nzs as ..:unq~o
aouum.m annao.«« aoqoo.«« --m. oaaam. nuuqa.« :O)uw.s a“ .13u
n-:~.:n~ ona~0.oan~au cosou.mun~ carnn.n o.n99.:m oqaga.oo aoaam.no @ «wan
c.0m..as oos~z.mm canso.aq a.gin.a _mssm.l as..n.m oqqsm. o MN m
youn:.:o- ooa~:.mna- aoqqm.«om og:m«.r« n-~n.m~ nonun.~m qa:;:.q ~ «n.7qm
mmJno.ouo~u coon:.a\mnau oaaoo.mm:« so:n¢.u~ m:mro.ma aqa.n.~m nqaao.na o oaunuwwu
:onn;.oc:o onsa~.os:.nu annum..ns. mkaoz.s~ swarm.m~ as,ni.m. son.,.,~ a sun”

, n-««.umo osq~o.oso cacao.n~ ~xuim.a ~ozo,.n anou,.q~ aaq.q.o : n-quq~_
q-n~.«nn~ onoc~.~0no~ aaoo~.ano ~unno.o« sandm.on ogan:.~o assas. n amomm:
sonom.-mo~ oos:a.~o-~ aoas~.ooo« ouota.nn --0.m, assoz.im ossnn.s ~ ~azu.~z
«amo~.¢nmn °°°-.n~o¢o cocon.ma- m4.~o.qa onos~.,m noes:.no ,os.o.,n « aim”.

. 2.”: some «macaam 23m zoap4asmo zcuz was.» Saga) .0: swans
mzo-<~>w3 to tan Ox<OZ¢nm tataxdz tarasz wandnxqa
ouuaaam .zam

musn<~x¢. ouz~oumz¢mp zo newsmapcpm - ¢ m 4 5 ¢ p .
omz~¢hm¢ 02¢ a amuaoyo - came mum.o no «mars:
aaaoao.nm oaasoo.ak noosoo.,aa o,q.,o.«
n« N“ “a «a
psaauua uNHm gazaaoxo prrno
uuuo.~.~o accuse. .oooou.~n cooaon.m« nausea..o cocoon. soowao.m~ coonsa.m~ season.om
a o s a m : n N a
gang u~nm au_zu« ohcnuxa~ chum asamzax. swoon: «o.n<x» alma>
aux"..u¢ umau has
nkxooz\-a zo oupquxu .m~.> n“ oz« mumqu - no wise zOHqumxa Minzam
.. .mumpuqza zowmwmrouy c-1mwrnm
.1 nnmm..H~adrUN
23:3 So: 324.; m xHEénac.
a moat mm.:n.ou wrxh .mawpqazd znunwuroum oumnrum

APPENDIX F

TABULATED DATA

252

 

 

 

 

—.-—- —-

253

 

 

 

 

 

 

 

.:..SE 228...; 235:. .23 25:8 2

~02. 5.53302»; «0 guano \H

.33 .5230 do scam 3.350 £25m co 53035: do once so 13:3...“ \M

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.3101

14‘.

. 1......

JIJOI wxwl.

$10..

L

 

 

 

 

 

 

 

 

on? 33233 9:. 26:3 .3 22288 .33 .518 ..o 3333. “53.22% \m
......mn.assesses...e... Maine...“ fishes...

1-m~o»:mnem1~fi.m..43. _¢44.J...~..83J: .0 ”WWW... ..0....d.. m. 01.0.4.5...Nom h.m-¢;®w. me...erl.1.o.......o.:..«.....WF1MMMHN. fl
....1. «23......wW1... .0_¢........\.w\ .501 1n.-. 0% m....m..0.. Q 1w ..xu..$w ......1 +4.0.m m... Mm. 0&1:
..15H1cofluc o .1 mm... 1. mdqj 0 Sent adm. .013014ML...\N.%_ \._~»W0_.07« $.0th mimw m: we»
.. 01.921820 0m. 3 fimij 92.1.. 0 . Fwd. mam..- mnwj mafim. mafia Q..Q.4 1.23“. term. $9.0
. 6:25:20 am. .. 011.... £213- 0 (Auto. «Tm. mum... 010%.”..11b. mini 1Tb”. him... 12...»...
...... : w. .3... a. 2% m... .25... .12.. 1...... 121... 3.1. M...» s-..
: . :2... . s3: .......-.s was ea. 9.... 2:..- 1......1._-m._o_-_ E... 3... ......
.. 61:51:... ..-. Mal... m3] 0 7.0%. 1.5m. 0:414 0.90m. Mil». 1.0.... @512 1.11.0. 1.6...
. _ .282. wm .. R., .44 03.... 0 1.4.0. 11.....- 1.....1.-10..31.n.&$ ...,0 . 0&2.» P240. 1.. m.
_ _ £71.12... mm 1. $1-14 3314 0 0.6.x... 6.0.- ov.03. TWm. WNW... $3.0.-. Mimm. m huh. wish“
.. ..S:E....m.w ... No.41. 0mg.-. 1 0.0%. .010. Sham. 0.640. 10%. 0.0%. $6. taxi. mass;
.5371. £5.37» 3 mm}; Ni... _ 510%. big. atom: 1122..., 0.10% We. p.20 _ W20- 0.901
.. 1.3... m0 __ Tm... ._ WE. 1 0200. m0. 020m. aims. F53. W10.-. may mimm... 0.501
: .12....9. : me. : m2: 1 max... be him. PE. #3.. 2m... stem. 9.9 _ .231
:18 ..1. 5...... mm... Mb... :1... 1 mfiu. 02.0. 1.5.3.. 0.00;».m? 0...1.--..o..w\_ 020.. 1.231
11:12.50»: 0%.... $4.. 1 afar“. 0.0.. 0..mm.1.imd. 0..o..w_ mum... 1.01. 0.. m. _riofi

.... ......1 .2. m... __ m1... .... 1.2.-.. 1_ .52... To. sir ...»...m. 2.59 1.1... 8.4.. ....C) his.
_ _c..17.5:.om._.. bk. 1. 02.... o h..\.@ v.1? 9.0%... 520% @106. die: tied. htww. 9.6+“
_ .12.... 8. : my... 2:1. mus... .2... has... mine miss. 1.6-. 93. .....2... 06.1
_ 1 03rd 0d. 1. kw. __ wow—....(h \ixlo. hid. miém. 10.6: filmm. m..0: w.mm. w m . Q..¢m1
_ 33:11.. .n . ... ..._ .. at]; .2130 his: 1:18.. 0.1.2. min. 1:0... .....ms x...» . mhL
_ ...1.,§a..1....?. .8112. $11.11 P; sis- 5.6m. as... 9.3. 6.9.. 8.2.. ......n. ”.31
.._.. 11......1........_.\. .....1._....._1.1t... «mm... m5. ...... ...m........._.... EVAN. .11.... .._..__._. J12... ..1..._m. $5.... .11....»
_ fill: X X X aux .X \Al XI .- nx . --I X

M xHozmnE< sumo oeuwflsnme

 

 

 

 

BIBLIOGRAPHY

 

254

BIBLIOGRAPHY

Adelman, I., and Taft, M. Society, Politics and Economic
Deve10pment. Baltimore, Md.: 1967.

 

 

Adelman, I., and Morris, I. "An Anatomy of Income Distri-
bution Patterns." Development Digest, Vol. IX
(October, 1971).

 

Ahmed, Bashir. "Farm Mechanization and Agricultural
Development: A Case Study of the Pakistan Punjab."
Ph.D. dissertation, Michigan State University,
1970.

Barber, W. J. "A Critique of Aggregate Accounting Concepts
in Underdeveloped Areas." Bulletin of the Oxford
University Institute of Economics and StatistICS
(November, 1963).

 

 

Bartsch, William H. Employment Effects of Alternative
Technologies and Techniques in Agriculture.
Unpublished draft c0py, 1972. (Mimeographed.)

 

 

Bator, F. The Question of Government Spending. Collier,
1960. ,

 

Becker, Gary S. Human Capital. New York: Columbia Uni-
versity Press, 1964.

 

Becker, Gary 8., and Chiswick, Barry R. "Education and
the Distribution of Earnings." American Economic
Review (May, 1960).

 

 

Boulding, Kenneth. The Image. Ann Arbor, Mich.: Ann
Arbor Paperback, 1956.

Boulding, Kenneth, and Singh, Pritam. "The Role of the
Prize Structure in Economic Deve10pment." Ameri-

Economic Review (May, 1962).

 

Bowley, A. L. "The National Income in the U.K."
Economica, 1933.

 

Brown, Lester. The Social Impact of the Green Revolution.
International Conciliation, 1971.

255

 

 

 

256

Byerlee, D., and Eicher, Carl K. Rural Employment, Migra-
tion, and Economic Development: Theoretical
Issues and Empirical Evidence from Africa. East
Lansing: Department of Agricultural Economics,
Michigan State University, 1972.

 

 

Cannon, E. "Division of Income." Quarterly Journal of
Economics (1905).

 

 

Castle, Emery, and Youmans, Russel. "Economics in
Regional Water Research and Policy. American
Journal of Agricultural Economics, Vol. 50
(December, 1968).

 

 

Central Bank of Ceylon. Annual Reports. Colombo: Central
Bank.

 

. Bulletins. Colombo: Central Bank.

 

. Survey on Cost of Production of Paddy. Colombo:
Economic Research Department, Central Bank, 1969.

 

. Survey of Defaults in the Repayment of New
Agricultural Loans. Colombo: Department of Econ-
omic Research, Central Bank, 1972.

 

 

 

Clower, R. W., et. al. Growth Without Development.
Evanston, Ill.: 1966.

 

Coward, Walter E., and Schutjer, Wayne. "The Green Revo—
lution: Initiating and Sustaining Change."
Civilizations, Vol. 20, No. 4 (1970).

 

Department of Census and Statistics. Socio-Economic
Survey, 1969-70, First Round. Colombo: 1971.

 

 

. Socio—Economic Survey. Colombo: 1972.

 

 

. Census of Agriculture. Colombo: 1962.

 

m Statistical Abstract of Ceylon, 1969. Colombo:
1970.

 

 

. Statistical Pocketbook of Ceylon. Colombo:
1971.

 

 

Department of Inland Revenue. Tax Concessions for Agri-
culture. Colombo: 1966.

*—

 

 

 

257

Director of Agriculture. Administration Report 1967-8.
Colombo: Department of Agriculture, 1970.

 

Dorner, Peter. "Discussions on Equity and Productivity."
American Journal of Agricultural Economics
(December, 1970).

 

Duesenberry, J. S. Income, Savings and the Theory of
Consumer Behavior. Cambridge, Mass.: Harvard
University Press, 1949.

 

Eicher, Carl K., et. a1. Employment Generation in
African Agriculture. East Lansing: Institute
of International Agriculture, Research Report
No. 9, Department of Agricultural Economics,
Michigan State University, 1970.

 

 

Falcon, Walter. "The Green Revolution: Generations of
Problems." American Journal of Agricultural
Economics (December, 1970).

 

 

Fisher, Irvin. Elementary Principles of Economics.
New York: Macmillan, 1912.

 

Food and Agriculture Organization. Production Yearbook
1971. Rome: FAO.

 

. The State of Food and Agriculture. Rome:
FAO, 1970.

 

. "Tea Trends and Prospects." Commodity Bulletin
Series. Rome: FAO, 1960.

Friedman, Milton. A Theory of the Consumption Function.
New York: National Bureau of Economic Research,
Inc., 1955.

 

"Choice, Chance and Personal Distribution of
Income." Journal of Political Economy (August,
1953).

 

Galbraith, John K. The New Industrial State. Boston:
Houghton Mifflin, 1967.

 

Gotsch, D. H. "Technical Change and the Distribution of
Income in RuralAreas." Embnomic Development
Report, No. 205. Cambridge, Mass.: Centre for
International Affairs, Harvard University, 1971.

 

Griliches, Zvi. "Hybrid Corn and the Economics of Innova-
tion." Science, Vol. 132, No. 3422 (July 29, 1960).

 

 

258

Haq, M. "Employment in the 1970's: A New Perspective."
International Deve10pment Review. Ottawa:
Society for International Development, December,
1971.

 

Havens, Eugene A., and Rogers, Everett M. "Adoption of
Hybrid Corn: Profitability and Interaction Effect."
Rural Sociology, Vol. 26, No. 4 (December, 1961).

 

Hayami, Yujiro, and Ruttan, Vernon W. "Sources of Agri—
cultural Productivity Differences Among Countries."
Agricultural Deve10pment: An International Per-
spective. Baltimore: John Hopkins Press, 1971.

 

Heady, E. 0. Economics of Agricultural Production and
Resource Use. New York: Prentice Hall, Inc.,
1952.

 

 

. "Changes in Income Distribution in Agriculture
with Special Reference to Technological Change."
Journal of Farm Economics, XXVI (3).

 

Hicks, J. R. "Distribution and Economic Progress: A
Revised Version." The Review of Economic Studies,
Vol. IV (1).

 

Hill, E. F., and Mosher, A. T. Organizing for Agricul-
tural Development. New York: Agricultural

 

 

Development Council, 1963. (Mimeographed.)
International Labor Organization. "Matching Employment
Opportunities and Expectations." Report.

Geneva: ILO, 1971.

"Matching Employment Opportunities and Expec-
tations." Technical Papers. Geneva: ILO, 1971.

 

Towards Full Employment: A Programme for
Columbia. Geneva: ILO, 1970.

 

 

 

Jogaratnam, T., and Schickele, Rainer. Practical Guide-
lines to Agricultural Deve10pment Policies in
Ceylon. Peradeniya: Agricultural Economics
Research Unit (AERU), University of Sri Lanka,
1970.

Johnston, B. F., and Mellor, J. W. "The Role of Agricul-
ture in Economic Development." American Economic

 

Review, Vol. 51, No. 4 (September, 1961).

4.1

 

259
Krishna, Raj. "Agricultural Price Policy and Economic
Development." Agricultural Development and Econ-

 

omic Growth. Edited by Herman M. Southworth and
Bruce F. Johnston. Ithaca, N.Y.: Cornell Uni-
versity Press, 1967.

 

Kuznets, Simon. Modern Economic Growth. New Haven, Conn.:
Yale University Press, 1966.

 

. Six Lectures on Economic Growth. New York:
The Free Press, 1959.

 

Kwok, Min-hsioh. "Farmer's Associations and Their Con-
tributions toward Agricultural and Rural Development
in Taiwan!’ Bangkok: FAO, ECAFE, 1966.

Leff, N. H. "Marginal Savings Rates in the Development
Process: The Brazilian Experience." Economic
Journal (September, 1968).

Lewis, W. A. "Is Economic Growth Desirable?" Theory of
Economic Growth. Homewood, 111.: Richard D.
Irwin, Inc.

 

 

MacNamara, Robert S. "The Environmental Dilemma." War on
Hunger. Washington, D.C.: U.S. Department of
State, December, 1971.

Marchal, J., and Ducros, eds. The Distribution of National
Income. New York: Macmillan, 1968.

 

Marshall, Alfred, Principles of Economics. 8th ed.
New York: Macmillan, 1961.

 

McHale, Thomas R. "Ecological Analysis and Differential
Economic Growth Rate." Human Organization, Vol.
21, No. 1 (Spring, 1962).

 

McKean, R. N. Public Spending. New York: McGraw Hill,

 

Mellor, John. The Economics of Agricultural Development.
Ithaca, N.Y.: Cornell University Press, 1970.

 

. "The Functions of Agricultural Prices in

 

 

Economic Development." Indian Journal of Agri-
cultural Economics, Vol. XXIII, No. 1 (Jan.-Mar.,
'1‘9‘68).

Ministry of Agriculture and Lands. Implementation Pro-
gramme 1973. Colombo: The Ministry, 1972.

 

 

 

 

260

Ministry of Planning and Employment. The Five Year Plan.
Colombo: The Ministry, 1971.

 

Moomaw, J. C., and Curfs, H. P. Some General and Special
Aspects of Rice Soil Tillage. (AGS/7l/11—1).
Netherlands: FAO, July, 1971.

 

 

Mosher, Arthur T. Getting Agriculture Moving. New York:
Fredrick A. Praeger, 1966.

 

. Creating a Progressive Rural Structure. New
York: Agricultural Development Council, 1969.

 

 

Myrdal, G. Asian Drama. New York: 1968.

 

. "The Inequality Issue." The Challenge of World
Poverty. New York: 1970.

 

 

 

Nair, Kusum. The Lonely Furrow. Ann Arbor: University -
of Michigan Press, 1969. pg
Naseem, M. "Small Farms and Agricultural Transformation
in Pakistan Punjab." Ph.D. dissertation, Univer-

sity of California, 1972.

National Planning Council. The Ten Year Plan. Colombo:
Planning Secretariat, 1959.

 

Niehoff, Arthur H. A Casebook of Social Change. Chicago:
Aldine Publishing Co., 1966.

 

 

Nulty, Leslie. The Green Revolution in West Pakistan:
Implications of Technological Change. New York:
Fredrick A. Praeger, 1972.

 

 

Oshima, Harry T. "The International Comparison of Size
Distribution of Family Incomes with Special
Reference to Asia." Review of Economics and Sta—
tistics (November, 1962).

 

Owen, Wyn F. "The Double Development Squeeze on Agri-
culture." American Economic Review, Vol. LVI,
No. 1 (March, 1966).

 

Pathirana, W. Administration Report of the Cooperative
Movement in Ceylon, 1968—69. Colombo: Ceylon
Government Press, 1970.

 

 

Peacock, D. L. "The Adoption of New Agricultural Practices
in Northeast Brazil: An Examination of Farmer
Decision—Making." Ph.D. dissertation, Michigan
State University, 1972.

261

Pen, Jan. Income Distribution. London: Allen Lane
Penguin Press, 1971.

 

Perkins, M., and Witt, Lawrence. "Capital Formation:
Past and Present." Journal of Farm Economics,
Vol. 43, No. 2 (May, 1961).

 

Raj, K. N. "Some Questions Concerning Growth, Transforma-
tion and Planning of Agriculture in Deve10ping
Countries." Journal of Development Planning, No.
l (1969).

 

Finn,
"Mechanization of Agriculture in India and
Sri Lanka. International Labor Review, Vol.
106, No. 6 (October 12).

9'11: F

 

Rao, V. K. R. V. "Redistribution of Income and Economic
Growth in Underdeveloped Countries." Income and _L
Wealth, Series X (1964). ,j “

é'

 

 

Rebello, N. S. "Agricultural Program Planning to Reflect
Potentials of Small Farms, with Application to
Mysore State, India." Ph.D. dissertation,
University of Tennessee, 1972.

Robbins, Lionel. An Essay on the NatureenuiSignificance
of Economic Science. 2nd ed. (1952). London:
Macmillan, 1935.

 

 

Rochin, R. I. "Microeconomic Analysis of Smallholder
Response to High-Yielding Varieties of Wheat in
West Pakistan." Ph.D. dissertation, Michigan
State University, 1972.

Rogers, Carl. ‘"Client—Centered Therapy." Theories of
Personality. Edited by H. S. Hall and G. Lindzey.
New York: John Wiley and Sons, 1970. .

 

 

Samuel, S. N. "An Analysis of Institutional Modifications
for Effective Farmer Credit in Ceylon." Unpub-
lished M.S. thesis, Department of Agricultural
Economics, Michigan State University, 1971.

Samuels, Warren. Welfare Economics, Power and Property.
East Lansing: Department of Economics, Michigan
State University. (Mimeographed.)

 

Schickele, Rainer. "Improving Income Distribution as a
Development Goal." Staff Paper P72—17. Depart-
ment of Agricultural Economics, University of
Minnesota, 1972.

 

“k

262

Schickele, Rainer. Ceylon Papers 1967—70. New York:
A/D/C, 1972.

 

. Agrarian Revolution and Economic Progress.
New York: Fredrick A. Praeger, 1968.

 

Schluter, M., and Mellor, John W. New Seed Varieties and
the Small Farm, Cornell International Agricultural
Development Reprint 52. Ithaca, N.Y.: Department
of Agricultural Economics, Cornell University,
June, 1972.

 

 

 

Schmid, A. Allan. "Economic Analysis of Water Resource
Problems." American Economic Review (May, 1967).

 

Schultz, T. W. Transforming Traditional Agriculture.
New Haven, Conn.: Yale University Press, 1964.

 

Schultze, Charles L. The PoliticsenxiEconomics of Public
Spending. Washington, D.C.: The Brookings Insti—
tution, 1968.

 

Scitovsky, Tibor. "The State of Welfare Economics."
American Economic Review, No. 41.

 

Seers, Dudley. "The Meaning of Development." Communica-
tions Series, Br. 44. Sussex: Institute of
Development, 1970.

 

 

. "Income Distribution and Employment: Some
Issues Raised by the Columbia Report." Bulletin
of the Institute of Development Studies, Vol. 2,
No. 4 (1970), Brighton England.

 

 

. "New Approaches Suggested by the Columbia
Employment Programme." International Labor Review
(October, 1970).

Shaffer, James D. Property, Market Structure and Effi-
ciency. East Lansing: Agricultural Economics
Miscellaneous Report 1966-9, Department of Agri-
cultural Economics, Michigan State University.

 

Shaw, Robert A. "Jobs and Agricultural Deve10pment."
Development Digest, Vol. IX, No. 1 (January,
1971).

Singh, R., and Leagans, J.P. Communication and Socio—
Economic Variables Related to Fertilizer Use by
Indian Farmers. Department of Education,
New York State College of Agriculture, 1972.

 

 

263

 

Staehle, H. "Short Period Variations in the Distribution
of Incomes." Reveiw of Economic Statistics (1937).
Stevens, Robert D. "Rural Development Programs for

Adaption from Comilla, Bangladesh. Agricultural
Economics Report No. 215. East Lansing: Michigan
State University, June, 1972.

 

 

Thambyahpillay, George. "Climatological Research and
Agricultural Development in the Dry Zone."
Development of Dry Zone Agriculture. Edited by
O. S. Perera. Colombo: Ceylon Association for
the Advancement of Science, 1967.

 

 

Thiesenhusen, William C. "Latin America's Employment
Problem." Science, Vol. 171 (March 5, 1971).

Tobin, J. "Liquidity Preference as Behavior Towards
Risk." Review of Economic Studies, 25(2):65-86
(1958).

Tweeten, Luther. Foundations of Farm Policy. Lincoln:

 

University of Nebraska Press, 1970.

United Nations. Report on the Second Development Decade.
U.N. Sales No. E71, 11, A.2.

 

United Nations Research Institute for Social Development
(UNRISD). Cooperatives and Planned Change in Asian
Rural Communities. Edited by Inayatullah.
Geneva: 1970.

 

U.S. Department of Agriculture. Rice Situation. USDA,
ERS RS—15 (March, 1970).

 

Weeks, John. Employment and the Distribution of Income.
Ms. IDS. Brighton: 1972. (Mimeographed.)

Wharton, C. R. "The Green Revolution: CornuCOpia or
Pandora's Box." Foreign Affairs (April, 1969).

 

"Risk, Uncertainty and the Subsistence Farmer."
Development Digest, Vol. III, No. 2. Washington,
D.C.: Agency for International Development, U.S.
Department of State, April, 1969.

 

Willis, I. R. "The Implications of the Green Revolution
for Future Production Income and Employment in
Agriculture in W. Uttar Pradesh, India." Ph.D.
dissertation, University of Illinois, 1970.

WOnnacott, Ronald J., and Thomas H. Econometrics. Toronto:
John Wiley and Sons, 1970.

 

 

MICHIGAN STATE UNIV. LIBRQRIES
ll“WNW”WWWWHIIWIMIWII
31293101214116