SMALLHOLDER AGRICULTURE IN THE CONTEXT OF INCREASING
POPULATION DENSITIES IN RURAL KENYA
By
Milu C. Muyanga

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Agricultural, Food and Resource Economics – Doctor of Philosophy
2013
 

ABSTRACT

SMALLHOLDER AGRICULTURE IN THE CONTEXT OF INCREASING POPULATION
DENSITIES IN RURAL KENYA
By
Milu C. Muyanga
Smallholder farmers constitute the bulk of agricultural producers in sub-Saharan Africa
and majority of them are poor. Based on sub-Saharan Africa’s land endowment, an agriculturalled growth strategy has been touted as solution for reductions in poverty in this region. However,
recent studies cast doubts on the land abundance hypothesis. Each day, the Africa smallholders’
landholding and access shrink as population density rises. Despite these mounting population
related challenges, analysts have pointed out that sub-Saharan Africa still has the potential to
revitalize smallholder agricultural productivity for reduced poverty and hunger if appropriate
policies are pursued. They cite the example of the smallholder-led broad-based Asian green
revolution that contributed greatly to rural poverty reduction in that region to argue their case.
The first essay investigates the potential for similar forms of inclusive agricultural growth
using Kenya as a case study. The study specifically investigate whether formal and informal land
institutions in Kenya are making it possible for a broad-based smallholder led agricultural
growth process as enjoyed in much of Asia. In Kenya, there has been a policy thinking that
agricultural and land reforms supported by adequate government budget allocation have the
potential to underpin a revitalized system of smallholder production especially in areas where
land sizes have become too small. This Essay uses three sources of data, namely: a panel data
spanning 13 years, cross-sectional medium scale farmers’ survey data, and qualitative data from
focus group discussions.

Several consistent findings are as follows: First, diminishing land sizes have become a
binding agricultural production constraint in the densely populated regions of the Kenya. Second,
the majority of medium-scale farmers, defined as using between 5-50 hectares for agricultural
purposes, owned on average over two times more land than they were using for agriculture,
implying a high degree of land owned for speculative purposes and/or an inability these farmers
to make productive use of the land they owned. Majority of medium scale farmers are either
current or former public sector employees; and acquired their land from savings from non-farm,
largely urban jobs; only a minority were primarily engaged in agriculture prior to achieving
medium-scale farming status.
The second essay examines how rising population pressure affects smallholders’
production, commercialization and household incomes. Using data from five panel surveys on
1,169 small-scale farms, econometric techniques are used to determine how increasing rural
population density is affecting farm household behavior and livelihoods. The estimation strategy
deals with the potential endogeneity of population density in input demand and output supply
equations using a two-stage control function approach.
The overall picture emerging from this essay is that land is becoming an increasingly
constraining factor of production and that smallholder agriculture farming practices in the areas
of high population density are distinctly more land-intensive. Inputs and output agricultural
intensification, household incomes and smallholder commercialization rise with population
2

density up to about 600-700 persons per km ; beyond this threshold, rising population density is
associated with sharp declines in agricultural intensification and commercialization.

Copyright by
MILU C. MUYANGA
2013

 

 
 
 
 
 
 
 
 

For my family
Esther, Annette and Tedd
For my mother
Katuvee wa Muyanga
In the memory of my late father
Muyanga wa Mbindyo
 

v
 

ACKNOWLEDGMENTS

Many people contributed to the success of this work, but owing to space limitations, I
will only mention a few. First, I deeply thank my major professor and my dissertation committee
chair, Thom Jayne. Thom has not only been my professor, but a mentor and friend. You added
the final and the most important weft in the weaving of my academic basket. Using Isaac
Newton’s words I can only say this: "if I have seen further it is by standing on the shoulders of
giants". I also admire your sixth sense. You know when the student needs a push, needs a break,
and above all you are able to guide students to discover and develop their potential. I would also
like to acknowledge the contributions and feedback from the rest of my examination committee
members, Eric Crawford, Robert Myers, Songqing Jin and the late David Campbell. Dr.
Campbell was able to go through my drafts, sometimes from the hospital bed, and to participate
in my dissertation defense regardless of his failing health. Dr. Campbell died a week after my
defense. Dr. Jennifer Olson, his colleague in Geography Department, told me that my defense
was the last thing he did professionally. What an honor? David, you confirmed to me that a true
warrior does not give up what he loves. You fought a heroic fight; you served as a role model for
how to walk a lonely path with style, grace, dignity and honor. R.I.P David.
Thanks are due also to Scott Loveridge, Scott Swinton, and Debbie Conway, for
assistance in matters of administration and to Linda Beck who usually came to my rescue
whenever my laptop slowed down or failed. Many thanks to the land research group, Jake
Ricker-Gilbert, Rui Benfica, David Mather, Nick Sitko, Derek Headey and Jordan Chamberlin
from whom I learned a great deal. Jordan you did a wonderful job in extracting the GIS data and

vi
 

generation of maps used in this dissertation. To Margaret Beaver and the Tegemeo Institute
(Egerton University) family, thanks for your help with the survey data collection and handling.
I had been warned that a PhD journey would be long and lonely, but this wasn’t the case
with friends and colleagues like Adam Lovgren, Miltone Ayieko, Jordan Chamberlin, Ramzi
Adjao, Tim Hodge, Keneilwe Kgosikoma, Nathalie Me-Nsope, Rie Muraoka, Chewe Nkonde,
Assistou Ouedraogo, Vivek Pandey, Alexandra Peralta, Christina Plerhoples, Fang Xia, Dale Yi,
Helder Zavale and Raul Pitoro. I acknowledge the gifts of time that the Kenyan friends and
families in Lansing bestowed on me and my family. The Ayiekos, the Chotis, the Onyanchas,
and the Ododas, just but to mention a few. Thank you for your support and encouragement.
Special thanks to my mother, Jemimah Katuvee. She took the responsibility of raising us
singlehandedly when our father passed on to glory in October 1991. She relentlessly toiled the
family farm to produce food and a surplus, Kambaland’s scorching sun and unreliable rainfalls
notwithstanding, to cater for our education needs. I would like to thank my brothers and sisters,
Ngala, Kataa, Monica, Mwanzia, Mwikali and Aphia, for their continued support. Many thanks
to my in-laws Ben Kavale and Susan Mwanzia, and family for your love, prayers and
encouragement. To my late sister, Janet Kasumuni, the journey that you initiated in January 1978
when enrollment me as a first grader in Kyatune Primary School has come to an end, barely four
monthly after you unexpectedly departed. I wish God gave you some more days to witness the
end of the journey. For those who did not receive a mention, I say ASANTENI sana.
My wife Esther, my daughter Annette, and son Tedd deserve and merit my deepest
gratitude. For the five years of the PhD journey I had been an absent husband and father but their
affection to me never vanished. Thanks for your endless love, support and understanding.

vii
 

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................................... x
LIST OF FIGURES ..................................................................................................................... xiii
KEY TO ABBREVIATIONS ....................................................................................................... xv
CHAPTER 1: BACKGROUND ..................................................................................................... 1
1.1 Introduction ...................................................................................................................... 1
1.2 Key findings ................................................................................................................... 10
1.2.1 Smallholder Land Access in Kenya: Are smallholders farming themselves out
of near-landlessness? ......................................................................................... 10
1.2.2 Effects of population density on smallholder agricultural production and
Commercialization in rural Kenya ..................................................................... 12
APPENDIX ............................................................................................................................ 14
REFERENCES ....................................................................................................................... 18
CHAPTER 2: SMALLHOLDER LAND ACCESS IN KENYA: ARE SMALLHOLDERS
FARMING THEMSELVES OUT OF NEAR-LANDLESSNESS?..................... 23
2.1 Introduction .................................................................................................................... 23
2.2 Historical background of land issues and policies in Kenya .......................................... 28
2.2.1 The history of the land question in Kenya ......................................................... 28
2.2.2 The post-colonial land policies and administration ........................................... 33
2.2.3 The National Land Policy .................................................................................. 36
2.3 Conceptual framework ................................................................................................... 42
2.4 Methods and data sources .............................................................................................. 47
2.4.1 Empirical models and estimation strategy ......................................................... 47
2.4.2 Data sources ....................................................................................................... 51
2.5 Results ............................................................................................................................ 54
2.5.1 Descriptive results .............................................................................................. 54
2.5.2 Econometric results............................................................................................ 84
2.6 Conclusions and policy implications.............................................................................. 96
APPENDIX .......................................................................................................................... 101
REFERENCES ..................................................................................................................... 110
CHAPTER 3: EFFECTS OF POPULATION DENSITY ON SMALLHOLDER
AGRICULTURAL PRODUCTION AND COMMERCIALIZATION IN RURAL
KENYA............................................................................................................... 119
3.1 Background .................................................................................................................. 119
3.2 Overview of the effects of population density on agricultural production .................. 124
3.3 Overview of effects of population density on agricultural commercialization ............ 131
3.4 Conceptual framework ................................................................................................. 136
viii
 

3.4.1 Agricultural household model.......................................................................... 136
3.4.2 Drivers of human population growth ............................................................... 140
3.4.3 Effects of population density on household landholding ................................. 142
3.4.4 Effects of population density on land rental prices .......................................... 144
3.4.5 Effects of population density on agricultural wages rate ................................. 145
3.4.6 Effects of population density on agricultural output prices ............................. 147
3.5 Data sources ................................................................................................................. 149
3.5.1 Household longitudinal data ............................................................................ 149
3.5.2 Geographic information systems (GIS) data.................................................... 150
3.5.3 Regional market maize price and weather data .............................................. 151
3.6 Data analysis methods .................................................................................................. 152
3.6.1 Descriptive analysis ......................................................................................... 152
3.6.2 Econometric models......................................................................................... 153
3.7 Results and discussions ................................................................................................ 159
3.7.1 Descriptive results ............................................................................................ 159
3.7.2 Econometric results .......................................................................................... 176
3.8 Conclusions and policy implications............................................................................ 199
APPENDIX .......................................................................................................................... 202
REFERENCES ..................................................................................................................... 210

ix
 

LIST OF TABLES

Table A1.1: Kenya's population density by province and districts............................................... 15
Table A1.2: The 20 percent most densely populated districts in Kenya (new districts) .............. 17
Table 2.1: Households mean landholding sizes (hectares) ........................................................... 55
Table 2.2: Gini indices of households’ landholding by agro-regional zones ............................... 60
Table 2.3: Gini indices of households’ land access by agro-regional zones ................................ 60
Table 2.4: Landholding by entry pathway into medium-scale farming ........................................ 66
Table 2.5: Household demographic characteristics and historical information ............................ 68
Table 2.6: Household landholding attributes ................................................................................ 70
Table 2.7: Medium-Scale Farmers’ landholding and productivity ............................................... 72
Table 2.8: Smallholder farmers’ landholding and productivity by landholding categories ......... 82
Table 2.9: Household income and assets ...................................................................................... 84
Table 2.10: Selection equation results-- pathways into medium scale farming ........................... 86
Table 2.11: Results of the endogenous switching model for medium-scale farmers’ land access 89
Table 2.12: Results of the endogenous switching model for area under crop .............................. 90
Table 2.13: Results of the endogenous switching model for intensity of fertilizer use ................ 93

x
 

Table 2.14: Results of the endogenous switching model for crop production per ha owned ....... 94
Table 2.15: Results of the endogenous switching model for crop production/ha cultivated ........ 95
Table A2.1: Distribution of government land by use in square kilometers ................................ 102
Table A2.2: Distribution of trust land that is not available for smallholders ............................. 102
Table A2.3: Classification of land in Kenya ............................................................................... 103
Table A2.4: Medium and large scale farmers survey sample ..................................................... 107
Table A2.5: Focus group discussion sites ................................................................................... 109
Table 3.1: Farming practices and factor intensities, by pop. density quintile ........................... 160
Table 3.2: Household income and wealth trends, by population density quintile ..................... 169
Table 3.3: Smallholder commercialization indices by population density quintiles .................. 171
Table 3.4: Smallholder commercialization by population density quintiles............................... 175
Table 3.5: OLS estimation results for population density .......................................................... 177
Table 3.6: CRE estimation results for producer farm gate maize prices .................................... 179
Table 3.7: CRE estimation results for agricultural wage ............................................................ 181
Table 3.8: CRE estimation results for land rental rates .............................................................. 182
Table 3.9: CRE estimation results for household landholding ................................................... 184
Table 3.10: CRE estimation results for intensity of cash input use per hectare cultivated......... 185
xi
 

Table 3.11: CRE estimation results for intensity of fertilizer use (kgs/ha cultivated) ................ 186
Table 3.12: CRE estimation results for net crop production per hectare owned ........................ 188
Table 3.13: CRE estimation results for net crop production per hectare cultivated ................... 189
Table 3.14 CRE estimation results for gross farm production per hectare owned ..................... 190
Table 3.15: CRE estimation results for net farm production per hectare owned ........................ 191
Table 3.16: CRE estimation results for net farm production per hectare cultivated................... 192
Table 3.17: CRE estimation results for household aggregate income per adult equivalent ....... 194
Table 3.18: CRE estimation results for household off-farm income per adult equivalent ......... 195
Table 3.19: Fractional probit estimation results of land allocated to non-maize crop ................ 197
Table 3.20: Fractional probit estimation results for household crop commercialization ........... 198
Table A3.1: Responses to population pressure and the conditioning variables .......................... 203
Table A3.2: Indices of agricultural commercialization .............................................................. 204
Table A3.3: Processes in agricultural commercialization........................................................... 206 
 

xii
 

LIST OF FIGURES	

Figure 1.1: Population growth rates in Kenya ................................................................................ 8
Figure 1.2: Population density in Kenya......................................................................................... 9
Figure 2.1a: Area cultivated by land size –overall ....................................................................... 73
Figure 2.1b: Area cultivated by land size –by growth pathways .................................................. 73
Figure 2.2a: Proportion of area cultivated--overall....................................................................... 74
Figure 2.2b: Proportion of area cultivated—by growth pathways ................................................ 74
Figure 2.3a: Fertilizer use per hectare cultivated – overall ........................................................... 76
Figure 2.3b: Fertilizer use per hectare cultivated -- by growth pathway ...................................... 76
Figure 2.4a: Value of crop production -- overall .......................................................................... 78
Figure 2.4b: Value of crop production – by growth pathway ...................................................... 78
Figure 2.5a: Value of crop production per hectare owned-- overall ............................................. 79
Figure 2.5b: Crop of production per hectare owned – by growth pathway .................................. 79
Figure 2.6a: Value of crop production per hectare cultivated -- overall ....................................... 81
Figure 2.6b: Crop production per hectare cultivated -- by growth pathway ................................. 81
Figure A2.1: Paths of access to land ........................................................................................... 105

xiii
 

Figure A2.2: Modified Fisher-Ideal Quantity Indices ................................................................ 106
Figure A2.3: Medium and large scale farmers survey sample.................................................... 108
Figure 3.1a: Fertilizer expenditure per hectare-overall............................................................... 162
Figure 3.1b: Fertilizer expenditure per hectare-by survey year .................................................. 162
Figure 3.5a: Owned land-labor ratio --overall ............................................................................ 166
Figure 3.5b: Owned land-labor ratio--by survey year ................................................................ 166
Figure 3.6a: Cultivated land-labor --overall ............................................................................... 167
Figure 3.6b: Cultivated land-labor ratio--by survey year ........................................................... 167
Figure 3.7: Proportion of area allocated to non-maize by population density ............................ 172
Figure 3.8: Household commercialization index (HCCI) by population density ....................... 172
Figure 3.9: Household dairy commercialization index (HDCI) by population density .............. 173
Figure 3.10: Livestock products commercialization index (LPCI) by population density ......... 173
Figure A3.1: Photos contrasting Kalama Hills, Machakos District, Kenya ............................... 204
Figure A3.2: Determinants and consequences of smallholder commercialization ..................... 205
Figure A3.3: Effects of population density on smallholder production and income .................. 207
Figure A3.4: Egerton University/Tegemeo Institute panel data households’ locations ............. 208
Figure A3.5: Modified Fisher-Ideal Quantity Indices ................................................................ 209 
xiv
 

KEY TO ABBREVIATIONS

2SLS

Two stage least square

ADC

Agricultural Development Corporation

AEZ

Agro-ecological zones

APE

Average partial effect

CAN

Calcium Ammonium Nitrate

CF

Control Function

CI

Confidence Interval

CIPEV

Commission of Inquiry into the Post Election Violence

CLB

Community Land Board

CPI

Consumer price index

CRE

Correlated random effects

DAP

Di-Ammonium Phosphate

DHS

Demographic and Health Surveys

DLB

District Land Board

FAO

Food and Agriculture Organization

FE

Fixed effects

FEWS

Famine Early Warning System

FGD

Focus group discussion

FIML

Full information maximum likelihood

FIQI

Fisher-Ideal Quantity Index

xv
 

GIS

Geographic Information System

GPS

Geographic Positioning System

GRUMP

Global Rural-Urban Mapping Project

ha

Hectare

HCCI

Household crop commercialization index

HDCI

Household dairy commercialization index

HLPCI

Household livestock products commercialization index

HPMZ

High potential maize zone

IV

Instrumental variable

kg

kilogram

km

kilometer

KNBS

Kenya National Bureau of Statistics

KSh

Kenya Shilling

LCB

Land Control Board

LDSB

Development and Settlement Board

LGP

Length of the growing period

LQI

Laspeyres Quantity Index

MSU

Michigan State University

NCPB

National Cereal and Produce Board

NLC

National Land Commission

NLP

National Land Policy

NPP

Net primary productivity

xvi
 

OLS

Ordinary least square

PQI

Modified Paasche Quantity Index

RE

Random effects

SE

Standard error

SSA

Sub-Saharan Africa

 
 

 

xvii
 

CHAPTER 1: BACKGROUND
1.1

Introduction
Reducing poverty and hunger have been a critical policy concern in most of the sub-

Saharan African countries for the past half-century. Governments and development agencies
have experimented with a series of alternative approaches for addressing poverty. Yet, poverty
still remains pervasive. In 2005, more than 40 percent of Sub-Saharan Africa’s population was
estimated to be below the poverty line (World Bank, 2006). Based on sub-Saharan Africa’s
natural resource endowment, an agricultural-led growth strategy has been touted as the only way
for rapid and sustained reductions in poverty in this region. Africa has for a very long time been
considered a continent with vast open lands where population pressures are rarely felt
(Bilsborrow, 1987; Wood, 2003). As early as 1900, the population density of sub-Saharan Africa
was estimated at 4.4 persons per Sq. Km, contrasted with 38.2 for South Asia, and 62.9 for
Europe (Herbst, 2000). This implied that the continent is characterized by high land-to-labor
1

ratios (Fenske, 2010). These historical facts have persuaded the proponents of agricultural-led
growth to argue that sub-Saharan Africa should pursue a development strategy more like that of
a “land-abundant” America than that of a “land-scarce” Asia (Wood and Mayer, 2001).
This early literature on land abundant Africa continues to inspire the thinking that land is
easily and cheaply accessible in this region. Recently, Fischer and Shah (2010) reported that subSaharan Africa has about 202 million hectares of the 446 million hectares of uncultivated arable
                                                            
1

Hopkins (1973) argued that because of the high land-labor ratios, wealth and power in Africa
was traditionally measured in terms of “men” rather than in “acres”. Consequently, the precolonial authorities were keen to attract more people with whom to subdue nature; strangers
could acquire land indefinitely for token payments (Austin, 2009).
1
 

land in the world. Similarly, the region is reported to possess an enormous yield gap in staple
grains (Fischer et al., 2009; Deininger and Byerlee, 2011; Deininger et al., 2011). Consequently,
this land-abundant-Africa thesis has set a platform for a renewed interest in what has been
referred to as the “unutilized” land in sub-Saharan Africa triggered by the rising demand for food
and fuel as well as the prevailing extreme weather events (Hertel, 2011).
However, recent turn of events cast doubts about the land abundance hypothesis and the
viability of an agriculture-led development strategy in Africa. First, population densities in this
region are much higher than they were some two centuries ago. Evidence exist showing that
there has been a gradual and a steady decline in mean farm size as rural population growth has
outstripped the growth in arable land over the past 50 years. A substantial fraction of Africa’s
rural population lives in relatively densely populated areas where land scarcity is likely to
preclude an extensive dimension of agricultural growth (Jayne and Muyanga, 2012). Data from
Columbia University’s Global Rural-Urban Mapping Project indicate that the proportion of the
2

rural population living in areas exceeding 250 persons per km is of similar or greater magnitude
in Nigeria, Rwanda, Burundi, Uganda, and Malawi, which together with Kenya account for
roughly 35 percent of sub-Saharan Africa’s total population. As a result of the growing
population densities, about half or more of Africa’s smallholder farms are estimated at or below
1.5 hectares in size with limited or no potential for area expansion (Jayne et al., 2003).
Second, even in countries characterized by low population densities and high aggregate
land endowments, there are inequalities in land access at the household level. For example,
studies have found that despite high aggregate land endowment in Northern Mozambique, land
access inequalities at the household level persist (Marule, 1998; Bruck and Schindler, 2009). The
region has the highest and fastest-rising rate of population growth and the distribution of people
2
 

across the continent is quite uneven. Consequently, to infer that the region has no problems of
rural land shortage would be quite wrong (Shipton and Goheen, 1992).
The third way in which times have arguably changed since the nineteenth century is that
a high proportion of farmers perceive that it is not possible for them to acquire more land
through customary land allocation procedures (Jayne et al., 2009). For example, in the densely
populated areas of Kenya, a considerable proportion of young men are start their families
without inheriting any land from their parents, forcing them to either commit themselves to offfarm employment, buy or lease land (Yamano et al., 2009). Landholding sizes have become so
small to an extent that further subdivision is not feasible.
What do such small landholding sizes in the context of increased inaccessibility to land
and limited off-farm employment opportunities mean for feasible poverty reduction strategy in
Africa? First, evidence shows the role of smallholder-led agricultural strategy in households’
food insecurity and poverty reduction. For example, the Asian green revolution was a small farm
phenomenon; over 80 percent of farms in India, Bangladesh, Indonesia, China, Japan and Viet
Nam are less than two hectares (Johnston and Kilby, 1975; Mellor, 1995). This revolution was
broad based and contributed greatly to rural poverty reduction in these Asian countries. For
agricultural-led growth process to substantially reduce poverty, it must be inclusive such that a
large percentage of the smallholder population is able to participate in the process in order to
achieve the linkage effects/multipliers associated with structural transformation (Jayne et al.,
2003; Vollrath, 2007). Smallholders tend to spend their incomes on locally produced goods and
services, therefore stimulating the rural non-farm economy and creating additional jobs. A
fundamental element of the structural transformation process is smallholder commercialization -a transition from subsistence to market-oriented patterns of production and input use.
3
 

Smallholder commercialization refers to a cycle in which farmers intensify their use of
productivity-enhancing technologies on their farms, achieve greater output per unit of land and
labor expended, produce greater farm surpluses, expand their participation in markets, and
ultimately raise their incomes and living standards. While there has been general consensus of
the need to improve smallholder agricultural production in Africa, little attention has been
devoted to understanding whether agricultural and land institutions and policies in this region are
compatible with land allocation patterns capable of generating broad based inclusive agricultural
growth as was achieved in much of Asia.
Second, evidence demonstrates the efficiency advantages of smallholder family owned
farms over large farms (Schultz, 1964; Hayami and Otsuka, 1993; Binswanger et al., 1995;
Vollrath, 2007; Hazell, 2011). The defining feature of family farms is the reliance on family
labor instead of hired labor. Small farm owners reside on the farm, manage the farm themselves,
and are aided by other family members who do not need a lot of supervision to work
(Binswanger-Mkhize et al., 2009). Consequently, of low cost of family labor and labor
supervision, family farms usually exhibit labor-intensive production practices (Hazell et al.,
2010). This is the main reason why family farms are considered more efficient than large farms.
Even though it is more difficult for family farms to access input and output markets, financing,
technical assistance, and information compared to larger farms, empirical studies have shown
that such disadvantages are offset by the advantages in terms of labor incentives (BinswangerMkhize et al., 2009). Moreover, such disadvantages can be countered if small farmers coordinate
their efforts through marketing and credit cooperatives (World Bank, 2005). Nevertheless, in the
recent years there are some pessimist undertones that still question the feasibility of smallholderled agricultural growth in Africa (Collier, 2008, 2009; Collier and Dercon, 2009).
4
 

Currently, it is not at all clear how a smallholder-led agricultural strategy must be adapted
to address the limitations of very small and declining farm sizes and increasingly reduced
prospects of land accessibility in the densely populated areas that are dependent on rain-fed
production systems with only one growing season per year. While the African smallholder
farmer has endured for years with unpredictable markets and weather conditions, there is no
greater challenge to his ingenuity and resilience than these challenges associated with the
mounting human population pressure. While there is broad acceptance of the need for improved
smallholder agricultural productivity and commercialization in sub-Saharan Africa, there is
significant debate about how to most effectively achieve it. Agricultural growth can be achieved
through either the intensification of production on existing land, or by bringing more land into
cultivation, or some combination of the two. The question then becomes, which should be the
priority focus for governments and development partners?
This dissertation is motivated by the need to identify and evaluate the available
development strategies for the increasingly densely populated countries of sub-Saharan Africa.
Policy issues to be addressed revolve around investigating how smallholders are coping with
shrinking farms sizes; identifying potential ways of improving smallholder access to land;
examining whether most farms are becoming, or have already become, “too small” to generate
meaningful production surpluses and participate in broad-based inclusive agricultural growth
processes given existing on-shelf production technologies; and examining whether there is scope
for agricultural intensification.
The dissertation is divided into two essays as follows. The first essay is entitled Smallholder
Land Access in Kenya: Are smallholders farming themselves out of near-landlessness? This
essay seeks to examine how, in the absence of major non-farm employment growth and limited
5
 

outmigration to urban areas, the current formal and informal land institutions are playing out in
facilitating smallholder access to land. The essay analyses the current and potential modes of
land acquisition in Kenya to assess the feasibility of achieving smallholder productivity growth
through improved land access. The study is inspired by two factors. First, the recent literature
indicating that Africa is characterized by landholding inequalities -- manifesting as land
underutilization on large farms and major constraints on farm income stemming from land
constraints in densely populated smallholder areas -- making redistributive land reforms a
potentially attractive strategy for improved smallholder land access. The second factor
motivating this study is the increasing number of medium scale (emergent) farmers in many
African countries over the last decade. Yet, the processes behind their growth have remained
unclear. Is this growth driven by farmers who began their farming careers as smallholder now
transitioning to a larger scale of production through the capital and assets accumulation; a
precursor to the inclusive agricultural-led structural transformation? Or is the growth driven by
land institutions and policies that encourage investment in land acquisitions by individuals from
non-agricultural employment sector signaling elite land capture? All these questions are
addressed in this study.
The second essay is entitled Effects of human population density on smallholder agricultural
production and commercialization. This essay assesses the impact of rising population densities
on input demand, output supply, household income, and smallholder commercialization. The
overarching objective of this study was to examine how rising population pressure affects
smallholders’ production, commercialization and household incomes. The study is motivated by
the need to understand the nature and magnitude of emerging land constraints in African
agriculture using Kenya as a case study. Using data from five panel surveys on 1,169 small-scale
6
 

farms over the 1997–2010 period, we use panel econometric techniques to determine how
increasing rural population density is affecting farm household behavior and livelihoods. The
estimation strategy deals with the potential endogeneity of population density in input demand
and output supply equations using a two-stage control function approach.
To my knowledge, there has been very little recognition of the potential challenges
associated with increasingly densely populated and land-constrained areas of rural Africa. Nor
has there been sufficient discussion of how institutions and policies relating to land access would
need to be modified to achieve inclusive smallholder-led agricultural growth leading to rural
poverty reduction.
Kenya is a useful case study to examine these issues, given that it is one of the more densely
populated countries in the region and may therefore provide an advance picture of the dynamics
that other countries in the region are likely to be experiencing in the not too distant future. The
2

country covers an area of approximately 582,646km , comprising 97.8 percent land and 2.2
percent water surface (Republic of Kenya, 2009). Of the land area, only 16 percent is classified
as medium to high potential. The remaining land is mainly arid or semi-arid. Figure 1.1 presents
the population growth while Table A1.1 shows the population densities in the original 41
districts in Kenya. While population growth rate has declined and stabilized at about three
percent (Figure 1.1), the population density has more than tripled increasing from an average of
19 persons per square kilometer in 1969 to about 66 persons per square kilometer in 2009 (Table
A1.1). Table A1.2 presents the 20 percent most densely populated districts in the country.
It is important to note that some of the districts, especially in the agricultural high potential
regions, are more densely populated than Nairobi was in 1969. Over half of Kenya’s rural
population lives in areas exceeding 250 persons per square kilometre (Figure 1.2). Cultivated
7
 

average land per person in agriculture has declined from 0.462 hectares in the 1960s to 0.219
hectares in the 2000-08 period. A similar picture emerges from comparisons in mean farm size
within the small-scale farming sector over time. A nationally representative survey of Kenya’s
small-scale farm sector in 1977 carried out by the Central Bureau of Statistics reports mean farm
size ranging across provinces from 2.10 to 3.48 hectares. By contrast, mean farm size in Egerton
University’s nationwide surveys from 1997 to 2010 show mean farm size to be 1.86 hectares per
farm; these longitudinal surveys show a decline in farm size even within that 13-year period.
 

Figure 1.1: Population growth rates in Kenya

 

Note: For interpretation of the references to color in this and all other figures, the reader is
referred to the electronic version of this dissertation
Data sources: [A] World Bank: http://data.worldbank.org/indicator/SP.POP.GROW ; [B] World
Bank: http://data.worldbank.org/indicator/SP.RUR.TOTL.ZG?page=5; and [C] United Nations,
Department of Economic and Social Affairs: http://esa.un.org/unpd/wup/index.htm

8
 

Figure 1.2: Population density in Kenya

 

Source: Longabaugh, S. 2008. LandScan High Resolution Global Population Data Set
copyrighted by UT-Battelle, LLC, Oak Ridge National Laboratory.

9
 

1.2

Key findings

1.2.1

Smallholder Land Access in Kenya: Are smallholders farming themselves out of
near-landlessness?
The results point to four key findings. First, the qualitative results from the focus group

discussions indicate that access to land is becoming a binding agricultural production constraint
in the densely populated regions of the Kenya. Even in the areas considered relatively land
abundant like in the lowlands regions, land sizes are swiftly declining due to mounting
population pressure. Not only are household farms shrinking in densely populated regions, soil
fertility is also rapidly deteriorating due to nutrients mining and degradation associated with
reduced fallows. Land conflicts among neighbors and siblings over boundaries and contested
inheritances are on the rise due to increasing land scarcity. Increasing land scarcity is also
triggering increased incidences of fraudulent land deals as a result of poorly drawn land sale
agreements and proliferation of fake land title deeds consequently increasing transaction costs in
land markets.
Second, the qualitative data show that the customary land inheritance from parents and
land sales and rental markets still remain the most important ways through which smallholder
farmers get access land in rural Kenya. The results indicated that there are not unallocated lands
and/or common grazing lands in both the low and highly densely populated areas. These
customary land transfer practices from parents to male children have led to land subdivisions
resulting in tiny landholding especially in the land constrained regions of the country. The land
subdivision problem is likely to be compounded by the new constitutional requirement providing

10
 

for equal treatment of children regardless of gender as far as inheritance of family assets is
concerned.
Third, while migration out of densely populated areas is considered a potential avenue to
ease land pressure in the land constrained regions, the results show outmigration trends are very
low. Rural to rural migration requires financial ability which the smallholder farmers facing land
constraints are lacking. Migration is also being inhibited ethnicity and cultural factors. While
attachment to ancestral lands limits outmigration, access to land in most of the regions in Kenya
is tied on one’s ethnic identity. The ethnicity problem was further aggravated by the 2008 postelection tribal land conflicts. Limited non-farm employment opportunities also impede rural to
urban migration in Kenya. It is important to mention that in most cases the migrating individuals
still retain hold of the land they own in the previous locations. Consequently, outmigration is not
easing land scarcity pressure in the densely populated areas of the country.
Fourth, the empirical results from the medium scale farmers’ survey show that majority
of them used lateral entry into medium-scale farming status. They attained their current farming
status by acquiring land from savings from non-farm, largely urban jobs; only a minority was
primarily engaged in agriculture prior to achieving medium-scale farming status. A big
proportion of them owned on average over two times more land than they were using for
agriculture, implying a high degree of land owned for speculative purposes and/or an inability of
farmers in this size category to make productive use of the land they owned. In terms of crop
productivity, the agricultural entry group seems to be more productive than the lateral entry
counterpart. Another interesting finding is that these emergent farmers are generally more
productive in terms of total production and production per hectare compared to the smallholder
farmers. This study, therefore, suggests that it is primarily individuals with the economic and
11
 

political capital conferred through public sector jobs that are able to navigate the land
administration system to access land.

1.2.2

Effects of population density on smallholder agricultural production and
Commercialization in rural Kenya
The overall picture emerging from these results so far is that land is becoming an

increasingly constraining factor of production and that smallholder agriculture farming practices
in the areas of high population density are distinctly more land-intensive. Consequently, there is
a rising strain on rural livelihoods in the densely populated rural areas due to land pressures and
declining farm sizes. Inputs and output agricultural intensification, household incomes and
smallholder commercialization rise with population density up to about 600-700 persons per
2

km ; beyond this threshold, rising population density is associated with sharp declines in
agricultural intensification and commercialization. For example, the intensity of purchased inputs
2

use increase with population density up to about 600 persons per km and declines after that
point. Crop and farm production intensification increase with population density up to about 700
2

persons per km and thereafter fall. Household aggregate income and non-farm income increase
2.

with population density reaching a maximum level at 600 persons per km

The empirical results also show that smallholder farmers allocate more of their shrinking
land to non-maize crop and sell a greater proportion of their production as population density
increases. The proportion of land allocated to non-maize crop increases with population density
2

up to about 870 persons per km and declines afterwards. Household crop commercialization as
12
 

measured by the proportion of marketed crop output also increases with population density
2

reaching a maximum at about 620 persons per km and declines thereafter. It seems farmers alter
production patterns to make the best out of their shrinking land resource by switching to high
value enterprises such as production of fresh fruits and vegetables, dairy and poultry products.
These results indicate that smallholder landholding sizes are gradually declining in Kenya
as in much of sub-Saharan Africa. Currently about 14 percent of Kenya’s rural population
2

resides in areas exceeding the 600 persons km population density threshold. Another 20 percent
of the rural population is fast approaching this threshold. The results also show that increased
access to input markets, passable roads and other physical infrastructural facilities considerably
influence the degree of smallholder production and commercialization.

 
 
 
 
 
 
 
 
 
 

 

13
 

 
 
 
 
 
 
 
 
 
 
 
 
 

APPENDIX

14
 

Table A1.1: Kenya's population density by province and districts
Density (population/sq. Km)
1969
1979
1989
1999
19
27
45
49
KENYA
Old District
Province
733
1,184
1,906
3,083
Nairobi
127
178
294
283
Central
Kiambu
151
218
378
442
Kirinyaga
147
197
330
309
Muranga
227
331
543
376
Nyandarua
55
72
125
148
Nyeri
108
146
215
198
11
16
29
30
Coast
Mombasa
1,129
1,559
2,992
3,039
Kilifi
24
34
59
65
Kwale
25
35
56
60
Lamu
4
7
12
12
Taita Taveta
6
9
15
14
Tana River
1
2
5
5
12
18
27
30
Eastern
Embu
63
93
76
159
Kitui
11
15
27
27
Machakos
50
72
120
118
Meru
65
91
122
155
Marsabit
1
1
2
2
Isiolo
2
3
6
6
2
3
4
8
North
Eastern
Garissa
1
3
4
9
Mandera
4
4
5
10
Wajir
2
2
3
6
15
 

2009
66

% change in density since 1969
1979
1989
1999
2009
41
141
162
253

4,515
333
569
357
397
184
208
40
4,292
88
79
16
17
6
37
183
33
139
177
4
13
18

62
40
44
34
46
32
35
42
38
40
40
89
33
82
43
47
35
45
39
87
44
52

160
131
150
125
139
130
99
152
165
142
126
231
124
280
115
20
137
141
87
190
224
84

321
122
192
111
65
171
83
163
169
169
141
224
123
257
143
151
139
137
136
239
235
292

516
162
276
143
75
237
92
252
280
261
216
353
157
374
197
189
195
180
171
548
588
840

14
39
12

100
11
62

143
46
82

508
164
270

866
980
668

Table A1.1 (cont’d)

Nyanza
Kisumu
Siaya
South Nyanza
Kisii
Rift Valley
Baringo
Keiyo/Marak
wet
Kajiado
Kericho
Laikipia
Nakuru
Nandi
Narok
Samburu
Trans Nzoia
Turkana
Uasin Gishu
West Pokot

1969
168
192
151
115
305
12
15
53

Density (population/sq. Km)
1979
1989
1999
218
365
348
231
405
386
188
315
284
159
249
245
392
687
654
18
35
38
19
33
37
49
89
94

2009
432
464
333
325
790
55
50
122

% change in density since 1969
1979
1989
1999
2009
29
117
107
156
20
111
101
142
24
108
88
120
38
117
114
184
29
126
115
159
47
190
216
353
26
122
149
243
(7)
69
79
132

4
7
17
19
31
73
345
373
97
128
221
172
299
32
128
78
7
14
30
34
42
102
331
384
39
70
160
158
214
80
312
308
72
104
197
201
261
43
172
177
7
12
28
30
47
68
303
328
3
4
7
7
11
11
102
106
50
104
210
231
328
109
320
363
2
2
3
7
12
(14)
33
173
57
90
173
186
267
57
203
226
9
17
32
68
56
92
260
658
160
221
387
404
522
38
142
153
Western
Bungoma
114
166
312
334
453
46
174
193
Busia
118
176
315
326
439
49
167
175
Kakamega
218
288
486
501
618
32
122
129
Source: Republic of Kenya. Kenya National Bureau of Statistics, Kenya Population and Housing Census Reports (various).
16
 

700
209
500
451
260
580
222
558
418
368
522
226
298
271
183

Table A1.2: The 20 percent most densely populated districts in Kenya (new districts)
Province
New District
Rural
Rural area
Density
Rural area
population
(sq. km)
prop.
Western
Emuhaya
135,723
134
1,011
0.77
Western
Hamisi
148,259
156
948
1.00
Western
Vihiga
96,535
104
931
0.52
Nyanza
Kisii Central
283,117
336
844
0.93
Nyanza
Gucha
364,460
444
821
0.96
Nyanza
Manga
87,859
111
789
1.00
Nyanza
Nyamira
263,201
338
779
0.85
Central
Githunguri
128,643
167
772
0.96
Nyanza
Gucha South
146,307
193
760
0.94
Nyanza
Masaba
142,987
193
739
0.64
Western
Kakamega South
104,669
144
729
1.00
Nyanza
Kisii South
54,969
78
701
0.62
Western
Butere
229,635
345
667
0.97
Western
Mumias
241,072
429
562
0.73
Western
Kakamega Central
203,513
371
549
0.88
Western
Bungoma South
314,145
579
543
0.87
Western
Bungoma West
209,286
388
540
0.87
Western
Bungoma North
222,573
421
529
0.75
Western
Bungoma East
168,063
327
513
0.81
Western
Kakamega North
200,276
424
473
0.99
Central
Kiambu East
45,074
100
449
0.53
Central
Gatundu
208,784
477
438
1.00
Western
Busia
283,514
650
436
0.95
Nyanza
Kuria East
81,833
188
435
1.00
Western
Lugari
271,700
642
423
0.96
Central
Murang'a South
321,310
763
421
0.74
Western
Teso South
111,762
278
403
0.93
Western
Teso North
72,818
183
399
0.70
Nyanza
Kisumu West
137,975
352
392
0.98
Rift Valley Trans Nzoia West
238,854
622
384
0.83
Data source: Republic of Kenya, 2009 National Population Census Data

  

17
 

REFERENCES

18
 

REFERENCES

Austin, G. 2009. Factor markets in Nieboer conditions: Pre-colonial West Africa, c.1500-c.
1900. Continuity and Change, 24(01):23-53.
Bezemer, Dirk and D. Headey. 2008. Agriculture, Development, and Urban Bias. World
Development 36.8: 1342–64.
Bilsborrow, R. E. 1987. Population pressures and agricultural development in developing
countries: A conceptual framework and recent evidence. World Development, 15(2): 183203.
Binswanger-Mkhize, H.P, Camille Bourguignon, Rogier van den Brink. 2009. Agricultural Land
Redistribution: Toward Greater Consensus, Washington D.C.: World Bank.
Binswanger, H.P., K. Deininger, and G. Feder. 1995. Power distortions revolt and reform in
agricultural land relations. In: Behrman, J. and Srinivasan, T.N. (eds). Handbook of
development economics, Volume III. Amsterdam, the Netherlands: Elsevier Science B.V.
Binswanger, H. P. and J. McIntire. 1987. Behavioral and material determinants of production
relations in land abundant tropical agriculture. Economic Development and Culture
Change 35 (1): 73–99.
Binswanger, Hans P. and Vernon W. Ruttan. 1978. Induced Innovation: Technology, Institutions
and Development. Baltimore: Johns Hopkins University Press.
Bruck, Tilman and Kati Schindler. 2009. Smallholder Land Access in Post-War Northern
Mozambique. World Development, 36(8): 1379-1389.
Byerlee, Derek and Alain de Janvry. 2009. The Politics of Hunger. Letter to the Editor. Foreign
Affairs (March/April).
Bryceson, Deborah Fahy and Vali Jamal editors. 1997. Farewell to Farms: De-Agrarianisation
and Employment in Africa. Aldershot, Hampshire: Ashgate Publishing.
Christiaensen, L., L. Demery and J. Kühl. 2011. The (Evolving) Role of Agriculture in Poverty
Reduction: An Empirical Perspective, Journal of Development Economics, 96(2): 239254.
Collier, Paul. 2008. The Politics of Hunger. Foreign Affairs November/December.
Collier, Paul. 2009. Africa’s Organic Peasantry: Beyond Romanticism. Harvard International
Review Summer: 62-65.
19
 

Collier, P. and S. Dercon. 2009. African Agriculture in 50 Years: Smallholders in a Rapidly
Changing World. Paper presented at the expert meeting on How to Feed the World in
2050, FAO: Rome, Italy, 24-26 June 2009.
Deininger, K. and D. Byerlee. 2011. The Rise of Large Farms in Land Abundant Countries: Do
They Have A Future? Policy Research Working Paper No. 5588, World Bank,
Washington, DC.
Deininger, K. and D. Byerlee, with J. Lindsay, A. Norton, H. Selod, M. Stickler. 2011. Rising
Global Interest in Farmland: Can it Yield Sustainable and Equitable Benefits? World
Bank, Washington, DC.
Eastwood, R., M. Lipton, and A. Newell. 2010. ‘Farm size’, in P.L.Pingali and R.E. Evenson
(eds) Handbook of Agricultural Economics, Elsevier: Amsterdam.
Ellis, F. 2005. Small Farms, Livelihood Diversification, and Rural–Urban Transitions: Strategic
Issues in Sub-Saharan Africa. Paper presented at The Future of Small Farms, 26-29 June.
Wye College, London.
Fenske, James. 2010. Does Land Abundance Explain African Institutions? Unpublished Ph.D.
Dissertation, Department of Economics, Yale University, New Haven, CT. Available at
http://mpra.ub.uni-muenchen.de/35294/2/FenskeLandAbundance.pdf [accessed January
28, 2012].
Fischer, R.A., D. Byerlee, and G.O. Edmeades. 2009. Can Technology Deliver on the Yield
Challenge in 2050? Report prepared for the Expert Meeting on How to Feed the World in
2050, Food and Agricultural Organization of the United Nations, Economic and Social
Development Department.
Fischer, G and M. Shah. 2010. Farmland Investments and Food Security: Statistical Annex.
Report prepared under the World Bank and International Institute for Applied System
Analysis (IIASA) Contract, Luxembourg, Austria.
Haggblade, Steven. 2009. Forgotten Farmers. Harvard International Review Winter: 7.
Hayami, Yujiro and Keijuro Otsuka. 1993. The Economics of Contract Choice: An Agrarian
Perspective, Oxford University Press.
Hayami, Yujiro and Vernon W. Ruttan. 1985. Agricultural Development. 2nd ed. Baltimore:
Johns Hopkins University Press.
Herbst, J. 2000. States and Power in Africa: Comparative Lessons in Authority and Control.
Princeton University Press.
20
 

Hertel. T. 2011. The Global Supply and Demand for Agricultural Land in 2050: A Perfect Storm
in the Making? American Journal of Agricultural Economics 93(2):259-275.
Hazell, Peter. 2011. Five Big Questions about Five Hundred Million Small Farms. Paper
presented at the IFAD Conference on New Directions for Smallholder Agriculture 24-25
January, 2011, IFAD Headquarters, Rome, Italy. Available at
http://www.ifad.org/events/agriculture/doc/papers/hazell.pdf [accessed April 9, 2012].
Hazell, P., C. Poulton, S. Wiggins, and A. Dorward. 2010. The Future of Small Farms:
Trajectories and Policy Priorities. World Development, 38(10): 1349-1361.
Hazell, P., A. Dorward, C. Poulton, and S. Wiggins. 2007. The Future of Small Farms for
Poverty Reduction and Growth. International Food Policy Research Institute 2020
Discussion Paper No. 42. Washington, D.C.: IFPRI.
Headey, D., D. Bezemer, and P. Hazell. 2010. Agricultural Employment Trends in Asia and
Africa: Too Fast or Too Slow? World Bank Research Observer 25.1: 57-89.
Hopkins, A. G. 1973. An economic history of West Africa. Columbia University Press, New
York.
Jayne, T.S. and Milu Muyanga. 2012. Land constraints in Kenya’s densely populated rural
areas: implications for food policy and institutional reform. Food Security, 4(3): 399421.
Jayne, T.S., B. Zulu, G. Kajoba, and M. Weber. 2009. Access to Land, and Poverty Reduction in
Rural Zambia: Connecting the Policy Issues. Working Paper 34. Food Security Research
Project, Lusaka, Zambia.
Jayne, T.S. T. Yamano, M. Weber, D. Tschirley, R. Benfica, A. Chapoto, and B. Zulu. 2003.
Smallholder Income and Land Distribution in Africa: Implications for Poverty Reduction
Strategies. Food Policy, 28(3): 253-275.
Johnston, B. F. and P. Kilby. 1975. Agriculture and Structural Transformation: Economic
Strategies in Late Developing Countries. New York: Oxford University Press.
Lipton, Michael. 2009. Land Reform in Developing Countries: Property Rights and Property
Wrongs. Abingdon: Routledge.
Lipton, M. 2005. The Family Farm in a Globalizing World: The Role of Crop Science in
Alleviating Poverty. IFPRI 2020 Discussion Paper No. 40. Washington, D.C.: IFPRI.
Marule, H. F. 1998. Land-poor in a ‘‘land-abundant‘‘ setting: Unravelling a paradox in
Mozambique. Unpublished MSc. Research Paper submitted to Michigan State University.
21
 

Available at http://ageconsearch.umn.edu/bitstream/11089/1/pb98de01.pdf [Accessed on
February 18, 2013].
Mayer, Jörg, and Adrian Wood. 2001. South Asia’s Export Structure in a Comparative
Perspective. Oxford Development Studies, 29 (1): 5-29.
Mellor, J. W. 1995. Agriculture on the Road to Industrialization. International Food Policy
Institute, Johns Hopkins University Press: Baltimore, MD.
Republic of Kenya. 2009. Sessional Paper No. 3 of 2009 on National Land Policy. Ministry of
Lands. Government Printer: Nairobi.
Schultz, T.W. 1964. Transforming Traditional Agriculture. New Haven, CT: Yale University.
Press.
Shipton, Parker and Mitzi Goheen. 1992. Understanding African Land-Holding: Power, Wealth,
and Meaning. Journal of the International African Institute, 62(3): pp. 307-325.
Vollrath, D. 2007. Land Distribution and International Agricultural Productivity. American
Journal of Agricultural Economics 89(1): 202-216.
Wiggins, S., J. Kirsten, and L. Llambi. 2010. The Future of Small Farms, World Development,
38: 1341-48.
Wood, Adrian. 2003. Could Africa be like America? In B Pleskovic, N Stern (eds) Annual Bank
Conference on Development Economics, World Bank and OUP.
World Bank. 2007. World Development Report 2008: Agriculture for Development. Washington
DC: World Bank.
World Bank. 2006. World Development Report 2006: Equity and Development. Washington,
D.C.: World Bank.
World Bank. 2005. World Development Report 2006: Equity and Development. Washington,
DC.
Yamano, Takashi, F. Place, W. Nyangena, J. Wanjiku and K. Otsuka. 2009. Efficiency and
Equity Impacts of Land Markets in Kenya. In The Emergence of Land Markets in Africa,
eds. Stein T. Holden, Keijiro Otsuka, and Frank Place. Washington DC: Resources for
the Future.
 

22
 

CHAPTER 2: SMALLHOLDER LAND ACCESS IN KENYA: ARE SMALLHOLDERS
FARMING THEMSELVES OUT OF NEAR-LANDLESSNESS?
2.1

Introduction
Smallholder farmers constitute the bulk of agricultural producers in sub-Saharan Africa.

They constitute more than 70 percent of the farming community in this region and majority of
them are poor (World Bank, 2007; Wiggins et al., 2010). Based on sub-Saharan Africa’s natural
resource endowment, an agricultural-led growth strategy has been touted as solution for
reductions in poverty in this region. For example, Africa has been considered a continent with
vast open lands (Bilsborrow, 1987; Wood, 2003). However, recent studies cast doubts on the
land abundance hypothesis (Jayne and Muyanga, 2012). Each day, the Africa smallholder
landholding shrinks as population density rises (Ellis, 2005). The land resource is not only
shrinking but also degrading as a result of reduced fallows (Niasse, 2011).
The rising population growth represents both production opportunity and constraint to
smallholder farmers in this region. On one hand, the increasing population signifies growing
demand for food and high food prices thus creating “pull” for increased smallholder production.
On the other hand, it results in diminishing farm sizes that may render many smallholders
incapable of responding to the expanding food markets. Evidence shows that there has been a
gradual decline in mean farm size as a result of population growth that has outstripped the
growth in arable land over the past 50 years (Jayne and Muyanga, 2012). As a result, a
substantial fraction of Africa’s rural population lives in relatively densely populated areas where
land scarcity is likely to preclude an extensive dimension of agricultural growth.

23
 

Despite these mounting population related challenges, analysts have pointed out that all is
not lost. It is still believed that the sub-Saharan Africa region has the potential to revitalize
smallholder agricultural productivity for reduced poverty and hunger if the appropriate policies
are pursued. The Asian green revolution, for example, was a small farm phenomenon; over 80
percent of farms in India, Bangladesh, Indonesia, China, Japan and Viet Nam are less than two
hectares. This broad based agricultural growth in these countries contributed greatly to rural
poverty reduction (Johnston and Kilby, 1975; Mellor, 1995). While there has been general
consensus of the need to improve smallholder agricultural production in Africa, little attention
has been devoted to understanding whether agricultural and land institutions and policies in this
region are compatible with land allocation patterns capable of generating broad based inclusive
agricultural growth as was achieved in much of Asia.
This study therefore investigates the potential for similar forms of inclusive agricultural
growth using Kenya as a case study. The study specifically investigates whether land institutions
and policies in Kenya are making it possible for a broad-based smallholder led agricultural
growth process as enjoyed in much of Asia. In Kenya, there has been a policy thinking that
agricultural and land reforms accompanied by increased government budget support have the
potential to underpin a revitalized system of smallholder production especially in areas where
land sizes have become too small. Thus, the Vision 2030 (Republic of Kenya, 2008) and the
National Land Policy (Republic of Kenya, 2009) proposes a range of agricultural and land
reforms aimed at revitalizing smallholder agriculture in Kenya.
The study is inspired by two factors. First, the recent literature indicating that Africa is
characterized by landholding inequalities -- manifesting as land underutilization on large farms
and major constraints on farm income stemming from land constraints in densely populated
24
 

smallholder areas -- making redistributive land reforms a potentially attractive strategy for
improved smallholder land access (World Bank, 2002). van de Brink et al. (2006) points out that
countries that have been least successful in reducing rural poverty (e.g. Brazil, Colombia,
Guatemala, and South Africa) are characterized by highly unequal landownership and substantial
public investments in large-scale farming. Evidence shows that two comparable countries in all
other aspects can lead to dramatically different development trajectories if they started with
initial inequality in household landholdings (Deininger and Squire, 1998; Deininger and Olinto,
2000; Acemoglu et al., 2001; Deininger, 2003). For that reason, in countries where farmland is
unequally distributed, land redistribution from the most landed and inefficient users, in some
cases the state, to the land-poor groups is been touted as a feasible strategy in combating food
insecurity and poverty (Lipton, 2009). However, as Binswanger-Mkhize et al. (2009) observes,
disagreements usually persist concerning land redistribution approach: should it be state-led or
beneficiary-led? The mechanisms for land acquisition -- should it be confiscation, expropriation,
negotiation, or market purchase? Who should be the beneficiaries of land redistribution? Who
should pay for the reforms? In this connection, the study aims at examining as to whether the
current institutions and policies have been successful in facilitating smallholders’ access to land,
and if not, how they can be improved. Are the intergenerational land transfers, land markets, and
government land allocation practices enabling smallholders’ access to land?
The second factor motivating this study is the increase in the number of medium-scale
(emergent) farmers in many African countries over the last decade. This refers to farmers
cultivating between 5 to 20 hectares of land. In Zambia for example, while the overall population
of smallholders has increased by 33.5 percent, the number of farmers cultivating between 10 and
20 hectares has grown by 103 percent (Sitko et al., 2013). Another encouraging feature of this
25
 

group is that their overall production is also increasing. Yet, the processes behind the medium
scale farmers’ growth have remained unclear in many countries. Is this growth driven by farmers
who began their farming careers as small-scale farmers, cultivating less than 5 hectares of land
and now transitioning to a larger scale of production through the capital and assets
accumulation? If this is the case, is this a precursor to the inclusive agricultural-led structural
transformation that Johnston and Kilby (1975) and Mellor (1976) hypothesized? Or is the growth
driven by land institutions, policies and public spending patterns that encourage investment in
land acquisitions by individuals from non-agricultural employment sector? The factors driving
the growth of this class of farmers and their implications is a pertinent policy puzzle that this
study intents to unravel.
Consequently, the study attempts to provide answers to the following questions:
1. How are households in densely populated areas responding to shrinking farm
sizes? What are the potential modes of increasing smallholder land access in these
areas and what are the challenges associated with these modes if any?
2. How do land policies influence land distribution through direct government land
allocations and operations of the land markets? Do land allocations and land
markets help address the emerging smallholder land constraints or do they
facilitate the landed group acquire more land leading to landholding concentration
among the large scale farmers?
3. Are households facing diminishing landholding sizes seeking to, and in a position
to, migrate to relatively more land abundant regions, urban areas or quitting
farming altogether? If so, what are the facilitating factors and if not what are the

26
 

impediments? Are “push” or “pull” factors the more important impetus for
emigration from rural areas?
4. Are households facing diminishing landholding sizes perceive agriculture as a
viable livelihood option in the future and especially for their children? If farming
is not an option in the future, what are their envisaged livelihood strategies?
5. What are the characteristics of the medium-scale farmers in Kenya? What modes
of land acquisition did they embrace to build up their landholding scale to the
current levels? Is their growth driven by a process of smallholder capital
accumulation and investment into area expansion? Or is driven by land
acquisitions by individuals from non-agricultural employment sector? How is the
medium-scale farmers using their land and how is their land productivity
compared to that of the smallholders?
By understanding the various processes by which these households acquired their land,
we can gain insights into the future distributional effects of existing land allocation policies in
Kenya and their impacts on various national policy objectives, such as poverty reduction,
equitable rural development, and household food security.
The rest of the essay is organized as follows: the second section presents an overview of
land issues in Kenya; the third section presents the conceptual framework while the fourth
section presents data sources. Data analysis methods are presented in section five. Section six
presents the results while the seventh section discusses policy implications and concludes.

27
 

2.2

Historical background of land issues and policies in Kenya
This section reviews literature on land issues and policies that are relevant to smallholder

farming system in Kenya. First we discuss the history of the land issues, popularly known to as
the land question, post- and pre-colonial land policies and the challenges encountered under
these policy regimes up to the introduction of the new National Land Policy (NLP) in 2009.
Then we conclude by discussing the elements of the NLP.

2.2.1 The history of the land question in Kenya
The land question in Kenya stems from the colonial administration’s expropriation of
land for establishment of settlement schemes for the settlers at the beginning of the twentieth
century. The colonial administration alienated land, often through compulsory acquisitions
(Kanyinga, 2009). Before the arrival of the settlers, a stable and flexible structure of access and
control of land was in place. Since the expropriation and alienation of land had to be based on
law, a legal framework was established to promote further alienation and protect what had
already been acquired. The colonial administration introduced the Crown Lands legislation in
1902 which stipulated that the Crown had original title to land that had been acquired (OkothOgendo, 1991; Syagga, 2010). This legislation also established the “scheduled” areas, the fertile
highlands, for white settlers and reserves for “natives”. The reserves were mostly in the marginal
and relatively non-productive areas. Mamdani (1996) refers to this action as the creation of
“citizens” (settlers) and “subjects” (natives). The effect was that Africans became tenants of the
Crown, with no more than temporary occupation rights to land (Okoth-Ogendo, 1991).
As Kanyinga (2009) explains, the White Highlands were about three Million hectares of
which about half of that constituted high potential land suitable for cash crop farming and the
28
 

remaining was appropriate for large scale livestock farming. Farm sizes averaged between 400
and 800 hectares in the highlands and above that in the pastoral areas. The White Highlands
occupied about six per cent of the country and shared between 3600 white families. Since about
68 per cent of the country is remote and unsuited for farming, the remaining 26 per cent of the
country’s land that is considered arable was shared by about six million Africans.
This new development changed completely the African indigenous systems of land
acquisition and control. Rather than the clan or kinship networks, individual families evolved as
the important medium of acquiring land (Okoth-Ogendo, 1991; Bruce and Migot-Adhola, 1994;
Verma, 2001). As Kanyinga (2009) explains, the Native Reserves were characterized by
congestion and landlessness as human population continued to grow. This new land system and
the boundaries introduced by the colonial administration could not allow free rural-to-rural
migration to ease population congestion, a practice that was common before the arrival of the
settlers. Land productivity also reduced significantly due to overuse and overgrazing. The
colonial administration paid no attention to the congestion in the reserves. This is because the
congestion triggered mass migration into the highlands by the landless in search for wage labor
thereby providing the settler economy with cheap labor.
The continued neglect of the African reserves and the harsh economic conditions
prevailing in these areas resulted in political unrests that were later addressed in a reform
program introduced in 1956. The colonial government came up with a “Plan to Intensify the
Development of African Agriculture in Kenya”, popularly known as the Swynnerton Plan,
named after then Assistant Director of Agriculture who designed it (Swynnerton, 1955). The
Plan introduced land consolidation, adjudication and registration process. The plan sought to
change the system from the collective control of land to a more individualized system by giving
29
 

individuals control of their individual holdings. It removed the legal racial barriers in ownership
of agricultural land and undertook to promote land purchase by Africans. The Swynnerton Plan
was formalized by the Native Land Tenure Rules of 1956.
Even though on the surface the reforms looked like a plan to increase the native Africans’
access to land, the colonial government had different underlying objectives altogether
(Wasserman, 1976. ; Okoth-Ogendo, 1991; Bruce and Migot-Adhola, 1994; Kanyinga, 2000 and
2009; Anderson, 2005). First, the plan aimed making the natives busy in their farms and thus
divert their attention from participating in the growing peasant uprising, the Mau Mau rebellion.
Second, the plan aimed at creating a stable new class of wealthy African farmers and political
leadership that would cooperate with the colonial governments in opposing the peasant uprising.
To ensure that the land rights granted through the Native Land Tenure Rules of 1956 were not
disturbed, the African Courts (Suspension of Land Suits) Ordinance was passed in 1957 to bar all
litigation regarding land (Syagga, 2010). These rules were later incorporated into the Native
Lands Registration Ordinance of 1959. A new clause was included in these rules that declared
that the first registration was not to be challenged even if the land had been obtained
fraudulently. At the household level, only five persons could be registered as owners of any
parcel of land and holding trust for the other members of the family (Syagga, 2010). In most
cases only the male head of household was registered as owner of the land. As a result of this,
the male headed could mortgage or even sell the land without recourse to other members of the
family. Women and children were effectively excluded from controlling land.
Numerous settlement schemes and land purchase programs were initiated but none of
them was successful addressing the problem of the native Africans’ landlessness (AtienoOdhiambo, 1995; Ogot, 1995; Anderson, 2005; Syagga, 2010). For example, in 1960 a Land
30
 

Development and Settlement Board (LDSB) was established to oversee resettlement schemes for
some 20,000 families through a 7.5 million sterling pounds facility from the World Bank
(Syagga, 2010). The scheme further offered credit facilities to Africans wishing to purchase
farmland in the White Highlands. It intended to buy some 240,000 acres in the White Highlands,
subdivide them into 100-acre parcels and distribute them to a select group of Africans who
would farm alongside the whites (Kanyinga, 2009; Syagga, 2010). During the independence
negotiations, this resettlement scheme program was transformed to the Million-Acre Settlement
Scheme funded by the World Bank and the British government aimed at a rapid transfer of the
departing European settlers farms to the emerging class of African elites. The transfers were on a
willing-seller/willing-buyer basis and the loans could only be given to those who were
financially able to pay on cash basis or in installments (Leo, 1989).
The Million-Acre Settlement Scheme came to a close in 1971. By this time, about 1.25
million acres had been used in resettlement exercise (Syagga, 2010). Since then, no other
settlement program of such a scale has been initiated to address the unequal distribution of land.
As Syagga (2010) notes, politicians and the colonial government loyalists, as well as wealthy
businessmen managed to acquire thousands of acres through these settlement programs
consequently creating a new class of African elites. As all these studies show, these programs led
to elite capture helping only the already landed class to acquire more land. Besides the
beneficiaries of the Million-Acre Settlement Scheme, there were some individuals who bought
land through either tribal-based land-buying companies or through private transactions mainly by
former European farms’ laborers and forest workers who used their savings to buy land from
departing Europeans settlers (Leo, 1989).

31
 

The post-independence years witnessed the grabbing of public land by politically wellconnected personalities as documented in the Report of the Commission of Inquiry into the
Illegal/Irregular Allocation of Public Land, popularly known as the Ndung’u Report (Republic of
Kenya, 2004).Granting of rights to public land tended to favor the political elite and was
basically in exchange of political support. This saw a lot of public land, both alienated and
unalienated, and trust land unlawfully shift into the hands of elites and members of politically
influential ethnic groups (Republic of Kenya, 2004). Many large government alienated land, for
example the Agricultural Development Corporation (ADC) farms that were previously used for
seed production and livestock breeding, were sub-divided and allocated to the political elites. As
the Ndung’u Report (Republic of Kenya, 2004) elucidates, about 90 percent of the originally
60,000 acres of land owned by ADC was allocated to politically well-connected people such as
politicians, judges, lawyers, and influential businessmen. The land was first illegally converted
into settlement schemes under the guise of settling the landless and subsequently illegally
allocated to the economic and political elites. Most of these allocations were undertaken after the
introduction of multiparty democracy in 1991 and thus were basically meant for buying political
patronage. The interests of the landless and the near landless smallholder farmers were ostensibly
ignored in favor of the so called “politically correct” individuals’ and their networks (Republic of
Kenya, 2004). The Ndung’u Commission conservatively estimated the public land that was
allocated illegally at about 246,964.79 hectares. Since the post-independence Constitution of
Kenya has a provision aimed at protecting private property rights, this provision was ironically
extended to protect lands acquired unlawfully.
Both pre- and post-independence skewed land allocation and land grabbing processes
resulted in a number of individuals and families owning large tracts of idle land amidst
32
 

dwindling smallholder landholding. The processes also fueled ethnic tension particularly against
the Kikuyu community that significantly benefitted from the settlement scheme programs and the
tribal-based land buying companies. For example, the Judicial Commission of Inquiry into Tribal
Clashes in Kenya (Akiwumi Commission) report (Republic of Kenya, 1999) and the
Commission of Inquiry into the Post Election Violence (CIPEV) (Waki Commission) report
(Republic of Kenya, 2008) indicate that the Kalenjins in many occasions, and especially during
general elections years, have attempted to flush out Kikuyus from what they regard as their
(Kalenjin) ancestral land. While the Constitution of Kenya provides for the liberty of individuals
to own land anywhere in the republic, such liberties do not seem to exist on the ground.

2.2.2

The post-colonial land policies and administration
The colonial legacy continued to cast its shadow over the post-colonial Kenya. The

immediate independence governments retained the same colonial land policies and institutions
despite recognizing inequitable distribution of land issue as a major post-independence
challenge. There was no motivation to change the colonial land policies. These policies
seamlessly served the selfish interests of the new landed Kenyan ruling elites, just as they did to
the departing white settlers. Legal framework governing land was a collection of very many
incoherent colonial statutes and the land administration was highly centralized. The result was a
long history of inefficient and ineffective land administration and governance system.
Consequently, this led to insecurity of tenure, excessive land fragmentation, disparities in land
ownership, injustices in land distribution, landlessness, environmental degradation, proliferation
of urban informal settlements, and poverty among others (Republic of Kenya, 2009).

33
 

Before the enactment of the new land policy in 2009, land ownership was classified into
three major categories: private (freehold) land, government (public) land, and trust (communal)
land. The private land made up about 20 percent of the country’s land. The land under this
category was either individually or collectively owned and was registered in the name of an
individual or a company. Most of the high-potential agricultural land has already been
adjudicated and registered as freehold. Collective freeholds include group ranches established
under the Land (Group Representatives) Act of 1968. Private land was created from either
government land or trust land through registration following the laid down legal procedures.
Approximately 10 percent of Kenya’s land is under government ownership (Table A2.1). It
comprises two sub-categories, unalienated land (land which had not been leased or allocated)
and alienated land (land which had been leased to a private individual or corporate body, or
which had been reserved for the use of a government entity, or which has been set aside for
another public purpose). Under the old land policies, it was only the President who had the right
to allocate unalienated government lands; the President was allowed by law to delegate the
powers to the Commissioner of Lands. The President or the Commissioner of Lands had no
powers to allocate alienated land. The remaining 70 percent were trust lands (Table A2.2). Trust
lands were held by County Councils on behalf of local communities, groups, families and
individuals in accordance with applicable African Customary Law. Trust land could only be
moved from the communal to private ownership through the formal adjudication processes.
The inefficient land administration coupled with declining and disparities in landholding
and the perceived large farm inefficiencies brought the land issues into sharp focus and
subsequent calls for land reforms during the 1990s. The land administration and management
framework was highly centralized, complex, and bureaucratic. The framework was characterized
34
 

by limited access to land information due to poor record keeping; lack of transparency in its
operations; and weak and ineffective mechanisms for fair, timely and affordable land disputes
resolution system. Among the most critical land reform issue were (and still remains) the
inequitable distribution of land rooted in what has been popularly referred to as “historical
injustices”. While the modes of land access were primarily through inheritance and the market,
access to public land through direct government allocation became a major political patronage
instrument (Republic of Kenya, 1999; Republic of Kenya, 2008).
The perceived large farm inefficiencies, also contributed to the call for land reforms in
Kenya. The land reforms protagonists highlighted underutilization of land, particularly the large
farms, leading to low agricultural productivity in the context of a high poverty and food insecure
country. Most of these large farms are held under the 999-year leases granted to some political
elites immediately after independence in 1963 (Bruce, 2009).
The efficiency of the land sales and rental markets also came under the spot light. It has
been argued that efficiently functioning land markets have the potential to provide land access to
households owning little or no land. Persistence of insecure land tenure systems coupled with
information asymmetry, political interference, bureaucratic inefficiencies, corruption and land
speculation by the political elite was perceived to slow down land market operations (Republic
of Kenya, 2009). As mentioned at the beginning of this section, most of the land in Kenya is held
under the trust lands (customary) system. The land reforms activists argued that economic
growth required the transformation of customary land tenure to private ownership system; either
individual or group ownership system. This would in turn permit the development of a large
impersonal land market that would eventually lead to distributive efficiency in land use. To
prove their point, they made reference to the impressive post-independence intensification and
35
 

commercialization of agriculture in the highlands that they attributed to land registration that
took place in the early independence days. However, doubts have been raised concerning the
distributive efficiency of land markets in Kenya. The country is seemingly highly ethnically
fractured and that land conflicts and tenure insecurity have persisted even in areas where land
registration was accomplished many years ago (Republic of Kenya, 1999; Republic of Kenya,
2008). While the “official” land registration is supposed to confer undisputable land ownership
rights, the authenticity of the past land titling process and the issued title deeds has remained
questionable; paradoxically within the government itself. The Standard, a local daily newspaper,
quotes a Kenya Government Cabinet Minister in charge of Lands and Settlement, Amos
Kimunya, publicly proclaiming that “a title deed is just but mere piece of paper” (Mutua, 2005).
The unavailability of agricultural land has also linked to the existence of the unproductive
marginal lands in the northern parts of the country. These areas lack the requisite infrastructural
facilities and services to boost agricultural productivity and thus have rendered unattractive. The
land reforms debate centered around putting in place an enabling infrastructure and services for
agriculture and livestock development such as passable roads, research, extension services,
finance, marketing, agro processing, rural electrification and farmers’ training centers.
 

2.2.3 The National Land Policy
To address the concerns raised in the foregoing sub-section, the government came up
with the National Land Policy (NLP) (Republic of Kenya, 2009). The NLP, nested in the new
constitution (Republic of Kenya, 2010), provides a new overall framework and defines the key
measures required to address the land administration and access issues. The new land policy
designates all land in Kenya in three distinct categories: public land, community land, and
36
 

private land. A detailed description of what land lies under each category and how it is supposed
to be administered is found in Table A2.3. The policy advocates for the following landholding
overarching principles: (1) equity; (2) non-discrimination; (3) efficiency; (4) productivity; and
(5) sustainability. These include measures to eliminate discrimination, ethnic and gender, in land
access; land taxation to curb inefficient land use and idle land holding and hoarding tendencies;
resolution of historical injustices; improvement of tenure security; efficient utilization of
marginal lands; and efficient land markets.
To implement the National Land Policy, the Government has set up three key land
management institutions: the National Land Commission (NLC), the District Land Boards
(DLBs) and Community Land Boards (CLBs) (Republic of Kenya, 2009; Daily Nation, 2013).
The NLC is an independent constitutional body responsible for land administration functions and
serves as the manager of all public land. The NLC primary task is to provide the public
infrastructure and support for property rights and markets in land. The mandates of the NLC
include, holding title to and manage public land on behalf of the Government; establish and
maintain a register of all public, private and community land in the country; exercise the powers
of compulsory acquisition and development control on behalf of the government and local
authorities; and to collect and manage all land tax revenues except those that are collected by
district-based authorities. The NLC is mandated to conduct investigations into the “historical
land injustices” and recommend appropriate redress. The NLC will execute its mandates through
DLBs, and CLBs. The DLBs are agents of the NLC at the district level and are accountable to
the NLC in the performance of their functions. They consist of democratically elected
community representatives and supported by officers appointed by NLC. Community Land
Boards (CLBs) constitute the third tier of a devolved land administration and management. As a
37
 

break from the past where the community lands (customary lands system then) were held and
managed by County Councils on behalf of local communities, the CLBs are mandated to
administer and manage land held under the community land category. Just as the DLBs, the
CLBs are composed of democratically elected community representatives and supported by
officers appointed by NLC. The elected members of CLBs are people ordinarily resident in an
area as determined by the DLBs in consultation with the affected communities and membership
criteria is expected to consider ethnic diversity, gender and socio-political dynamics.  
Next, we discuss some of the key features of the new land policy and some of the likely
challenges the policy is bound to encounter. While the new policy has reformed the way
community lands are managed and administered, it is not clear whether the policy will allow the
community land ownership system to evolve over time to private (individual, family or group)
ownership (Bruce, 2009). The policy is also unclear about the need to reform the community
(customary) land system, a system that is known to be bedeviled by corruption (authoritarianism
and non-transparency) and gender and ethnic based land access discrimination.
The policy calls for the recovery of public land acquired irregularly and establishment of
a Land Title Tribunal to determine the legitimate ownership. Article 68 of the new Constitution
of Kenya requires Parliament to enact a law that will “enable the review of all grants or
dispositions of public land to establish their propriety or legality”. It proposes establishment of a
legal and institutional framework to facilitate the returning land back to the original owners or,
where this is not possible, providing original owners with compensation. Historical land
injustices are to be resolved using mechanisms such as restitution, resettlement, reparation, and
compensation of historical injustices and claims. A number of issues have been raised
concerning the tackling of the historical land injustices. It is important to note that land initially
38
 

acquired illegally may have changed hands and now in possession of innocent buyers; in some
cases, smallholder farmers. There are fears that what constitutes a redress of first generation
injustice to one group may constitute a second generation injustice to another group.
Concerning the resettlement of the landless, the policy does not suggest how to overcome
challenges encounted in the past resettlement programs. As discussed earlier on, the past
resettlement programs ended up “settling” the political elites and their networks instead of the
deserving landless or near landless groups (Ogot, 1995; Anderson, 2005; Syagga, 2010). It is
important to note that empirical evidence indicates that well-targeted land redistribution
programs have a direct impact on poverty reduction (Binswanger-Mkhize et al., 2009). As Bruce
(2009) suggest, the government should also consider the option of setting up a fund to facilitate
the landless or those in need of resettlement to purchase land from the open markets. This will
lead to scattered patterns of resettlement rather than concentrated patterns that may lead to ethnic
tensions and land conflicts later. The concentrated patterns of resettlements of one community in
what is considered other ethnic group’s territory to a large extent contributed to the 2008 postelection violence in Kenya (Kniss, 2010). In Uganda, similar consequences resulted when the
Bakiga migrants from densely populated southwestern Uganda purchased land in Buganda as a
group (Muhanguzi, 2008; Kato, 2009).
The new policy proposes to facilitate the commercialization of land rights to make land
market operations more efficient and effective and to institute a regulatory framework for land
rental markets. Land registries will be decentralized to facilitate and promote land market
operations in community land ownership system. Increased security of land tenure and the
development of land market is hypothesized to allow those who have more land than they need
to lease it to others thus increasing distributive efficiency. However, as discussed in the next
39
 

section, land markets rarely lead to redistribution of land access from large to smallholders in
countries where there is a highly unequal distribution of land (Deininger et al., 2004). Land
markets alone cannot be counted on to redistribute land from large landowners to the landless or
near landless smallholders (Binswanger-Mkhize et al., 2009).
To address land fragmentation problem, the policy suggests setting up of economically
viable minimum landholding sizes for various agroecological zones and regulation of the manner
in which any land may be converted from one category to another. According to this policy,
measures will be put in place to determine the appropriate land sizes according to use and
productivity in each zone and to provide incentives to stimulate voluntary readjustment of
landholding sizes. However, the implementation of this proposal will not be easy task. First, a
blanket minimum land size for all agricultural enterprises in a given zone is not practical. For
example, land considered too small for maize production may be enough for large scale poultry
or vegetables production. Secondly, this requirement may drive the already small parcels out of
the registry especially when subdivided to sizes way below the set standards.
With regard to access to public land, the policy suggests a comprehensive review of
public land allocation procedures to determine the appropriate amount of land to be allocated to
individuals depending on intended use and agro- ecological zone. Concerning the active long
land leases, especially the 999-year leases, mechanisms will be established for the surrender of
leases going beyond 99 years in exchange for shorter leaseholds. Closely connected to this, the
new policy prohibits non-citizens from holding freehold interests in land. Several questions still
linger about this long term lease conditions. Are they meant to curb the pervasive unproductive
use of land under leaseholds? What is the purpose of the long term leases? Are they meant to
foster foreign direct investments and/or technology transfer? If so, then the appropriate lease
40
 

length should be enterprise specific depending on the level of investment and the duration the
investors need to recoup their costs and earn returns on their investment. If the land policy
concern is improving smallholder access to land, then the government should consider allocating
the land released from the long term leases to the landless and near landless groups. The policy
should also put in place incentives to entice leaseholder who have more land than they need to
sub-lease some of their land to others.
To facilitate the efficient utilization of land and land-based resources, the policy calls for
the establishment of appropriate land taxation system to discourage land speculation. The goal is
the provision of appropriate incentives and sanctions to ensure that land owners use their land
productively and sustainably. The main shortcoming to this proposal is that it fails to
conceptually define what constitutes “productive and sustainable” use in various agroecological
zones and enterprises.
The policy puts in place measures to curb land allocation malpractices and elite capture.
It also takes away some powers from some state organs such as the Presidency and the Ministry
of Lands and transfers them to others bodies such as the National Land Commission and
Communal Land Boards. The policy points out clearly that that it will not be possible for every
person to own land and that the government will facilitate secure access to land for the citizens
through other means, such as leasehold mechanisms. However, the policy certainly leaves many
hard decisions to legislators and implementers preparing key legislations to support the
implementation of the policy and the Land and Environment chapter of the new constitution.

41
 

2.3

Conceptual framework
In this section we develop a framework to analyze the processes behind the smallholder

and medium scale farmers’ land access in Kenya. Distributions of land and land access in any
society in Africa are the outcome of complex historical and geographical processes, and hence, it
should not be surprising if the processes fail to conform to any known economic theory. Farm
households acquire and exercise rights to obtain and maintain access to land or they lose these
rights through multiple paths: as members of social groups or networks, through labor and
investment, and through purchase (Shipton and Goheen, 1992; Berry, 1993). To explore the
various avenues that households use to access land and how successful they have been in
improving smallholder access to land, a framework proposed by de Janvry and Sadoulet (2001)
is used. According to this framework, pathways through which households can access land fall
into four categories: (1) Intra-family transfers such as inheritances, inter-vivo transfers, and
allocation of plots to specific family members; (2) access through community membership and
community open access resource; (3) access through land sales and rental markets; and (4)
access through non-coercive policy interventions such colonization schemes, decollectivization
and devolution, and land market-assisted land reform (Figure A2.1). As de Janvry and Sadoulet
(2001) point out, each of these avenues is a potential object of policy interventions to

improve land access and use efficiency for the smallholders facing shrinking farm sizes.
Historically, family access to land depended on their membership and good standing
within a particular group exerting control over the land. Kinship and ethnic adherence along with
status, gender and seniority determined access and use rights (Berry, 1989; Migot-Adholla and
Bruce, 1994). However, with growing human population densities and the ensuing land squeeze
land access through within community transfers is becoming less and less feasible.
42
 

Most farm households in sub-Saharan Africa gain access to land through intra-family
transfers. Even in countries where land markets are well developed, access to land through
inheritance remains fundamental. The transfer of land from parents to children takes place when
children marry (inter-vivo) and form new productive units or when parents age and eventually
die (post-mortem) (de Janvry and Sadoulet, 2001; Muyanga et al., 2010). Land transfers from
parents to children depend on the number of potential recipients of the parents’ assets. For this
reason, factors such as the number of male (female) children of the parent’s family in patrilineal
(matrilineal) inheritance systems may influence current households’ landholding. In polygamous
families, land is shared equally between wives regardless of their number of children.
Some groups, notably women and children, have in many communities been excluded
from land inheritances processes in sub-Saharan Africa (Cooper, 2008). Females land user rights
most often come through their ties to kin and husbands and are contingent on marriage (Berry,
1993; Smucker, 2002). A senior wife may have stronger rights than a junior wife. A woman's
rights may increase with the length of marriage or with number of children and the rights may
end with divorce, with widowhood, with failure to have sons (Gray and Kevane, 1999). As Gray
and Kevane (1999) further point out, transfer rights are usually limited for women -- they cannot
designate an heir, sell land, or lend land to others. Traditionally, females have no claim on
parental assets (Garg and Morduch, 1998; Morduch, 2000). Unmarried daughters stayed with the
parents but are not entitled to land allocation. In some cases unmarried daughters are allocated
land for farming purposes only, not to own; and are required to surrender the land back to the
parents or brothers upon marriage. Since in most African societies individuals are considered
adult after marriage, unmarried daughters remain “children” and thus required to continue to “eat

43
 

from their mother’s pot”. However, nowadays girls born without male brothers are being
allocated the parents’ land in some communities.
With diminishing land sizes, access to land via family lineage and community
membership is giving way to individualization of property rights and to access to land via land
sales and land rental markets (Pinckney and Kimuyu, 1994; Place and Migot-Adholla, 1998;
Holden, et al. 2009). Land markets have the potential to improve households land access and
production efficiency by equilibrating land and non-land factor ratios across farms (Deininger,
2003). However, in practice the outcomes of land markets in achieving these objects remain
under researched especially in sub-Saharan Africa. The achievement of the positive land markets
outcomes may be thwarted by credit market imperfections, high transaction costs, government
regulations, and ethnicity and cultural related constraints. If land markets are characterized by
high transaction costs and credit markets imperfections, farmers’ ability to access land from
participating in the land markets is directly correlated with their wealth and land endowment
(Deininger and Jin, 2008). In such cases, land markets lead to transfer of land from landconstrained to land-abundant households (Andre and Platteau, 1998; Zimmerman and Carter,
1999; Kranton and Swamy, 1999).
There have been relatively few studies of the performance and impact of land markets in
sub-Saharan Africa (Holden et al., 2009). Land markets, and particularly land rental markets,
have recently been found to be widespread in sub-Saharan Africa, especially in the more densely
populated areas where land is relatively scarce and highly fragmented. In some countries such as
Ethiopia and Mozambique, land sales are prohibited by the law while in Kenya, both land sales
and rental markets are legal and active. In Kenya, land sales transacting parties are required to
seek approval from the Land Control Board (LCB) for private owned land (registered) and
44
 

approval from the government provincial administrators (chiefs) for customary owned land (land
held without title deeds). Evidence shows that between 10 and 20 percent of households in rural
Kenya rent-in land (Wangila et al., 1999; Yamano et al., 2009; Jin and Jayne, 2013). As Jin and
Jayne (2013) observe, land purchases require a much greater up-front payment than renting land
and thus may not work for credit constrained smallholder farmers. However, the practice of
deferring a part or full rent payment until after harvest facilitates smallholders’ participation in
land rental markets (Jin and Jayne, 2013).
A study by Yamano et al. (2009) found out that both land sales and land rental markets
improve equity and efficiency among smallholders in Kenya. More recently, Jin and Jayne
(2013) established that land rental markets transfer land from less efficient to more efficient
smallholders and also improve access to land for households with relatively small farms.
However, they found that the overall income gains accruing to households with the smallest
farms from renting were insufficient to pull a substantial proportion of rural households out of
poverty. In general, there is still little evidence Africa on how land rental and sales markets play
out in Africa and whether facilitate land transfer from the landed group to the land-poor groups.
Are the smallholders able to generate surpluses, accumulate more land and farm themselves out
of smallholder farming and poverty into medium scale and subsequently into large commercial
farming? This is a puzzle this study is set to unravel.
Socio-political capital and networks are also important factors explaining access to land
in certain situations. Access to public and community land or the ability to navigate through
bureaucratic land market systems is concentrated in the hands of those with socio-political
capital and networks conferred through public sector employment. For example, Jayne et al.
(2008) show that households in which the male head is related by blood to the local headman
45
 

have an average of 0.4 hectares more land than other households in Zambia’s smallholder
farming sector. In other situations, elites manipulate the decision-making arena and agenda to
obtain most benefits (Beard and Phakphian, 2009; Bruck and Schindler, 2009). Elite capture is a
function of four factors, namely: disparate access to economic resources, asymmetrical social
positions, varying levels of knowledge of political protocols, different education attainment and
employment status (Platteau, 2004). As earlier pointed out, the pre-and post-independence
settlement schemes meant to settle the landless and public land grabbing in Kenya ended up
benefiting the political elites such as chiefs, politicians, and provincial administrators and their
associates (Republic of Kenya, 2004; Kanyinga, 2009; Syagga, 2010).
Governments can improve smallholder access to land in a number of ways. In some other
situations, governments have intervened to improve the functioning of land markets but
obviously with mixed results. Instruments such as taxes on idle land have the potential to shift
land from the landed but less efficient users to the more efficient but landless or near landless
groups. In post-apartheid South Africa, market-based land reforms implemented by the
government to address the racial unequal distribution of agricultural land on “willing-seller
willing-buyer” basis inadvertently ended up promoting new black elite entry into commercial
farming at the expense of the land-poor smallholders (Lahiff and Cousins, 2005). Lahiff and
Cousins (2005) point out that this approach is likely to succeed if accompanied by
complementary interventions such as investment in infrastructure, availability of new
technologies, reorientation of agricultural research and extension services towards the
smallholders, and establishment of smallholder credit programs. Binswanger-Mkhize et al.
(2009) argues that when poor people are given good farmland and adequate post- settlement
support, they are likely to lift themselves out of poverty.
46
 

In situations where the state owns land that is unutilized or underutilized, the government
can institute transparent and orderly procedures for land allocation to land-poor smallholders
(Munshifwa, 2002; Stambuli, 2002). In situations where state-sponsored redistributive land
reforms have been carried out, the results have been mixed. For example, in Zimbabwe, while
studies show that 68.2 percent of the land went to land-poor, development commentators and
media argue that the process was marked by gross violation of individuals’ property rights and
that the beneficiaries of the process were the political elites (Scoones et al., 2010).
The government can also raise value of underutilized or unutilized but agriculturally
potential land improved public investments in physical infrastructure. This approach was
successfully pursued by southern Rhodesia and Zimbabwe starting in the 1970s with its “growth
point” strategy in the Gokwe area, once cleared of tse tse flies (Govereh, 1999). As Govereh
explains, key public investments in this once abandoned but agro-ecologically productive area
induced rapid migration into Gokwe from heavily populated rural areas, leading to the “white
gold rush” of smallholder cotton production in the 1980s. Investments that increase land
economic value include investments in road infrastructure, schools, health care facilities,
electrification and water supply, and other public goods required to induce migration, settlement,
and investment in these currently under-utilized areas.

2.4

Methods and data sources

2.4.1

Empirical models and estimation strategy

In this section, we present econometric models to be estimated. The objectives of this
exercise are twofold: first, to understand the factors determining the entry pathway into the
medium-scale farming status, and second to understand the correlates of the medium-scale
47
 

farmers’ input use intensities and productivity. In evaluating the determinants of medium
farmers’ outcomes and the influence pathways into medium scale farming have on agricultural
productivity, a model that can be employed is the following:

ln y  X '   I  

(2.1)

where y is the medium scale farmers’ agricultural production outcome (landholding, and input
use and output supply intensities) of interest, X is a vector of exogenous characteristics, and I
is a dummy variable (I = 1 if the individual followed agricultural led pathway in to medium scale
farming and 0 if entered laterally). This approach, however, might yield to biased estimates
because it assumes that pathways into medium scale farming are exogenously determined. Some
unobserved characteristics that influence the probability to follow a particular pathway could
also influence farmer’s agricultural production outcome. Pathway into medium scale farming
may be based on individual self-selection. Farmers that follow agricultural pathway may have
systematically different characteristics from the farmers that entered into medium scale farming
laterally. Unobservable characteristics of farmer and their farm may affect the pathway choice
and the production outcome, resulting in inconsistent estimates in specification (2.1). Similarly,
pathway choice into medium scale farming should not be assumed to have an average impact
over the entire sample of medium scale farmers by way of an intercept shift only. Pathways into
medium scale farming should be allowed to influence other covariates’ coefficients in the
farmers’ production outcome equations.
This motivates the use of an endogenous switching regression model (Lee, 1978;
Maddala, 1983; Willis and Rosen, 1979). Endogenous switching model accounts for both
endogeneity and sample selection issues and allows interactions between pathway into medium
scale farming and other covariates in the input demand and output supply functions. The model
48
 

consists of pathway into medium scale farming equation and medium scale farming outcome of
interest equation. The pathway part is estimated using a standard limited dependent variable
method while the production outcome equation is estimated separately for each group
(agricultural and lateral entry) conditional on pathway into medium scale farming. The
endogenous switching model is specified as follows:

I i*  Z i    i ,

1
I i*  0
I i   if
otherwise
0

ln y1i  X i 1  1i

if I i  1

ln y 2i  X i  2   2i if I i  0
where

I*

Selection equation

(2.2)

Regime 1

(2.3)

Regime 2

(2.4)

is latent variable for pathways into medium scale farming and

I

is its observable

counterpart (equals one if household used agricultural pathway and zero if entered medium scale
faming laterally),

Z is a vector of observed farm and non-farm characteristics determining

pathway into medium scale farming, and



is random error term. The vector

Z

contains unit

as its first element and variables that are hypothesized to influence pathways into medium scale
farming status namely, household demographic factors such as the age, gender and education
attainment of the household head, household size; landholding when the family started farming,
and number of years in the current location. As shown in the conceptual framework, family
history is a good predictor of channels that households follow to increase landholding and
consequently build their production scale. The family history factors relate to the characteristics
of the father to the initial household head. They include the amount of land he owned before sub49
 

division, his level of education, whether he’s alive or not and whether he held community
leadership position. The type of non-farm job also matter. Individual holding public offices or
community leadership position have both the social capital and political clout that enable them to
acquire land (Bruck and Schindler, 2009). In this model, household’s social capital is captured
by the household’s duration in the current location and participation in public as well as private
employment. Land attributes variables include the distance from the plot to the nearest water and
motorable road; land location (in the region whether household head was born or outside); and
mode of land acquisition and land tenure status. Variables extracted from various geographic
information system sources were plunged in to control for land quality in the location where the
household is situated. These include the length of the growing period, net primary productivity,
elevation and slope. District dummies are also included to capture remaining local fixed effects
as well as the survey respondent dummies.
Separate endogenous switching models (equations 2.2 to 2.4) are estimated depending on
how the dependent variable in equations (2.3) and (2.4) is defined. First,

y

is defined as the

household land access (land owned and land rented in) and the area under crops; second,
intensity of fertilizer application per hectare cultivated; and third as crop production per hectare
cultivated and controlled (owned and rented-in). It is important to mention that medium-scale
famers grow an array of crops on one land plot each season and the crop grown vary widely
across agroecological zones. Consequently, it becomes imperative to aggregate the crop outputs
in some manner. To do this, a modification of the Fisher-Ideal index (Figure A2.2) suggested by
Mason (2011) is used. The vector of explanatory variable ( X ) for the landholding estimation
contains unit as the first element, maize price, fertilizer prices, distance to infrastructure facilities
and markets, household demographic variables, household land attributes and agricultural
50
 

potential of the location where the household is located. Data on other output and input prices
(such as wages and land rental rates) was not collected during the survey.
An efficient method to estimate endogenous switching regression models is by full
information maximum likelihood (FIML) estimation (Lee and Trost, 1978; Lokshin and Sajaia,
2004). The FIML method simultaneously estimates the selection equation and the regression
equations to yield consistent standard errors. For the endogenous switching model identification,
there should be at least one explanatory variable in the first stage selection equation that is not
included in the second stage regression (Maddala, 1983). In this case, some variables that are
hypothesized to influence the pathways into medium scale farmers include households’ head
participation in non-farm employment activities and some household historical factors.

2.4.2

Data sources
The study uses data from various sources. The first data set came from a survey involving

about 200 medium scale farmers in Kenya. Medium scale farmers in this study are defined as
farmers cultivating 5 to 50 hectares of land. Based on field experience, this group has little in
common with large-scale commercial farmers in terms of farm size, access to finance, input
application rates, and farm management strategies. The aim of the survey was to get insights into
the characteristics of the medium scale farmers, how they achieved their current scale of
operation, and how efficiently they are utilizing their land. With help of Ministry of Agriculture
officials, the Kenya 47 counties were first stratified into five broad categories depending on
farming systems and agroecological characteristics. Since the focus of the survey was medium scale farmers, four counties (Bungoma, Trans Nzoia, Uasin Gishu, and Nakuru) were randomly
selected from the containing the highest concentration of this type of farmers. With help of local
51
 

Ministry of Agriculture officials in the selected counties, lists of farmers operating between five
and 50 hectares in each county was complied. Owing to costs and logistic reasons, only two to
four divisions in each county were included. The 200 households for interviews were selected
2

randomly from the selected divisions (Table A2.4). The survey instrument captured information
both at the household and land parcel level including, household characteristics and history,
farming patterns, agricultural production, household incomes and asset, and land parcel features.
The 200 medium-scale farmers’ survey instrument captured the geographic positioning
system (GPS) coordinates of each household (Figure A2.3). This made it possible to compliment
the survey data with land and soil quality variables data extracted from various geographic
information systems (GIS) sources. These variables extracted from the GIS sources included the
length of the growing period (LGP) (Fischer et al., 2000); net primary productivity (NPP) ( Zhao
et al., 2005); and elevation (meters above the sea level) and slope (measure of steepness -degrees), both from Wilson et al. (2007). The LGP and NPP are useful indicators of agricultural
potential.
The second data source is the nationwide Egerton University/Tegemeo Institute Rural
Household Survey, a panel dataset tracking roughly 1,300 small-scale farm households in 5
survey waves over the 13-year period from 1997 to 2010. The sampling frame for the panel was
prepared in consultation with the Kenya National Bureau of Statistics (KNBS) in 1997.
Subsequent surveys were conducted in June of 2000, 2004, 2007 and 2010. Over these 5 panel
surveys, 1243 household were able to be consistently located and surveyed. For this analysis,
households in the coastal region of the country were excluded because farming is found to
                                                            
2

Survey instrument available here:
http://aec.msu.edu/fs2/kenya/Land_Access_in_Kenya_Study_Instrument_Final_2012.pdf
52
 

account for a relatively small share of household incomes. This leaves a balanced panel of 1146
households surveyed consistently in each of the five years. The surveys collect information on
demographic changes, movements of family members in and out of the household since the prior
survey, landholding size, land transactions and renting, farming practices, the production and
3

marketing of farm products, and off-farm income-earning activities . This data set was used for
three main purposes. First, the data is used analyze how smallholder landholding sizes evolve
over the panel period. Second, the data is used to examine if land markets lead to equitable land
distribution in rural Kenya. The household landholding Gini index is computed for each panel
survey to examine whether inequality in landholding is reduced through smallholder households’
participation in land sales and rental markets during the 13 years panel period. Third, the 2010
survey data from the districts where the medium-scale farmers’ survey was conducted is used to
compare the medium scale farmers and smallholders in terms of landholding, land use and farm
productivity.
The household level data is complemented by focus group discussions (FGDs)
information collected in the densely populated and land constrained regions of the country
(Table A2.5). The objective of the FGDs was to get insights into smallholders’ perceptions of the
increasing land constraints, their coping mechanisms, emerging nonfarm income earning
4

opportunities, outmigration trends, and households’ future outlook . The FGDs consisted of
between 12 and 15 participants. All the FGDs were similar in composition and included male
and females, young and old farmers and local leaders.
                                                            
3

Each of these surveys instruments, which contain the details of the types of information
collected and used in this study, can be viewed and downloaded at
http://www.aec.msu.edu/fs2/kenya/index.htm.
4
Checklist available here: http://aec.msu.edu/fs2/kenya/FGDs_Checklist.pdf
53
 

2.5

Results

2.5.1

Descriptive results

As a prelude to the econometric models, descriptive results from the medium-scale farmers’
survey and the smallholder panel survey data are reported. The analysis is enriched by qualitative
data from the focus group discussions. The descriptive results are organized in four major
categories reflecting the major study questions, namely: (i) Emerging land constraints with
increasing population densities; (ii) Modes of land access and associated challenges; (iii) Out
migration trends in the densely populated areas; (iv) Future prospects of the households in the
densely populated areas; and (v) Characteristics of the emerging medium-scale farmers.
(i) Emerging landholding constraints with increasing population densities: Access to land
emerged as a key factor limiting agricultural production in the densely populated regions of the
Kenya. Generally, household landholding sizes have declined across the country. Table 2.1
shows that the average landholding in rural Kenya stands at about 2.6 hectares which translates
to about 05 hectares per adult equivalent. Household landholding sizes range from a high of 4.62
hectares in the high potential maize zone to a low of 1.08 hectares in the densely populated
Western Highlands. It was reported in the FGDs that it is impossible for persons starting families
in the densely populated highland areas to access more land either through inheritance from their
parents, or through land markets. Even in the areas considered relatively land abundant like in
the lowlands, land sizes are swiftly declining due to mounting population pressure. Besides the
growing population densities and the ensuing land subdivisions, distress land sales as a result of
pressing family financial needs (e.g. education and medical costs) and conversion of agricultural

54
 

land at close proximity to urban centers to residential plots are also gradually contributing to
declining agricultural land problem.

Table 2.1: Households mean landholding sizes (hectares)
Survey year
2000
2004
Eastern Lowlands
2.30
2.31
Western Lowlands
1.64
1.64
Western Transitional
2.87
2.89
High Potential Maize
4.62
4.64
Western Highlands
1.08
1.08
Central Highlands
1.24
1.24
Marginal Rain Shadow
2.22
2.20
Overall
2.59
2.60
Source: Tegemeo Institute Rural Household Surveys.

2007
2.36
1.66
2.89
4.67
1.09
1.24
2.20
2.62

2010
2.40
1.68
2.91
4.69
1.10
1.26
2.23
2.64

Not only are household farms shrinking in densely populated regions, a decline in soil
fertility has also been experienced. It emerged from the FGDs that the declining soil quality is
attributable to soil nutrients mining without adequate replenishment due to mono cropping, over
cropping and relay cropping; soil erosion/degradation occasioned by heavy rains; deforestation;
and use of unsuitable fertilizers as well as poor soil conservation practices among households.
The use of wrong fertilizers and continued use of non-organic fertilizer types have exacerbated
the soil quality problems in some areas. It was reported that most of the farm households do not
carry out soil testing to establish the suitable fertilizer types for their soils. Continued use of nonorganic fertilizer types was associated with increasing soil acidity. Application of Di-Ammonium
Phosphate (DAP) and Calcium Ammonium Nitrate (CAN) fertilizers for a long period
contributes to increased acidity resulting in stifled crop yields.

55
 

Results from the FGDs also showed that land conflicts and disputes among neighbors and
siblings are on the rise with declining land sizes. Land conflicts relate to boundary disputes with
neighbors and contested land inheritances among siblings. Inheritance related disputes are
common especially in situations where land parcels are not registered or are registered in the
name of a deceased parent or grandparent. Cases of land conflicts occasioned by returning
siblings who have been away in search of non-farm employment but now would want to revert to
farming are also on the raise as a result of increasing non-farm employment opportunities
scarcity. Land scarcity is also associated with increased incidences of fraudulent land deals due
to poorly drawn land sale agreements and fake land title deeds.
(ii) Modes of land access and their success in fostering increased smallholder
landholding: Two major land access channels where reported, namely: (1) inheritance from
parents; and (2) land sales and rental markets. It emerged from the FGDs that, unlike about three
decades ago, there are no unallocated lands and/or common grazing lands available in both the
low and highly densely populated areas. Some cases of households accessing land through
government allocations were reported in the North-Rift region. Most of this land was trust land,
including settlement scheme land purchased by government at independence for the settlement of
smallholder farmers; land acquired through illegal excisions of forestland, national parks, game
reserves, wetlands, riparian reserves and other protected lands; and Agricultural Development
Corporation (ADC) land. Results from the FGDs indicated these land allocations only benefited
economic and political elites and not the landless or the near landless smallholder farmers.
We next discuss the major land access modes in more detail and how they have evolved
in the context of the increasing human population densities next. Land inheritances from parents
emerged as the most important way in which household heads acquired land. Parents are
56
 

required by the African customary laws to bequeath sons who get married and are starting
families with land regardless of landholding sizes. These customary land transfer practices have
led to land subdivisions resulting in tiny landholding especially in the land constrained regions of
the country. The land subdivision problem is likely to be aggravated by two new developments.
First, unlike in the traditional African society where only sons inherited land from their parents,
some communities are now allocating land to unmarried daughters. Second, the new Land Policy
(Republic of Kenya, 2009) provides for equal treatment of children regardless of gender as far as
inheritance of family assets is concerned. This provision is also entrenched in the new
constitution promulgated in August 2010 (Republic of Kenya, 2010). Consequently, women are
now entitled to land inheritance from their parents, married or not. It was reported that the
implementation of the new law is bound to introduce a new dimension of family land conflicts as
married women return to their parents to claim their land entitlements.
Even after parents to children land transfers, it was reported in the FGDs that the title
deeds in most cases remain in the parents’ names due to a bureaucratic and costly land transfer
process prevailing in the country. Cases in which the land title deeds are in the names of
deceased parents or grandparents were also reported. It is important to note that transfer of title
deeds involving deceased persons is even more complicated, costly, and time consuming process
in Kenya. In some other situations, to circumvent the costly land transfers process, siblings agree
to register their land inheritance in one of the sibling’s name. This practice has however resulted
in increased cases of land conflicts among siblings and/or siblings’ offsprings with time as land
scarcity becomes a binding constraint.
The other major avenue through which households acquire land is through participating
in land sales and land rental markets. Results from the focus group discussions indicate that land
57
 

rental and sales markets are available in the low and medium densely populated areas. The
markets are even becoming more vibrant with land scarcity as population density increases. With
land scarcity, land has acquired value and is systematically replacing livestock as the primary
indicator of wealth in rural Kenya. Consequently, land values have steadily increased over time
due to high demand arising mounting population pressure. Land sale prices are almost three fold
in relatively densely populated than in medium and low populated regions. For example, a
hectare of high quality land costs KSh900 thousand in Bungoma County (medium density
county) and about KSh3 million in Vihiga County (high density county).
Like in other countries, land markets are characterized by high transaction costs in
Kenya. Transaction costs include the search cost of finding a willing seller, the right type of land,
and at the right location. For example, cases of individual looking for land to buy and could not
find the quality of land they were interested in were reported in the FGDs. Transactions costs are
also being amplified by the proliferation of fraudulent land sales and fake title deeds as buyers
have to incur additional costs to verify the authenticity of the land registration documents.
While it emerged that inter-community land transactions were allowed as long as the
sellers and buyers agree and the land parcel was dispute-free, side interviews with opinion
leaders revealed that some communities are less receptive to buyers from other communities. It
emerged that people facing land constraints are now becoming less inclined to buying land
outside what is considered “their ancestral territory”. This situation was exacerbated by the
ethnic-based land conflicts that rocked some parts of the country following the 2008 election
outcome. It was mentioned in some FGDs that background checks on the buyers are carried out
to avoid allowing in criminals elements and outdated/outlawed cultural practices such as illicit

58
 

liquor brewing and female circumcision. Ethnicity related land markets limitations were also
reported in the demand (buyers’) side too.
Land rental markets are active in all regions with agriculturally arable land. Just like the
land buying markets, rental prices have been on the increase in the last ten years. It emerged
from the FGD that a hectare of good quality land ranges between KSh12000 and KSh25000 in
the highly densely populated regions and between KSh7000 and KSh15000 in the medium
densely populated regions. Households seeking to rent in land in most cases are required to pay
the rent upfront before the start of the farming season. The demand for rental land in the high and
medium populated areas is so high that in some areas (e.g. Nakuru County), farmers interested in
renting in land are required to pay two to three years before the cropping season. While there are
no ethnic-based discriminations in land rental markets, in some situations, the government
administrators (chiefs) are involved in overseeing land rental agreements.
How successful are the land sales and rental markets in reducing landholding inequalities
among the smallholder farmers? To answer this question, we use the 13 year Tegemeo Institute
Panel data. We compute the Gini index of smallholder landholding after first factoring in land
sales and buying transactions only (Table 2.2) and second when we factor in land sales and
buying transaction and land rented in (2.3) over the panel period. The results generally show that
land selling and buying does not significantly reduce landholding inequality among the
smallholders. Landholding inequality as indicated by the Gini index actually increased 0.51 to
0.52 even though the difference is not statistically significant at 95 percent level (Table 2.2).

59
 

Table 2.2: Gini indices of households’ landholding by agro-regional zones
Survey year
2000
2004
Gini
[95% CI]
Gini
[95% CI]
Gini
Eastern Lowlands
0.38
[0.33 0.43]
0.38
[0.33 0.43]
0.38
Western Lowlands
0.36
[0.32 0.40]
0.37
[0.32 0.41]
0.38
Western Transitional
0.38
[0.33 0.42]
0.38
[0.33 0.43]
0.38
High Potential Maize
0.52
[0.48 0.55]
0.52
[0.48 0.55]
0.52
Western Highlands
0.37
[0.34 0.41]
0.37
[0.33 0.41]
0.38
Central Highlands
0.39
[0.36 0.42]
0.39
[0.36 0.42]
0.39
Marginal Rain Shadow
0.43
[0.33 0.53]
0.43
[0.33 0.53]
0.43
Overall
0.51
[0.49 0.54]
0.51
[0.49 0.54]
0.52
Source: Author’s calculation from the Tegemeo Institute Rural Household Surveys.

2007
[95% CI]
[0.33 0.43]
[0.33 0.42]
[0.33 0.42]
[0.48 0.56]
[0.34 0.42]
[0.36 0.42]
[0.33 0.53]
[0.49 0.54]

Gini
0.38
0.38
0.38
0.52
0.38
0.39
0.43
0.52

2010
[95% CI]
[0.34 0.43]
[0.34 0.43]
[0.33 0.42]
[0.48 0.56]
[0.34 0.42]
[0.36 0.42]
[0.34 0.53]
[0.49 0.54]

Gini
0.38
0.38
0.37
0.51
0.37
0.39
0.40
0.51

2010
[95% CI]
[0.33 0.43]
[0.33 0.43]
[0.33 0.42]
[0.48 0.55]
[0.33 0.41]
[0.36 0.41]
[0.31 0.49]
[0.48 0.53]

 

Table 2.3: Gini indices of households’ land access by agro-regional zones
Survey year
2000
2004
Gini
[95% CI]
Gini
[95% CI]
Gini
Eastern Lowlands
0.38
[0.33 0.42]
0.38
[0.34 0.43]
0.38
Western Lowlands
0.36
[0.32 0.41]
0.37
[0.33 0.41]
0.37
Western Transitional
0.39
[0.34 0.44]
0.38
[0.33 0.42]
0.37
High Potential Maize
0.55
[0.50 0.60]
0.51
[0.47 0.55]
0.51
Western Highlands
0.38
[0.34 0.42]
0.38
[0.34 0.42]
0.38
Central Highlands
0.38
[0.35 0.41]
0.42
[0.35 0.48]
0.38
Marginal Rain Shadow
0.42
[0.32 0.52]
0.42
[0.31 0.52]
0.41
Overall
0.53
[0.50 0.57]
0.51
[0.48 0.53]
0.51
Source: Author’s calculation from the Tegemeo Institute Rural Household Surveys.

60
 

2007
[95% CI]
[0.34 0.43]
[0.32 0.41]
[0.32 0.42]
[0.47 0.55]
[0.34 0.42]
[0.35 0.41]
[0.31 0.50]
[0.48 0.53]

Similarly, smallholder landholding inequality does reduce when we further factor in land
rentals by households (Table 2.3). Information on land rented out by households in each survey
was not available. However, since the sample consists mainly of smallholder farmers who are
assumed to be land constrained and thus not engaged in renting out land activities. Table 2.3
shows that even land rental markets appear to have increased land inequality in 2000; they
somehow improved smallholder land access equality in 2007 and 2010.
The overall land inequality declines from a Gini index of 0.52 to 0.51 in these two years.
Consequently, the general conclusion is that land sales and rental markets are not reducing
landholding inequality among the smallholder farmers. These findings are confirmed by
information elicited from the FGDs that showed the following: first, it is mostly individuals with
non-farm employment who want to get into farming/settle or medium/large scale farmers
interested in expanding their scale of production who can afford to buy or lease land at the
prevailing rates. Second, increasing cases of distress land sales and rentals especially in the
densely populated areas. Some households are selling or renting out their much needed land to
meet pressing family financial needs such as educations fees and medical expenses. All these
factors combined indicate that land sales and rental markets may not only be unsupportive of the
land constrained smallholder but also exacerbating their situation.
(iii) Out-migration trends in the densely populated areas: Out-migration is perceived to be
a potential avenue to ease land pressure in densely populated and land constrained regions.
While no permanent outmigration was reported in the low densely populated areas, results from
the FGDs indicate the high densely populated areas are experiencing low but growing trends in
out-migration. The village out-migration rate of people aged 15 years and over was estimated at
61
 

about one percent. In Vihiga and Bungoma counties for example, some households were
reported to have relocated permanently to the neighboring but relatively land abundant counties
of Busia and Trans Nzoia. A few cases of households migrating permanently to the coastal
lowlands and the neighboring country, Uganda, were also reported. Seasonal out-migration
trends are also insignificant in the high densely populated regions.
What factors force/facilitate households or households’ members to consider migrating to
other areas? FGDs identified diminishing land sizes followed closely by land related conflicts as
as the major “push” factors. In some cases, households with exceedingly small landholding sizes
were reported to have sold their land and moved out in search of bigger land parcels in relatively
5

land abundant areas . The most important out-migration facilitating factor was financial ability
while the limiting factors include attachment to ancestral lands and insecurity and intercommunity land related conflicts, especially following the 2008 post-election violence. With
regard to the 2008 post-election violence, scholars and development commentators associate its
occurrence with longstanding inter-community disputes over land rights and land inequalities
(Kniss, 2010). Essentially, most of the out-migration cases reported in the densely populated
areas were basically intra-community and not inter-community.
Besides the land search related permanent migration, there is also the rural-urban
migration especially by the youth in search of non-farm employment. However, as reported
earlier in this section, due scarcity in non-farm employment opportunities in the urban centers,
individuals facing land constraints are opting to remain in the rural areas and to derive their
livelihoods from the tiny pieces of land they inherited from parents.
                                                            
5

It was reported in some FGDs that some households commit a large portion of the proceeds
from their land sale to other family needs and eventually end up buying even smaller land pieces
than the ones they sold or even end up landless.
62
 

It is important to mention that most of the time the migrating households or people still
retain the land parcels owned in the previous locations. This is because of number of reasons.
First, in Africa people revert to farming upon retirement from non-farm employment. In some
cases, holders of salaried employment either continue farming their land or lease it out. For
example, an urban consumption surveys conducted by Tegemeo Institute in 2003 and 2009
indicated that over 50 percent of the sample reported to have received farm produce from their
own farm located in the rural areas in the month prior the day of the interview. As mentioned
elsewhere in this section, with proliferation of fraudulent land transactions some individuals
engaged non-farm employment opt not to lease their lands. Second, the fear of the unknown and
the desire to be buried close to ones kinsmen when they eventually die. As one of the
respondents put it, “conditions allowing, our people (kinsmen) prefer staying close to one
another while alive and even in death”. Consequently, to a large extent these out-migrations,
whether permanent or temporary, are not helping in easing land scarcity pressure in the densely
populated areas.
The FGDs results indicate that areas near growing urban centers (districts and divisional
headquarters) are also experiencing high in-migration by the upper and middle income urban
elites. This group is buying agricultural land and converting it into urban plots for building
residential homes. The field survey team came across a farmer who owns 1500 acres of
agricultural land along Eldoret to Kitale road but is now sub-dividing the land into one acre
residential plots for sale. The conversion of agricultural land to residential plots has been on the
rise following the continued sub-division provincial administration units and creation of new
headquarters that started in 2002.

63
 

The low densely populated areas are experiencing low in-migration. The in-migrating
people consist mainly of the urban middle income group looking for relatively cheap land to
settle upon retirement from non-farm jobs. These include the nurses, teachers and other
government employees working in these areas. These people have worked in these areas for long
and hence have built the necessary social capital to enable them live comfortably with the locals.
Very few of them are coming from the agricultural land deficit regions. The low population
density regions are not agriculturally high potential areas and have poor infrastructural facilities
thus not attractive to individuals in searching for land for agricultural production purposes.
(iv) Future prospects of the households in the densely populated areas: How do we
expect farming practices and livelihoods to evolve in the next decades especially in the densely
populated areas? In the extremely high densely populated regions, most of the people predict that
there will be no land to farm in the next 10-20 years. There will be more kitchen gardening and
more residential plots. Some people expect a possibility of land consolidation. As land sizes
become extremely small, siblings may consider selling their tiny pieces to one of them and
relocating (conditionings allowing) to land abundant regions or urban centers. Some respondents
predicted a possibility of mass exodus by the landless to small and large urban centers. Already
in some areas the landless have moved and settled in urban slums in the nearby small urban
centers.
Land extensive crops such as sugar cane are gradually disappearing in the medium
populated areas. Respondents expect increased land intensive enterprises such as horticulture
(high value crops grown in green houses) and dairy production (zero grazing). As land sizes
diminish, people expect a shift towards short maturing and high yielding crops; relay cropping;
and increased use of modern technologies in agricultural production.
64
 

It was also reported that offsprings of smallholders in the extremely densely populated areas
are unlikely to have livelihoods as good as those led by their parents if they rely solely on
farming. This is likely to be caused by land scarcity and emerging climatic challenges. For the
future generation to lead lives as good as the current generation, the current generation will need
to invest more in education of their offsprings to enable them secure non-farm employment. The
future generation must embrace modern technology farming-- actually, a consensus emerged in
all FGDs that education for the children is not only important for securing good paying non-farm
jobs but also necessary for adoption of modern intensive farming technologies.
While crop and livestock farming continue to be the main economic activities in the rural
areas, there are some emerging off-farm economic activities in the rural and small urban centers.
These activities include sale of second hand clothes (mitumba) within the villages; cellphone
money transfer services (M-Pesa); cellphone repairing and battery charging; buying of old
household items -- scrap metal; bicycle passenger transport (boda-boda) services; repairing and
hawking of household utensils (mali-mali); haircut and salon services; artisan jobs/jua kali –
welding, carpentry and masonry; and washing of clothes for other families. It is important to note
that these emerging informal non-farm income generating activities are basically “poverty jobs”
that individuals, and especially the youth, are being pushed into by increasing land constraints.
These jobs are characterized by insufficient social protection and thus increased vulnerability to
poverty.
(v) Characteristics of emerging medium-scale farmers: In this section, we present the
descriptive analysis of the medium scale household survey data. The aim is to highlight the
distinguishing features between households that followed agricultural-led path and those that
pursued lateral entry path into medium-scale farming status. In the survey instrument,
65
 

households were asked to indicate the primary source of capital used to enter into medium-scale
farming status. Based on the responses from this question, the sample was divided into two
mutually exclusive analytical groups, namely: (1) Those that used capital from non-farm sources
to attain medium-scale farming status, hereafter referred to as “lateral entry”; and (2) Those that
used capital generated from farming as the main vehicle for expanding their scale of production
into medium scale farming, hereafter referred to as “agricultural entry”.
The results show that majority (59%) of the households used lateral entry into mediumscale farming status (Table 2.4). This group starting farming owning about 14 hectares of land
and have managed to build up their landholding threefold to about 38 hectares. About half of the
lateral entry group started farming with landholding of about 5 hectares or less. The agricultural
entry pathway group owned about 18 hectares of land when they started farming. This group has
managed to more than double their mean landholding sizes to 42 hectares. Out of this
agricultural entry group, only about 38 percent of them started farming with 5 hectares or less.
 

Table 2.4: Landholding by entry pathway into medium-scale farming
Entry into
Started farming
Started farming
medium-scale
owning <=5ha
with > 5ha
farming
Mean
N
% Mean
N
%
Agricultural entry Initial (ha)
1.89
31 16 27.62 51
26
Current (ha) 26.83
54.86
Lateral entry
Initial (ha)
2.03
58 28 25.89 60
30
Current (ha) 33.11
42.91
Total
Initial (ha)
1.98
89 44 26.69 111 56
Current (ha) 30.27
48.60
Source: MSU/Tegemeo Institute Medium Scale Farmers’ Household Survey.

Overall
Mean
17.89
42.40
14.16
38.33
15.69
40.22

N
82

%
41

118

59

200

100

The results show that households that pursued the lateral entry pathway were headed by
relatively younger individuals who had higher education attainment and were less likely
66
 

polygamous compared to the households that followed agricultural entry pathway. The average
age of heads in the lateral entry group was 62 years compared to 65 years in the agricultural
entry group. A sizeable proportion of households that used lateral entry pathway were headed by
persons who had post-secondary education (47%) compared to only 18 percent among the
households that followed agricultural entry pathway.
Households that followed agricultural entry pathway into medium scale farming have
spent relatively more years in farming compared to those that followed a lateral entry pathway
(Table 2.5). The results also show that just a few household heads were born in their current
location of residence. Only about six percent of the lateral and 15 percent of the agricultural-led
entry household heads were born in the location in which they are current farming. This finding
suggests that migration may be a prerequisite to improved access to land and production scale
expansion. As mentioned earlier on, some migrating households still maintain land, especially
land acquired through inheritance, in areas where they were born. The results show that about 17
and 26 percent of the agricultural and lateral entry groups, respectively, still own land in the
areas where they were born. This is an indication that sometimes migration may not be helpful in
easing land pressure in densely populated areas.
The survey also elicited information on whether the household heads were engaged in
any other non-farm income earning opportunities such as business or salaried non-farm
employed. The results also show that a large proportion of the medium-scale household heads
were/had been involved in some form of non-farm business (Table 2.5). This included about 52
percent among the agricultural entry category and 42 percent in the lateral entry group. A major
distinguishing feature between households that followed agricultural entry and those that entered
medium-scale laterally was participation in non-farm employment.
67
 

 

Table 2.5: Household demographic characteristics and historical information
Agricultural-led Growth
Lateral Entry in Emergent
Pathway
Farming
Variables
mean
[95% C.I]
mean
[95% C.I]
Household head age (years)
65.29
[63.01 67.58]
62.34
[59.21 65.46]
Male household heads (prop.)
0.77
[0.68 0.86]
0.81
[0.73 0.88]
Household size
6.41
[5.73 7.10]
6.92
[6.31 7.52]
Head’s marital status (prop.)
single
0.26
[0.18 0.34]
0.26
[0.16 0.35]
monogamous
0.60
[0.51 0.69]
0.70
[0.59 0.80]
polygamous
0.14
[0.07 0.20]
0.05
[0.00 0.10]
Household head’s level of
education (prop.)
informal
0.12
[0.05 0.19]
0.07
[0.02 0.11]
primary
0.43
[0.32 0.54]
0.24
[0.16 0.32]
secondary
0.27
[0.17 0.37]
0.22
[0.14 0.30]
post-secondary
0.18
[0.10 0.27]
0.47
[0.38 0.57]
Prop. of heads born in the area
0.15
[0.07 0.22]
0.06
[0.02 0.10]
Prop. owning land where they
0.17
[0.09 0.25]
0.26
[0.18 0.34]
migrated from
Years in farming
34.76
[32.07 37.44]
27.78
[18.98 35.26]
Prop. of heads have/had business
0.52
[0.41 0.63]
0.42
[0.33 0.51]
Prop. of heads have/had non-farm
0.17
[0.09 0.25]
0.84
[0.77 0.91]
job
Non-farm job employer (%)
no other job
0.83
[0.75 0.91]
0.16
[0.09 0.23]
civil servant
0.12
[0.05 0.19]
0.57
[0.48 0.66]
private sector
0.05
[0.00 0.10]
0.27
[0.19 0.35]
Duration in non-farm job (years)
3.27
[1.39 5.15]
17.14
[15.03 19.25]
Father to initial household head
attributes:
Had non-farm job (prop.)
0.33
[0.23 0.43]
0.38
[0.29 0.47]
Had some formal education
0.35
[0.25 0.46]
0.40
[0.31 0.49]
(prop.)
Still alive (prop.)
0.06
[0.01 0.11]
0.14
[0.07 0.20]
Land before subdivision (ha)
94.68
[40.85 148.50]
45.06
[22.89 67.23]
Source: MSU/Tegemeo Institute Medium Scale Farmers’ Household Survey.

68
 

A large proportion (84%) of the lateral entry group was/had been in non-farm salaried
employment. On average, this group had spent about 17 years in non-farm employment. Out of
this group, about 68 percent of them held government-related jobs. Conversely, only 14 heads
out of 82 (17%) in the agricultural entry pathway group was/had been in salaried employment.
Table 2.5 also presents retrospective information about the characteristics of the father to
the initial household head. The results show that a sizeable proportion of the households were
headed by persons whose fathers had some formal education, held non-farm jobs, and had a large
landholding before subdivision. Essentially, the amount of land held by the father to the initial
household head was a major distinguishing feature between households entry into medium-scale
farming pathway. Fathers to the household heads who had pursued agricultural-led pathway had
more than double (95 hectares) landholding sizes compared to fathers to lateral entry household
heads (45 hectares). A sizeable proportion of households whose fathers to the initial household
head was still alive appeared to have followed the lateral entry pathway.
In Table 2.6, some more descriptive features on landholding attributes by entry to
medium-scale farming pathway. Most of the medium-scale farmers acquired their land through
purchases from land markets. The lateral entry group significantly obtained much more land
(85%) through purchases compared to the agricultural-led group (64%). Conversely, the
agricultural entry category obtained comparably more land from inheritances from parents
compared to the lateral entry group. Households were asked to indicate the proportion of their
land that is located in the district where they were born and the proportion that is owned with
title deed (Table 2.6). While most of the sampled households had more land in their district of
birth, the agricultural led group had comparably much more (90%).

69
 

Table 2.6: Household landholding attributes
Agricultural-led Growth
Lateral Entry in Emergent
Pathway
Farming
Variable
mea
[95% CI]
mean
[95% CI]
n
Prop.of land acquired through
0.64
[0.54 0.74]
0.85
[0.77 0.92]
purchases
Prop. of land acquired through
0.36
[0.26 0.46]
0.15
[0.08 0.23]
inheritance
Prop. of land in the district of birth
0.90
[0.85 0.95]
0.75
[0.72 0.78]
Proportion of land owned with title
0.59
[0.49 0.70]
0.79
[0.72 0.86]
Proportion of land acquisition by
decade:
1969 or earlier
0.29
[0.21 0.37]
0.06
[-0.04 0.16]
1970 through 1979
0.24
[0.17 0.31]
0.18
[0.09 0.27]
1980 through 1989
0.20
[0.13 0.27]
0.20
[0.12 0.28]
1990 through 2000
0.18
[0.11 0.25]
0.32
[0.25 0.39]
2000 or later
0.09
[0.05 0.13]
0.25
[0.19 0.31]
Source: MSU/Tegemeo Institute Medium Scale Farmers’ Household Survey.

Perhaps the most interesting observation is that a sizable proportion (25%) of the lateral
entry group had managed to acquire land outside the region where they were born. Similarly,
most of the sampled households owned their land with title deeds. The lateral entry group held
relatively a larger proportion (79%) of their land under title deeds compared to 59 percent among
the agricultural entry pathway group. This implies that the lateral entry group managed to either
access land that was already registered or overcome land registration bureaucracies to get their
land parcels registered.
The results also show that while the agricultural-led pathway group accumulated much of
their land before 1980 while the lateral entry pathway group accumulated their land gradually
and much of it from year 1990 onwards (Table 2.6). This indicates that the agricultural entry
group perhaps used the land markets in the early years when land was still abundant or benefited,
either directly or indirectly, from the immediate post-independence government land allocation
70
 

schemes. The lateral entry group land accumulation seems to have driven by their high social,
economic and political capital derived from higher education attainment and non-farm
employment jobs especially in the public service sector. Also, it is apparent that this group
accumulated most of their land in the 1990s period. This coincides with the introduction of
competitive politics (multiparty democracy) and thus there is a possibility this group benefited,
directly or indirectly, from political patronage related government land allocation discussed in
section 2.2. The lateral entry group land accumulation can also be explained by the settling
public service employees who were retrenched following the structural adjustment programs
(SAPs) related public service reforms in the 1990s period.
Next, we discuss land use and crops productivity characteristics of the two groups
(agricultural-led and lateral entry). Agricultural crop productivity is defined and the net value of
crop outputs (after netting out fertilizer, seed and, land preparation, and hired labor costs) per
unit of land. We break each of the two groups into two categories based on landholding: 5 to 25,
and 25 to 50 hectares (Table 2.7). The results show that area under crop among the medium scale
farmers is an increasing function of landholding with the two medium-scale entry groups putting
almost the same proportion of their land under crop. The proportion of land under crop to total
household landholding appears to be a decreasing function of landholding. A more complete
presentation of these relationships is revealed when we look at the non-parametric regressions
results. Figures 2.1, (a) and (b), show the non-parametric regression relationship between
cultivated land on the y-axis and household landholding on the x-axis. The results show that area
under crop is an increasing function of the household landholding. Figures 2.2, (a) and (b), show
the relationship between the proportion of land cultivated over total landholding on the y-axis
and household landholding on the x-axis.
71
 

Table 2.7: Medium-Scale Farmers’ landholding and productivity
agricultural entry
lateral entry
Overall
5-25 ha 25-50ha mean
5-25 ha 25-50ha mean
5-25 ha 25-50ha mean
N
56
26
82
95
23
118
151
49
200
Landholding (ha)
18.62
69.99
42.40
14.39
81.84
38.33
16.17
75.53
40.22
Area under crop (ha)
11.53
35.33
22.55
8.62
38.95
19.38
9.84
37.02
20.85
Prop. of area under crop
0.62
0.50
0.53
0.60
0.48
0.51
0.61
0.49
0.52
Crop production ('000KSh)/ ha owned
82.32
78.20
80.41
65.04
96.55
76.22
72.30
86.79
78.17
Crop production ('000KSh)/ ha cultivated 132.03 136.71 134.20 100.50 161.82 122.26 113.75 148.47 127.82
Fertilizer use ('000KSh)/ha cultivated
12.13
12.16
12.14
11.91
11.27
11.68
12.01
11.74
11.90
Fertilizer use (kg)/ha cultivated
261.49 265.69 263.43 257.79 243.99 252.89 259.35 255.53 257.80
*
0.42
0.89
0.64
0.57
1.25
0.81
0.51
1.06
0.73
Landholding (ha) / tropical cow
Livestock value ('000KSh)/ha owned
97.41
76.10
87.55
145.70
44.51
109.79 125.40
61.31
99.44
Source: MSU/Tegemeo Institute Medium Scale Farmers’ Household Survey. Note: * Conversion using livestock grazing equivalent
scales (FAO, 1987).

72
 

Figure 2.1a: Area cultivated by land size –overall

--overall--

60
40
0

20

hectares

80

.

0

25

50

75

100

125

100

125

landholding--hectares
Figure 2.1b: Area cultivated by land size –by growth pathways

100

--by growth pathways--

60
40
0

20

hectares

80

.

0

25

50

75

landholding--hectares
agricultural entry
73
 

lateral entry

Figure 2.2a: Proportion of area cultivated--overall

--overall--

.55
.5
.45

proportion

.6

.

0

20

40

60

80

100

120

100

120

hectares
Figure 2.2b: Proportion of area cultivated—by growth pathways

--by growth pathway--

.4

proportion
.5
.6

.7

.8

.

0

20

40

60

80

hectares
agricultural entry

74
 

lateral entry

The results show that the proportion under crop declines with household land holding up
th

to about 15 hectares (25 percentile) and thereafter increases reaching a maximum at around 75
hectares (Figure 2.2a). However, when we break down the analysis by the pathways used to
achieve medium-scale status a different picture emerges (Figure 2.2b).
The proportion of land under cultivation is an increasing function of household
landholding sizes for the agricultural entry group. For the lateral entry group, the relationship
between proportion under crop and landholding is a non-linear cubic function. Proportion of area
th

under crop decreases with landholding up to about 25 hectares (75 percentile), thereafter, it
increases reaching a maximum at about 75 hectares. The proportion of land under crop never
goes beyond 0.50 in this group. It seems while the agricultural entry group is keen to bring most
of the land they access under cultivation, the lateral entry group is either holding land for
speculative motive and/or for use at a later date.
While fertilizer application rates per hectare cultivated do not seem to vary much
household landholding sizes in Table 2.7, the non-parametric regression results reveal a different
picture (Figures 2.3 a & b). The intensity of fertilizer uses (y-axis) increase with household
landholding size (x-axis) among the medium-scale farmers up to 30 hectares (after the 75

th

percentile) (Figure 2.3b). Beyond this landholding, the relationship almost flattens out. The same
relationship is replicated when we do the analysis by the pathways into medium-scale farming
status (Figure 2.3b). The intensity of fertilizer use increases, but at a decreasing rate, with
landholding in both groups with the lateral entry group application rates being consistently below
th

the agricultural entry group up to the 99 percentile of the distribution.
75
 

Figure 2.3a: Fertilizer use per hectare cultivated – overall

--overall--

200

220

kg/ha

240

260

.

0

20

40

60

80

100

120

hectares
Figure 2.3b: Fertilizer use per hectare cultivated -- by growth pathway
28
0

--by growth pathway--

24
0
18
0

20
0

22
0

kg/ha

26
0

.

0

20

40

60

80

100

hectares
agricultural entry

76
 

lateral entry

120

In terms of crop productivity, the results also show that net crop production per unit of
land is somehow an increasing function of landholding sizes (Table 2.7). This result is
surprising. However, it should be noted that this is a bivariate analysis finding and that crop
productivity will be a subject of further analysis using econometric methods. The agricultural
entry group seems to be more productive than the lateral entry counterpart.
The agricultural entry group is producing on average KSh80 thousand per hectare owned
(KSh134 thousand/ per hectare cultivated) while the lateral entry group is producing KSh76
thousand (KSh122 thousand). A more complete presentation of these relationships is revealed
when we look at the non-parametric regressions results. Figures 2.4, (a) and (b), show the nonparametric regression relationship between values of crop production on the y-axis and
household landholding on the x-axis. The results show that crop production is an increasing
function of the household landholding in the entire sample and by the entry into medium-scale
farming pathways subsamples. Figures 2.5, (a) and (b), show the relationship between crop
production values per hectare owned y-axis and household landholding on the x-axis. The results
show that the crop productivity first declines with landholding size reaching a minimum at about
10 hectares and increases thereafter (Figure 2.5a). A different picture emerges when we break
down the sample by the pathways into medium-scale farming status (Figure 2.5b). While crop
productivity increases with land among the agricultural entry group, among the lateral entry
th

group, it first declines reaching a minimum at about 30 hectares (after the 75 percentile) and
thereafter goes up.

77
 

Figure 2.4a: Value of crop production -- overall

--overall--

10
0

5

m
illion KSh

15

.

0

20

40

60

80

100

120

hectares
Figure 2.4b: Value of crop production – by growth pathway

--by growth pathway--

15
10
0

5

m
illion KSh

20

.

0

20

40

60

80

100

hectares
agricultural entry

78
 

lateral entry

120

Figure 2.5a: Value of crop production per hectare owned-- overall

--overall--

65
55

60

'000KSh/ha

70

75

.

0

20

40

60

hectares
Figure 2.5b: Crop of production per hectare owned – by growth pathway

120

--by growth pathway--

80
60
20

40

'000KSh/ha

100

.

0

20

40

60

hectares
agricultural entry

79
 

lateral entry

Figures 2.6, (a) and (b), show the relationship between crop production values per hectare
cultivated y-axis and household landholding on the x-axis. The results show that the crop
productivity first decline with landholding size reaching a minimum at about 35 hectares and
thereafter increases (Figure 2.6a). While crop productivity increases gradually with land among
the agricultural entry group, among the lateral entry group, it first declines reaching a minimum
th

at about 30 hectares (after the 75 percentile) and thereafter gradually picks up.
Even though data on livestock production (milk, meat, hides/skins, eggs, etc.) was not
collected, data on household livestock holding and the estimated value was collected. Using the
scales developed by the Food and Agriculture Organization’s Tropical Livestock Unit (FAO,
1987), all livestock are converted into tropical cow equivalent based on livestock grazing
equivalence. This is meant to help in comparing how the agricultural and lateral entry groups are
using their uncultivated land in livestock production. Subsequently, in terms of livestock
holding, the results show that the agricultural entry group is using on average 0.64 hectares per
tropical cow compared to 0.81 hectares by the lateral entry group. In terms of livestock values,
the agricultural-led group is supporting lower livestock value (KSh88000) per uncultivated
hectare compared to the lateral entry group (KSh109000) (Table 2.7).
To examine the distinguishing features between medium and smallholder farmers in
terms of land use and agricultural productivity, we present some smallholder agricultural
production characteristics in Table 2.8. Smallholders are defined as households farming five
hectares and below. As mentioned in the data sources section, the smallholder data came from
the same locations as the medium-scale data. While the medium-scale farmers’ data was
collected in 2012, the smallholder data was collected in 2010.

80
 

Figure 2.6a: Value of crop production per hectare cultivated -- overall

--overall--

140
130
110

12
0

'000 S
K h/ha

150

160

.

0

20

40

60

80

hectares
Figure 2.6b: Crop production per hectare cultivated -- by growth pathway

250

--by growth pathway--

150
50

100

'000KSh/ha

200

.

0

20

40

60

hectares
agricultural entry

81
 

lateral entry

80

Thus, the smallholder crop productivity figures are inflated using the consumer price
index (CPI) developed by Kenya National Bureau of Statistics to make them comparable to the
medium-scale farmers’ figures. The smallholder sample is also broken down four categories
based on landholding (0 – 1; 1 – 2; 2 – 3; and 3 – 5 hectares) to examine how land productivity
changes with landholding sizes (Table 2.8). In terms of area under crop, compared to mediumscale farmers, the smallholder farmers are putting a larger proportion (0.69) of their land under
crop. Just as in the medium-scale farmer case, proportion of land under crop is a declining
function of landholding with households holding less than one hectare putting all their land under
crop.

Table 2.8: Smallholder farmers’ landholding and productivity by landholding categories
0<ha<=1 1<ha<=2 2<ha<=3 3<ha<=5 overall
N
66
83
51
36
236
Landholding (ha)
0.59
1.36
2.37
3.99
1.76
Area under crop (ha)
0.55
1.15
1.51
2.23
1.22
Prop. of area under crop
0.93
0.84
0.64
0.56
0.69
Crop production ('000KSh)/ ha owned
104.46
59.76
53.22
64.20
63.55
Crop production ('000KSh)/ ha cultivated 112.32
71.18
83.50
114.87
91.77
Fertilizer use ('000KSh)/ha cultivated
9.40
9.10
9.58
8.93
9.22
Fertilizer use (kg)/ha cultivated
159.08
153.63
163.10
152.22
156.46
*
0.02
0.11
0.36
0.75
0.26
Landholding (ha) / tropical cow
Livestock value ('000KSh)/ha owned
1,420.64 299.94
139.13
76.56
158.17
Source: Tegemeo Institute Rural Household Surveys. Note: * Conversion using livestock
grazing equivalent scales (FAO, 1987).

In terms of crop productivity, surprising the smallholders are less productive than the
medium-scale farmers. The smallholders are producing on average KSh64 thousand per hectare
owned (KSh92 thousand/ per hectare cultivated) compared to KSh78 thousand (KSh127
thousand) among the medium-scale farmers. The results also show that crop productivity is a
decreasing non-linear function of landholding sizes. Crop productivity seems to be very high
82
 

among households owning one hectare and below. Crop productivity per hectare owned declines
with landholding reaching a minimum at about 2-3 hectares and increases thereafter. Similarly,
productivity per hectare cultivated declines with landholding reaching a minimum at about 1-2
hectares and increases thereafter.
In terms of fertilizer application rates, the smallholder farmers seem to applying
comparably more fertilizer per hectare cultivated. The smallholders apply about 258 kilograms
of fertilizer per hectare compared to 157 kilograms by the medium-scale farmers. Likewise, it
seems smallholders are spending more (KSh12000) on fertilizer per hectare cultivated compared
to the medium-scale farmers (KSh9000). This difference could be attributable to higher fertilizer
application rates among the smallholders and economies of scale experienced by medium-scale
farmers due to bulk fertilizer purchases.
Turning to livestock production, results show that the smallholders are using their
uncultivated land more productively compared to the medium-scale farmers. The smallholders
are using on average 0.26 hectares (Table 2.8) per tropical cow compared to 0.73 hectares by the
medium-scale farmers (Table 2.7). It is clear from Table 2.8 that the use of uncultivated land on
livestock productivity declines as household land increases.
In Table 2.9, we present household descriptive statistics on income and asset wealth
tabulated by the two broad entry pathways into medium scale farming. Even though there appear
to be some differences between these two aggregate group’s incomes and asset wealthy, the
difference are not statistically significant at 95 percent level. The only distinguishing feature
between these two groups is in terms of the main source of incomes. While all the households in
the sample derive most of their incomes from crop income, the households that followed lateral
entry pathway are deriving a sizeable proportion (34%) of their incomes from non-farm sources.
83
 

The households that used agricultural pathway are only obtaining about 19 percent of their
income from non-farm sources. When you consider incomes derived from salaried employment,
the households that used lateral entry pathway are obtaining more than double from salary
compared to those households that pursued agricultural led entry pathway. It is important to
mention that data on livestock income was not captured in this survey.

Table 2.9: Household income and assets
Agricultural-led Growth
Lateral Entry in Emergent
Pathway
Farming
Variable
mean
[95% CI]
mean
[95% CI]
Aggregate income-‘million KSh
2.84
[1.77 3.91]
2.62
[1.17 4.07]
Crop income-‘millions KSh
2.62
[1.57 3.68]
2.30
[0.89 3.71]
Non-crop income-‘000KSh
217.18 [147.98 286.38] 323.11 [236.20 410.02]
Income from business-‘000KSh
134.36
[74.66 194.07] 123.42
[70.82 176.03]
Income from salary-‘000KSh
82.81
[45.67 119.96] 199.68 [129.94 269.43]
Crop income/total income prop.
0.81
[0.72 0.84]
0.66
[0.64 0.74]
Non-farm/total income prop.
0.19
[0.16 0.28]
0.34
[0.26 0.36]
Total assets- 'million KSh
30.5
[21.70 39.30]
32.00
[22.40 41.60]
Land- ‘million KSh
25.7
[17.70 33.60]
28.00
[18.90 37.00]
Non-land assets--'million KSh
4.84
[3.64 6.04]
4.05
[3.15 4.94]
Agric. equipment and other assets4.06
[3.04 5.09]
3.38
[2.59 4.17]
‘million KSh
Livestock- 'million KSh
0.78
[0.54 1.01]
0.66
[0.50 0.82]
Land asset/total assets prop.
0.81
[0.78 0.84]
0.83
[0.81 0.86]
Livestock/total assets prop.
0.03
[0.03 0.04]
0.03
[0.03 0.04]
Ag. equipment/total assets prop.
0.15
[0.13 0.18]
0.14
[0.12 0.16]
Source: MSU/Tegemeo Institute Medium Scale Farmers’ Household Survey.

2.5.2

Econometric results
This section summarizes the econometric estimation results from the methods described

in section 2.4.1. While the model for each outcome variable is estimated jointly using FIML
estimator, the selection equation estimation results are presented first followed by the outcome
equations results. Given that the same explanatory variables were used in the estimation of the
84
 

selection equation in all endogenous switching models and that there were no marked differences
in the selection equations’ results in all models, only the result from the estimation of the first
outcome of interest (land access) are reported. It is important to mention at the onset that the
correlation coefficients (  ) between the pathways into medium scale farming and the outcome
of interest in the agricultural and lateral entry equations in all the models are both positive and
statistically significant (Tables 2.11-2.15). This result confirms that an endogenous switching
regression model is the appropriate model for this analysis.
(i) Pathways into medium-scale farming selection equation estimation results: The
estimates of the correlates of pathways into medium scale farming are presented in Table 2.10.
The dependent variable is a binary variable: agricultural entry is represented by “one” while
lateral entry is represented by “zero”. Only variables’ marginal effects are presented. The results
show that the following characteristics influence pathways into medium scale farming: (1)
household demographics; (2) household head participation in non-farm employment; (3)
household historical factors (the initial household head father’s attributes); and (4) land holding
characteristics and soil quality.
The age of the household head and family size influence pathways into medium scale
farming. An increase in the head’s age by an additional year increases the probability of
agricultural entry into medium scale farming by 0.03. Similarly, an increase in the number of
adult members of the household by one person increases the probability of agricultural entry by
about 0.17. While increased household head’s age may be capturing farming experience,
increased adult household members is possibly capturing labor availability in the family, a
variable that was not controlled for directly in the models due to lack of data. Farming

85
 

experience and labor availability are key factors in agricultural production and thus their role in
facilitating agricultural led entry into medium scale farming result is not surprising.
 

Table 2.10: Selection equation results-- pathways into medium scale farming
Dep. var.:1=agricultural entry; 0=lateral entry
Marginal effects Std. Err.
Household demographic variables
Household head age
-0.033
0.013
Household head sex (1=male; 0=female)
0.408
0.392
Number of adult members in the household
-0.166
0.050
Household head education (base=no educ.)
_primary
-0.211
0.468
_secondary
0.489
0.595
_post-secondary
0.025
0.607
Participation in non-farm activities
Head’s years in salaried job
0.007
0.021
Employer type (base=no salary job)
_private sector
-0.265
0.056
_civil service
-0.321
0.064
Household’s years in non-farm business
0.029
0.014
Household’s history
Landholding when started farming (ha)
0.016
0.035
Household’s years in the current settlement
0.499
0.485
Father to head’s landholding (acres)
0.109
0.056
Father to head’s education (1=educated; 0=informal)
0.219
0.303
Father to head is alive (1=alive; 0=dead)
-0.228
0.068
Father to head was community leader (1=yes; 0=no)
0.404
0.411
Land characteristics and quality
Location (1=region; 0=outside)
0.142
0.052
Acquisition (1=purchase; 0=inherit)
0.186
0.305
Tenure (1=title; 0=w/out title)
-0.333
0.348
0.041
0.025
Length of growing period
-0.005
0.004
Net primary productivity
0.004
0.001
Elevation: meters above sea level
0.023
0.013
Slope/steepness -- degrees
Number of obs.sigma_1
200
Log pseudolikelihood
-246.247

86
 

P>z
0.01
0.30
0.00
0.65
0.41
0.97
0.72
0.00
0.00
0.04
0.08
0.30
0.05
0.47
0.00
0.33
0.01
0.54
0.34
0.09
0.17
0.00
0.08

Participation in non-farm employment influences entry into medium scale farming. Using
households that relied solely on farm employment as the base, the results show that a change to
private sector employment increases the probability of lateral entry by 0.27 while a change to
public sector increases it by about 0.32. It is interesting to note that that what mattered most was
the non-farm employer type and not the number of years in non-farm employment. Similarly,
participation in off-farm business activities facilitates agricultural led entry into medium scale
farming. Each additional year in off-farm business participation increases the probability of the
household’s agricultural entry into medium scale farming by 0.03.
Family history was also found to influence entry into medium scale farming pathways. If
the land owned by the father to the initial household head before subdivision increased by one
hectare, the probability of the current household to enter into medium farming via the
agricultural led pathway increased by 0.11. Similarly, the death of the initial household head’s
father increases the current household’s probability of agricultural entry into medium scale
farming by 0.23. This result can be explained in two ways. First, after the death of the father to
the initial household head, his land is likely to be inherited by his sons thereby making the
beneficiaries agricultural entry pathway into medium scale farming possible. Second, fathers to
initial households are assumed to possess high social capital and many networks that they
accumulate with age. Social capital and networks are helpful in lowering land market transaction
costs (e.g. search costs, access to quality land, and land deals negotiations) and thereby
facilitating the current household head’s, especially those in non-farm employment, entry into
medium scale farming possible through a lateral entry. The results also indicate that the size of
initial land owned when the family started farming positively influences the probability of the
household’s entry into medium scale farming via agricultural led pathway.
87
 

Households that followed agricultural led entry accessed most of their land in the location
where the household head born while those that entered medium scale farming laterally accessed
most of their land outside the locations where the heads were born. Relocation from the area
where the household head was born increased the probability of household’s lateral entry into
medium scale farming by 0.14. This finding implies that for a household to enter medium scale
farming laterally, it has to look for land beyond the areas where the household head was born.
The results also show that the agricultural entry group somehow had high quality land in terms of
length of growing period and land elevation.
(ii) Determinants of medium-scale farming access and area under crop: The results on the
determinants of medium scale farmers’ land access (owned and rented in) and areas under crops
from the FIML estimation are presented in Tables 2.11 and 2.12, respectively. The most
important determinants of household land access among the famers who entered medium scale
farming through agriculture include the age of the household head; the size of land owned when
the household started farming; and the location of the household. An increase in the household
head’s age by one year increased household land access by about two percent while an increase
in the initial landholding by one hectare increased current household land access by three
percent.
Among the factors influencing land access in lateral entry subsample include: education
level of the household head; landholding when the household started farming, and the distance
from the household to the nearest motorable road. Household landholding is a function of the
household head’s education attainment. For example, a switch from no education to primary
education increases household land access by eight percent, while a switch to secondary and
post-secondary increases land access by 11 and 14 percent, respectively.
88
 

Table 2.11: Results of the endogenous switching model for medium-scale farmers’ land
access
Agricultural entry
Lateral entry
Variables
Coef.
P>z
Coef.
P>z
Household demographic variables
Household head age
0.017
0.04
-0.004
0.47
Household head sex (1=male; 0=female)
0.262
0.24
-0.104
0.53
Household head education (base=no educ.)
_primary
-0.255
0.25
0.083
0.00
_secondary
-0.008
0.98
0.111
0.00
_post-secondary
0.114
0.60
0.143
0.00
Number of adult members in the household
0.038
0.18
0.016
0.56
Family history
Landholding when started farming (ha)
0.032
0.01
0.009
0.00
Household’s years in the current settlement
-0.044
0.88
-0.166
0.41
Distance to infrastructure
Distance to nearest water source
-0.052
0.19
-0.006
0.89
Distance to nearest motorable road
-0.048
0.41
-0.112
0.04
Land characteristics and quality
Location (1=region; 0=outside)
-0.694
0.08
-0.981
0.05
Acquisition (1=purchase; 0=inherit)
-0.126
0.57
0.050
0.76
Tenure (1=title; 0=w/out title)
0.284
0.18
0.544
0.00
0.050
0.78
0.070
0.53
Length of growing period
0.002
0.60
0.003
0.22
Net primary productivity
0.048
0.58
-0.014
0.01
Elevation: meters above sea level
-0.028
0.00
-0.029
0.00
Slope/steepness -- degrees
County (base=Nakuru)
_Bungoma
0.477
0.46
0.171
0.49
_Uasin Gishu
1.283
0.00
0.906
0.00
_Trans-Nzoia
0.853
0.13
0.757
0.01
_cons
-2.814
0.52
2.976
0.38
Number of obs.
82
118

0.051
0.00
0.344
0.00
 

89
 

Table 2.12: Results of the endogenous switching model for area under crop
Agricultural entry
Lateral entry
Coef.
P>z
Coef.
P>z
Household demographic variables
Household head age
-0.014
0.03
0.006
0.58
Household head sex (1=male; 0=female)
0.023
0.92
-0.080
0.76
Household head education (base=no
educ.)
_primary
-0.186
0.58
0.025
0.51
_secondary
-0.143
0.72
0.103
0.02
_post-secondary
0.034
0.93
0.094
0.04
Number of adult members in the
0.059
0.03
-0.006
0.87
household
Landholding and assets
Current landholding (ha)
0.014
0.00
0.016
0.00
Own tractor (1=yes; 0=no)
0.603
0.01
0.622
0.00
Input/output Prices
Price of maize (KSh/kg)
0.023
0.17
0.014
0.05
Price of DAP (KSh/kg)
-0.156
0.53
0.157
0.35
Price of CAN (KSh/kg)
0.253
0.28
-0.184
0.19
Distances to infrastructure and
markets
Distance to nearest ag. input dealer (km) -0.326
0.00
0.049
0.50
Distance to nearest water source (km)
-0.003
0.96
0.051
0.20
Distance to nearest motorable road (km)
0.068
0.35
-0.098
0.20
Land characteristics and weather
Acquisition (1=purchase; 0=inherit)
-0.082
0.75
-0.150
0.42
Tenure (1=title; 0=w/out title)
0.404
0.18
0.155
0.47
0.024
0.22
-0.010
0.96
Length of growing period
0.024
0.12
0.012
0.04
Net primary productivity
0.003
0.52
0.005
0.04
Elevation: meters above sea level
0.030
0.01
-0.008
0.96
Slope/steepness -- degrees
-0.066
0.23
-0.031
0.38
Rainfall (mm)
-0.301
0.42
-0.367
0.21
Rainfall stress
_cons
-7.606
0.31
5.669
0.28
Number of obs.
82
118
0.304
0.00
0.552
0.02


90
 

Similarly, households that entered medium scale farming through a lateral entry had
access to registered land and land at close proximity to motorable roads. This implies that the
lateral entry group had the social economic capital and connections that enabled them to access
land close to infrastructural facilities and to either access registered land or navigate the land
registration bureaucracy to get their land registered. Generally, the medium scale farmers were
more likely to access land in locations outside the areas where the households’ heads was born.
Next, we examine the determinants of medium-scale farming area cultivated (Table
2.12). Only two variables seem to be common determinants of area under crops in the two
regression results presented in Table 2.12, namely household land access and tractor ownership.
Access to additional hectare of land increases land under crops by about two percent while
tractor ownership increases area under crop by about 60 percent. In the agricultural entry group,
the other factors that influence area under crop include, the age of the household; household size;
distance to the nearest input dealer; and land slope in the areas where the household is situated. A
reduction in the distance to the nearest input dealer by one kilometer increases area under crops
by about 33 percent in the agricultural pathway subsample. In the lateral entry group, the only
other factors influence area under crops include households’ heads education attainment; price of
maize prevailing in the regional markets are the planting time; and the net primary productivity
and land slope in the area where the household is located. A switch in the household head’s
education from no education to secondary education or post-secondary education increases land
under cultivation by about ten and nine percent, respectively. Increase in the price of maize price
by a shilling per kilogram increases area under crop by one percent in the lateral entry group.
 

91
 

(iii) Determinants of medium-scale farmers’ fertilizer use intensities and crop
productivity: The results on determinants of fertilizer use intensity (kilograms applied per hectare
cultivated) are presented in Table 2.13. Generally, only ownership of a tractor and maize and
fertilizer prices, influence the intensity of fertilizer application. An increase in the price of maize
by one shilling per kilogram increases fertilizer use intensity by about two percent in the two
pathways into medium-scale farming. Ownership of a tractor increase fertilizer use by about 26
percent among the households that followed agricultural-led pathway and by 27 percent in the
lateral entry group. The results also show that an increase in the price of DAP by a shilling per
kilogram reduces the fertilizer application intensity by three percent in the lateral entry group.
Next, we turn to crop productivity. Agricultural crop productivity is loosely defined and
the net value of crop outputs (after netting out fertilizer, seed and, land preparation, and hired
labor costs) per unit of land. The results show that the only common factor influencing medium
scale farming crop productivity per hectare owned in the two regressions reported in Table 2.14
are the maize prices prevailing in the regional markets at the planting time and the distances to
the nearest agricultural input dealers. An increase in maize price by a shilling per kilogram
increases medium scale farmers’ crop productivity by about three percent while an increase by
one kilometer to the nearest input dealer reduces productivity by about 15 percent. The other
factors that influence medium farmers’ crop productivity are the price of fertilizer for the lateral
entry group and households’ location land elevation (meters above the sea level) for the
agricultural entry group. An increase in DAP price by one shilling per kilogram reduces crop
productivity by about three percent.
Similarly, the only common factors influencing medium scale farmers’ crop productivity
per hectare cultivated in the two regressions are maize prices and distances to agricultural input
92
 

dealers (Table 2.15). An increase in maize price by a shilling per kilogram increases crop
production by four percent.

Table 2.13: Results of the endogenous switching model for intensity of fertilizer use
Agricultural entry
Lateral entry
Coef.
P>z
Coef.
P>z
Household demographic variables
Household head age
-0.002
0.75
0.006
0.29
Household head sex (1=male;
-0.087
0.58
0.265
0.08
0=female)
Household head education
(base=no educ.)
_primary
0.027
0.89
-0.104
0.65
_secondary
-0.047
0.88
-0.007
0.98
_post-secondary
0.200
0.66
-0.158
0.52
Number of adult members in the
-0.016
0.48
0.016
0.55
household
Landholding and assets
Current landholding (ha)
0.032
0.81
-0.003
0.19
Own tractor (1=yes; 0=no)
0.255
0.04
0.274
0.05
Input/output Prices
Price of maize (KSh/kg)
0.016
0.02
0.021
0.01
Price of DAP (KSh/kg)
0.039
0.53
-0.026
0.04
Price of CAN (KSh/kg)
0.004
0.94
-0.016
0.12
Land quality and weather
-0.037
0.35
0.012
0.20
Slope/steepness -- degrees
0.020
0.52
-0.012
0.62
Rainfall (mm)
0.950
0.68
-0.462
0.80
Rainfall stress
_cons
1.497
0.67
-3.443
0.18
Number of obs.
82
118
0.221
0.00
0.545
0.01


However, an increase in the distance to the nearest input dealer by a kilometer reduces
crop productivity by about two percent. The other factors that influence crop productivity per
hectare cultivated include land tenure, land access, the amount of rainfall and rainfall stress.
While crop productivity appears to be higher in registered land parcels in the agricultural entry
93
 

subsample, crop production is a decreasing function of household land access in the lateral entry
subsample.  
 

Table 2.14: Results of the endogenous switching model for crop production per ha owned
Agricultural entry
Lateral entry
Variables
Coef.
P>t
Coef.
P>t
Household demographic variables
Household head age
-0.006
0.44
0.002
0.76
Household head sex (1=male; 0=female)
0.025
0.92
0.203
0.22
Household head education (base=no
educ.)
_primary
0.024
0.91
0.029
0.91
_secondary
0.000
1.00
0.127
0.63
_post-secondary
0.231
0.47
0.362
0.17
Number of adult members in the
-0.004
0.90
-0.002
0.92
household
Input/output Prices
Price of maize (KSh/kg)
0.029
0.01
0.034
0.01
Price of DAP (KSh/kg)
-0.002
0.99
-0.026
0.04
Price of CAN (KSh/kg)
0.054
0.63
-0.033
0.11
Distances to infrastructure and
markets
Distance to nearest ag. input dealer (km)
-0.157
0.10
-0.151
0.04
Distance to nearest motorable road (km)
0.435
0.68
0.318
0.77
Distance to nearest water source (km)
-0.089
0.19
-0.011
0.93
Land characteristics
Current landholding (ha)
0.130
0.42
0.148
0.28
Tenure (1=title; 0=w/out title)
0.169
0.36
0.210
0.27
Own tractor (1=yes; 0=no)
-0.082
0.49
0.012
0.95
Land quality and weather
0.004
0.75
-0.026
0.19
Length of growing period
0.001
0.55
0.000
0.53
Net primary productivity
0.005
0.02
-0.003
0.30
Elevation: meters above sea level
-0.005
0.95
0.012
0.27
Slope/steepness -- degrees
0.017
0.69
0.042
0.15
Rainfall (mm)
-2.955
0.42
-2.773
0.18
Rainfall stress
_cons
1.857
0.78
11.708
0.08
Number of obs.
82
118
0.350
0.00
0.016
0.01

94
 

Table 2.15: Results of the endogenous switching model for crop production/ha cultivated
Agricultural entry
Lateral entry
Variables
Coef.
P>t
Coef.
P>t
Household demographic variables
Household head age
-0.119
0.30
-0.004
0.65
Household head sex (1=male; 0=female)
-0.002
0.99
0.074
0.72
Household head educ. (base=no educ.)
_primary
-0.234
0.46
0.170
0.62
_secondary
0.015
0.97
0.240
0.54
_post-secondary
0.405
0.41
0.513
0.16
Number of adult members in the
-0.027
0.47
0.022
0.65
household
Input/output Prices
Price of maize (KSh/kg)
0.037
0.02
0.036
0.02
Price of DAP (KSh/kg)
-0.290
0.17
-0.395
0.18
Price of CAN (KSh/kg)
-0.218
0.18
-0.288
0.12
Distances to infrastructure and
markets
Distance to nearest ag. input dealer (km)
-0.017
0.09
-0.192
0.04
Distance to nearest motorable road (km)
-0.729
0.62
1.434
0.31
Distance to nearest water source (km)
0.156
0.19
-0.034
0.81
Land characteristics
Current landholding (ha)
-0.037
0.14
-0.053
0.01
Tenure (1=title; 0=w/out title)
0.062
0.02
0.077
0.77
Own tractor (1=yes; 0=no)
-0.001
0.67
-0.001
0.85
Land quality and weather
0.017
0.25
-0.036
0.14
Length of growing period
0.019
0.17
-0.010
0.24
Net primary productivity
0.006
0.10
0.002
0.54
Elevation: meters above sea level
0.007
0.50
-0.003
0.83
Slope/steepness -- degrees
0.002
0.67
0.095
0.03
Rainfall (10 mm)
-0.197
0.66
-0.617
0.06
Rainfall stress
_cons
9.433
0.27
16.789
0.04
Number of obs.
82
118
0.386
0.00
0.188
0.01

 

95
 

The expected rainfall increases crop productivity but the expected rainfall shock
6

somehow reduces it among the farmers that entered medium scale farming laterally . Increase in
the expected rainfall by 10mm increase crop productivity per hectare cultivated by about ten
percent. While increased expected rainfall shocks reduce crop productivity, the coefficient is
only significant at ten percent level. 
 

2.6

Conclusions and policy implications
Smallholder farmers constitute the bulk of agricultural producers in sub-Saharan Africa

and majority of them are poor. Based on sub-Saharan Africa’s land endowment, an agriculturalled growth strategy has been touted as solution for reductions in poverty in this region. However,
recent studies cast doubts on the land abundance hypothesis. Each day, the Africa smallholder
farming land holding and access shrinks as population density rises. Despite these mounting
population related challenges, analysts have pointed out that sub-Saharan Africa still has the
potential to revitalize smallholder agricultural productivity for reduced poverty and hunger if
appropriate policies are pursued. They cite the example of the smallholder-led broad-based Asian
green revolution that contributed greatly to rural poverty reduction in that region.
This study investigates the potential for similar forms of inclusive agricultural growth
using Kenya as a case study. The study specifically investigate whether land institutions and
policies in Kenya are making it possible for a broad-based smallholder led agricultural growth
process as enjoyed in much of Asia. In Kenya, there has been a policy thinking that agricultural
                                                            
6

Expected rainfall is defined as a six-year moving average of rainfall prior to the main growing
season in survey year, while expected rainfall shock is a six-year moving average of the
percentage of 20-day periods during the main growing season with less than 40 mm of rainfall. 
96
 

and land reforms supported by adequate government budget allocation have the potential to
underpin a revitalized system of smallholder production especially in areas where land sizes have
become too small.
The study is inspired by two factors. First, the recent literature indicating that Africa is
characterized by landholding inequalities making redistributive land reforms an attractive
strategy for improved smallholder land access. Using a panel data spanning 13 years and
qualitative data from focus group discussions in the land scarce densely populated regions we
examine how smallholders are coping with shrinking farm sizes and whether land rental and
sales markets and other land access modes have been successful in reducing inequalities in land
holding. The second factor motivating this study is the increasing number of medium scale
(emergent) farmers in many African countries over the last decade. Yet, the processes behind
their growth have remained unclear. Is this growth driven by farmers who began their farming
careers as smallholder now transitioning to a larger scale of production through the capital and
assets accumulation; a precursor to the inclusive agricultural-led structural transformation? Or is
the growth driven by land institutions and policies that encourage investment in land acquisitions
by individuals from non-agricultural employment sector signaling elite land capture? Using a
survey conducted with a random sample of medium scale farmers in Kenya, we explore the rise
in medium-scale farming operations over the 1980-2010 period, how the medium scale farmers
achieved their current scale of operation, the underlying interpretation of these findings about the
priorities and motivations of governments, and the implications for prospects of broadly-based
agricultural development strategies.
The results point to four key findings. First, the qualitative results from the focus group
discussions indicate that access to land is becoming a binding agricultural production constraint
97
 

in the densely populated regions of the Kenya. Even in the areas considered relatively land
abundant like in the lowlands regions, land sizes are swiftly declining due to mounting
population pressure. Not only are household farms shrinking in densely populated regions, soil
fertility is also rapidly deteriorating due to nutrients mining and degradation. Land conflicts
among neighbors and siblings over boundaries and contested inheritances are on the rise due to
increasing land scarcity. Increasing land scarcity is also triggering increased incidences of
fraudulent land deals as a result of poorly drawn land sale agreements and proliferation of fake
land title deeds consequently increasing transaction costs in land markets.
Second, the qualitative data show that the customary land inheritance from parents and
land sales and rental markets still remain the most important ways through which smallholder
farmers gain access land in rural Kenya. The results indicated that there are not unallocated lands
and/or common grazing lands in both the low and highly densely populated areas. These
customary land transfer practices from parents to male children have led to land subdivisions
resulting in tiny landholding especially in the land constrained regions of the country. The land
subdivision problem is likely to be compounded by the new constitutional requirement providing
for equal treatment of children regardless of gender in family assets inheritance.
Third, while migration out of densely populated areas is considered a potential avenue to
ease land pressure in the land constrained regions, the results show outmigration trends are very
low. Rural to rural migration requires financial ability which the smallholder farmers facing land
constraints are lack. Migration is also being inhibited ethnicity and cultural factors. While
attachment to ancestral lands limits outmigration, access to land in most of the regions in Kenya
is tied on one’s ethnic identity. Following the 2008 post-election tribal land conflicts, ethnicity is
slowly becoming a binding land access constrain in rural Kenya. Limited non-farm employment
98
 

opportunities are impeding rural to urban migration in Kenya. It is important to mention that in
most cases the migrating individuals still retain hold of the land they own in the previous
locations. Consequently, outmigration is not helping in easing land scarcity pressure in the
densely populated areas of the country.
Fourth, the empirical results revealed that land sales and rental markets do not reduce
landholding inequality among smallholder farmers in rural Kenya. With increasing land scarcity,
land sale and rental prices have been bid high beyond the reach of the land-poor smallholders. It
is mostly the medium and large scale farmers interested in expanding their scale of production
and as well as individuals transitioning from non-farm employment into farming who can afford
such prices and who have the political and economic capital to navigate through the
bureaucracies of land markets to access land.
Fifth, the empirical results from the medium scale farmers’ survey show that majority of
them used lateral entry into medium-scale farming status. They attained their current farming
status by acquiring land from savings from non-farm, largely urban jobs; only a minority was
primarily engaged in agriculture prior to achieving medium-scale farming status. A big
proportion of them owned on average over two times more land than they were using for
agriculture, implying a high degree of land owned for speculative purposes and/or an inability of
farmers in this size category to make productive use of the land they owned. In terms of crop
productivity, the agricultural entry group seems to be more productive than the lateral entry
counterpart. Another interesting finding is that these emergent farmers are generally more
productive in terms of total production and production per hectare compared to the smallholders.
This study, therefore, suggests that land is becoming a binding agricultural production
constraint especially in the densely populated regions of the country. However, the current land
99
 

policies and institutions and by extension land markets are not facilitating smallholders facing
increasing land constraints access to land. The policies are primarily working for individuals
with economic and political capital conferred through off-farm jobs that are able to navigate the
land markets systems and land administration bureaucracies.
The new national land policy in Kenya spells out a wide range of measures aimed at
eliminating land access discrimination based on gender and ethnicity; land taxation to curb
inefficient land use and idle land holding and hoarding tendencies; resolution of historical land
injustices; improvement of tenure security; efficient utilization of marginal lands; setting up of
economically viable minimum landholding sizes; review of public land allocation procedures;
and commercialization of land rights to make land market operations more efficient. The policy
however leaves very many hard decisions to legislators and implementers preparing key
legislations to support the implementation of the policy. The policy does not suggest how to
overcome challenges encounters in the past initiatives aimed at improving land access to the
smallholders facing landlessness. Empirical studies have shown that well-targeted land
redistribution programs can a positive impact on smallholder land access and poverty reduction.
While some smallholder farming systems “pessimists” have argued for smallholders whose
farms have become too small to quit farming and pursue non-farm income options, Hazell (2011)
and Lipton (2009) opines that the exit of small farms is not a “driver” but a “consequence” of 
economic growth. Hazell (2011) plainly says: “When economies grow, many small farmers (or
their children) leave farming because they can find better paying jobs elsewhere. But
consolidating land and pushing small farmers off the land before there are better jobs available
simply leads to worsening poverty and unwanted levels of rural-urban migration”, (Hazell,
2011).
100
 

 
 
 

APPENDIX

101
 

Table A2.1: Distribution of government land by use in square kilometers
Region
Forest
Townships Alienated land Unalienated land National parks Open water
Kenya
9,116
2,831
38,546
28,598
24,067
10,960
Nairobi
21
93
225
16
117
Central
2,541
156
1,505
28
900
3
Coast
454
838
15,202
19,979
15,065
563
Eastern
1,289
227
3,148
8,397
7,721
4,131
N. Eastern
Nyanza
179
113
1
3,480
R. Valley
4,195
1,338
18,353
177
262
2,646
Western
616
2
137
Source: Syagga (2009)

Others
2,136
77
155
1,022
452
23
404
3

Total
116,254
549
5,288
53,123
25,365
3,796
27,375
758

Others
492
163
52
244
5
21
59,625

Total
59,625
368
5,897
8,913
3,740
587
39,731
389

 

Table A2.2: Distribution of trust land that is not available for smallholders
Region
Forest
Townships Alienated land Unalienated land National parks Open water
Kenya
7,084
1,812
33,397
13,810
3,030
Nairobi
Central
9
94
102
Coast
63
214
3,881
1,687
Eastern
789
546
647
4,203
2,484
N. Eastern
396
202
3,142
Nyanza
1
285
177
119
R.Valley
5,964
154
28,387
4,659
546
Western
258
123
1
Source: Syagga (2009)
 

102
 

Table A2.3: Classification of land in Kenya
Public Land
Community Land
(1) Community land shall vest in and be
(1) Public land is—
held by communities identified on the
(a) unalienated government land;
(b) land lawfully held, used or occupied by basis of ethnicity, culture or similar
community of interest.
any State organ;
(2) Community land consists of—
(c) land transferred to the State by way of
(a) land lawfully registered in the name of
sale, reversion or surrender;
group representatives under the provisions
(d) land in respect of which no individual
of any law;
or community
(b) land lawfully transferred to a specific
ownership can be established;
(e) land in respect of which no heir can be community by any process of law;
(c) any other land declared to be
identified;
community land by an Act of
(f) all minerals and mineral oils;
Parliament; and
(g) government forests;
(d) land that is—
(h) all roads and thoroughfares;
(i) all rivers, lakes and other water bodies; (i) lawfully held, managed or used by
specific communities as community
(j) the territorial sea, the exclusive
forests, grazing areas or shrines;
economic zone and the sea bed;
(ii) ancestral lands and lands traditionally
(k) the continental shelf;
(l) all land between the high and low water occupied by
hunter-gatherer communities; or
marks;
(iii) lawfully held as trust land by the
(m) any land not classified as private or
county governments, but not including any
community land under
public land held in trust by the county
this Constitution; and
government.
(n) any other land declared to be public
(3) Any unregistered community land
land by an Act of Parliament—
(2) Public land under clause (a) clause (1) shall be held in trust by county
(a), (c), (d) or (e), and clause (1) (b), other governments on behalf of the
communities for which it is held.
than land held, used or occupied by a

103
 

Private Land
I. Private land consists of —
(a) registered land held by any person
under any freehold
tenure;
(b) land held by any person under
leasehold tenure; and
(c) any other land declared private land
under an Act of Parliament.
(1) A person who is not a citizen may hold
land on the basis of leasehold tenure only,
and any such lease, however granted, shall
not exceed ninety-nine years.
(2) For purposes of this Article—
(a) a body corporate shall be regarded as a
citizen only if the body corporate is wholly
owned by one or more citizens; and
II (1) There is established the National
Land Commission.
(2) The functions of the National Land
Commission are to —
(a) manage public land on behalf of the
national and county governments;
(b) recommend a national land policy to the
national government;
(c) advise the national government on a
comprehensive programme for the
registration of title in land throughout
Kenya;

Table A2.3 (cont’d)
Public Land
Community Land
national State organ shall vest in and be
held by a county government in trust for
(4) Community land shall not be disposed
the people resident in the county, and shall of or otherwise used except in terms of
be administered on their behalf by the
legislation specifying the nature and extent
National Land Commission.
of the rights of members of each
(3) Public land classified under clause (1) community individually and collectively.
(f) to (m) shall vest in and be held by the
(5) Parliament shall enact legislation to
national government in trust for the
give effect to this
people of Kenya and shall be administered Article.
on their behalf by the National Land
Commission.
(4) Public land shall not be disposed of or
otherwise used except in terms of an Act
of Parliament specifying the nature and
terms of that disposal or use.
Source: Condensed by the author from Republic of Kenya (2010)

104
 

Private Land
(d) conduct research related to land and the
use of natural resources;
(e) initiate investigations, on its own
initiative or on a complaint, into present or
historical land injustices, and recommend
appropriate redress;
(f) encourage the application of traditional
dispute resolution
mechanisms in land conflicts;
(g) assess tax on land and premiums on
immovable property in any area designated
by law; and
(h) monitor and have oversight
responsibilities over land use planning
throughout the country.

Figure A2.1: Paths of access to land

 

Source: de Janvry and Sadoulet (2001) with author’s modification

105
 

7

Figure A2.2: Modified Fisher-Ideal Quantity Indices
To aggregate crop production across multiple commodities, we use a modification of the
Fisher-Ideal index by Mason (2011), which uses information on the individual household
production (kg) and national-level prices of each crop in the crop group. The Fisher-Ideal (FI)
index is a combination of two indices, the Modified Laspeyres Quantity Index (ML) and the
Modified Paasche Quantity Index (MP) (Diewert 1992; Diewert 1993).
For each crop j  1 to J , we use the national median production quantity as the base
quantity in the denominator of both the ML and MP indices. We use the median national-level
price in the first year of the Tegemeo panel household dataset (1997) as the base year price,
p j,base. Thus, changes in the ML index are driven by changes in quantities of each commodity

produced over time, as prices do not vary from the base year, nor across households. For p j in
the MP index, we use the national median price for each year. Thus, the MP index allows price
variation by year but not across households.
J

Modified Laspeyres Quantity Index
(LQI*)

LQI i*,t 

q
j 1
J

q
j 1

Modified Paasche Quantity Index

i , j ,t

p j ,base

*
j ,base

p j ,base

J

(PQI*)

PQI i*,t 

q
j 1
J

i , j ,t

p j ,t

q *,base p j ,t
 j
j 1

Fisher-Ideal Quantity Index (FIQI*)

FIQI i*,t  ( LQI i*,t  PQI i*,t

                                                            
7

Adapted from Mather and Jayne (2011).
106

 

Table A2.4: Medium and large scale farmers survey sample
County
Division
Location
Bungoma
Tongaren
Ndalu / Mitua
Tongaren
Ndalu / Mitua
Central
Milima/Mukuyuni
Trans Nzoia
Saboti
Kinyoro /Saboti
Waitaluk
Waitaluk
Kwanza
Kaisagat
Kaplamai
Ngonyek farm
Kaplamai
Kaplamai/Sibanga
Uasin Gishu
Soy
Soy
Soy
Soy
Moiben
Moiben
Moiben
Karuna
Moiben
Karuna
Nakuru
Rongai
Rongai
Rongai
Rongai
Rongai
Rongai
Ngata
Ngata
 

 
 

107
 

Number of households
10
15
15
15
15
15
13
7
15
15
15
14
6
10
5
5
10

Figure A2.3: Medium and large scale farmers survey sample

 

108
 

Table A2.5: Focus group discussion sites
District
Division
Location
Vihiga
Sabatia
North Maragoli
Kisii
Marani
Mwamunari
Bungoma
Kimilili
Kamukunywa
Kakamega
Mumias
Etenje
Bungoma
Kanduyi
Bukembe
Nyeri
Mukurweini
Muhito
Nyeri
Mukurweini
Muhito
Machakos
Mwala
Mwala
Kitui
Chuluni
Mbitini
Uasin Gishu
Ainabkoi
Olare
Uasin Gishu
Moiben
Sergoit
Nakuru
Njoro
Ngata
Nakuru
Njoro
Ngata

Sub location
Mulundu
Rioma
Nabikoto
Musanda
North Sangalo
Gatura
Gaturia
Myanyani
Katwala
Kapkeno
Kelji
Kirobon
Ngecha

 

109
 

Village
Kigaro
Emanyi
Esikweya
Mukhwenje
Muluhu
Ikiyu
Gachiriro
Kandumbo
Nzovia
Gaiti
Kapsking
Kapkatet
Ngecha B

density
911
792
786
659
616
487
804
171
167
161
161
146
146

Landholding (acres)
2
3
4
4
7
3
2
9
10
5
7
14
14

  
 
 

 
 
 
 
 
 

REFERENCES

110
 

REFERENCES

Acemoglu, D., S. Johnson and J.A. Robinson. 2001. The Colonial Origins of Comparative
Development: An empirical investigation, The American Economic Review, 91(5): 1369–
401.
Anderson, David. 2005. Histories of the Hanged: The Dirty War in Kenya and the End of
Empire, London: Weidenfeld and Nicolson.
Andre, C. and Jean-Philippe Platteau. 1998. Land Relations under Unbearable Stress: Rwanda
Caught in the Malthusian Trap. Journal of Economic Behavior and Organization 34 (1):
1-47.
Atieno-Odhiambo, E.S. 1995. The Formative Years: 1945–55. In Ogot, Bethwell Alan and W.R.
Ochieng', Decolonization and Independence in Kenya, 1940–93. Oxford: James Currey.
Beard, V. and S. Phakphian. 2009. Community-based Planning in Chiang Mai, Thailand: Social
Capital, Collective Action and Elite Capture’. Paper presented at Dialogical Conference
‘Social Capital and Civic Engagement in Asia’, University of Toronto, 7–10 May.
Berry, Sara. 1993. No condition is permanent: the social dynamics of agrarian change in subSaharan Africa. Madison: University of Wisconsin Press.
Berry, Sara. 1989. Social institutions and access to resources in African agriculture. Africa vol.
59(1): 41-55.
Bilsborrow, R. E. 1987. Population pressures and agricultural development in developing
countries: A conceptual framework and recent evidence. World Development, 15(2): 183203.
Binswanger-Mkhize, H.P, Camille Bourguignon, Rogier van den Brink. 2009. Agricultural Land
Redistribution: Toward Greater Consensus, Washington D.C.: World Bank.
Bruce, John. 2009. Kenya Land Policy: Analysis and Recommendations. Report prepared for the
United States Agency for International Development. Available at
http://pdf.usaid.gov/pdf_docs/PNADP494.pdf [accessed March 25, 2012].
Bruce, J. W. and S. Migot-Adholla. 1994. Searching for land tenure security in Africa.
Washington D.C.: World Bank.
Bruck, Tilman and Kati Schindler. 2009. Smallholder Land Access in Post-War Northern
Mozambique. World Development, 36(8): 1379-1389.
111
 

Cooper, Elizabeth. 2008. Inheritance Practices and the Intergenerational Transmission of
Poverty in Africa. CPRC Working Paper No. 116. Manchester: Chronic Poverty Research
Centre.
Daily Nation. 2013. Kibaki finally gazettes land commission. Daily Nation, viewed Thursday,
February 21 2013 http://www.nation.co.ke/News/politics/Kibaki-finally-gazettes-landcommission/-/1064/1699954/-/14o7x6xz/-/index.html.
Deininger, Klaus and Derek Byerlee. 2011. Rising Global Interest in Farmland: Can it yield
sustainable and equitable benefits? World Bank Group. Available at
http://siteresources.worldbank.org/INTARD/Resources/ESW_Sept7_final_final.pdf
[accessed April 10, 2012].
Deininger, Klaus, and Songqing Jin. 2008. Land Sales and Rental Markets in Transition:
Evidence from Rural Vietnam.” Oxford Bulletin of Economics and Statistics (70): 67101.
Deininger, K. and D. Byerlee, with J. Lindsay, A. Norton, H. Selod, M. Stickler. 2011. Rising
Global Interest in Farmland: Can it Yield Sustainable and Equitable Benefits? World
Bank, Washington, DC.
Deininger, K., R. Castagnini, and M. González. 2004. Comparing Land Reform and Land
Markets in Colombia: Impacts on Equity and Efficiency. Policy Research Working Paper
3258. World Bank, Washington, DC.
Deininger, Klaus. 2003. Land Policies for Growth and Poverty Reduction. World Bank Policy
Research Report. Oxford, UK:World Bank and Oxford University Press.
Deininger, Klaus, and Pedro Olinto. 2000. Asset Distribution, Inequality, and Growth. Policy
Research Working Paper 2375. World Bank, Washington, DC.
Deininger, Klaus, and Lyn Squire. 1998. New Ways of Looking at Old Issues. Journal of
Development Economics 57: 259–87.
Diewert, W.E. 1993. Index Numbers. In Essays in Index Number Theory, ed. W. Erwin Diewert
and Alice O. Nakamura. New York: North Holland.
Diewert, W.E. 1992. Fisher Ideal Output, Input, and Productivity Indexes Revisited. The Journal
of Productivity Analysis 3.3: 211-48.
Ellis, F. 2005. Small Farms, Livelihood Diversification, and Rural–Urban Transitions: Strategic
Issues in Sub-Saharan Africa. Paper presented at The Future of Small Farms, 26-29 June.
Wye College, London.
FAO. 1987. ICS-data, Statistical Division, FAO, Rome.
112
 

Feder, G. and T. Onchan. 1987. Land ownership security and farm investment in Thailand.
American Journal of Agricultural Economics, 69(2): 311–320.
Fischer, G., Harrij van Velthuizen and F. O. Nachtergaele. 2000. Global Agro-Ecological Zones
Assessment: Methodology and Results. Interim Report IR-00-064. International Institute
for Applied Systems Analysis and The Food and Agricultural Organization of the United
Nations. The data are available from: http://geodata.grid.unep.ch/.
Fischer, R.A., D. Byerlee, and G.O. Edmeades. 2009. Can Technology Deliver on the Yield
Challenge in 2050? Report prepared for the Expert Meeting on How to Feed the World in
2050, Food and Agricultural Organization of the United Nations, Economic and Social
Development Department.
Garg, A. and J. Morduch. 1998. Sibling Rivalry and the Gender Gap: Evidence from Child
Health Outcomes in Ghana. Journal of Population Economics 11.4: 471-93.
Gray, L. and M. Kevane. 1999. Diminished access, diverted exclusion: women and land tenure in
sub-Saharan. African Studies Review 42( 2): 15-39.
Govereh, J. 1999. Impacts of Tsetse Control on Immigration and Household Accumulation of
Capital: Zambezi Valley, Zimbabwe. Unpublished PhD Dissertation, Michigan State
University, East Lansing.
Holden, S. T., K. Otsuka, and F. Place. 2009. The Emergence of Land Markets in Africa: Impacts
on Poverty, Equity and Efficiency. Washington D.C.: Resources For the Future Press.
de Janvry, Alain and Elisabeth Sadoulet. 2001 . Access to Land and Land Policy Reforms. UNU
World Institute for Development Economics Research (UNU/WIDER), Helsinki,
Finland.
Jayne, T.S. and Milu Muyanga. 2012. Land constraints in Kenya’s densely populated rural
areas: implications for food policy and institutional reform. Food Security, 4(3): 399421.
Jayne, T.S. T. Yamano, M. Weber, D. Tschirley, R. Benfica, A. Chapoto, and B. Zulu. 2003.
Smallholder Income and Land Distribution in Africa: Implications for Poverty Reduction
Strategies. Food Policy, 28(3): 253-275.
Jayne, T.S., Zulu, B., Kajoba, G., and Weber, M. 2008. Access to Land and Poverty Reduction in
Rural Zambia: Connecting the Policy Issues. Working Paper No. 34, Food Security
Research Project, Lusaka, Zambia.
Jin, S. and T.S. Jayne 2013. Land Rental Markets in Kenya: Implications for Efficiency, Equity,
Household Income, and Poverty. Land Economics, 89 (2): 246-271.
113
 

Johnston, B. F. and P. Kilby. 1975. Agriculture and Structural Transformation: Economic
Strategies in Late Developing Countries. New York: Oxford University Press.
Kanyinga, Karuti. 2009. Land redistribution in Kenya.in Binswanger-Mkhize, H.P, Camille
Bourguignon, Rogier van den Brink. 2009, Eds. Agricultural Land Redistribution:
Toward Greater Consensus, Washington D.C.: World Bank.
Kanyinga, Karuti. 2000. Redistribution from above: The Politics of Land Rights and Squatting in
Coastal Kenya. Uppsala: Nordic Africa Institute.
Kato, Joshua. 2009. Our problem is land not Bakiga, say Banyoro. New Vision, Uganda daily
newspaper, Uganda. Available at http://www.newvision.co.ug/PA/9/183/692039
[accessed March 1, 2012].
Kniss, Michael. 2010. Walking Kenya Back from the Brink: A Micro-Level Study of Horizontal
Inequity and Civil Conflict Prevention. Center for International and Security Studies
Working Paper, School of Public Policy, University of Maryland. Available at
http://www.cissm.umd.edu/papers/files/walking_kenya_back_from_the_brink.pdf
[accessed September 27, 2012].
Kranton, Rachel E., and Anand V. Swamy. 1999. The Hazards of Piecemeal Reform: British
Civil Courts and the Credit Market in Colonial India. Journal of Development Economics
58 (1): 1-24.
Lahiff, Edward and Ben Cousins. 2005. Smallholder Agriculture and Land Reform in South
Africa. IDS Bulletin, Vol. 36(No 2). Institute of Development Studies, Brighton: UK
Lee, L. 1978. Unionism and Wage Rates: A Simultaneous Equation Model with Qualitative and
Limited Dependent Variables. International Economic Review, 19(2):415-33.
Lee, L.F., and R.P. Trost. 1978. Estimation of Some Limited Dependent Variable Models with
Application to Housing Demand. Journal of Econometrics, 8(3):357-382.
Leo, C. 1989. Land and Class in Kenya, Harare: Nehanda Publishers.
Lipton, Michael. 2009. Land Reform in Developing Countries: Property Rights and Property
Wrongs. Abingdon: Routledge.
Lokshin, M. and Sajaia, Z. 2004. Maximum Likelihood Estimation of Endogenous Switching
Regression Models. Stata Journal, 4(3): 282-289.
Place, F., and S. Migot-Adholla. 1998. The Economic Effects of Land Registration on
Smallholder Farms in Kenya: Evidence from Nyeri and Kakamega Districts. Land
Economics 74(3):360–73.
114
 

Maddala, G. S. 1983. Limited Dependent and Qualitative Variables in Econometrics. New York:
Cambridge University Press.
Mamdani, M. 1996. Citizens and Subjects: Contemporary Africa and the Legacy of Late
Colonialism. Princeton, NJ: Princeton University Press.
Mason, N. 2011. The Effects of the Zambian Food Reserve Agency and Government Fertilizer
Programs on Smallholder Farm Household Fertilizer Use and Crop Production.
Unpublished Ph.D. dissertation, Agricultural, Food and Resource Economics
Department, Michigan State University, East Lansing, Michigan.
Mather, David and T.S. Jayne. 2011. The Impact of State Marketing Board Operations on
Smallholder Behavior and Incomes: The Case of Kenya. International Development
Working Paper 119, Department of Agricultural, Food, and Resource Economics,
Michigan State University, East Lansing, Michigan.
Mellor, J. W. 1995. Agriculture on the Road to Industrialization. International Food Policy
Institute, Johns Hopkins University Press: Baltimore, MD.
Morduch, Jonathan. 2000. Sibling Rivalry in Africa. American Economic Review 90.2: 405-409.
Muhanguzi, Ganzi. 2008. Bakiga settlers in Kalangala evicted. New Vision, Uganda daily
newspaper, Uganda. Available at http://www.newvision.co.ug/PA/8/13/648215 [accessed
March 1, 2012].
Munshifwa, E. 2002. Rural Land Management and Productivity in Rural Zambia: The Need for
Institutional and Land Tenure Reforms. Paper presented at the Surveyor’s Institute of
Zambia Seminar, July 2002, Oxfam.
Mutua, Martin. 2005. MPs to lead squatters back to farms. The Standard, Thursday July 7, 2005.
Publication of the Standard Group Limited, Kenya: Nairobi.
Muyanga, M., T.S. Jayne and William J. Burke. 2010. Pathways into and out of Poverty: A
Study of Rural Households Wealth Dynamics in Kenya. Tegemeo Institute of
Agricultural Policy and Developmemt Working Paper No. 39/2010.
Niasse, M. 2011. Access to land and water for the rural poor in a context of growing resource
scarcity. Paper presented at the IFAD Conference on New Directions for Smallholder
Agriculture 24-25 January, 2011, IFAD Headquarters, Rome, Italy. Available at
http://www.ifad.org/events/agriculture/doc/papers/niasse.pdf [accessed April 9, 2012].
Ogot, Bethwell Alan. 1995. The Decisive Years: 1956-63. In Ogot, B.A. and W.R. Ochieng',
Decolonization and Independence in Kenya, 1940–93. Oxford: James Currey. pp. 48–82.

115
 

Okoth-Ogendo, W.H.O. 1991. Tenants of the Crown: Evolution of Agrarian Law and Institutions
in Kenya. Nairobi: ACTS Press.
Pinckney, Thomas C. and P.K. Kimuyu. 1994. Land Tenure Reform in East Africa: Good, Bad
or Unimportant? Journal of African Economies, 3(1), 1–28.
Platteau, J. P. 2004. Monitoring Elite Capture in Community-Driven Development. Development
and Change, 35(2): 223–46.
Republic of Kenya. 2010. Constitution of Kenya. Government Printer, Nairobi.
Republic of Kenya. 2008. Report of the Commission of Inquiry into the Post Election Violence
(CIPEV). Available at http://kenyastockholm.files.wordpress.com/2008/10/the-wakireport.pdf [accessed April 22, 2013].
Republic of Kenya. 2009. Sessional Paper No. 3 of 2009 on National Land Policy. Ministry of
Lands. Government Printer: Nairobi.
Republic of Kenya. 2004. The Report of the Commission of Inquiry into the Illegal/Irregular
Allocation of Public Land—the Ndung’u Report. Government Printer: Nairobi.
Republic of Kenya. 1999. Report of the Judicial Commission of Inquiry into Tribal Clashes in
Kenya. Available at
http://www.marsgroupkenya.org/pdfs/crisis/2008/01/Judicial_Commission_Report_On_T
ribal_Clashes_In_Kenya/Introduction.pdf [accessed April 22, 2013].
Scoones, I., Marongwe, N., Mavedzenge, B., Murimbarimba, F., Mahenehene, J and Sukume, C.
2010. Zimbabwe’s Land Reform: Myths and Realities. Oxford: James Currey.
Shipton, P. and M. Goheen. 1992. Understanding African Land-Holding: Power, Wealth, and
Meaning. Journal of the International African Institute, 62(3): 307-325.
Sitko, N., C. Nkonde, and T.S. Jayne. 2013. Pathways for becoming an Emergent Farmer in
Zambia: Can current land and agricultural policies encourage smallholder
extensification? Working Paper, Food Security Research Project, Lusaka, Zambia.
Smucker, Thomas. 2002. Land tenure reform and changes in land-use and land management in
semi-arid Tharaka, Kenya. The Land Use Change, Impacts and Dynamics (LUCID)
Project Working Paper Series Number 11, International Livestock Research Institute,
Nairobi, Kenya.
Stambuli, K. 2002. Elitist food and agricultural policies and the food problem in Malawi.
Journal of Malawi Society - Historical & Scientific, 55(2).

116
 

Swynnerton, R.J.M. 1955. The Swynnerton Report: A plan to intensify the development of
African agriculture in Kenya. Nairobi: Government Printer.
Syagga, Paul. 2010. Public land, historical land injustices and the new Constitution. Society for
International Development Constitution Working Paper No. 9, Nairobi.
Thurow, Roger. 2010. The fertile continent. Foreign Affairs, 89(6).
Verma, Ritu. 2001. Gender, Land, Livelihoods in East Africa: Through Farmers’ Eyes.
International Development Research Centre, Canada: Ottawa.
Wangila, J., R. Rommelse, and J. De Wolf. 1999. An insight into the socioeconomics of
agroforestry client-farmers in western Kenya. Pilot Project Report No. 12. Nairobi,
Kenya: International Centre for Research in Agroforestry.
Wasserman, G. 1976. Politics of Decolonization: Kenya Europeans and the land issue 1960–
1965, African Study Series No 17, London: Cambridge University Press.
Willis, R.J. and S. Rosen. 1979. Education and Self-Selection. Journal of Political Economy,
87(5):S7–S36.
Wilson, M.F.J., B. O'Connell, C. Brown, J.C. Guinan and A.J. Grehan. 2007. Multiscale terrain
analysis of multibeam bathymetry data for habitat mapping on the continental slope.
Marine Geodesy 30: 3-35. Available from: http://srtm.csi.cgiar.org/.
Wiggins, Steve, Johann Kirsten, Luis Llambi. 2010. The Future of Small Farms. World
Development, 38(10): 1341-1348.
Wood, Adrian. 2003. Could Africa be like America? In B Pleskovic, N Stern (eds) Annual Bank
Conference on Development Economics, World Bank and OUP.
Wooldridge, J. 2010. Econometric Analysis of Cross-Section and Panel Data. 2nd Edition.
Massachusetts: MIT Press.
World Bank. 2007. World Development Report 2008: Agriculture for Development, World Bank:
Washington DC.
World Bank. 2002. World Development Report 2003: Sustainable Development in a Dynamic
World. Washington, D.C.: World Bank.
van de Brink, R., G. Thomas, H. Binswanger, J. Burce, and F. Byamugisha. 2006. Consensus,
Confusion, and Controversy: Selected Land Reform Issues in Sub-Saharan Africa. World
Bank Working Paper No. 71. Washington, D.C.: World Bank.

117
 

Yamano, Takashi, F. Place, W. Nyangena, J. Wanjiku and K. Otsuka. 2009. Efficiency and
Equity Impacts of Land Markets in Kenya. In The Emergence of Land Markets in Africa,
eds. Stein T. Holden, Keijiro Otsuka, and Frank Place. Washington DC: Resources for
the Future.
Zhao, M., F. A. Heinsch, R. R. Nemani, and S. W. Running. 2005. Improvements of the MODIS
terrestrial gross and net primary production global data set. Remote Sensing of
Environment. 95: 164.176. Dataset is available from:
ftp://firecenter.umt.edu/pub/MODIS/Mirror/MOD17A3.LATEST/ .
Zimmerman, F.J. and M.R. Carter. 1999. A Dynamic Option Value for Institutional Change:
Marketable Property Rights in the Sahel. American Journal of Agricultural Economics
81(2): 467-78.
 
 

118
 

CHAPTER 3: EFFECTS OF POPULATION DENSITY ON SMALLHOLDER
AGRICULTURAL PRODUCTION AND COMMERCIALIZATION IN RURAL KENYA

3.1

Background

Reducing poverty and hunger have been a critical policy concern in most of the subSaharan African countries for the past half-century. Governments and development agencies
have experimented with a series of alternative approaches for addressing poverty. Yet, poverty
still remains pervasive. In 2005, more than 40 percent of Sub-Saharan Africa’s population was
estimated to be below the poverty line (World Bank, 2006). More than 70 percent of the poor
live in the rural areas and derive their livelihood from smallholder farming. For this reason,
broad based agricultural growth has been widely understood to be the most powerful vehicle for
reducing rural poverty and kick-starting broader structural transformation processes (Johnston
and Kilby, 1975; Mellor, 1995). Based on green revolution in Asia, growth starting among
smallholder farmers has been viewed as having far higher growth ‘linkages’ than growth in any
other sector (Mellor, 1995; Lipton, 2005; Binswanger-Mkhize, 2012). A major feature of the
structural transformation processes achieved in other parts of the world such as Asia was small
farm-led and thus broad-based (Johnston and Kilby, 1975; Mellor, 1995). Smallholders tend to
spend their incomes on locally produced goods and services, therefore stimulating the rural nonfarm economy and creating additional jobs that would lead to diversification out of agriculture
and rural-urban transition (Hazell et al., 2010; Bryceson and Jamal, 1997; Ellis, 2005).
Consequently, a smallholder-led growth strategy has been touted as having the brightest
prospects for rapid and sustained reductions in poverty and hunger in sub-Saharan Africa
(Lipton, 2005; World Bank, 2007; Hazell et al., 2007; Bezemer and Headey, 2008; Byerlee and
119
 

de Janvry, 2009; Haggblade, 2009; Christiaensen et al., 2010; Eastwood et al., 2010; Headey,
Bezemer, and Hazell, 2010; Wiggins et al., 2010).
However, the widely held view that agricultural development and structural
transformation in sub-Saharan Africa can be achieved by largely replicating the smallholder-led
growth processes in Asia have never fully be juxtaposed with salient features of farm structure
and trends in land productivity in much of Africa. Evidence for Africa tends to show quite
limited land productivity growth in response to rising population density. Unlike in Asia, where
land productivity growth was achieved with the aid of extensive irrigation/water control and
improved seed varieties, which made high application rates of fertilizer use very profitable, by
contrast most of Africa relies on rain-fed production where land productivity growth is
challenging because the economics of fertilizer use are much different. As a result of this, most
of the agricultural growth in sub-Saharan Africa (SSA) has been based on area expansion, not
yield growth, but in many areas of Africa, smallholder farmers are facing increased land
constraints.
While land is relatively abundant in some areas, this tend to exist simultaneously with
land shortages in more densely populated smallholder areas, reflecting failures in land and labor
markets and impediments on migration. Survey evidence consistently shows that farm structure
in SSA is highly concentrated. While mean farm size tends to be around 2-3 hectares in most of
the region, at least 40 percent of farms are under one hectare, and the marketed surplus and
dynamism in smallholder agriculture consistently seems to be confined to 20 percent of the
smallholder households or less (Jayne et al., 2003). Increasing rural population density will only
lead to smaller farm sizes, and without interventions capable of achieving major and broadly
based growth in land productivity, it is not clear how a smallholder-led strategy can kick-start the
120
 

broad based growth processes required to achieve structural transformation as was experienced
in Asia.
Currently, it is not at all clear how a smallholder-led agricultural strategy must be adapted
to address the limitations of small and declining farm sizes and increasingly reduced prospects of
land accessibility in the densely populated areas that are dependent on rain-fed production
systems with only one growing season per year. The overarching issue to be investigated is
whether most farms are becoming, or have already become, “too small” to generate meaningful
production surpluses and to enable smallholders to participate in broad-based inclusive
agricultural growth processes given existing on-shelf production technologies; and whether there
is scope for agricultural intensification. Land poverty and landlessness may pose bigger potential
problems in African countries because of lack of developed industrial sector to absorb persons
squeezed off their lands. In the long run, diminishing land sizes are likely to contribute to
hunger, poverty, and perhaps political instability and violence (Gladwin, 1990).
The overarching objective of this study is to model how rising population pressure affects
smallholder households’ behavior and outcomes, in particular agricultural incomes, productivity,
and commercialization. The study consists of two parts. The first part models how the increasing
population density affects smallholder input demand and output supply. It is hypothesized that
farm households in the relatively densely populated areas will exhibit declining farm sizes over
time that will constrain farm intensification, and thus lower farm productivity and household
incomes. The research questions to be explored in this study include:
1. How are the farming and livelihood systems of densely populated and land-constrained
rural areas evolving differently than those in relatively land abundant areas?

121
 

2. How is increasing population density affecting behavior through (i) diminished
landholding sizes; (ii) increasing labor availability per unit of land cultivated; (ii)
increasing land sales and rental prices; and (iii) increasing food demand?
3. What is the impact of raising population densities on agricultural productivity?
4. Do the marginal products of fertilizer and other purchased modern inputs vary as a
function of population density?
The second part examines the impact of emerging land constraints and rising population
density on smallholder agricultural commercialization and the factors that condition these
relationships. It is hypothesized that rational farmers in modern agricultural system always try to
maximize their net agricultural income from their limited resources. Thus, as land size
diminishes due to human population pressure, farmers are expected to use their resources in the
best way possible to maximize net returns. For example, with low and stable food crops prices,
farmers are expected to rely more on the markets for their food needs and thus able to put a
larger proportion of their land entitlement on high value enterprises.
Consequently, the second part of this study examines smallholder diversification in the
context of how farmers allocate land among competing crops. Commercialization is examined in
the context of farmers’ participation in input and output markets. Consequently, we examine how
allocation of farmland among alternative crops is adjusting to emerging constraints and
opportunities created by increasing population holding other factors constant. Next, we
investigate the extent to which increasing human population density influences the degree of
smallholder commercialization in rural Kenya. While declining landholding sizes might arguably
encourage smallholders to put a larger proportion of their land in food crop production for
household subsistence, stable food prices and improved access to markets may by contrast
122
 

trigger a switch towards agricultural commercialization. When food prices become more stable,
smallholder farmers can be hypothesized to follow a comparative advantage production
orientation that implies greater reliance on markets for food needs. While increased human
population density result in increased food demand, improved infrastructural facilities and
market access are hypothesized to enhance smallholder commercialization by triggering
production of high value and perishable crops. Consequently, whether and to what extent
population growth affects smallholder cropping patterns and commercialization is ultimately an
empirical issue, which requires careful econometric analysis.
Since the two study parts derive from the same theory and use the same data and
methods, we discuss the conceptual framework, data sources, and econometric methods jointly
but present the study results separately. We conclude this section by discussing the study’s
contributions to the discipline as well as policy implications. With regard to the contribution to
the discipline, the study provides an explicit modeling approach for determining the factors
explaining farm productivity growth (or lack thereof) and smallholder commercialization within
the context of endogenous variables. The incorporation of population density into prices and
distances to infrastructure is new. In most of the earlier studies examining the impact of
population density on agricultural production in the region, population density, prices and
distances to infrastructural facilities are taken as an exogenous (Benin, 2006; Pender and
Gebremedhin, 2006; Pender et al., 2006). This study thus shows how to explicitly model input
demand and output supply functions where the dependent variables are either continuous or
fractional in nature (restricted between zero and one) and in the context of endogenous
explanatory variables. The potentially endogenous covariate necessitates the use of the threestage correlated random effect and control function approaches.
123
 

From the policy front, the research comes up with useful policy insights about the future
of smallholder agriculture and the viability of a smallholder-led development strategy for densely
populated areas with limited potential for area expansion. As the land frontier closes in many
countries in many sub-Saharan Africa, smallholder farming areas are confronted by a challenge
of increasing agricultural productivity through technical innovation. The results shed light on the
validity of the structural transformation process; whether there is a threshold of population
density beyond which some other paradigm is needed; and the range of policy parameters that
need to be tweaked for it to be successful in Kenya. The study will provide insights into the state
of production in the densely populated areas, how productions trends are likely to evolve in these
areas in the future, and suggest some institutional and policy measures to be put in place for a
feasible smallholder-led development.

3.2

Overview of the effects of population density on agricultural production

The early literature on households’ responses to increasing human population pressures
came from demographers and anthropologists. This literature posited that households would
recognize the need to change their demographic behavior with increasing population pressure.
Consequently, households are hypothesized to adopt behavioral changes such as higher age at
marriage, increased celibacy, increased contraceptives use and abstinence, and increased
abortions of unwanted conceptions (Bilsborrow, 1987). These behavioral changes were
congruent with the preventive checks suggested by Malthus (1798). In his book, The Principle of
Population, Malthus had earlier on pointed out that population growth had the potential to
outpace agricultural growth leading to starvation. Later literature found this demographic
approach overly narrow in that it neglected the possibility of any economic responses. The
124
 

framework somewhat assumed fixed supply of natural resources, especially land access, and
unchanging methods of production.
The earliest economic conceptualization of the effects of rising population density came
from Boserup (1965). Boserup argued that as arable land becomes scarcer relative to population,
land is used more intensively. As population grows, societies move through the following stages:
(1) forest or long fallow; (2) bush fallow; (3) short fallow; (4) annual cropping; and (5) multiple
cropping. In Boserup’s framework, agricultural intensification is defined in terms of cropping
frequency. The framework postulates a linear switch from long fallows, to shorter fallows,
annual cropping and ultimately multiple cropping.
The post Boserup literature introduces a broader conceptual framework to analyze
responses to population pressure that combines both demographic and economic factors
(Bilsborrow, 1987). Increased population pressure is associated with decreased land holding as a
result of fragmentation, spreading the agricultural work thinly resulting in high labor inputs per
land unit, reduced fallows, and soil nutrient depletion as a result of high plant nutrients extraction
beyond the natural regeneration rates (Drechsel et al., 2001). In response to the increasing
population pressure, households innovate to maintain their accustomed standards of living. They
not only recognize the need to alter their demographic behavior, but also embrace new
production tactics and take advantage of emerging economic opportunities. Bilsborrow (1987)
hypothesizes three possible types of responses to human population pressure, namely, (1)
demographic, (2) economic, and (3) “demographic-economic”. Demographic responses include
changes in nuptiality and declines in marital fertility as suggested by Boserup (1965) and
Malthus (1798). Increased population pressure within a constrained land environment forces
farmers to alter economic production methods, often demanding greater inputs. The economic
125
 

responses constitute actions such as switching from land-intensive to labor-intensive methods of
production; changes in land use patterns, for instance switch from traditional crops to high value
crops potentially leading to agricultural commercialization; adoption of modern output
enhancing inputs such as inorganic fertilizers and hybrid seeds; and increase in land under
irrigation, water availability allowing. By “demographic-economic” response, Bilsborrow (1987)
means actions that have both economic and demographic implications such population
outmigration, seasonal or permanent, from densely populated areas to less densely populated
areas or migration out of agriculture to urban centers.
The theory does not offer much guidance as to which types of responses are more likely
in which situations and whether the responses are mutually exclusive -- occurrence of one
response precludes the occurrence of others responses. Hopkins (1973) points out that, different
responses can emerge in regions which do not show marked variations. Communities with
similar demographic characteristics may respond differently to population pressures depending
on differences in conditioning variables such as technological elasticities due to differences in
soils and climates; differences in work habits; and differences in external influences (Darity,
1980). The conditioning variables help in predicting the likely responses in different situations.
Bilsborrow (1987) does an excellent literature review on the probable response types to
population pressure, the conditioning variables, and country examples where such responses
happened. A summary is presented in Table A3.1. In the remaining part of this section I discuss
in greater detail the economic related responses and their conditioning variables.
While a land extensive technique is more productive with low labor input, a land
intensive technique is more productive with higher levels of labor (Robinson and Schutjer,
1984). A shift in agricultural production methods is hypothesized to occur gradually as
126
 

population pressure increases. Lack of opportunities for colonizing new land spur technological
change that make possible increases in per capita output. As arable land becomes scarce, it is
used more intensively with the help of land-saving inputs courtesy of science and technology,
thereby raising agricultural output per unit of land. As Smucker (2002) points out, intensification
of agricultural production has been defined and measured in a number of ways. These include
total output per unit of land, frequency of cultivation, use of agricultural technologies and
implements, and investments in labor-intensive soil and water conservation measures.
Binswanger and Ruttan (1978), Hayami and Ruttan (1985), Binswanger and McIntire (1987) and
8

Pingali et al. (1987) explain the agricultural intensification process using the induced innovation
model of new technologies adoption in response to factor scarcity. Human ingenuity is

hypothesized to optimally resolve resource constraints imposed by population pressure if the free
markets are allowed to operate freely and appropriate price signals are transmitted to producers
(Simon, 2000). Rising population density induces changes in the relative prices of the factors of
production. The price of the scarce factor, land, increases relative to the price of the more
abundant factor, labor. Prices serve as scarcity signals as well as innovation signals. Farmers
facing shrinking landholding innovate by adopting the already on-shelf technologies or
developing new technologies to raise land productivity. Innovations may also involve bringing
the factors of production together in new ways. Farmers adopt high-yielding and quick-maturing
seeds, inorganic fertilizers, pesticides and herbicides as well as reconsidering crop choices.
Intensification also includes crops irrigation, water availability permitting (Zilverber et
al., 2010). Adoption of high‐value crops or of inputs and irrigation imply labor-intensive
                                                            
8

Although it was Hicks in The Theory of Wages (1932) who first discussed the induced
innovation, the concept has been used by analysts to explain agricultural development within the
context of increasing population pressure.
127
 

cultivation techniques that raise output per unit of scarce land. Factor inputs are applied in
quantitatively and qualitatively new ways; a new 'curve' of production is developed (Brookfield,
1972). As Blaikie and Brookfield (1987) observes, agricultural intensification in not a linear
function; intensification may be limited by environmental thresholds beyond which additional
application of inputs to the production process are not tenable. Also, if the hypothesized landsaving induced innovation process is subject to extremely long time lags (McMillan et al., 2011)
or does not occur (Krautkraemer, 1994) due to institutional bottlenecks and markets
inefficiencies, population pressure is likely to lead to a fall in per capita consumption and lower
household well-being; a result that has a Malthusian effect flavor.
Research based on in-depth case studies has highlighted situations where population
growth has been accompanied by agricultural intensification and improved soil fertility. For
example, Tiffen et al. (1994) presents a case study of agricultural intensification in the semi-arid
district of Machakos, Kenya. Using several photo pairs (1936 versus 1989) of different parts of
the district, Tiffen et al. (1994) contrast the denudated and sparsely populated lands of 1930s
with vegetated, intensively farmed and densely populated lands of the 1990s (Figure A3.1). The
agricultural intensification occurred alongside a fivefold increase in population density.
However, Tiffen et al. (1994) critics argue that the observed changes were facilitated by other
factors such as availability of non-farm income and markets due to the district’s proximity to
Nairobi (the capital city of Kenya), land tenure security (Murton, 1999), and other favorable
9

government policies (Zaal and Oostendorp, 2000; Wiggins, 2012 ).
                                                            
9

“… Masaku (Machakos) benefited from higher coffee prices in the 1950s. But why was that?
Part is down to the 1954 removal of the pernicious restrictions on small farmers growing coffee,
part down to rising world prices; but also very probably owing to public investment in roads that
transmitted prices back to the farm gate.” Steve Wiggins -- personal communications.
128

 

The feasibility of land intensification is dependent on a number of factors. These include
the household endowments (physical, human, natural, financial, and social capital); agricultural
inputs and outputs pricing policy (price control, subsidies and taxes); agricultural research and
development (availability of appropriate technologies); access to programs and services
(technical assistance and credit); access to markets and transportation networks (input and output
distribution networks); the agricultural potential of the land (soil quality and rainfall); and
property rights (Kelly, 2006; Pender et al., 2006). Also, the returns to new technologies and
crops must be high; returns required to induce adoption must be relatively higher than those
required to sustain the use of current technologies (Dercon and Zeitlin, 2009). Households’
demographic characteristics such as education attainment, age and gender of the household head
as well community belief systems may also influence technologies adoption

10

(Feder et al.,

1985). For example, unresolved policy issues and religious belief systems have constrained the
adoption of genetically modified organisms seeds (GMOs) (Hertel, 2011). Household
idiosyncratic shocks (such as illnesses and adult mortality) also condition agricultural
intensification.
Intensive farming requires new approaches to research and extension, as well as an
enabling policy environment. Government policies can make adoption of new technologies a less
costly and risky affair particularly for early adopters. While formal land titles are hypothesized to
facilitate access to credit and help to prevent land disputes, the impact of land tenure security on
land intensification still remains unresolved in the literature (Besley, 1995; Haugerud, 1989;
Atwood, 1990; Migot-Adholla et al., 1991; Bruce and Migot-Adholla, 1994). Weather related
                                                            
10

However, on the other hand, more educated households may be less inclined to invest in
inputs-intensive land investments because the opportunity costs of their labor may be increased
by education.
129
 

challenges and especially rainfall quantities and its variability are also hypothesized to influence
agricultural intensification especially in the rural Africa region where agriculture is rain-fed
(McCown et al., 1992).
Believing that there is a link between population density and agricultural intensity does
not necessarily imply population density causes agricultural intensification (Darity, 1980). There
could be some possible feedback effects prompting potential endogeneity that needs to be
addressed in econometric modeling. As Allen (2001) mentions, agricultural intensification
caused by other factor may allow the system to support more people thus leading to high
population. For example, in Papua New Guinea, Brookfield (1972) encountered intensive
practices in situations where there was no population pressure and extensive practices in areas
where land was short. Brookfield (1972) points out that innovations may be triggered not only by
physical resource scarcities but also by opportunities and government interventions that impinge
upon farmers’ production decisions. In fact, the need to hasten the process of agricultural
intensification through altering market signals has often been the motivation for government
interventions. Farmers may change their land use in response to market signals; market
incentives can induce farmers to intensify production in the absence of population growth and
land shortage. Other factors that are hypothesized to lead to intensified agricultural production
include risk management and production for social

11

purposes (Stone, 2001). Even without

population pressure, intensive production techniques may be adopted to cope with risk of crop
failure and social demands.

                                                            
11

Stone (2001) defines social production as production of goods produced for use by others in
ceremonies and rituals such as births, marriages, funerals, and other ritual gathering; a
production that he says may seem weirdly “uneconomical” but which earn real social dividends.
130

 

3.3

Overview of effects of population density on agricultural commercialization

The structural transformation process has long been considered the main route through
which poverty and hunger in Africa would be overcome. A fundamental element of the structural
transformation process is smallholder commercialization. What constitutes agricultural
commercialization has been a subject of a long standing debate. Smallholder commercialization
is defined in terms of smallholder participation in commercial input and output markets, type of
crops grown, and the goals of smallholder farmers (Braun et al., 1994; Moti el al., 2009). It is
attained when household product choice and input use decisions are made based on the principles
of profit maximization (Pingali, 1997). Thus, smallholder commercialization commonly refers to
the transition from subsistence to market-oriented patterns of production and input use. The term
refers to a cycle in which farmers intensify their use of productivity-enhancing technologies on
their farms, achieve greater output per unit of land and labor expended, produce greater farm
surpluses, expand their participation in markets, and ultimately raise their incomes and living
standards. Agricultural commercialization is also associated with agricultural diversification,
because market-oriented crop or livestock production represent diversification away from
production of basic food for home consumption.
Commercialization of smallholder agriculture has been touted as an indispensable
pathway towards inclusive economic growth, food security and poverty reduction in sub-Saharan
Africa (von Braun, 1995; Pingali and Rosegrant, 1995; Timmer, 1997; Moti et al., 2009). This
thinking is inspired by a number of reasons. First, since the majority of Africa’s population
remains engaged primarily in smallholder agriculture, it is difficult to envision any other sector
that could trigger mass movement out of poverty other than agriculture. Growth starting among
smallholders is suggested to have far higher growth ‘linkages’ than growth in any other sector
131
 

(Mellor, 1995). For the last almost five decades, African policymakers and multilateral
institutions such as the World Bank have focused almost exclusively on improved agricultural
productivity as the way to reduce wide spread rural poverty in sub-Saharan countries. Second,
empirical evidence exists demonstrating that households’ income increases as farm resources are
shifted from subsistence to commercial agriculture (Kennedy and Cogill, 1987; Bouis and
Haddad 1990; von Braun et al., 1994; Dorsey, 1999). Third, with continuously shrinking farm
sizes due to human population pressure, subsistence agriculture may not be sustainable in the
long run. Consequently, smallholder farmers have to intensify to achieve greater output per unit
of land or migrate out of agriculture altogether.
Various indicators of agricultural commercialization have been developed. These
indicators reflect how differently authors construe the concept of commercialization. Moti et al.
(2009) does any excellent review of these indicators and a summary is available in Table A3.2.
At household level, von Braun et al. (1994) suggested three approaches to analyze smallholder
commercialization: output and input side commercialization; commercialization of the rural
economy; and degree of a household’s integration into the cash economy. The first approach
measures proportion of agricultural output sold to the market and input acquired from market to
the total value of agricultural production. The second approach measures the ratio of the value of
goods and services acquired through market transactions to total household income while the
third measures the ratio of the value of goods and services acquired by cash transaction to the
total household income. Govereh et al. (1999) and Strasberg et al. (1999) developed another
indicator of crop commercialization which is a ratio of the gross value of all crop sales per
household per year to the gross value of all crop production. Since agricultural
commercialization can also be interpreted in terms of how smallholder households diversify
132
 

away from production of basic food for home consumption to production of market-oriented
output, the share of land allocated to non-food crops is also used as a measure of smallholder
agricultural commercialization.
Analysts have suggested various factors that are thought to influence smallholder
agriculture commercialization (von Braun et al., 1991; Pingali and Rosegrant, 1995; Pender et
al., 2006). These factors are grouped into two categories, namely: those that are within the
farmer’s control (internal) and those that are beyond his control (external). Factors that are
beyond the smallholder’s control include population growth and demographic changes,
technological change, development of infrastructure and market institutions, development of the
non-farm sector, rising labor opportunity costs, macroeconomic, trade and sectoral policies
affecting prices, development of input and output markets, institutions like property rights and
land tenure, market regulations, cultural and social factors affecting consumption preferences,
production and market opportunities and constraints, agro-climatic conditions, and production
and market related risks. Factors within the smallholder’s control include land, labor, physical
capital, and human capital. A diagrammatic presentation of the determinants and consequences
of smallholder commercialization as per von Braun et al. (1994) is presented in Figure A3.2.
While these factors are highly interlinked, in this study, we are interested in examining how
changes in human population densities are affecting smallholder commercialization.
While it is postulated that agricultural diversification and commercialization is
fundamental part of evolving agricultural development, mounting population pressure is
hypothesized to affect it in a number of ways. First, increasing population pressure leads to
shrinking farm sizes. Land differs from other agricultural inputs in that its supply is essentially
constant and thus its quantity cannot change as price changes (Binkley and McKinzie, 1984).
133
 

With limited and diminishing landholding sizes, the existence of various crop production
possibilities creates a conflict in land allocation among competing crops. Increased use of land
for one crop obviously comes at the expense of others. Early studies argued that as land sizes
diminish, farmers allot a major proportion of their land entitlement to food production for selfconsumption for family food security (Narain, 1965). Production for self-consumption refers to
the production of food grains. Smallholders tend to be highly vulnerable to food price volatility,
thus undertake food crop production to cope with risks inherent in agricultural and labor markets
(Chavas and Holt, 1990; Pingali et al., 1997; Wehbe et al., 2006; Lin and Dismukes, 2007).
Besides, agricultural production in SSA is subject to price and yield uncertainties due to
fluctuations in output prices and unpredictable weather conditions. Farming in this region is
mostly rain-fed and hence any irregularity in weather has adverse production implications. The
production risks are real in SSA because of inexistence of crop insurance schemes to mollify
their severity (Dercon, 2002). Consequently, if the risk to food insecurity is high, smallholders
are better off devoting a large share of their shrinking land to grow their own food rather than
rely on uncertain markets (von Braun et al. 1994; Govereh et al. 1999).
However, if the process of market development contributes to more stable food prices
and reduces the wedge between producer and consumer prices, then smallholder may
increasingly adopt a more commercialized orientation, involving increased reliance on markets
for inputs and outputs. Profit maximization becomes the farm household’s new objective
function and optimal land allocation depends upon the revenues and costs in its various uses
(Binkley and McKinzie, 1984; Pingali and Rosegrant, 1995). Such processes could increase the
value of output per unit of land and partially or fully offset the trend toward declining farm size.
Thus, the extent to which the shrinking land sizes are affecting cropping patterns and how
134
 

stabilized/volatile food prices are conditioning the relationship is an empirical question that this
study seeks to answer.
Second, the population pressure represents growing demand for food and this effect is
likely to be reflected in increasing relative food prices. Changes in these factors are likely to lead
to changes in cropping patterns. A high relative price of any crop motivates farmers to grow that
particular crop, holding all other factors constant including technology. Since the output prices
are observed at a later date, after the production decisions and investments are made, earlier
studies have shown that it is the expectations about the future prices that condition farmers’ crop
choices (Nerlove, 1956; Tegene et al., 1988). Thus, price expectation models, such as the
Nerlovian adaptive price expectations model, have been used in agricultural supply response
analyses studies. Also, as De (2005) notes, changing output prices may not bring about
significant changes in production patterns in the short run. This is associated with biological lags
that are typical in agricultural production.
Third, high population density is associated with increased agricultural
commercialization due to increased urbanization and improved access to markets. High
population density is also associated with urbanization. Urbanization in turn increases demand
for marketed agricultural products which tend to increase commodity prices and stimulate
agricultural production for the market. High population density is also associated with improved
economic infrastructures owing to reduced average cost of infrastructure development (Glover
and Simon, 1975; Frederiksen, 1981; Chu, 1997). Improved infrastructure represents enhanced
access to agricultural inputs and output markets. Consequently, increased demand for
agricultural products and improved access to markets is hypothesized to induce not only
agricultural production in general, but the production of high value and perishable products,
135
 

other factors such as technology and environmental conditions allowing. The high value usually
refers to fresh fruits and vegetables and dairy. Subsequently, the extent to which the
opportunities presented by the increasing population density, shrinking land sizes
notwithstanding, are driving smallholders’ commercialization is an empirical issue being pursued
in this study.
Fourth, increased population pressures results in land degradation and soil nutrient
depletion (Drechsel et al., 2001; Pender and Alemu, 2007). If left unattended, soil degradation
will lead to lower agricultural productivity. Households facing shrinking land sizes are expected
to increase the use of modern productivity enhancing inputs obtainable from the markets leading
to increased smallholder commercialization from the inputs perspective. Consequently, the
intensity of use of purchased agricultural inputs as population density increases by the
smallholders is an empirical issue that this study examines.

3.4

Conceptual framework

3.4.1 Agricultural household model

Smallholder agriculture systems in sub-Saharan Africa are characterized by semicommercial farms that produce multiple crops. These systems combine two fundamental units of
microeconomic analysis, the household and the firm, that are highly interdependent. As opposed
to the purely subsistence systems, in semi-commercial systems some farm inputs are purchased
and some outputs are sold in the markets (Table A3.3). To analyze the semi-commercial systems,
Singh et al. (1986) proposed a theoretical framework popularly known as the agricultural
household model that captures the farm household’s consumption and production
interdependences in a theoretically coherent manner. In this framework, the objective of farm
136
 

households is assumed to be maximization of expected household utility subject to budget and
other resource constraints. Agricultural production either contributes to household’s resource
constraint through consumption or through cash generation if farm output is sold at market.
Thus, agricultural production is incorporated as part of the household’s budget constraints. Later,
de Janvry et al. (1991) extended the original Singh et al. (1986) agricultural household model to
include market failures while Omamo (1998) incorporated transactions costs.
In the extended agricultural household model, the household problem is to maximize its
utility:

max U  U ( X a , X m , X l )
where the commodities are agricultural goods

leisure

(3.1)

( X a ) , market-purchased goods ( X m ) , and

( X l ) . Utility is maximized subject to several constraints, among them: a cash

constraint, production technologies for own-farming and nonfarm self-employment activities;
exogenous effective prices for tradables; an equilibrium condition for self-sufficiency of farm
production; and an equilibrium condition for family labor. First-order conditions of this model
give a system of factor supply and demand functions, which in turn allows the estimation of
factor inputs and supply functions. Depending on the objective of the study, the dependent
variable could defined in terms of quantity of inputs and outputs or the income shares coming
from different crops or enterprises.
At the minimum, the theory posits that the desired supply is a function of the expected
output price, and supply shift variables such a vector of input prices, and the expected output and
input prices of other production possibilities. Since the objective of this study is to examine how
137
 

human population density affects smallholder agricultural production and income, the immediate
next task is to conceptualize how this variable of interest enters the input demand, output supply
and income functions. A diagrammatic presentation of how population density influences these
smallholder production outcomes is presented in Figure A3.3. Assuming markets are allowed to
operate freely and the appropriate price signals are transmitted to producers, escalating
population density is hypothesized to affect agricultural production through three fronts, namely,
decreasing land holding sizes, increasing labor supply, and increasing demand for food. From the
first two fronts, the decreasing land sizes are hypothesized to trigger changes in relative factor
prices, consequently triggering changes in the land-labor ratio. The price for the scarce factor
(land) is bid high while the price for more abundant factor (labor) declines. From the third front,
population growth directly affects the demand for agricultural products by shifting the demand
curve for food crops outwards. This shift exerts pressure on food prices thereby inducing a
supply response, other factors held constant including internal and external trade, thus putting
more pressure on the factor prices.
Relative prices determine how various factors of production are combined in the
production process by profit maximizing or cost minimization producers. Prices serve as scarcity
signals as well as innovation and adoption incentives for new technologies. According to the
“induced innovation” theory, a change in the relative price of factors influence the factor use
proportions (Hicks, 1932). A change in the relative prices of factors of production spur
innovations aimed at economizing the use of a factor which has become relatively expensive.
Since changes in population density influences input demand and output supply indirectly
through prices, this process suggests a first stage reduced form regression of output prices on
population density variable among other relevant covariates.
138
 

However, in a world characterized by market imperfections, factor prices may not fully
pick up the effects of the increasing population pressure (Ghebru and Holden, 2009). For
example, in many African countries Kenya included, land sales markets are characterized by
information asymmetry, enormous transaction costs, government bureaucracy, and ethnicity and
cultural constraints. If this is the case, increasing population density may affect landholding sizes
and input demand and supply functions in ways that are not fully reflected in market prices.
Thus, the existence of inefficient markets suggest a first stage reduced form regression of
household landholding on population density among other relevant covariates and the inclusion
of population density variable in the estimation of input demand and output supply functions.
As mentioned in Section 3.2, the hypothesized relationship between population density
and agricultural intensity does not necessarily imply population density causes agricultural
intensification. There could be some possible reverse causality signaling potential endogeneity of
the population density variable. Population density drives food production (Boserup, 1965),
while food production could also drive population density (Malthus, 1798). Besides, current
indicators of land intensification potential and farm size may influence migration which will in
turn affect future population density. Households in areas with low potential and/or declining
land access may choose to migrate to areas which they perceive to be of relatively higher
potential and/or better land access. When confronted by potentially endogenous variable, two
methods are available to get around the problem. First is the usual two stage least square (2SLS)
instrumental variable method while the second is the control function approach (Wooldridge,
2010). To implement either of the two methods, it requires the availability of at least one
instrumental variable (IV); a variable that is correlated with population density but uncorrelated
with indicators of land intensification. Consequently, to appropriately model how population
139
 

density influences agricultural intensification, we will need to estimate a reduced form
population density equation using covariates that are hypothesized to influence population
growth and at least one plausible IV.
In the remaining part of this section, we conceptualize the drivers of population growth
with aim of identifying plausible instruments to be used in the first stage population density
reduced form regression followed by conceptualizing how the population density variable enters
the prices and landholding equations.

3.4.2

Drivers of human population growth

Economics as a discipline has paid little attention to population studies. The earliest
conceptualizations of the drivers of population growth come from demographers and
sociologists. Most of these studies have used demographic transitions theory to explain the
causes and mechanisms behind human population changes over time (Notestein, 1945). The term
demographic transition has two primary meanings (Johnson-Hanks, 2008). First, it refers to the
historically specific change from high to low rates of fertility and mortality that human
populations have undergone since 1750, and second, to a set of theories regarding the causes and
mechanisms behind that change. The complementary approach of economists conceptualizes
population growth through the various factors influence the demand for children (Bilsborrow and
Winegarden, 1985). Economists hypothesize that the demand for children will increase with an
increase in the family income and a reduction in the costs of raising them. The supply side is
influenced by factors such as the knowledge and use of fertility-regulating methods. These
theories are basically household level based and are therefore inadequate in explaining the
drivers in aggregate changes in populations densities over time.
140
 

Consequently, in this study we will be guided by the demographic transitions theory.
According to the demographic transitions theory, as living standards rise and health conditions
improve, first mortality rates decline and then, somewhat later, fertility rates decline (Kirk,
1996). People would naturally deploy the advantages of modernity to reduce death rates, but that
fertility rates would be stalled by cultural factors that would only slowly give way (Notestein,
1945). Population growth is fuelled by two components: the “demographic momentum”, which
is built into the age composition of current populations, and changes in reproductive behavior,
mortality and migration. As Fischer and Heilig (1997) explain, the ‘echo effect’ of a highfertility period in the past creates a demographic momentum that counters reproductive control
measures that favor smaller families. No significant changes in fertility are expected to counter
the demographic momentum since most developing countries have large rural populations whose
fertility is driven by deep-rooted cultural norms and values, and religious beliefs. Fertility rate is
loosely defined as the average number of children that would be born to a woman over her
lifetime. Other determinants of fertility include age at marriage (or beginning of sexual activity),
prevalence and effectiveness of contraception, prevalence of induced abortion, and duration of
postpartum infecundability, especially due to breast feeding (Lutz and Qiang, 2002). Speakers of
the same language are generally expected to experience fertility transition at the same time
(Johnson-Hanks, 2008).
Changes in reproductive behavior include changes in nuptiality and declines in marital
fertility as suggested by Boserup (1965) and Malthus (1798). Changes in mother’s literacy and
infant mortality rates greatly influence fertility rates. Studies have found inverse relationship
between mothers’ education attainment and family sizes (Fischer and Heilig, 1997). However,
better health status of women on the other hand increases fertility rates. The decline in mortality
141
 

among infants resulting from the spread of modern hygiene and medicine is associated with
decreased fertility. Adult mortality declines due to increased life expectancy as a result of
diseases eradication or control with the advancement in modern medical technology.
Next, we examine the role of migration in rural population redistribution. As Bilsborrow
(2001) explains, resource scarcity or depletion drives human migration. Migration, be it rural to
urban or rural to rural, decrease rural density in one area while raising it in others, as rural
populations leave areas with scarce resources in search of resources and opportunities elsewhere.
At the macro-level, perhaps the leading determinants of migration and displacement of people is
caused by scarcity of water and land, conflicts over natural resources, natural hazards and natural
disasters (Naude, 2010). Inter-ethnic and political conflicts also influence migration. At the
micro-level, factors influencing people's decisions to migrate include differences in economic
opportunities and living conditions between origin and destination areas, people’s awareness of
those differences, distance to the potential destination, and their ability to move (Bilsborrow and
Winegarden, 1985). Under these broad categories come educational levels, and differences in
wage rates and living conditions. Cultural factors such as psychological or emotional
attachments to family, friends, and community stop people from migrating as well as the
government policies such as land tenure system.

3.4.3

Effects of population density on household landholding

In most studies, landholding sizes are generally assumed to be fixed (Zhang et al., 2009).
However, the factors that affect landholding size and how landholding size changes with socioeconomic circumstances are important issues that have not been fully investigated. Previous
studies relating to the size of landholdings have often been centered on community well-being
142
 

(Gilles and Dalecki, 1988; Swanson, 1988; Labao and Michael, 1991). Only a handful of them
have explicitly examined the impact of increasing population densities on farm land size.
Land in the African context is not just like any other economic commodity (Shipton and
Goheen, 1992). Landholding in this region is an outcome of complex historical and geographical
processes that may fail to conform to any known economic theory. Consequently, we use
previous studies to conceptualize how population density affects household landholding.
Households acquire rights to obtain and maintain access to land or they lose these rights through
multiple paths such as social groups or networks membership, kinship ties, and through purchase
(Shipton and Goheen, 1992; Berry, 1993).
Even in countries where land markets are well developed, access to land through
inheritance remains fundamental. Land transfers from parents to children depend on the parent’s
landholding sizes and the number of potential recipients of the parents’ assets. In polygamous
families, land is shared equally between wives regardless of their number of children. Females
have lesser claim on parental resources (Garg and Morduch, 1998; Morduch, 2000). Females’
land access comes through ties to husbands (Berry, 1993; Smucker, 2002). A senior wife has
stronger rights than a junior wife. A woman's rights increase with the length of marriage or with
number of children and the rights may end with divorce, with widowhood, with failure to have
sons (Gray and Kevane, 1999). As Gray and Kevane (1999), transfer rights are usually limited
for women -- they cannot designate an heir, sell land, or lend land to others. The growing human
population density triggers land subdivisions resulting in tiny landholding pieces. Aging
population results in land transfers and parcelization (Gobster and Rickenbach, 2003). Other
factors influencing landholding at the household level include agricultural technology and
education levels (Shipton and Goheen, 1992).
143
 

Historically, family access to land depended on their membership and good standing
within a particular group exerting control over the land. Kinship and ethnic (tribe) adherence
along with status, gender and seniority determined access and use rights (Berry, 1989; MigotAdholla and Bruce, 1994). Social capital is also an important factor determining access to land
and other assets. Jayne et al. (2008) show that households in which the male head is related by
blood to the local headman have more land than other households in Zambia. However, with
mounting population densities and the ensuing land squeeze, land access through within
community transfers is becoming less and less feasible option.
Land sales markets are becoming crucial in the re-allocation of land (Pinckney and
Kimuyu, 1994; Place and Migot-Adholla, 1998; Holden, et al. 2009). If markets are functioning
efficiently, access to land though this channel is depended on land sales prices among other
factors. With the swelling population densities, land prices have been bid so high rendering this
land access channel inaccessible to the land poor. As mentioned earlier, land sales markets in
SSA are characterized by many imperfections as a result of factors such transaction costs and
ethnicity related constraints. This calls for a direct inclusion of the population density variable in
the reduced form household landholding model.

3.4.4

Effects of population density on land rental prices

Though land rent literature is distinct from land price literature, studies have found that
farmland price and rent movements are highly correlated and to larger extent land rent is the
most widely accepted factor driving farmland price Falk (1991). In this section, we explore the
process through which land sales and rental prices are determined and the role increasing
population density plays. Just like in other markets, equilibrium prices in the land markets are
144
 

determined by the forces of demand and supply. In this study, we assume that what we observe
and collect data on are the equilibrium prices after the demand for land available for sale or
rental purposes is equated with supply and the households’ solve their profit maximization
problem. Therefore, we model the observed equilibrium prices as a function of the variables that
drive the supply as well as demand side of the land sales and rental markets to equilibrium.
Increasing population density directly puts pressure on the supply side of the land
markets thereby pushing up prices (Weerahewa et al., 2008). Indirectly, increasing population
density represents increasing demand for food and thus mounting pressure on factors of
production; land being one of them. From the Ricardian rent theory, land prices are highly
depended on land productivity (Ricardo, 1821). The higher the quality of land available for sale
or rental purposes, the higher the price. Thus, from the supply side of the market, land pieces that
are more fertile and closer to the market represent lower cost of production and thus command
relatively higher prices. From the demand side, Du et al. (2007) shows land rental rates are a
function of the expected market prices. For example, the rental price an individual is willing to
pay for a piece of land at the planting time is determined by the expected profitability of the land
as reflected in the output’s expected market prices at the time of harvest. Land prices are also
greatly influenced by development pressure emanating from the nearest urban centers (Shi et al.,
1997; Livanis et al., 2006).

3.4.5

Effects of population density on agricultural wages rate

The effects of population density on agricultural wages depend on how population
pressure affects the determinants wages, in other words, the demand and supply conditions in
agricultural labor market. In this study, we follow previous studies that have attempted to model
145
 

how agricultural wage rates respond to certain economic and demographic changes. It is assumed
that a substantial proportion of the rural population is primarily is dependent on agricultural
wage employment, and labor is hardly organized for group bargaining or covered by any wage
protection law by the government. Just like the land price case, we assume that what we observe
and collect data on are the equilibrium wage after the demand for labor equates with the labor
supply. Consequently, we model the equilibrium wage rate as a function of the variables that
drive the supply and the demand side of the labor markets to equilibrium.
On the supply side, increasing population density increases the pool of the potential
workers. Most of the earlier studies have argued that an increase in the price of agricultural
outputs increase the demand for agricultural labor. Khan (1984) explains that increase in
agricultural productivity increases wages by stimulating the demand for labor. The effect of
increase in crop prices, however, can go either way. On one hand, higher crop prices encourage
production and increase the labor demand; on the other hand, these higher prices increase
workers’ consumption expenditure on food and therefore, induce higher labor supply especially
among the landless workers who are net-buyers of food (Boyce and Ravallion, 1991; Ravallion,
1990; Hani, 1996). Thus, the ultimate effect of increase in crop prices is an empirical issue.
The other important nonwage determinants of labor demand are identified as agricultural
productivity (in terms of yield per unit of land); the important nonwage determinants of labor
supply are food prices, prices of other non-food items that workers consume and their alternative
nonfarm sources of income (Ahmed, 1981; Bardhan, 1984; Hossain, 2004; Khan, 1984).
Technological factors such as use of labor-saving or land-saving technologies (e.g. fertilizers and
tractors) also affect wage rates. The expansion of alternative means of livelihood, including
nonfarm employment, which reduces the workers’ dependence on land, also increases their real
146
 

wages (Hossain, 2004). Proximity to urban centers and reduced distances to infrastructural
facilities also influence wage rate in two different ways. First, reduced distances to urban centers
and infrastructural facilities exposes workers to other nonfarm income earning opportunities thus
bidding high the agricultural wage. Thus, trends in labor wage rates may be influenced greatly by
wages in the non-farm sector. Second, reduced distances represents reduced agricultural
production costs and probably reduced demand for agricultural labor. Thus, the impact of
reduced distances on labor is an empirical question.

3.4.6

Effects of population density on agricultural output prices

The theory posits that the input demand and the desired supply are a function of the
expected output price and demand and supply shift variables such a population density.
Population growth directly affects the demand for agricultural products by shifting the demand
curve for food crops outwards. This shift exerts pressure on food prices thereby inducing a
supply response that puts pressure on the factor prices, other factors held constant. Due to
biological lags in the production, output prices are however not observed at the time the
production decisions are taken. It is important to note that we can only observe and collect data
on the market equilibrium prices after they are determined by the demand and supply forces.
Consequently, we conceptualize how to model the equilibrium price as a function of the
variables that drive the supply and the demand of the crops market to equilibrium.
Smallholder production systems in sub-Saharan Africa are highly diversified and are
partially integrated into the markets. This situation presents two challenges in econometric
modeling. First, farmers grow a wide array of crops on one land plot each season and crop
enterprises vary across agroecological zones making it difficult to obtain a balanced panel data
147
 

on crop production and the respective prices. Second, prices data is rarely reported for crops
produced basically for home consumption without marketed surplus. Since most of the
smallholders are partially integrated into markets, it is assumed that their main objective is
production for household consumption. If the smallholders can be assured of low and stable
maize prices then they can diversify their cropping patterns and produce for the markets; else
they produce staples for self-sufficiency. Consequently, it is assumed that the expected price of
maize, the main staple crop, is the most important output price determining production decisions
among smallholders in rural Kenya.
Following Mason (2011), we use an approach somewhat similar to quasi-rational
expectations (Nerlove and Fornari, 1998) to estimate the expected subjective maize price values.
We model expected farm-gate maize prices as a function of variables observed by the farmer at
planting time. These include wholesale market prices of maize from the nearest regional market,
effective NCPB pan-territorial prices, and household and village level characteristics that might
affect the maize sale price received by a given household. We do not replicate a price prediction
model for each of the non-maize crops as we due to data limitations. Instead, following Mather
and Jayne (2012), we use a naïve price expectation for each crop, which is the wholesale price at
the nearest regional market prevailing during the planting season. However, because most
outputs are regionally tradable and Kenya is at least a semi-open economy, we hypothesize that
population density may not have as great an impact on output prices as population density would
on prices of non-tradables, such as farmland and labor.

148
 

3.5

Data sources

3.5.1 Household longitudinal data

The study mainly draws from the nationwide Egerton University/Tegemeo Institute Rural
Household Survey, a panel dataset tracking roughly 1,300 small-scale farm households in 5
survey waves over the 13-year period from 1997 to 2010 (Figure A3.4). The sampling frame for
the panel was prepared in consultation with the Kenya National Bureau of Statistics (KNBS) in
1997. Twenty four (24) districts were purposively chosen to represent the broad range of agroecological zones (AEZs) and agricultural production systems in Kenya. Next, all non-urban
divisions in the selected districts were assigned to one or more AEZs based on agronomic
information from secondary data. Third, proportional to population across AEZs, divisions were
selected from each AEZ. Fourth, within each division, villages and households in that order were
randomly selected. In the initial survey in 1997, a total of 1,500 households were surveyed in
109 villages in 24 districts within eight agriculturally-oriented provinces of the country.
Subsequent surveys were conducted in June of 2000, 2004, 2007 and 2010. Over these 5
panel surveys, 1243 household were able to be consistently located and surveyed. For this
analysis, households in the coastal region of the country were excluded because farming is found
to account for a relatively small share of household incomes. This leaves a balanced panel of
1146 households surveyed consistently in each of the five years. The surveys collect information
on demographic changes, movements of family members in and out of the household since the

149
 

prior survey, landholding size, land transactions and renting, farming practices, the production
12

and marketing of farm products, and off-farm income-earning activities .
Attrition bias is a potential problem in panel data estimations. The average attrition rate
between any two consecutive rounds is about five per cent. Jin and Jayne (2013) estimated reinterview models to assess the degree to which attrition could be a problem in this panel data.
While the results are not presented here due to space limitation, the authors find that the
observed attrition is largely random. Accordingly, no need to worry about selection biases
caused by attrition, although efficiency is somehow lost because of a reduced sample size.

3.5.2

13

Geographic information systems (GIS) data

The household panel survey instrument captured the geographic positioning system
(GPS) coordinates of each household. This made it possible to compliment the survey data with
geographic information systems (GIS) data on soil quality and more disaggregated data on
current and historical population numbers at the villages where the panel households are located.
Data on population densities was extracted from the Global Rural-Urban Mapping Project
14

2

(GRUMP) . To obtain a density (persons/km ), the GRUMP population estimate was divided
by land area suitable for agriculture. The population estimate is the average population count per
pixel for all pixels within a 10km radius.

                                                            
12

Each of these surveys instruments, which contain the details of the types of information
collected and used in this study, can be viewed and downloaded at
http://www.aec.msu.edu/fs2/kenya/index.htm.
13
Jordan Chamberlin, a colleague and fellow PhD candidate did the GIS data extraction.
14

 

See http://sedac.ciesin.columbia.edu/gpw/docs/UR_paper_webdraft1.pdf.
150

The other variables extracted from the GIS sources included the length of the growing
period (LGP) (Fischer et al., 2000); net primary productivity (NPP) ( Zhao et al., 2005); and
elevation (meters above the sea level) and slope (measure of steepness -- degrees), both from
Wilson et al. (2007). The LGP is a useful indicator of agricultural potential. It combines
information on temperature and available moisture to determine the length of time crops are able
to grow. That is, periods which are too cold or too dry (or both) for crop growth are excluded.
LGP is often measured in terms of the number of days experiencing temperatures > 5°C when
moisture conditions are adequate for plant growth. The NPP refers to the standing biomass in a
given area at a given time. It is typically measured in units of mass / area / time. NPP is
sometimes used as an alternative measure of agricultural potential, since it reflects the amount of
vegetative biomass growing in a given area (whether cultivated or not).

3.5.3

Regional market maize price and weather data

The study also drew from the monthly wholesale price data for maize and for each of the
main food and cash crops collected from regional wholesale markets across Kenya by the Market
Research and Information Department of the Ministry of Agriculture. Data on rainfall estimates
comes from the Famine Early Warning System (FEWS), which was produced at the level of
every 0.1 degree latitude and 0.1 degree longitude. This data interpolates rainfall estimates based
on data from rain stations as well as satellite data (such as on cloud cover and cloud top
temperatures). The FEWS rainfall estimates were then matched to Tegemeo panel survey
households using the GPS coordinate.
 
 

151
 

3.6

Data analysis methods

3.6.1

Descriptive analysis

As a prelude to the econometric analysis, bivariate descriptive analysis are carried out to
examine how various household groupings based on population density of the villages they are
located evolve over the panel period in terms of demographic trends, farming patterns, farm
production, agricultural commercialization and household incomes. We superimposed the
2

longitude-latitude coordinates of the 109 villages in the Tegemeo survey on the 10km pixel
population density estimates from the Global Urban-Rural Mapping Project database for 2009, to
obtain population density estimates for each village. Population densities in the sample ranged
2

2

from 44 persons per km in the case of Laikipia West to 965 persons per km in Vihiga District.
We then sort these 109 villages by population density and next stratified them into five equal
population density groups, or quintiles. Population densities range from 30 to 147 persons per
2

km in the lowest quintile, 148 to 313 in the second quintile, 315 to 470 in the third quintile, 475
2

to 655 in the fourth quintile, and 659 to 1135 persons per km in the highest quintile. We then
examine how the five groups are evolving differently over the 1997-2010 period in terms of two
main features:
i.

Farming patterns: changes in farm size, land values, rental rates, land-to-labor ratios,
input intensity per unit of land cultivated and cropping patterns. The 2007 survey also
contains a module exploring household members’ inheritance of land and the amount of
land controlled by their parents.

ii.

Farm production, assets and household incomes: changes in incomes from crops, animal
production, and non-farm income as well as household asset wealth.
152

 

iii.

Indicators of smallholder commercialization: Land allocation to non-maize crops, nonmaize crop production, intensity of purchased inputs use, and value of crops and
livestock product sales.

3.6.2 Econometric models

We are interested in measuring the effect of population density on smallholder input
demand and output supply, commercialization, and household income. As mentioned in the
conceptual framework and summarized in Figure A3.3, three issues must be considered when
specifying the econometric models to be estimated: First, population density is likely to influence
these variables either directly or indirectly through its effects on input factor prices and output
prices. Second, population density is potentially endogenous in the factor demand and output
supply models. Third, while the expected output prices directly influence the output supply, they
also influence output supply indirectly through effects on input factor prices. As mentioned
earlier on, we use the control function approach suggested by Wooldridge (2010) to circumvent
the endogeneity problem.
Considering all these three issues suggest the following estimation strategy: (i) first stage
estimation of population density model on a vector of covariates where at least one of them is a
plausible instrumental variable; (ii) second stage estimation of maize prices model whereby
population density enters as a covariate while controlling for its potential endogeneity; third
stage estimation of input factor and output supply functions where population density and its
residuals from the first stage regression, the expected maize price (predicted maize values from
the second stage regression) and input factor prices enters the vector of the covariates.
Consequently, we estimate the following models:
153
 

First stage: Population density equation

Dit  H i   it

(3.2)

Second stage: Maize price equation

ˆ
Pit   01 Dit  Z 0,it 0   0 it  c0,i   0,it

(3.3)

Third stage: Input factor prices, farm size, input demand and output supply equations

ˆ
ˆ
Fit  11 Dit   11 Pit  Z1,it1  1 it  c1,i  1,it

(3.4)

ˆ
ˆ
Lit   21 Dit   21 Pit  Z 2,it 2   2 it  c 2,i   2,it

(3.5)

ˆ
X it  31Dit   31Pit  Z3,it3  Fit ( Dit )3   3 Lit ( Dit ) 

(3.6)

ˆ
3 it  c3,i  3,it
ˆ
Qit   41 Dit   41 Pit  Z 4 ,it 4  Fit ( Dit )4   4 Lit ( Dit ) 

(3.7)

ˆ
 4 it  c4 ,i   4 ,it
In the first stage, we estimate the population density equation (3.2). The dependent
variable, D , is measured at the village level where the household is located and is defined as the
number of persons per square kilometer of the potentially arable land rather than the standard
total surface area; H is a vector of the covariates that includes unity as its first element and
other variables that influence population growth (population growth momentum, fertility and
mortality rates) as identified in the conceptual framework section. The variables include land

154
 

quality variables (length of growing period, net primary productivity, elevation and slope),
village population size in 1950, literacy rates, distances to water source, religious affiliations,
contraceptive use, age at first marriage and first intercourse, child mortality, and ethnicity. Most
of these covariates are hypothesized to be correlated with population growth variable but
uncorrelated with agroecological potential where these households are located, thus plausible
instrumental variables.
In the second stage, the maize price model (3.3) equation is estimated. The dependent
variable price vector P is the price of maize per kilogram. The vector

Z 0 includes unit as its

first element, household ownership of means of transport (truck and/or bicycle), distances from
the homestead to the nearest infrastructural faculties, level of investment in in storage facilities,
maize buyer type, regional maize price at planting time, NCPB maize prices in the previous year,
and demographic characteristics of the household head (gender, age, and level of education). In
this model and those in third stage, we include population density variable, and its square if
necessary, to test for its direct effects on the dependent variable. The null hypothesis for the
absence of direct effect is

ˆ
ˆ
  0 . The residuals (  ) from the first stage population density

estimation are also included. The inclusion of the residuals (

ˆ
 ) from the first stage population

density reduced form regression into the second stage regressions, help in breaking the
endogeneity link between the population density variable and the error terms (  ) in the second
and third stage models. The null hypothesis

ˆ  0

tests the exogeneity of population density

variables in the second and third stage models. While
effects,



represents the unobserved time-varying effects.
155

 

c represents the unobserved time-constant

In the third stage, four set of models (3.4 to 3.7) are estimated. The dependent variable
price vector Fit includes land rental rates per hectare for a year, and agricultural wage rate per
day. We do not model land sales prices due to data limitations. Besides, as mentioned elsewhere
in this paper, land sales markets are characterized by high transaction costs and thus assumed to
be inefficient. Consequently, we model household landholding ( L ) directly. We also do not
model fertilizer prices since fertilizer prices are not determined locally but in the international
markets. The dependent variables X and Q are defined depending on the study objective. In
the first part of the study (effect of population density on smallholder input demand, output
supply and income), X represents the intensity of fertilizer and purchased input use per hectare
owned while Q represent crop production per hectare owned.
Smallholder production systems in sub-Saharan Africa are highly diversified and are
partially integrated into the markets. This situation presents a challenge for modeling in a
number of ways. First, farmers grow a wide array of crops on one land plot each season and crop
enterprises vary across agroecological zones making it difficult to obtain a balanced panel data
on crop production and the respective prices. Second, the wide array of crops produced implies
too few degrees of freedom for statistical modeling. Given these circumstances, it becomes
imperative to aggregate the outputs in some manner. To aggregate crop production across
multiple commodities, we convert crop production into monetary values using prices and
modification of the Fisher-Ideal index (Figure A3.5) suggested by Mason (2011). In the second a
part of this study, the dependent variables are the indicators of smallholder diversification and
commercialization. Consequently, in this study X represents the proportion of land allocated to
non-maize crops to the total land owned (input side) while Q captures the proportion of
156
 

marketed crop production to the total crop production (supply side); that is, the household crops
commercialization index (HCCI).
Next we discuss the explanatory variables used in the third stage models. Just as in the
estimation of model (3.3), population density variable and its residuals (

ˆ
 ) from the first stage

ˆ
regression are also included. We also include the predicted maize prices ( P ) from the second
stage maize model estimation as a proxy for the farmers’ output price expectations. It is
important to note that all the vectors Z include unity as the first element. The vector

Z1 include

the district mean land holding sizes, land quality variables, distances to infrastructural faculties,
naïve expectation of maize and beans prices—prices prevailing in the regional markets at the
planting time, survey year and agricultural zone dummies. The vector Z 2 include household
demographic variables, (gender, age, education attainment of household head and household
size), land holding of the household of the household head’s father before sub-division, land
holding of the spouse’s father before sub-division, household duration in the current location,
and tribe and survey year dummies. The vectors Z 3 and Z 4 include land productivity variables
(length of growing period, net primary productivity, elevation and slope); expected rainfall and
15

expected drought shocks ; fertilizer price; distance to the nearest motorable road; ownership of
radio (access to information); household demographic variables (gender, age, and level of
education of the household head and household size) among other variables.
                                                            
15

These two variables were computed by David Mather. Expected rainfall is defined as a sixyear moving average of rainfall prior to the main growing season in survey year, while expected
rainfall shock is a six-year moving average of the percentage of 20-day periods during the main
growing season with less than 40 mm of rainfall.
157

 

The third stage models are estimated as a system using the correlated random effects
(CRE) approach (Mundlak 1978; Chamberlain, 1984). The reason as to why we use CRE is
because some of the covariates in these models are time constant and thus drop out if fixed
16

effects (FE) estimation approach is used . The CRE approach involves the inclusion of the
long-term average of each time-varying variable in the model. However, in the second part of
this study, since it involves non-linear models, we estimate the models equation by equation. The
dependent variables are essentially fractional response variables; they are continuous strictly
between zero and one. Consequently, we use the Fractional Probit estimation method proposed
by Papke and Wooldridge (2008).
To compute the total partial effects of the population density on input demand and output
supply function we use the following method:

X it X it  X it dFit   X it dLit 

 
 F  dD    L  dD 
 

Dit  it
Dit
it 
it
it  
 
 
DIRECT

(3.8)

Qit
Qit  Qit dFit   Qit dLit 

 
 

 F  dD    L  dD 
Dit
Dit  it
it 
it
it  
 
 


DIRECT

(3.9)

INDIRECT

INDIRECT

It is also important to mention that since the estimation of the second and third stage
models involve generated regressors, standard errors generated by most econometric software for
the coefficients are bound to invalid since they ignore the sampling variation in the estimation of
                                                            
16

Random effects (RE) also allow the inclusion of time-constant variables. However, the
assumption that fixed effect factor is not correlated with the explanatory variables is often not
plausible.
158
 

the coefficients in the first two steps. Disregarding the sampling error in the generated regressors
is likely to underestimate the computed standard errors. Consequently, we use the bootstrap
approach with 500 replications to get a valid estimate of the standard errors. Inferences are also
made fully robust to arbitrary heteroskedasticity and serial correlation. The results from the input
demand and output supply models are presented in the next section.

3.7

Results and discussions

3.7.1

Descriptive results

This initial section discusses bivariate relationships as a prelude to the econometric
findings. Table 3.1 presents information on farm size and farming practices by village population
density quintiles over the four survey years. The results clearly show that landholding sizes and
areas under crop are decreasing functions of the population density. Landholding sizes among
the smallholders in the 20 percent most densely populated are roughly one third of those in the
20 percent low densely populated group. Over the 10-year panel period, landholding size in the
20 percent most densely populated villages averaged 1.35 hectares. In the same period,
landholding sizes in the 20 percent least densely populated villages averaged 4.25 hectares. The
same scenario is replicated when we look at landholding per adult equivalent across the
population density quintiles. The areas under cultivation have somehow declined across the
population density categories over the 10-year period. The area under crop in the highest density
quintile averaged 0.89 hectares and are about half of those in the lowest density quintile. The
proportion of farmland under fallow (uncultivated area) has also declined by about 20 percent
over time across all the quintiles.

159
 

Table 3.1: Farming practices and factor intensities, by pop. density quintile

Population
Survey year
Four survey panel
density
2000
2004
2007
2010
mean
95% CI
quintile
Family
5 [highest]
1.28
1.22
1.38
1.36
1.35
[1.16 1.55]
land
4
1.63
1.62
1.85
1.58
1.68
[1.58 1.78]
(hectares)
3
2.06
2.13
1.69
1.59
1.89
[1.72 2.05]
2
2.90
3.23
2.88
2.47
2.99
[2.71 3.27]
1 [lowest]
3.80
4.46
4.28
3.79
4.25
[3.94 4.55]
Land
5 [highest]
0.28
0.29
0.35
0.30
0.31
[0.27 0.34]
0.34
0.37
0.47
0.36
0.38
[0.36 0.41]
holding per 4
3
0.45
0.50
0.49
0.45
0.47
[0.44 0.51]
adult
0.55
0.61
0.56
0.63
0.59
[0.54 0.65]
equivalent 2
1 [lowest]
0.83
0.96
0.93
0.95
0.92
[0.85 0.99]
(hectares)
Area
5 [highest]
0.99
0.94
0.85
0.79
0.89
[0.85 0.94]
4
1.18
1.26
1.22
1.03
1.17
[1.10 1.24]
cultivated
1.54
1.36
1.16
0.99
1.27
[1.20 1.35]
in the main 3
2
1.73
1.79
1.54
1.30
1.58
[1.48 1.67]
season
1.98
1.87
1.74
1.59
1.80
[1.68 1.91]
(hectares ) 1 [lowest]
Labor per
5 [highest]
6.04
7.15
5.94
6.43
6.39
[5.84 6.94]
4
4.54
4.12
4.19
4.71
4.39
[4.12 4.67]
hectare
3
5.14
5.18
4.81
4.67
4.96
[4.47 5.46]
cultivated
2
3.10
3.19
3.65
3.57
4.49
[2.33 6.65]
1 [lowest]
3.06
3.11
3.34
3.15
3.16
[2.94 3.39]
Cost of
5 [highest]
12.53 12.85
11.32
13.25
12.49
[11.70 13.27]
4
15.76 17.68
14.72
18.23
16.60
[15.41 17.80]
purchased
3
11.19 13.09
10.67
14.57
12.38
[10.88 13.88]
inputs/ha
5.25
10.26
10.52
12.05
9.52
[8.45 10.58]
(‘000 KSh) 2
1 [lowest]
7.45
7.25
7.47
9.01
7.80
[6.53 9.06]
-- real
Land
5 [highest]
703.02 703.02 [541.27 864.78]
4
633.03 633.03 [359.66 906.40]
values
3
723.67 723.67 [479.64 967.70]
/hectare
626.00 626.00 [276.30 975.70]
(‘000KSh) 2
1 [lowest]
271.82 271.82 [103.76 439.87]
Hired
5 [highest]
54.50 57.34
56.45
65.68
58.49
[57.70 59.27]
65.50 77.74
74.07
88.24
76.39
[75.20 77.59]
agricultural 4
62.25 63.63
64.88
75.41
66.54
[65.04 68.04]
wage labor 3
2
63.83 76.97
74.49
85.98
75.31
[74.24 76.37]
rate (KSh/
76.47 80.99
81.41
80.03
79.73
[78.46 80.99]
day)-- real 1 [lowest]
Source: Tegemeo Institute Rural Household Surveys. Note: Population density quintiles are
defined by ranking all households in the surveys by village-level population density and dividing
them into five equal groups.
160
 

Next, we examine how major agricultural inputs and inputs prices vary across population
density quintiles. Family labor, defined as the number of adult equivalents, adjusted to the
number of months spent in the household, per hectares of land cultivated, has generally increased
over the 13-year period, and is highest in the 20 percent densely populated villages (Table 3.1).
Similarly, the 2010 land values were more than twice as high in the three highest population
density quintiles than in the lowest density quintile. Information on land values was only
collected in the 2010 survey. Results also show that agricultural wage rates in the lowest densely
populated villages are 30 percent higher than in the highest densely populated areas (Table 3.1).
Capital expenditure, defined as the cost of purchased inputs (cost of fertilizer, seed and land
preparation) per hectare seems to be an increasing but non-linear function of population density.
It increases with population density from the first (lowest) quintile up to the fourth but declines
in the fifth (highest) quintile.
Perhaps the relationships between inputs use and population density are more revealing
when we look at the non-parametric regression results. Figures 3.1a and 3.1b show the nonparametric regressions of fertilizer use per hectare cultivated on population density. The fertilizer
th

use intensity increases with population density up to the 75 percentile and declines thereafter.
Similarly, fertilizer use is seemingly an increasing function of household labor (Figures 3.2a and
3.2b).

161
 

Figure 3.1a: Fertilizer expenditure per hectare-overall

4

--overall--

3
2.5
2

KSh('000)/ha

3.5

-

0

250

500

750

1000

persons/sq km
Figure 3.1b: Fertilizer expenditure per hectare-by survey year

--by survey year--

4
2
0

KSh('000)/ha

6

.

0

250

500

750

1000

persons/sq km
1997
2004
2010

162
 

2000
2007

1250

Figure 3.2a: Fertilizer expenditure per unit of labor-overall

--overall--

4
3.5
2.5

3

KSh('000)/ha

4.5

5

.

0

2

4

6

8

10

adult equivalents
Figure 3.2b: Fertilizer expenditure per unit of labor-by year

--by survey year--

6
4
2
0

KSh('000)/ha

8

.

0

2

4

6

8

adult equivalents
1997
2004
2010

163
 

2000
2007

10

Figure 3.3a: Capital-labor ratio-overall

--overall--

5
4
3
2

capital ('000K
Sh)/labor

6

.

0

250

500

750

1000

persons/sq km
Figure 3.3b: Capital-labor ratio-by survey year

2

4

6

8

.

0

capital ('000K
Sh)/labor

--by survey year--

0

250

500

750

1000

persons/sq km
1997
2004
2010

164
 

2000
2007

Figure 3.4a: Capital-land ratio-overall

--overall--

6
5
4

capital ('000KSh)/land (ha)

7

.

0

250

500

750

1000

persons/sq km

--by survey year--

4

6

8

10

.

2

capital ('000K
Sh)/land (ha)

Figure 3.4b: Capital-land ratio-by survey year

0

250

500

750

1000

persons/sq km
1997
2004
2010

165
 

2000
2007

Figure 3.5a: Owned land-labor ratio --overall

2.5

--overall--

.5

la -la r ra
nd bo tio
1
1.5
2

.

0

250

500

750

1000

persons/sq km

Figure 3.5b: Owned land-labor ratio--by survey year

2.5

--by survey year--

0

land-labor ratio
.5
1 1.5 2

.

0

250

500

750

1000

persons/sq km
1997
2004
2010

166
 

2000
2007

Figure 3.6a: Cultivated land-labor --overall

1.2

--overall--

.4

land-labor ratio
.6
.8
1

.

0

250

500

750

1000

persons/sq km

Figure 3.6b: Cultivated land-labor ratio--by survey year

land-labor ratio
.5
1
1.5

--by survey year--

0

.

0

250

500

750

1000

persons/sq km
1997
2004
2010

167
 

2000
2007

The results also show that households use increasing units of capital per unit of labor as
th

population density increases but up to a certain threshold (75 percentile); thereafter, the capitallabor ratio declines (Figures 3.3a and 3.3b). Figures 3.4a and 3.4b present the non-parametric
regression results of capital-land ratio on population density. The results are very similar to those
obtained in the capital-labor ratio regressions. Capital-land ratio is an increasing but non-linear
function of population density. Last, we present the non-parametric regressions of land-labor on
population density. Generally, the results show that the proportion of land, owned or cultivated,
to family labor declines with population density (Figures 3.5a to 3.6b).
Table 3.2 presents trends in farm production, income, and asset wealth over the panel
period by village population density quintiles. The value of net crop income (gross crop income
minus input costs) per hectare or per unit of labor, a measure of partial land productivity,
increases with population density up to the fourth density quintile and declines thereafter. This
finding confirms the results presented in Table 3.1 and from non-parametric regressions showing
that the intensity of labor and purchased inputs use is an increasing function of the population
2

density up to a certain threshold, 531 to 678 persons per km , and declines thereafter. Similarly,
the value of net farm income (from crops and animal products) per hectare is also a non-linear
increasing function of population density. Table 3.2 also shows that off-farm income per adult
equivalent is slightly higher for households in the low density areas.
Next, we discuss the value of asset wealth per adult equivalent by population density
quintiles (Table 3.2). The list of productive assets consistently collected and valued in each of
the four surveys includes but not limited to ploughs, tractors, carts, trucks, spray pumps, water
tanks, stores, wheelbarrows, combine harvesters, donkeys and livestock.
168
 

Table 3.2:
‘000 KSh
-- real
Net crop
income
per hectare

Household income and wealth trends, by population density quintile
Pop den
Survey year
Five survey panel
quintile
1997 2000 2004 2007 2010
average
95% CI
5 [highest] 27.17 51.03 40.47 44.71 44.28
41.53
[37.76 45.30]
4
27.30 45.13 41.74 45.28 82.71
48.43
[43.65 53.21]
3
22.35 31.01 31.50 35.65 51.64
34.43
[31.23 37.62]
2
16.41 18.83 25.59 33.18 36.80
26.16
[23.01 29.31]
1 [lowest]
17.40 20.00 16.68 14.83 9.48
15.68
[14.02 17.34]
Net crop
5 [highest]
9.39 22.82 18.72 17.76 16.63
17.06
[14.89 19.23]
4
11.80 25.46 21.80 22.59 41.31
24.59
[21.55 27.64]
income
11.56 20.33 17.14 21.17 29.51
19.94
[17.55 22.34]
per unit of 3
2
15.73 16.63 18.28 26.15 31.39
21.64
[16.63 26.64]
labor
1 [lowest]
14.28 18.97 22.41 16.39 8.60
16.13
[14.28 17.98]
Net farm
5 [highest] 46.75 74.20 69.57 46.15 47.75
56.88
[49.65 64.12]
income
4
44.55 69.20 69.83 46.07 83.81
62.69
[57.63 67.74]
per hectare 3
30.71 40.70 45.28 36.33 53.24
41.25
[37.52 44.99]
owned
2
30.51 29.02 38.28 35.29 40.03
34.62
[30.92 38.31]
1 [lowest]
25.13 29.26 29.61 16.41 10.21
22.12
[20.05 24.19]
Net farm
5 [highest] 14.81 31.01 29.08 18.03 17.00
21.99
[18.75 25.23]
income
4
18.23 36.31 32.80 22.92 41.72
30.40
[27.19 33.61]
per unit of 3
15.10 25.55 23.85 21.45 30.18
23.23
[20.69 25.77]
labor
2
25.54 24.02 27.73 28.75 34.06
28.02
[22.59 33.44]
1 [lowest]
19.57 30.08 37.56 19.94 29.70
27.37
[14.52 20.23]
Value of
5 [highest]
7.84
8.45 11.11 10.75 13.24
10.28
[9.08 11.47]
off-farm
4
8.75 10.91 16.56 18.54 28.38
16.63
[13.73 19.54]
income
3
6.68
8.59 11.85 13.20 15.74
11.21
[9.56 12.85]
per adult
2
8.84
9.82 12.66 12.87 19.14
12.67
[10.73 14.61]
equivalent 1 [lowest]
7.88 12.50 13.17 15.95 17.80
13.46
[11.58 15.34]
Value of
5 [highest]
8.37
7.91
8.49 10.58 8.49
8.77
[7.64 9.89]
assets/wea 4
11.14 11.06 12.93 21.01 20.47
15.32
[12.47 18.17]
lth per
3
9.06
8.41 12.69 14.37 16.82
12.27
[10.61 13.92]
adult
2
19.16 13.11 15.82 15.09 20.83
16.80
[14.51 19.09]
equivalent 1 [lowest]
22.20 24.20 36.55 38.26 39.10
32.06
[27.68 36.43]
Household 5 [highest] 16.10 26.95 25.20 25.74 29.36
24.67
[22.56 26.74]
aggregate 4
19.10 31.83 38.33 40.09 64.27
38.72
[34.61 42.77]
annual
3
15.50 24.29 26.78 29.39 37.78
26.75
[24.17 29.25]
income
2
22.40 22.45 28.27 30.86 42.98
29.39
[26.04 32.70]
1 [lowest]
19.00 29.35 35.01 38.15 31.55
30.61
[27.19 34.07]
Source: Tegemeo Institute Rural Household Surveys.
Note: Population density quintiles are defined by ranking all households in the surveys by
village-level population density and dividing them into five equal groups.

169
 

Recent studies in the poverty literature (e.g. Carter and Barrett, 2006; Krishna et al, 2004)
argue that the value of assets more accurately measures wealth than income or consumption, as it
is less susceptible to random shocks, and is likely to be a more stable indicator of household
welfare. This is especially true in regions where rain-fed agriculture is a major source of annual
income and where households rely greatly on their physical assets for their livelihoods. For these
reasons, we consider asset holdings to be an important measure of household livelihood,
productive potential, and safety net. The results show that the households’ asset wealth per adult
equivalent has been consistently higher (more than twice) in households located in the low
population density areas (Table 3.2). Family size in adults and adult equivalents is almost the
same across all five population density quintiles, meaning that asset wealth per household is also
substantially higher on average in the low density areas. Conversely, aggregate household
income rise with population density up to the fourth population density quintile, and thereafter
starts to decline (Table 3.2).
Table 3.3 presents information on smallholder commercialization indicators by
population density quintiles over the five survey years. The results clearly show the proportion of
land allocated to non-maize crops and the proportion of non-maize crop production are both
increasing but non-linear functions of population density. These variables generally increase with
the population density up to the fourth population density quintile, and fall thereafter. A quick
calculation using the information given in Table 3.3 show that a hectare of land allocated to nonmaize crop yields on average about three and four times in the low densely and high densely
populated areas, respectively, compared to a hectare allocated to maize crop.

170
 

Table 3.3: Smallholder commercialization indices by population density quintiles

Pop den
quintile

Survey year
Panel
average
1997
2000
2004
2007
2010
Proportion of land 5 [highest]
0.58
0.64
0.50
0.55
0.56
0.57
allocated to non- 4
0.58
0.72
0.72
0.72
0.65
0.68
maize crop
3
0.44
0.54
0.48
0.50
0.46
0.49
2
0.39
0.34
0.40
0.40
0.41
0.39
1 [lowest]
0.48
0.53
0.46
0.40
0.44
0.46
Proportion of non- 5 [highest]
0.70
0.78
0.69
0.69
0.74
0.72
maize crop
4
0.71
0.82
0.81
0.77
0.81
0.78
production
3
0.58
0.68
0.66
0.62
0.68
0.64
2
0.43
0.50
0.53
0.50
0.61
0.51
1 [lowest]
0.50
0.62
0.52
0.50
0.65
0.56
Maize production 5 [highest]
218.85
213.81
114.49
134.05
105.54
156.65
(kgs) per hectare 4
209.49
280.59
202.13
200.68
175.63
213.53
allocated to maize 3
213.74
135.06
107.87
137.23
102.21
139.69
2
321.25
123.87
149.72
156.38
88.07
170.92
1 [lowest]
335.51
122.69
214.08
165.69
59.99
179.16
Non-maize
5 [highest]
635.53
698.25
495.18
315.70
420.06
513.85
production (kgs) 4
456.42
366.32
302.54
254.04
354.78
346.15
per hectare
3
592.23
425.81
916.92
327.65
827.72
621.76
allocated to non- 2
943.80
492.56
411.88
832.81
538.80
639.14
maize crops
1 [lowest]
640.37
638.27
818.48
458.04
636.19
638.27
 Source: Tegemeo Institute Rural Household Surveys. Note: Population density quintiles are
defined by ranking all households in the surveys by village-level population density and dividing
them into five equal groups.

The non-parametric regression results of proportion of land allocated to non-maize crops
on population density variable are presented in Figures 3.7 show similar pattern. The area
2

allocated to non-maize crop increases with population density up to about 750persons/km and
after somehow plateaus. The same scenario is replicated when we look at the non-parametric
regression results of the proportion of non-maize crop production on population density variable;
although results are not presented here to save on space.

171
 

Figure 3.7: Proportion of area allocated to non-maize by population density

--overall--

.45

proportion
.5
.55

.6

.

0

250

500

750

1000

persons/sq km

Figure 3.8: Household commercialization index (HCCI) by population density

--overall--

.35

.4

HCCI
.45

.5

.

0

250

500

750

persons/sq km

172
 

1000

Figure 3.9: Household dairy commercialization index (HDCI) by population density

--overall--

.245

.25

HDCI
.255

.26

.265

.

0

250

500

750

1000

persons/sq km
Figure 3.10: Livestock products commercialization index (LPCI) by population density

--overall--

.26

.265

HLPCI
.27

.275

.

0

250

500

persons/sq km
173
 

750

1000

In Table 3.4 results on household aggregate value of crop and livestock products sales per
hectare are presented. Livestock products include milk, eggs, skin and hides, honey, wool and
manure among others. We also present dairy products separately since milk constitutes the
largest portion of the livestock products. Generally, the results show that crop sales, aggregate
livestock products, and milk sales increase with population density up to the fourth population
density quintile, and decline thereafter. Similar results emerge when we look at how various
indices of smallholder output commercialization relate to population density. These indices
include household crop commercialization index (HCCI) defined as the proportion of crop sales
value to total crop production value; household livestock products commercialization index
(HLPCI) defined as the proportion of the value of livestock products sold to total livestock
products value; and household dairy commercialization index (HDCI) defined as the proportion
of dairy products sold to total dairy products. The results show that all these indices are positive
but non-linear functions of population densities (Table 3.4). The non-parametric regression
results show clearly the turning points. The HCCI increases with population density up to about
2

600 persons per km and drops thereafter (Figure 3.8). Similarly, HLPCI and HDCI are positive
2

but non-linear functions population density with a turning point at about 750 persons per km

(Figures 3.9 and 3.10). Clearly, crop commercialization hits its turning point at comparatively
lower population density compared to livestock products commercialization.

174
 

Table 3.4: Smallholder commercialization by population density quintiles
‘000KSh nominal
Pop. den.
Survey year
Panel
quintile
average
1997
2000
2004
2007
2010
Value of crops
5 [highest] 18.93
30.77
21.98
28.77
35.07
27.03
sales per hectare
4
18.81
34.31
33.89
44.76
83.84
43.63
'000KSh
3
15.82
24.59
23.75
29.93
45.12
28.22
2
11.20
9.98
18.62
26.37
32.93
20.13
1 [lowest]
13.00
15.57
10.90
9.00
4.45
10.59
Value of livestock 5 [highest]
11.17
16.76
10.73
28.65
16.86
product sales per
4
13.70
18.24
23.22
34.37
22.44
hectare '000KSh
3
4.08
4.54
9.78
16.22
8.63
2
4.35
7.30
9.03
12.78
8.46
1 [lowest]
6.66
7.60
9.82
12.96
9.23
Value of milk sold 5 [highest]
10.89
16.41
9.94
25.02
15.60
per hectare
4
13.37
17.79
22.90
33.74
22.02
'000KSh
3
3.74
4.13
9.68
15.32
8.19
2
3.76
5.54
7.41
9.65
6.66
1 [lowest]
4.72
7.02
8.59
12.71
8.21
Household crop
5 [highest]
0.33
0.27
0.33
0.27
0.31
0.30
commercialization 4
0.44
0.51
0.51
0.58
0.59
0.52
index (HCCI)
3
0.41
0.50
0.43
0.45
0.45
0.45
2
0.40
0.39
0.43
0.44
0.40
0.42
1 [lowest]
0.43
0.41
0.44
0.39
0.23
0.38
Household animal 5 [highest]
0.28
0.22
0.25
0.32
0.27
commercialization 4
0.39
0.43
0.48
0.50
0.45
index (HACI)
3
0.22
0.21
0.29
0.39
0.28
2
0.30
0.24
0.31
0.32
0.29
1 [lowest]
0.33
0.33
0.42
0.43
0.38
Household dairy
5 [highest]
0.27
0.21
0.24
0.30
0.25
commercialization 4
0.38
0.43
0.47
0.49
0.44
index (HDCI)
3
0.22
0.19
0.27
0.36
0.26
2
0.26
0.23
0.30
0.30
0.27
1 [lowest]
0.31
0.32
0.40
0.42
0.36
Source: Tegemeo Institute Rural Household Surveys.
Note: Population density quintiles are defined by ranking all households in the surveys by
village-level population density and dividing them into five equal groups.
HICI= value of inputs acquired from market/ gross value of crop production
HCCI= gross value of crop sales /gross value of crop production
HACI=gross value of animal product sales / gross value of animal products production
HDCI=gross milk sales / gross milk production
 

175
 

The overall picture emerging from these results so far is that land is becoming an
increasingly constraining factor of production and that smallholder agriculture farming practices
in the areas of high population density are distinctly more land-intensive. Smallholder farmers
allocate more of their shrinking land to non-maize crop and sell a greater proportion of their
production. It seems farmers alter production patterns to make the best out of their shrinking land
resource by switching to high value enterprises such as production of fresh fruits and vegetables,
dairy and poultry products. From another perspective, high population density has been shown to
lead to improved economic infrastructures and market access owing to reduced average cost of
infrastructure and urbanization (Glover and Simon, 1975). Improved market access enables
farmers and increased demand motivates farmers to produce a surplus for the market and to
engage in high value and perishable enterprises that are not possible with poor market access.
However, these increasing trends in production are experienced up to a certain population
density threshold. These bivariate relationships, while providing a fairly consistent picture, do
not control for the effects of other variables affecting farm productivity, incomes and asset
wealth. However, these relationships do lead to an important hypothesis for more rigorous
analysis in the next section.

3.7.2

Econometric results

In this section, we report the econometric results from the input demand, output supply,
households’ incomes, and smallholder commercialization models estimation. But before that, we
present and discuss briefly the results from the first stage population density, second stage maize

176
 

price estimations, and the key insights from the third stage input prices and landholding
regressions. The results on determinants of population density are presented in Table 3.5.

Table 3.5: OLS estimation results for population density
2

Log of persons/km of total
land
[standard]
Coef. Robust
P>t
SE
0.004
0.000
0.00
0.001
0.000
0.03
0.002
0.000
0.00
0.050
0.004
0.00

2

Log of persons/km of
arable land [GRUMP]
Coef.

Robust
SE
0.000
0.000
0.000
0.003

0.004
Length of Growing Period (LGP)
0.002
Net Primary Productivity (NPP)
0.001
Elevation: meters above sea level
0.047
Slope: measure of steepness -degrees
0.026
0.000
0.00
0.015
0.000
Estimated population count for
1950
0.082
0.00
-1.919
0.064
Literacy (1=literate; 0=illiterate) * -3.351
*
-0.006
0.001
0.00
-0.001
0.001
Distance to water source -km)
7.486
0.192
0.00
5.565
0.153
Religion (1=Catholic; 0=non*
Catholic)
0.061
0.00
-1.049
0.049
Women-average age at first birth* -1.898
*
-0.261
0.120
0.03
-2.602
0.086
Contraceptive use (1=yes; 0=no)
*
-0.503
0.035
0.00
-0.654
0.026
Women: age at first marriage
*
-0.300
0.007
0.00
-0.142
0.005
Women: age at first intercourse
-5.542
0.101
0.00
-1.544
0.084
Women: number of children
*
dead
Tribe dummies
_ethnic_1
6.782
0.144
0.00
2.188
0.113
_ethnic_2
4.159
0.116
0.00
0.974
0.092
_ethnic_3
5.081
0.113
0.00
1.986
0.085
_ethnic_4
4.340
0.083
0.00
1.421
0.064
_ethnic_5
3.179
0.066
0.00
0.835
0.050
_ethnic_6
3.404
0.059
0.00
0.337
0.058
_ethnic_7
5.452
0.153
0.00
2.149
0.117
_cons
38.411 0.931
0.00
15.985
0.771
Number of obs
5845
5845
R squared
0.88
0.89
Note: * b/1989 DHS available at http://www.measuredhs.com/data/available-datasets.cfm
177
 

P>t
0.00
0.00
0.00
0.00
0.00
0.00
0.17
0.00
0.00
0.00
0.00
0.00
0.00

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

We run two separate models. In the first model (I) using the standard population density
defined as the number of persons per square kilometer of total land, and its square if necessary,
and controls for soil quality and climatic conditions. In the second model (II), we use
2

population density defined as the total number of persons per kilometer km of arable land
(hereafter referred to as GRUMP population density). As the results shows, all the coefficients
are statistically significant and most of them bear the expected signs.
Table 3.6 present the second stage regression of the determinants of the maize prices. The
results show that maize prices are a decreasing but non-linear function of population density in
the two models. Maize price decrease with population density reaching a minimum at 490 and
2

787 persons per km in the first and the second models, respectively, and falls thereafter. The
average partial effects (APES) indicate that an increase in population density by 100 persons per
2

km reduces maize prices by about one percent in model I and by eight percent in model II.
Generally, it seems like maize prices are highest in the maize-deficit low (lowlands) and high
densely (highlands) populated areas and are lowest in the medium population density areas,
basically high potential maize zones. Even though we had hypothesized that population density
may not have as great an impact on maize prices due to trade effects, it seems there are price
differentials across regions due to delays in price adjustments, market inefficiencies and
transport costs. The other important correlates of maize price include type of maize buyers and
the maize prices prevailing at the regional markets at the planting time and National and Cereal
Produce Board (NCPB) previous year’s maize buying prices in the region. It is important to
mention that the significant coefficient of the residuals from the first stage population density
equation implies that population density variable is endogenous in the maize price equation.
178
 

Table 3.6: CRE estimation results for producer farm gate maize prices
Model I
Model II
Dependent variable: log of maize
Coef. Bootstrap P>z
Coef. Bootstrap
price/kg (KSh)
SE
SE
Population density (100 pp/sq km)
-0.063
0.017
0.00
--[Standard]
Population density square
0.006
0.001
0.00
--Density (100 pp/sq km) [GRUMP]
----0.139
0.025
GRUMP density square
---0.009
0.001
Distance to motorable road (km)
-0.002
0.003
0.41
-0.002
0.003
Distance to tarmac road (km)
0.030
0.093
0.75
0.053
0.092
Own a truck (1=yes; 0=no)
0.035
0.026
0.18
0.038
0.026
Own a bicycle (1=yes; 0=no)
-0.005
0.009
0.56
-0.005
0.009
Own a radio (1=yes; 0=no)
0.010
0.011
0.33
0.009
0.011
Storage facility-- estimated value
0.002
0.002
0.34
0.002
0.002
(KSh)
Maize buyer type (base=private
buyer)
NCPB
0.092
0.011
0.00
0.087
0.011
processor
0.136
0.019
0.00
0.158
0.020
other
0.077
0.009
0.00
0.074
0.009
Regional maize price/kg - planting
0.008
0.002
0.00
0.008
0.002
time
NCPB previous years buying price/kg 0.011
0.005
0.02
0.013
0.005
Sex of household head (1=male;
-0.004
0.014
0.76
-0.003
0.014
0=female)
Age of the household head (years)
-0.036
0.056
0.51
-0.046
0.056
Education attainment (# of years)
0.001
0.002
0.53
0.001
0.002
Survey year dummies (base=1997 )
_Iyear_2000
0.103
0.050
0.04
0.095
0.050
_Iyear_2004
0.160
0.067
0.02
0.188
0.067
_Iyear_2007
0.125
0.068
0.07
0.180
0.069
_Iyear_2010
0.547
0.088
0.00
0.617
0.088
First stage population density
-0.016
0.040
0.69
-0.212
0.058
residuals
_cons
3.367
0.125
0.00
3.106
0.107
Population den. direct effect (APEs)
-0.006
-0.076
Turning point: density (pp/sq km)
490
787
Observations
5845
5845
Number of households
1169
1169
R squared
0.53
0.52
Note: Time averages of time varying variables and region dummies included

179
 

P>z
--0.00
0.00
0.53
0.56
0.15
0.59
0.41
0.37

0.00
0.00
0.00
0.00
0.01
0.81
0.41
0.46
0.06
0.01
0.01
0.00
0.00
0.00

As expected, agricultural wage rate is a decreasing but somehow nonlinear function of
population density (Table 3.7). Wage rate decreases with population density up to about at 1025
2

and 1280 persons per km in the first and the second models, respectively, and rises thereafter. It
is important to note that these two turning points are way outside the sample population density
th

variable distribution 99 percentile. The APEs show that an increase in population density by
2

100 persons per km reduces wage rates by about 11 and 14 percent in model I and II,
respectively. The significant coefficient of the residuals from the first stage population density
equation implies that population density variable is endogenous in the wage equation. Other
variables that influence positively agricultural wage rates are reduced distances to infrastructural
to roads and infrastructural facilities, and high expected maize prices.
Conversely, land rental rates increase with population pressure (Table 3.8). Land rental
rates increase with population density reaching a maximum at about 1040 and 1138 persons per
2

km in the first and the second model, respectively, and drops thereafter. Again these land rental
th

rates turning points are way beyond the 99 percentile of the sample population density variable
2

distribution. The APEs indicate that an increase in population density by 100 persons per km
increases land rental rates by about two and six percent in model I and II, respectively. The

significant coefficient of the residuals from the first stage population density equation implies
that population density variable is endogenous in the land rental equation. Other variables that
seem to significantly influence land rates include household landholding as proxied by the
district landholding median sizes and the expected maize price. While land rental rates
understandably increase with diminishing land sizes, they rise with high expected maize prices.
180
 

It appears the rental equilibrium rates are higher if leasers and leaseholders expect high crop
prices. Weather conditions and quality of soils in the areas where the households are located as
proxied by length of growing period (LGP), net primary productivity (NPP) and elevation also
influence land rental rates.
 

Table 3.7: CRE estimation results for agricultural wage
Model I
Model II
Coef. Bootstrap P>z
Coef. Bootstrap P>z
SE
SE
Population density (100 pp/sq km)
-0.123
0.018
0.00
---[Standard]
Standard population density square
0.006
0.001
0.01
---Population density (100 pp/sq km)
----0.197
0.021
0.00
[GRUMP]
GRUMP population density square
---0.008
0.001
0.00
Distance to motorable road (km)
-0.008
0.004
0.04
-0.009
0.004
0.01
Distance to tarmac road (km)
-0.004
0.001
0.00
-0.004
0.001
0.00
Distance to health center (km)
-0.002
0.001
0.00
-0.002
0.001
0.01
Distance to electricity supply (km)
-0.004
0.002
0.04
-0.004
0.002
0.05
Expected maize price/kg
0.052
0.006
0.00
0.052
0.006
0.00
Survey year dummies (base=1997 )
_Iyear_2000
0.186
0.020
0.00
0.204
0.020
0.00
_Iyear_2004
0.335
0.023
0.00
0.416
0.025
0.00
_Iyear_2007
0.455
0.019
0.00
0.578
0.026
0.00
_Iyear_2010
0.404
0.058
0.00
0.561
0.058
0.00
First stage population den. residuals 0.020
0.006
0.00
-0.055
0.013
0.00
_cons
3.777
0.080
0.00
3.546
0.083
0.00
Population den. direct effect (APEs) -0.108
-0.138
Turning point: pop. den. (pp/sq km)
1025
1280
Observations
5845
5845
Number of households
1169
1169
R squared
0.60
0.62
Note: Time averages of time varying variables and survey years and region dummies included

181
 

Table 3.8: CRE estimation results for land rental rates
Model I
Coef. Bootstrap
SE
Population density (100 pp/sq km)
0.156
0.009
[Standard]
Standard population density square
-0.008
0.001
Population density (100 pp/sq km)
--[GRUMP]
GRUMP population density square
--District median landholding (ha)
-0.019
0.007
Distance to motorable road (km)
-0.003
0.003
Distance to tarmac road (km)
-0.005
0.001
Distance to water source (km)
0.001
0.001
Distance to health center (km)
-0.005
0.001
Distance to electricity supply (km)
-0.030
0.010
Expected maize price/kg
0.059
0.005
0.014
0.002
Length of Growing Period (LGP)
0.001
0.000
Net Primary Productivity (NPP)
0.004
0.000
Elevation: meters above sea level
-0.071
0.003
Slope -- degrees
Survey year dummies (base=1997 )
_Iyear_2000
-0.036
0.009
_Iyear_2004
-0.057
0.020
_Iyear_2007
0.214
0.019
_Iyear_2010
-0.067
0.048
Region dummies (base: central
highlands)
_Izone_3
-0.558
0.059
_Izone_4
-0.968
0.029
_Izone_5
-0.715
0.030
_Izone_6
-0.190
0.033
_Izone_7
-0.612
0.032
_Izone_9
-0.549
0.054
First stage population den. residuals 0.023
0.008
_cons
6.184
0.124
Population den. direct effect (APEs) 0.045
Turning point: pop. den. (pp/sq km)
1040
Observations
5845
Number of households
1169
R squared
0.71
Note: Time averages of time varying variables included
182
 

P>z

Coef.

0.00

--

Model II
Bootstrap
SE
--

0.00
--

-0.091

-0.018

-0.00

-0.01
0.33
0.55
0.84
0.00
0.00
0.00
0.00
0.00
0.00
0.00

-0.004
-0.030
-0.001
0.003
-0.004
-0.003
-0.031
0.048
0.014
0.001
0.005
-0.089

0.001
0.013
0.003
0.012
0.006
0.001
0.012
0.005
0.003
0.001
0.001
0.004

0.00
0.02
0.80
0.82
0.51
0.84
0.01
0.00
0.00
0.07
0.00
0.00

0.00
0.01
0.00
0.16

-0.002
0.014
0.301
0.130

0.011
0.026
0.025
0.060

0.88
0.58
0.00
0.03

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

-0.398
-0.984
-0.780
-0.195
-0.566
-0.646
-0.059
6.243

0.069
0.043
0.030
0.032
0.026
0.072
0.013
0.165
0.021
1138
5845
1169
0.61

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

P>z
--

 

Household landholding sizes decreases with population pressure (Table 3.9). If
2

population density increases by 100 persons per square km , household landholding declines by
about nine and 15 percent in the first and the second model. Besides population density,
household demographic variables, intergenerational factors, and ethnicity as captured by the
household head’s tribe also influence household landholding sizes. Household landholding
increases with the household head’s age and education attainment. Similarly, the more the
amount of land the father to the household head or the spouse had, the more the land the current
household controls. The significant coefficient of the residuals from the first stage population
density equation implies that population density variable is endogenous in the landholding
equation.
In the remaining part of this section, we present the key econometric results from the
input factor demand, output supply, households’ incomes, and smallholder commercialization
models estimation. For each model, we run three separate regressions depending on the
population density variable used: model I – the standard population density; model II – GRUMP
population density; model III – without population density variable. It is important to mention at
the onset that the population density variables were found to be endogenous in all the input
demand, output supply, household income, and smallholder commercialization models.
(i) Effects of population density on factor input demand: Results presented in Table 3.10
shows that the intensity of purchased inputs use per hectare cultivated increase with population
2

density up to about 400 persons per km and declines after that point. Using the standard
2

definition of population density, an increase in population density by 100 persons per km
183
 

increases the mean purchased inputs use per hectare directly by about six and indirectly by about
one percent.

Table 3.9: CRE estimation results for household landholding
Model I
Model II
log land owned (ha)
Coef. Bootstrap P>z
Coef. Bootstrap
S.E
S.E
Population density (100 pp/sq km)
-0.087
0.029
0.00
--[Standard]
Population density (100 pp/sq km)
----0.149
0.040
[GRUMP]
Household size
0.026
0.004
0.00
0.026
0.004
Head’s gender (1=male; 0=female)
0.045
0.035
0.20
0.045
0.035
Age of the household head (years)
0.006
0.001
0.00
0.006
0.001
Education attainment (# of years)
0.008
0.004
0.06
0.008
0.004
Land rental rates ('000KSh)
-0.016
0.009
0.07
-0.017
0.009
Landholding --father of initial head
0.005
0.001
0.00
0.005
0.001
(ha)
Landholding --father of the spouse
0.002
0.001
0.08
0.002
0.001
(ha)
# of years in the current location
0.006
0.003
0.06
0.006
0.003
Ethnic group dummies
_Itribe_1
-0.021
0.180
0.91
0.021
0.180
_Itribe_2
-0.551
0.203
0.01
-0.503
0.204
_Itribe_3
-0.814
0.286
0.00
-0.563
0.303
_Itribe_4
-0.438
0.178
0.01
-0.344
0.180
_Itribe_5
-1.456
0.272
0.00
-1.191
0.288
_Itribe_6
-1.172
0.274
0.00
-0.742
0.303
_Itribe_7
-0.736
0.198
0.00
-0.407
0.222
_Itribe_8
-0.724
0.239
0.00
-0.436
0.258
_Itribe_9
-0.236
0.286
0.41
0.049
0.296
First stage population den. residuals 0.001
0.000
0.00
0.001
0.000
_cons
-0.822
0.220
0.00
-0.834
0.219
Observations
5845
5845
Number of households
1169
1169
R squared
0.68
0.68
Note: Household ethnicity, time averages of time varying variables, and survey year and
agroecological zone dummies included

184
 

P>z
-0.00
0.00
0.20
0.00
0.06
0.06
0.00
0.09
0.07
0.91
0.01
0.06
0.06
0.00
0.01
0.07
0.09
0.87
0.00
0.00

Table 3.10: CRE estimation results for intensity of cash input use per hectare cultivated
Model I
Model II
Model III
Dep. var: Log purchased inputs (KSh) Coef.
P>z
Coef.
P>z
Coef.
P>z
Population density (100 pp/sq km)
0.157
0.00
----[Standard]
Standard population density square
-0.016
0.00
----Population density (100 pp/sq km)
--0.110
0.00
--[GRUMP]
GRUMP population density square
---0.013
0.00
--Expected maize price (KSh/kg )
0.052
0.00
0.399
0.03
0.057
0.00
Land owned (ha)
-0.078
0.00
-0.080
0.00
-0.079
0.00
Wage rate ('00Ksh/day)
-0.027
0.12
-0.035
0.04
-0.037
0.03
Land rental rates ('000Ksh/ha )
-0.017
0.24
-0.017
0.24
-0.011
0.44
DAP price (KSh/50kg )
-0.014
0.00
-0.013
0.00
-0.013
0.00
Distance to motorable road (km)
-0.010
0.14
-0.010
0.15
-0.012
0.09
Distance to tarmac road (km)
-0.010
0.00
-0.010
0.00
-0.010
0.00
Gender of head (1=male; 0=female)
0.058
0.13
0.055
0.14
0.058
0.12
Household size
-0.005
0.34
-0.006
0.28
-0.006
0.27
Age of head (years)
0.003
0.12
0.003
0.11
0.003
0.10
Education attainment of head (# of
0.011
0.01
0.011
0.02
0.010
0.02
years)
First stage population density
0.001
0.00
0.001
0.01
--residuals
_cons
6.321
0.00
5.609
0.00
5.747
0.00
Direct partial effects
0.060
0.021
-Indirect partial effects
0.006
-0.018
0.006
Total partial effects
0.066
0.003
0.006
496
439
-Turning point: pop. density
2
(persons/km )
Observations
5845
5845
5845
Number of households
1169
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and
survey year and agroecological zone dummies included

When we use the GRUMP population density, a similar increase in population density
increases the mean intensity of purchased inputs use by about two percent. However, this direct
effect is cancelled out by the negative indirect effect of the increasing population density on
purchased inputs use through input factor prices and landholding. Besides population density and
185
 

input factor prices, the other variables associated with increased farm inputs use intensification
are reduced distances to roads and high education attainment by the household heads.

Table 3.11: CRE estimation results for intensity of fertilizer use (kgs/ha cultivated)
Model I
Model II
Model III
Dep. var: Log fertilizer intensity (
Coef.
P>z
Coef.
P>z
Coef.
P>z
kgs/ha)
Population density (100 pp/sq km)
0.077
0.01
----[Standard]
Standard population density square
-0.010
0.00
----Population density (100 pp/sq km)
--0.073
0.05
--[GRUMP]
GRUMP population density square
---0.009
0.02
--Expected maize price (KSh/kg )
0.065
0.00
0.971
0.00
0.063
0.00
Land owned (ha)
-0.056
0.00
-0.057
0.00
-0.057
0.00
Wage rate ('00Ksh/day)
0.031
0.17
0.023
0.31
0.025
0.27
Land rental rates ('000Ksh/ha )
0.003
0.90
0.010
0.62
0.006
0.78
DAP price (KSh/50kg )
-0.008
0.04
-0.008
0.04
-0.008
0.04
Distance to motorable road (km)
-0.005
0.10
-0.005
0.10
-0.005
0.11
Gender of head (1=male; 0=female) -0.003
0.97
-0.002
0.97
-0.003
0.96
Household size
0.004
0.56
0.003
0.61
0.003
0.62
Age of head (years)
0.003
0.21
0.003
0.19
0.003
0.20
Education attainment of head (# of
0.010
0.11
0.009
0.15
0.009
0.12
years)
First stage population density
0.009
0.00
0.012
0.00
--residuals
_cons
0.425
0.57
-1.341
0.14
0.461
0.52
Direct partial effects
0.008
0.011
Indirect partial effects
0.004
-0.065
0.004
Total partial effects
0.012
-0.054
397
417
Turning point: pop. density
2
(persons/km )
Observations
5845
5845
5845
Number of households
1169
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and
survey year and agroecological zone dummies included

Similar results obtain when we consider households’ intensity of fertilizer use (Table
3.11). The intensity of fertilizer use per hectare cultivated increase is also a nonlinear increasing
186
 

2

function of population density reaching a maximum at about 400 persons per km and declining
thereafter. The APEs however show that while increasing population density increases fertilizer
use when we use the standard population density variable, using the GRUMP definition
population density actually reduces fertilizer intensification.
2

An increase in the standard population density by 100 persons per km increases fertilizer
application intensity both directly and indirectly by about one percent. While increasing
population density increases fertilizer use intensities directly when we use the GRUMP
population density, the negative indirect effects of the increasing population density through
other avenues such as input and output prices crowds out the positive direct effects.
(ii) Effects of population density on factor output supply: Crop intensification defined as
crop production per hectares owned increases with the population density up to 710 and 1164
2

persons per km in the first and second model, respectively, and thereafter drops (Table 3.12). In
2

the first model, an increase in population density by 100 persons per km increase mean crop
production per hectare by about 11 percent directly and by two percent indirectly through the
effects of increasing population density on prices and landholding sizes. A similar increase in
population density increases smallholder crop intensification by 19 percent, directly, and four
percent, indirectly, in the second model. A similar picture emerges when we consider
intensification defined as crop production per hectare cultivated (Table 3.13). The only
differences are that the crop intensification turning points occur at slightly lower population
density levels and the APEs are slightly smaller too in this case. Besides population density, and
input and output prices, other factors that influence crop production intensification include the

187
 

amount of rainfall received, ownership of a radio that proxy access to information and to a
certain extent capture households asset wealth, and household head’s education attainment.

Table 3.12: CRE estimation results for net crop production per hectare owned
model I
model II
model III
Dep. var: log crop output kg/ha owned
Coef.
P>z
Coef.
P>z
Coef.
P>z
Population density (100 pp/sq km)
0.276
0.00
----[Standard]
Standard population density square
-0.019
0.00
----Population density (100 pp/sq km)
--0.269
0.00
--[GRUMP]
GRUMP population density square
---0.012
0.00
--Expected maize price (KSh/kg )
0.036
0.04
0.052
0.00
0.059
0.00
Land owned (ha)
-0.259
0.00
-0.261
0.00
-0.260
0.00
Wage rate ('00Ksh/day)
0.010
0.53
-0.008
0.62
-0.004
0.81
Land rental rates ('000Ksh/ha )
-0.052
0.04
-0.045
0.08
-0.044
0.08
DAP price (KSh/50kg )
0.000
0.96
0.002
0.72
0.001
0.80
Rainfall '00mm
0.030
0.09
0.032
0.07
0.036
0.04
Rainfall stress
-0.110
0.49
-0.075
0.63
-0.053
0.74
Own radio (1=yes; 0=no)
0.147
0.00
0.145
0.00
0.146
0.00
Distance to motorable road (km)
-0.002
0.65
-0.002
0.58
-0.003
0.54
Gender of head (1=male; 0=female)
0.070
0.28
0.069
0.29
0.072
0.27
Household size
-0.009
0.24
-0.010
0.19
-0.010
0.19
Age of head (years)
0.002
0.49
0.002
0.49
0.002
0.45
Education attainment of head (# of years) 0.014
0.04
0.013
0.07
0.013
0.06
First stage population density residuals
0.027
0.27
-0.077
0.04
--_cons
6.145
0.00
6.258
0.00
3.539
0.00
Direct partial effects
0.105
0.188
Indirect partial effects
0.022
0.035
0.022
Total partial effects
0.127
0.223
2
710
1164
Turning point: pop. density (persons/km )

Observations
5845
5845
5845
Number of households
1169
1169
1169
R squared
0.65
0.65
0.65
Note: Time averages of time varying variables, naïve expectation, and survey year and
agroecological zone dummies included

188
 

Table 3.13: CRE estimation results for net crop production per hectare cultivated
model I
model II
model III
Dep. var: log crop output kg/ha
Coef. P>z
Coef. P>z
Coef. P>z
Population density (100 pp/sq km)
0.249
0.00
----[Standard]
Standard population density square
-0.018
0.00
----Population density (100 pp/sq km)
--0.233
0.00
--[GRUMP]
GRUMP population density square
---0.012
0.00
--Expected maize price (KSh/kg )
0.031
0.05
0.932
0.00
0.059
0.00
Land owned (ha)
-0.053
0.01
-0.054
0.01
-0.054
0.01
Wage rate ('00Ksh/day)
-0.027
0.05
0.013
0.39
-0.014
0.35
Land rental rates ('000Ksh/ha )
-0.052
0.03
-0.045
0.05
-0.045
0.04
DAP price (KSh/50kg )
0.002
0.69
0.003
0.59
0.003
0.56
Rainfall '00mm
0.028
0.10
0.030
0.08
0.034
0.05
Rainfall stress
-0.133
0.37
-0.095
0.52
-0.074
0.62
Own radio (1=yes; 0=no)
0.104
0.01
0.098
0.02
0.103
0.01
Distance to motorable road (km)
-0.002
0.65
-0.003
0.53
-0.003
0.51
Gender of head (1=male; 0=female)
0.011
0.85
0.013
0.84
0.013
0.83
Household size
0.005
0.48
0.004
0.55
0.004
0.58
Age of head (years)
0.003
0.30
0.003
0.24
0.003
0.26
Education attainment of head (# of years) 0.013
0.04
0.011
0.07
0.012
0.06
First stage population density residuals
0.024
0.22
-0.042
0.16
--_cons
6.946
0.00
5.123
0.00
4.657
0.00
Direct partial effects
0.089
0.149
Indirect partial effects
0.004
-0.062
0.004
Total partial effects
0.094
0.087
2
686
982
Turning point: pop. density (persons/km )

Observations
5845
5845
5845
Number of households
1169
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and
survey year and agroecological zone dummies included
 
The smallholder farm output intensification as measured by gross farm production (crop
and livestock) per hectare controlled is also a nonlinear but increasing function of population
2

density (Table 3.14). It increases with population density up to 670 and 1144 persons per km in
the first and the second model, respectively, and drops thereafter. Increase in population density
189
 

2

100 persons per km directly increases mean gross farm intensification by eight and three
percent directly and indirectly, respectively, in the first model. In the second model, a similar
increase in population density triggers a 14 percent (directly) and a four percent (indirectly)
growth in smallholder farm intensification.

Table 3.14 CRE estimation results for gross farm production per hectare owned
model I
model II
model III
Dep. var: log farm output KSh/ha owned
Coef.
P>z
Coef.
P>z
Coef.
P>z
Population density (100 pp/sq km)
0.233
0.00
----[Standard]
Standard population density square
-0.017
0.00
----Population density (100 pp/sq km)
--0.196
0.00
--[GRUMP]
GRUMP population density square
---0.009
0.01
--Expected maize price (KSh/kg )
0.032
0.04
0.040
0.01
0.045
0.00
Land owned (ha)
-0.311
0.00
-0.313
0.00
-0.312
0.00
Wage rate ('00Ksh/day)
-0.057
0.00
-0.040
0.01
-0.044
0.01
Land rental rates ('000Ksh/ha )
-0.010
0.63
-0.003
0.88
-0.003
0.88
DAP price (KSh/50kg )
-0.003
0.55
-0.001
0.80
-0.002
0.72
Rainfall '00mm
-0.005
1.00
0.020
0.88
0.052
0.70
Rainfall stress
-0.013
0.92
0.013
0.92
0.033
0.78
Own radio (1=yes; 0=no)
0.169
0.00
0.167
0.00
0.168
0.00
Distance to motorable road (km)
-0.003
0.75
-0.003
0.70
-0.005
0.58
Gender of head (1=male; 0=female)
0.031
0.56
0.030
0.57
0.032
0.54
Household size
-0.017
0.01
-0.018
0.01
-0.018
0.01
Age of head (years)
0.003
0.21
0.003
0.22
0.003
0.19
Education attainment of head (# of years) 0.011
0.08
0.010
0.12
0.010
0.11
First stage population density residuals
0.036
0.10
-0.005
0.09
0.040
0.00
_cons
7.351
0.00
7.352
0.00
5.395
0.00
Direct partial effects
0.080
0.136
Indirect partial effects
0.027
0.044
0.027
Total partial effects
0.107
0.179
2
670
1144
Turning point: pop. density (persons/km )

Observations
5845
5845
5845
Number of households
1169
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and
survey year and agroecological zone dummies included.
190
 

Table 3.15: CRE estimation results for net farm production per hectare owned
model I
model II
model III
Dep. var: log farm output KSh/ha owned Coef. P>z
Coef. P>z
Coef. P>z
Population density (100 pp/sq km)
0.257
0.00
----[Standard]
Standard population density square
-0.020
0.00
----Population density (100 pp/sq km)
--0.183
0.00
--[GRUMP]
GRUMP population density square
---0.008
0.03
--Expected maize price (KSh/kg )
0.036
0.04
0.049
0.01
0.053
0.00
Land owned (ha)
-0.288
0.00
-0.291
0.00
-0.290
0.00
Wage rate ('00Ksh/day)
-0.050
0.00
-0.033
0.04
-0.036
0.03
Land rental rates ('000Ksh/ha )
-0.022
0.37
-0.014
0.55
-0.014
0.57
DAP price (KSh/50kg )
-0.013
0.80
0.034
0.95
-0.010
0.98
Rainfall '00mm
-0.003
0.86
0.004
0.98
0.003
0.84
Rainfall stress
-0.215
0.13
-0.175
0.21
-0.161
0.25
Own radio (1=yes; 0=no)
0.122
0.00
0.121
0.00
0.122
0.00
Distance to motorable road (km)
0.004
0.67
0.005
0.60
0.007
0.50
Distance to tarmac road (Km)
-0.001
0.83
-0.001
0.74
-0.002
0.71
Gender of head (1=male; 0=female)
0.067
0.30
0.066
0.30
0.068
0.29
Household size
-0.010
0.17
-0.011
0.13
-0.011
0.13
Age of head (years)
0.004
0.16
0.004
0.15
0.004
0.14
Education attainment of head (# of years) 0.014
0.04
0.013
0.04
0.014
0.05
First stage population density residuals
0.031
0.19
-0.048
0.17
_cons
7.449
0.00
7.166
0.00
5.357
0.00
Direct partial effects
0.084
0.126
Indirect partial effects
0.025
0.040
0.025
Total partial effects
0.109
0.165
2
654
1128
Turning point: pop. density (persons/km )

Observations
5845
5845
5845
Number of households
1169
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and
survey year and agroecological zone dummies included.

Similar results obtain when we consider farm intensification defined as net farm output
per hectare owned (Table 3.15) and per hectare cultivated (Table 3.16). Smallholder farm
intensification attain maximum at nearly the same population density thresholds and the APEs
are comparable in terms of sign and magnitude. Other important factors that are associated with
191
 

high farm intensification are household’s ownership of radio, smaller household’s sizes, and
households headed by highly educated individuals.

Table 3.16: CRE estimation results for net farm production per hectare cultivated
model I
model II
model III
Dep. var: log farm output KSh/ha cultivated Coef. P>z
Coef. P>z
Coef. P>z
Population density (100 pp/sq km)
0.231
0.00
----[Standard]
Standard population density square
-0.018 0.00
----Population density (100 pp/sq km)
--0.129
0.00
--[GRUMP]
GRUMP population density square
---0.006 0.05
--Expected maize price (KSh/kg )
0.029
0.05
0.045
0.00
0.051
0.00
Land owned (ha)
-0.056 0.00 -0.058 0.00 -0.057 0.00
Wage rate ('00Ksh/day)
-0.071 0.00 -0.056 0.00 -0.058 0.00
Land rental rates ('000Ksh/ha )
-0.024 0.28 -0.018 0.43 -0.017 0.45
DAP price (KSh/50kg )
0.001
0.79
0.003
0.57
0.002
0.62
Rainfall '00mm
0.039
0.79 -0.004 0.98
0.019
0.90
Rainfall stress
-0.199 0.14 -0.158 0.24 -0.143 0.29
Own radio (1=yes; 0=no)
0.085
0.04
0.083
0.04
0.084
0.04
Distance to motorable road (Km)
-0.001 0.95 -0.001 0.84 -0.001 0.80
Gender of head (1=male; 0=female)
0.010
0.86
0.010
0.87
0.011
0.85
Household size
0.005
0.43
0.004
0.52
0.004
0.52
Age of head (years)
0.005
0.04
0.005
0.04
0.005
0.03
Education attainment of head (# of years)
0.013
0.04
0.012
0.06
0.012
0.05
First stage population density residuals
0.035
0.07
0.027
0.02
--_cons
8.231
0.00
7.669
0.00
6.443
0.00
Direct partial effects
0.070
0.083
Indirect partial effects
0.005
0.005
0.004
Total partial effects
0.074
0.088
2
630
994
Turning point: pop. density (persons/km )

Observations
5845
5845
5845
Number of households
1169
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and
survey year and agroecological zone dummies included.

192
 

(iii) Effects of population density on household income: The aggregate household income
per adult equivalent is an increasing but a nonlinear function of the standard population density
(Table 3.17). It increases with population density reaching a maximum level at 450 persons per
2

km and falls thereafter. The household aggregate income reaches its maximum point at a
relatively lower population density threshold compared to that of crop and farm output
thresholds. Looking at the APEs, increasing population density generally reduces aggregate
2

household income. An increase in population density by 100 persons per km increases mean
household income directly by less than one percent.
However, this positive effect is wiped out by the negative two percent indirectly
population density effect transmitted through the influence of population density on prices and
landholding. Conversely, household income is a straightforward declining function of the
2

GRUMP population density. An increase in GRUMP population density by 100 persons per km
reduces household’s mean income directly by four percent and indirectly by three percent.
There are also other important factors that determine the level of aggregate household
incomes. These include access to information, distances to input and output markets, and

household demographic variables. While ownership of a radio increases household mean income
by about 22 percent, increased distances to motorable roads reduce income by about two percent.
A switch from male to female headship reduces household income by 21 percent while an
additional year in the education attainment of the household head increases household income by
about two percent. Relatively smaller households have higher incomes compared to larger ones.

193
 

Table 3.17: CRE estimation results for household aggregate income per adult equivalent
model I
model II
model III
Dep. var: log income (KSh)/adult equivalent Coef. P>z
Coef. P>z
Coef. P>z
Population density (100 pp/sq km)
0.143
0.00
----[Standard]
Standard population density square
-0.016 0.00
----Population density (100 pp/sq km)
---0.037 0.02
--[GRUMP]
Expected maize price (KSh/kg )
0.022
0.18
0.021
0.20
0.022
0.17
Land owned (ha)
0.190
0.00
0.188
0.00
0.189
0.00
Wage rate ('00Ksh/day)
-0.021 0.14 -0.032 0.02 -0.030 0.03
Land rental rates ('000Ksh/ha )
-0.048 0.01 -0.043 0.03 -0.043 0.02
DAP price (KSh/50kg )
-0.009 0.06 -0.008 0.10 -0.008 0.10
Rainfall '00mm
-0.015 0.27 -0.010 0.45 -0.010 0.46
Rainfall stress
-0.057 0.63 -0.024 0.84 -0.016 0.89
Own radio (1=yes; 0=no)
0.218
0.00 -0.218 0.00
0.219
0.00
Distance to motorable road (km)
-0.018 0.04
0.019
0.02 -0.019 0.02
Gender of head (1=male; 0=female)
0.206
0.00
0.206
0.00
0.207
0.00
Household size
-0.100 0.00 -0.101 0.00 -0.101 0.00
Age of head (years)
0.022
0.93
0.029
0.91
0.032
0.90
Education attainment of head (# of years)
0.023
0.00
0.023
0.00
0.023
0.00
First stage population density residuals
0.068
0.00
0.018
0.59
0.012
0.44
_cons
9.338
0.00
8.738
0.00
9.236
0.00
Direct partial effects
0.003
-0.037
Indirect partial effects
-0.017
-0.030
-0.017
Total partial effects
-0.014
-0.067
2
450
Turning point: pop. density (persons/km )

Observations
5845
5845
5845
Number of households
1169
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and
survey year and agroecological zone dummies included.

Next we examine the effect of population density on off-farm income (Table 3.18).
Surprisingly, the results show that off-farm income per adult equivalents is not a function of
population density. Increasing population density only influences off-farm income indirectly
through its effects on household landholding sizes.
194
 

Table 3.18: CRE estimation results for household off-farm income per adult equivalent
model I
model II
model III
Dep. var: log off-farm income (KSh)/ae
Coef.
P>z
Coef.
P>z
Coef.
P>z
Population density (100 pp/sq km)
0.012
0.61
----[Standard]
Population density (100 pp/sq km)
--0.014
0.73
--[GRUMP]
Expected maize price (KSh/kg )
-0.015
0.49
-0.013
0.55
-0.013
0.54
Land owned (ha)
0.112
0.00
0.112
0.00
0.112
0.00
Wage rate ('00Ksh/day)
-0.017
0.61
-0.017
0.61
-0.018
0.60
Land rental rates ('000Ksh/ha )
-0.006
0.84
-0.006
0.85
-0.006
0.85
DAP price (KSh/50kg )
0.003
0.68
0.003
0.68
0.003
0.68
Rainfall '00mm
0.025
0.22
0.025
0.22
0.025
0.22
Rainfall stress
0.221
0.20
0.224
0.20
0.221
0.20
Own radio (1=yes; 0=no)
0.122
0.02
0.122
0.02
0.122
0.02
Distance to motorable road (km)
-0.020
0.11
-0.021
0.10
-0.020
0.10
Distance to tarmac road (Km)
-0.001
0.80
-0.001
0.79
-0.001
0.80
Gender of head (1=male; 0=female)
0.418
0.00
0.418
0.00
0.418
0.00
Household size
-0.067
0.00
-0.067
0.00
-0.067
0.00
Age of head (years)
-0.009
0.01
-0.009
0.01
-0.009
0.01
Education attainment of head (# of years) 0.022
0.02
0.022
0.02
0.022
0.02
First stage population density residuals
-0.014
0.62
-0.009
0.84
-0.004
0.85
_cons
10.082 0.00
9.991
0.00
9.796
0.00
Indirect partial effects
-0.010
-0.016
-0.010
Observations
5845
5845
5845
Number of households
1169
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and
survey year and agroecological zone dummies included.
2

An increase population density by 100 persons per km reduces household’s mean offfarm income indirectly by one percent in the first model and by two percent in the second model.
Generally, it seems household off-farm income is explained more significantly by household’s
access to information and demographic variables. Ownership of a radio increases household nonfarm income by about 12 percent while a switch from male to female headship reduces off-farm
income by about 42 percent in the first and second models. Similarly, relatively smaller
195
 

households, those headed by younger persons and by persons with high education attainment
earn more off-farm income.
While it was hypothesized that a variety of off-farm income earning opportunities would
emerge as land becomes a binding constraint in the densely populated areas, this doesn’t seem to
be the case. To the contrary, off-farm income declines with shrinking household farm sizes.
While qualitative evidence gathered during the fieldwork showed some evidence of emergence
of some new off-farm activities, these activities are very low paying “poverty jobs”. The
emerging non-farm activities include sale of second hand clothes (mitumba) within the villages;
cellphone money transfer services (M-Pesa); cellphone repairing and battery charging; buying of
old household items -- scrap metal; bicycle passenger transport (boda-boda) services; and
repairing and hawking of household utensils (mali-mali).
(iv) Effects of population density on smallholder commercialization: We conclude this
section by examining the effects of increasing population density on smallholder diversification
and commercialization. The results from the fractional probit regression of the proportion of land
allocated to non-maize crop are presented in Table 3.19. The results show that the proportion of
land allocated to non-maize crop is an increasing but nonlinear function of population density
(first model). The proportion increases with the GRUMP population density up to about 890
2

persons per km and declines afterwards. This turning point occurs almost at the 90th percentile
2

of the sample distribution. An increase in population density by 100 persons per km increases
the mean proportion of area allocated to non-maize crops directly by four percent and by two
percent indirectly through the effects of the increasing population density on input factor and
output prices and household landholding. A reduction in the expected maize prices by one
196
 

shilling per kilogram is found to increase the proportion of land allocated to maize by about
0.013 while reduced maize price variability increases share of land allocated to maize.

Table 3.19: Fractional probit estimation results of land allocated to non-maize crop
Dep. var: Proportion of land allocated
Model I
Model II
to non-maize crop
Coef.
APEs
Coef.
APEs
***
Population density (100 pp/sq km)
0.196
[0.046]
--[GRUMP]
GRUMP population density square
-0.011***
[-0.003]
--[-0.013]
-0.081***
[-0.019]
Expected maize price (KSh/kg)
-0.054***
[-0.002]
-0.010***
[-0.002]
Maize price variability
-0.008**
**
[-0.004]
Land owned (ha)
-0.018
[-0.004]
-0.018
***
**
[-0.006]
-0.385
[-0.091]
Wage rate ('00Ksh/day)
-0.026
[-0.003]
-0.181**
[-0.043]
Land rental rates ('000Ksh/ha)
-0.013*
**
**
[-0.002]
-0.011
[-0.003]
DAP price (KSh/50kg)
-0.010
-0.003
-0.003
Rainfall '00mm
*
-0.230
[-0.054]
-0.251**
[-0.059]
Rainfall stress
0.061
0.055
Own radio (1=yes; 0=no)
0.001
0.001
Distance to nearest fertilizer seller (km)
-0.009
-0.006
Distance to motorable road (km)
*
-0.012
[-0.003]
-0.011**
[-0.003]
Household size
-0.001
-0.001
Age of the household head (years)
0.002
0.003
Head education attainment (years)
**
First stage population density residuals 0.002
0.003**
0.553**
_cons
6.104**
Direct partial effects
[0.043]
Indirect partial effects
[0.019]
[0.019]
Total partial effects
[0.062]
890
Turning point: pop. density
2
(persons/km )
Observations
5845
5845
Number of households
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and region
and survey year dummies included.

Other variables that influence the proportion of land allocated to non-maize crop include
the price of DAP, expected rainfall shock as measured by a six-year moving average of the
197
 

percentage of 20-day periods during the main growing season with less than 40 mm of rainfall,
and household size. While increasing price of fertilizer and rainfall shocks reduce the proportion
of land allocated to non-maize crop, relatively large family sizes reduces this proportion.

Table 3.20: Fractional probit estimation results for household crop commercialization
Model I
Model II
Dependent variable: HCCI
Coef.
APEs
Coef.
APEs
***
Population density (100 pp/sq km) [GRUMP] 0.198
[0.047]
--***
GRUMP population density square
-0.016
[-0.004]
--***
[-0.014]
Expected maize price (KSh/kg)
0.011
-0.059
***
***
[-0.002]
-0.010
[-0.002]
Maize price variability
-0.009
[-0.007]
-0.026** [-0.006]
Land owned (ha)
-0.029**
[-0.007]
-0.156
Wage rate ('00Ksh/day)
-0.030***
**
[0.003]
-0.216** [-0.051]
Land rental rates ('000Ksh/ha)
-0.014
DAP price (KSh/50kg)
0.004
0.003
0.004
0.007
Rainfall '00mm
-0.039
0.032
Rainfall stress
0.040
0.038
Own radio (1=yes; 0=no)
***
-0.006
[-0.002]
-0.007*** [-0.002]
Distance to nearest fertilizer seller (km)
-0.005**
[-0.001]
-0.013** [-0.003]
Distance to motorable road (km)
-0.011
0.001
Household size
***
0.004
[0.001]
0.045***
[0.011]
Age of the household head (years)
**
0.006
[0.002]
0.007
Head education attainment (years)
**
First stage population density residuals
0.059
0.029***
-3.242***
_cons
-2.118**
Direct partial effects
[0.043]
Indirect partial effects
[0.004]
[0.004]
Total partial effects
[0.047]
2
619
Turning point: pop. density (persons/km )

Observations
5845
5845
Number of households
1169
1169
Note: Time averages of time varying variables, naïve expectation of other crop prices, and region
and survey year dummies included.

From the farm output side, the proportion of marketed crop output as measured by the
household crop commercialization index (HCCI) is also found to be a positive but nonlinear
198
 

function of population density (Table 3.20). The HCCI increases with population density
2

reaching a maximum at about 620 persons per km in the first model, and declines thereafter. An
2

increase in population density by 100 persons per km increases the mean HCCI directly by
about four percent and indirectly by less than one percent. The results also show that distances to
input and output markets, and infrastructural facilities, and the age and the level of education of
the household head influence smallholder crop commercialization. In the first model, crop
commercialization appears to be relatively higher among the households headed by older
persons, persons with high education attainment, and among households located at close
proximity to markets.
What these results imply is that when households expect low and stable maize prices,
they are able to put a greater proportion of their shrinking land to relatively high value. But this
can only occur up to a certain population density threshold beyond which increasing population
density has negative effects to smallholder commercialization. It should also be quickly pointed
out that maize is considered a cash crop especially in the high potential maize region of the
country.

3.8

Conclusions and policy implications

The overarching objective of this study was to examine how rising population pressure
affects smallholders’ production, commercialization and household incomes. Using data from
five panel surveys on 1,169 small-scale farms over the 1997–2010 period, we use panel
econometric techniques to determine how increasing rural population density is affecting farm
household behavior and livelihoods. The estimation strategy deals with the potential endogeneity
199
 

of population density in input demand and output supply equations using a two-stage control
function approach. The study is motivated by the need to understand the nature and magnitude of
emerging land constraints in African agriculture using Kenya as a case study.
The overall picture emerging from these results so far is that land is becoming an
increasingly constraining factor of production and that smallholder agriculture farming practices
in the areas of high population density are distinctly more land-intensive. Consequently, there is
a rising strain on rural livelihoods in the densely populated rural areas due to land pressures and
declining farm sizes. Inputs and output agricultural intensification, household incomes and
smallholder commercialization rise with population density up to a certain threshold; beyond this
point, rising population density is associated with sharp declines in agricultural input and output
intensification.
The empirical results also show that smallholder farmers allocate more of their shrinking
land to non-maize crop and sell a greater proportion of their production as population density
increases. However, this can only feasible up to a certain population density threshold, beyond
which increasing population density reduces smallholder commercialization. It seems farmers
alter production patterns to make the best out of their shrinking land resource by switching to
high value enterprises such as production of fresh fruits and vegetables, dairy and poultry
products.
These results indicate that smallholder landholding sizes are gradually declining in Kenya
as in much of sub-Saharan Africa. Currently about 14 percent of Kenya’s rural population
2

resides in areas exceeding the 600 persons per km population density threshold. Another 20
percent of the rural population is fast approaching this threshold.  The results also show that

200
 

increased access to input markets, passable roads and other physical infrastructural facilities
considerably influence the degree of smallholder production and commercialization.
Generally, these findings suggest that increased access to land will ensure that
smallholder facing land size constraints are able to generate agricultural surpluses and
consequently participate in agricultural output markets. Strategies to improve rural households’
access to land will need to be not only on the country’s land agenda, but also its food security
and poverty reduction agendas. There is also some scope for promoting equitable access to land
through land redistribution reforms to reduce landholding inequalities. A coordinated strategy of
public goods and services investments in road infrastructure, schools, health care facilities,
electrification and water supply would also be helpful in raising the economic value of arable
land in the country that is relatively remote and still unutilized. The study also calls for
redoubled public investment in the national agricultural research systems to focus on new farm
land-saving technologies and practices appropriate for one-hectare farms or smaller.  

201
 

 
 
 
 
 
 
 
 
 
 
 

APPENDIX

202
 

Table A3.1: Responses to population pressure and the conditioning variables
Response
Conditioning variables
Country
examples
Demographic responses
Availability of inexpensive modern
Puerto Rico;
(e.g. postponed marriage, contraceptives and sterilization; increased women China; Taiwan
increase celibacy,
education; women employment opportunities;
contraception use,
cultural factors; age at marriage.
sterilization, abortions)
Extensification (more
Availability of untapped cultivable land; land
Chile; Taiwan
land brought into
tenure systems; efficiency of land markets;
cultivation), land reforms existing infrastructure (transport and
communication); unequal land distribution
implying greater potential for land reforms;
cultural factors.
Intensification (hybrid
Existence of labor-intensive, land-saving
Nigeria; Chile;
seeds, fertilizer and
technologies; institutional structure (access to
China; Taiwan;
17
credit for purchasing seeds, fertilizer); supply of
Indonesia;
‘night soils’ use);
surface or underground water.
Kenya
irrigation
Switch from traditional
Soil and climatic conditions; availability of water; Senegal;
crops; commercialization proximity to markets; state of physical
Nigeria;
of agriculture
infrastructure.
Indonesia
Off-farm diversification
Availability of off-farm rural employment
Uganda;
and out-migration [rural- opportunities; favorable macroeconomic policies Nigeria; Chile;
rural; rural-urban; out of
for urban jobs creation; wide urban-rural wage
Puerto Rico
the country]
differential (pull factor); immigration policies;
information; education; cultural factors.
Source: Summarized by the author from Bilsborrow (1987).
 

                                                            
17

Night soil refers to the use of human excrement collected at night as a fertilizer in China.
203

 

Figure A3.1: Photos contrasting Kalama Hills, Machakos District, Kenya
[a] 1936
[b] 1989

Source: Norton-Griffiths (2008)

Source: Norton-Griffiths (2008)

Table A3.2: Indices of agricultural commercialization
Index
Commercialization = Value of agricultural sales in markets/
agricultural product value
of agriculture
(output side)
Commercialization = Value of inputs acquired from market/
agricultural product value
of agriculture (input
side)
Commercialization = Value of goods and services acquired through
of rural economy
market transactions/ Total income
= Value of goods and services acquired by cash
Degree of
transactions/ Total income
integration into the
cash economy
= Gross value of crop sales/ Gross value of all
Household crop
commercialization
crop production

Net-market position
(sales)
Net-market position
(purchases)
Specialization

= sales/(volume of commodities stored at the
beginning+ volume of commodities produced
during season)
= purchases/(volume of commodities stored at
the beginning+ volume of commodities
produced during season)
= Value of purchased agricultural products not
produced by household/ Gross value of
agricultural production

 

204
 

Source
von Braun et al. 1994;
Gabre-Madhin et al.
(2007)
von Braun et al. 1994
von Braun et al. 1994
von Braun et al. 1994
Govereh et al. (1999);
Strasberg et al. (1999);
Gabre-Madhin et al.
(2007)
Gabre-Madhin et al.
(2007)
Gabre-Madhin et al.
(2007)
Gabre-Madhin et al.
(2007)

Figure A3.2: Determinants and consequences of smallholder commercialization

Source: von Braun et al. (1994) with some modification by the author

205
 

Table A3.3: Processes in agricultural commercialization
Level of
Farmers
Source of inputs
Product mix
market
objective
orientation
Subsistence
Food selfHousehold
Wide range
system
sufficiency
generated (nontraded)
SemiSurplus
Mix of traded and Moderately
commercial
generation
non- traded inputs specialized
system
Commercial
Profit
Predominantly
Highly
system
maximization
traded inputs
specialized
Source: Pingali (2001).

206
 

Household
income source

Human nutrition

Soil fertility

Predominantly
agriculture

Predominantly
home produced

Farm yard manure

Agricultural and
non-agricultural

Home produced
and purchased

Pre-dominantly
non-agricultural

Predominantly
purchased

Farm yard manure
and chemical
fertilizers
Chemical
fertilizers

Figure A3.3: Effects of population density on smallholder production and income

Source: Chamberlin (2013)

207
 

Figure A3.4: Egerton University/Tegemeo Institute panel data households’ locations

Source: Suri, 2005 

208
 

18

Figure A3.5: Modified Fisher-Ideal Quantity Indices
To aggregate crop production across multiple commodities, we use a modification of the
Fisher-Ideal index by Mason (2011), which uses information on the individual household
production (kg) and national-level prices of each crop in the crop group. The Fisher-Ideal (FI)
index is a combination of two indices, the Modified Laspeyres Quantity Index (ML) and the
Modified Paasche Quantity Index (MP) (Diewert 1992; Diewert 1993).

For each crop j  1 to J , we use the national median production quantity as the base
quantity in the denominator of both the ML and MP indices. We use the median national-level
price in the first year of the Tegemeo panel household dataset (1997) as the base year price,
p j,base. Thus, changes in the ML index are driven by changes in quantities of each commodity
produced over time, as prices do not vary from the base year, nor across households. For p j in
the MP index, we use the national median price for each year. Thus, the MP index allows price
variation by year but not across households.
J

Modified Laspeyres Quantity Index
(LQI*)

LQI i*,t 

q
j 1
J

q
j 1

Modified Paasche Quantity Index

p j ,base

*
j ,base

p j ,base

J

(PQI*)

PQI i*,t 

q
j 1
J

q
j 1

Fisher-Ideal Quantity Index (FIQI*)

i , j ,t

p j ,t

*
j ,base

p j ,t

FIQI i*,t  ( LQI i*,t  PQI i*,t

 

                                                            
18

Adapted from Mather and Jayne (2011).
209

 

i , j ,t

 
 
 

REFERENCES

210
 

REFERENCES

Ahmed, I. 1981. Wage determination in Bangladesh agriculture. Oxford Economic Papers,
33(2): 298–322.
Allen, B. 2001. Boserup and Brookfield and the association between population density and
agricultural intensity in Papua New Guinea. Asia Pacific Viewpoint, 42(2/3):237-254.
Atwood, D. A. 1990. Land registration in Africa: The impact on agricultural production. World
Development 18 (5): 659–671.
Bardhan, P. K. 1984. Land, labor and rural poverty: Essays in development economics. New
York: Columbia University Press, Delhi: Oxford University Press.
Benin, S. 2006. Policies and Programs Affecting Land Management Practices, Input Use, and
Productivity in the Highlands of Amhara Region, Ethiopia. In Pender, John, Frank Place
and Simeon Ehui (ed), Strategies for Sustainable Land Management in the East African
Highlands, International Food Policy Research Institute, Washington DC, 217- 256.
Berry, Sara. 1993. No condition is permanent: the social dynamics of agrarian change in subSaharan Africa. Madison: University of Wisconsin Press.
Berry, Sara. 1989. Social institutions and access to resources in African agriculture. Africa vol.
59(1): 41-55.
Besley, T. 1995. Property rights and investment incentives: Theory and evidence from Ghana.
Journal of Political Economy, 103 (5): 903–937.
Bezemer, Dirk and Derek Headey. 2008. Agriculture, Development, and Urban Bias. World
Development 36.8: 1342–64.
Bilsborrow, R.E. 2001. Migration, population change, and the rural environment. Population,
environmental change, and security paper series, Issue 8, University of Michigan, Ann
Arbor.
Bilsborrow, R. E. 1987. Population pressures and agricultural development in developing
countries: A conceptual framework and recent evidence. World Development, 15(2): 183203.
Bilsborrow, R.E. and C.R. Winegarden. 1985. Landholding, Rural Fertility and Internal
Migration in Developing Countries: Econometric Evidence from Cross-National Data.
The Pakistan Development Review, XXIV(2).
211
 

Binswanger-Mkhize, H.P. 2012. India 1960-2010: Structural Change, the Rural Nonfarm Sector
and Prospects for Agriculture. Paper presented at Stanford Food Security and the
Environment Symposium, Palo Alto, May 10, 2012. Available at http://iisdb.stanford.edu/evnts/7078/Binswanger_5_10_12.pdf [accessed April 29, 2013].
Binswanger, H. P. and J. McIntire. 1987. Behavioral and material determinants of production
relations in land abundant tropical agriculture. Economic Development and Culture
Change 35 (1): 73–99.
Binswanger, Hans P. and Vernon W. Ruttan. 1978. Induced Innovation: Technology, Institutions
and Development. Baltimore: Johns Hopkins University Press.
Binkley, J.K. and L. McKinzie. 1984. Acreage Response Analysis with Multiple Equations.
North Central Journal of Agricultural Economics 6(2): 91-98.
Blaikie, P., and H. Brookfield. 1987. Land degradation and society. London, New York:
Methuen.
Brookfield, H.C. 1972. Intensification and disintensification in Pacific agriculture: A theoretical
approach, Pacific Viewpoint 13(1): 30–48.
Boserup, Ester. 1965. The Conditions of Agricultural Growth: The Economics of Agrarian
Change under Population Pressure. London: Allen & Unwin.
Bouis, H.E. and L.J. Haddad. 1990. Effects of agricultural commercialization on land tenure,
household resource allocation, and nutrition in the Philippines. Research Report 79.
IFPRI (International Food Policy Research Institute), Washington, DC, USA.
Boyce, J.K. and M. Ravallion. 1991. A dynamic econometric model of agricultural wage
determination in Bangladesh. Oxford Bulletin of Economics and Statistics, 53(4): 361376.
von Braun J. 1995. Agricultural commercialization: Impacts on income and nutrition and
implications for policy. Food Policy 20(3): 87–202.
von Braun, J., H. Bouis and E. Kennedy. 1994. Conceptual framework. In: von Braun J and
Kennedy E (eds), Agricultural commercialization, economic development, and nutrition.
Johns Hopkins University Press, Baltimore, Maryland, USA.
von Braun, J., H. de Haen and J. Blanken. 1991. Commercialization of agriculture under
population pressure: Effects on production, consumption and nutrition in Rwanda.
Research Report 85. FPRI (International Food Policy Research Institute), Washington,
DC, USA.

212
 

Bruce, J. W. and S. E. Migot-Adholla. 1994. Searching for land tenure security in Africa.
Dubuque, Iowa, U.S.A.: Kendall/Hunt.
Bryceson, Deborah Fahy and Vali Jamal editors. 1997. Farewell to Farms: De-Agrarianisation
and Employment in Africa. Aldershot, Hampshire: Ashgate Publishing.
Byerlee, Derek and Alain de Janvry. 2009. The Politics of Hunger. Letter to the Editor. Foreign
Affairs (March/April).
Carter, M. and C. Barrett. 2006. The economics of poverty traps and persistent poverty: An
asset-based approach. The Journal of Development Studies, 42(2): 178-199.
Chamberlain, G. 1984. Panel data. In Grilliches, Z., Intriligator, M.D., eds. Handbook of
Econometrics, Vol. 2 North Holland, Amsterdam, 1247-1318.
Chamberlin, J. 2013. Market Access and Smallholder Development in Kenya and Zambia.
Unpublished Ph.D. dissertation, Agricultural, Food and Resource Economics
Department, Michigan State University, East Lansing, Michigan.
Chavas, J. P. and M. T. Holt. 1990. Acreage decisions under risk: The case of corn and soybeans.
American Journal of Agricultural Economics 72: 529–38.
Christiaensen, L., L. Demery and J. Kühl. 2011. The (Evolving) Role of Agriculture in Poverty
Reduction: An Empirical Perspective, Journal of Development Economics, 96(2): 239254.
Chu, C. Y. Cyrus. 1997. Population Density and Infrastructure Development. Review of
Development Economics 1(3): 294–304.
Darity, W.A. Jr., 1980, The Boserup theory of agricultural growth, Journal of Development
Economics 7, 137-157.
De, Utpal Kumar and Manabendu Chattopadhyay. 2010. Crop diversification by poor peasants
and role of infrastructure: Evidence from West Bengal. Journal of Development and
Agricultural Economics 2(10): 340-350.
Dercon, S. and A. Zeitlin. 2009. Rethinking Agriculture and Growth in Ethiopia: A Conceptual
Discussion. Paper prepared as part of a study on Agriculture and Growth in Ethiopia.
Available at
http://www.economics.ox.ac.uk/members/Stefan.Dercon/ethiopia%20paper%202_v2.pdf
[accessed March 6, 2012].
Dercon, Stefan. 2002. Income risk, coping strategies and safety nets. The World Bank Research
Observer, 17(2): 141-166.
213
 

Dorsey, B. 1999. Agricultural intensification, diversification and commercial production among
smallholder coffee growers in Central Kenya. Economic Geography 75(2):178–195.
Drechsel, P., L. Gyiele, D. Kunze, and O. Cofie. 2001. Population density, soil nutrient
depletion, and economic growth in sub-Saharan Africa. Ecological Economics, 38: 251–
258.
Du, X., D.A.Hennessy and W.M. Edwards. 2007. Determinants of Iowa Cropland Cash Rental
Rates:Testing Ricardian Rent Theory Center for Agricultural and Rural Development
Working Paper 07-WP 454, Iowa State University, Ames, Iowa, USA.
Eastwood, R., M. Lipton, and A. Newell. 2010. ‘Farm size’, in P.L.Pingali and R.E. Evenson
(eds) Handbook of Agricultural Economics, Elsevier: Amsterdam.
Ellis, F. 2005. Small Farms, Livelihood Diversification, and Rural–Urban Transitions: Strategic
Issues in Sub-Saharan Africa. Paper presented at The Future of Small Farms, 26-29 June.
Wye College, London.
Falk, B. 1991. Formally Testing the Present Value Model of Farmland Prices. American Journal
of Agricultural Economics, 73:1–10.
Feder, G., R.E. Just and D. Zilberman. 1985. Adoption of agricultural innovations in developing
countries: A survey. Economic Development and Cultural Change, 33(1): 255-298.
Fischer, G., Harrij van Velthuizen and F. O. Nachtergaele. 2000. Global Agro-Ecological Zones
Assessment: Methodology and Results. Interim Report IR-00-064. International Institute
for Applied Systems Analysis and The Food and Agricultural Organization of the United
Nations. The data are available from: http://geodata.grid.unep.ch/.
Fischer, G. and G.K. Heilig. 1997. Population momentum and the demand on land and water
resources. Philosophical Transactions of the Royal Society of London, B 352: 869–889.
Frederiksen, Peter C. 1981. Further Evidence on the Relationship between Population Density
and Infrastructure: The Philippines and Electrification. Economic Development and
Cultural Change 29(4): 749-758.
Garg, A. and J. Morduch. 1998. Sibling Rivalrsy and the Gender Gap: Evidence from Child
Health Outcomes in Ghana. Journal of Population Economics, 11(4), pp. 471-93.
Ghebru, Hosaena and Stean T. Holden. 2009. Factor market imperfections and rural land rental
markets if Northern Ethiopia Highlands. In The Emergence of Land Markets in Africa,
eds. Stein T. Holden, Keijiro Otsuka, and Frank Place. Washington DC: Resources for
the Future.

214
 

Gilles, J.L. and M. Dalecki. 1988. Rural well-being and agricultural change in two farming
regions. Rural Sociology, 53(1):40-55.
Gladwin, C. 1990. African Women Farmers and Structural Adjustment. Gainesville, Fla.:
University of Florida Press.
Glover, D. R. and J.L. Simon. 1975. The Effect of Population Density on Infrastructure: The
Case of Road Building. Economic Development and Cultural Change, 23(3): 453-468.
Gobster, P.H. and M.G. Rickenbach. 2003. Private forestland parcelization and development in
Wisconsin’s Northwoods: perceptions of resource-oriented stakeholders. Landsc Urban
Plan 69(2/3):165–182.
Govereh J., T.S. Jayne and J. Nyoro. 1999. Smallholder commercialization, interlinked markets
and food crop productivity: Cross-country evidence in eastern and southern Africa.
Gray, L. and M. Kevane. 1999. Diminished access, diverted exclusion: women and land tenure in
sub-Saharan. African Studies Review 42( 2): 15-39.
Haggblade, Steven. 2009. Forgotten Farmers. Harvard International Review, Winter: 7.
Hani, A.B. 1996. Determinants of Agricultural Labor Wage in Jordan: A Comparative Study of
Expatriate versus Local Labor, Economic Development and Cultural Change, 44(2): 405411.
Haugerud, A. 1989. Land tenure and agrarian change in Kenya. Africa, 59 (1): 61–90.
Hayami, Yujiro and Vernon W. Ruttan. 1985. Agricultural Development. 2nd ed. Baltimore:
Johns Hopkins University Press.
Hazell, P., C. Poulton, S. Wiggins, and A. Dorward. 2010. The Future of Small Farms:
Trajectories and Policy Priorities. World Development, 38(10): 1349-1361.
Hazell, P., A. Dorward, C. Poulton, and S. Wiggins. 2007. The Future of Small Farms for
Poverty Reduction and Growth. International Food Policy Research Institute 2020
Discussion Paper No. 42. Washington, D.C.: IFPRI.
Hertel. T. 2011. The Global Supply and Demand for Agricultural Land in 2050: A Perfect Storm
in the Making? American Journal of Agricultural Economics 93(2):259-275.
Headey, D., D. Bezemer, and P. Hazell. 2010. Agricultural Employment Trends in Asia and
Africa: Too Fast or Too Slow? World Bank Research Observer 25.1: 57-89.
Hicks, John. 1932. The Theory of Wages. London: Macmillan.
215
 

Holden, Stein T., Keijiro Otsuka, and Frank Place (eds). 2009. The Emergence of Land Markets
in Africa. Washington DC: Resources for the Future.
Hopkins, A. G. 1973. An economic history of West Africa. Columbia University Press, New
York.
Hossain, M. 2004. Rural non-farm economy: Evidence from household surveys. Economic and
Political Weekly, 39(36): 4053–4058.
de Janvry, A., M. Fafchamps and E. Sadoulet. 1991. Peasant Household Behaviour with Missing
Markets: Some Paradoxes Explained. The Economic Journal, 101(409): 1400-1417.
Jayne, T.S., Zulu, B., Kajoba, G., and Weber, M. 2008. Access to Land and Poverty Reduction in
Rural Zambia: Connecting the Policy Issues. Working Paper No. 34, Food Security
Research Project, Lusaka, Zambia.
Jayne, T.S. T. Yamano, M. Weber, D. Tschirley, R. Benfica, A. Chapoto, and B. Zulu. 2003.
Smallholder Income and Land Distribution in Africa: Implications for Poverty Reduction
Strategies. Food Policy, 28(3): 253-275.
Jin, S. and T.S. Jayne 2013. Land Rental Markets in Kenya: Implications for Efficiency, Equity,
Household Income, and Poverty. Land Economics, 89 (2): 246-271.
Johnston, B. F. and P. Kilby. 1975. Agriculture and Structural Transformation: Economic
Strategies in Late Developing Countries. New York: Oxford University Press.
Johnson-Hanks, J. 2008. Demographic Transitions and Modernity. Annual Review of
Anthropology, Vol. 37: pp. 301-315.
Kelly, Valerie A. 2006. Factors Affecting Demand for Fertilizer in Sub-Saharan Africa.
Agriculture and Rural Development Discussion Paper 23, World Bank: Washington DC.
Kennedy, E. and B. Cogill. 1987. Income and nutritional effects of the commercialization of
agriculture in southwestern Kenya. Research Report No. 63. IFPRI (International Food
Policy Research Institute), Washington, DC, USA.
Khan, A. R. 1984. Real wages of agricultural workers in Bangladesh. In A. R. Khan and E. Lee
Ed., Poverty in rural Asia (pp. 185–203). Bangkok: ILO.
Kirk, D. 1996. Demographic Transition Theory. Population Studies: A Journal of Demography,
50:3, 361-387.
Krautkraemer, J. A. 1994. Population growth, soil fertility, and agricultural intensification.
Journal of Development Economics 44: 403-428.
216
 

Krishna, A., P. Kristjanson, M. Radeny, and W. Nindo. (2004) Escaping Poverty and Becoming
Poor in Twenty Kenyan Villages. Journal of Human Development, 5(2), pp. 211-26.
Labao, L.M. and D.S. Michael. 1991. Farming patterns, rural restructuring, and poverty: a
comparative regional analysis. Rural sociology, 56(4):565–602.
Lin, W. and R. Dismukes. 2007. Supply response under risk: Implications for counter-cyclical
payments’ production impact. Review of Agricultural Economics 29: 64–86.
Lipton, M. 2005. The Family Farm in a Globalizing World: The Role of Crop Science in
Alleviating Poverty. IFPRI 2020 Discussion Paper No. 40. Washington, D.C.: IFPRI.
Livanis, G., C.B. Moss, V.E. Breneman, and R. Nehring. 2006. Urban Sprawl and Farmland
Prices. American Journal of Agricultural Economics, 88:915–929.
Lutz, W. and R. Qiang. 2002. Determinants of human population growth. Philosophical
Transactions of the Royal Society of London, B 357: 1197–1210.
Mason, N. 2011. The Effects of the Zambian Food Reserve Agency and Government Fertilizer
Programs on Smallholder Farm Household Fertilizer Use and Crop Production.
Unpublished Ph.D. dissertation essay, Michigan State University.
Mather, David and T.S. Jayne. 2011. The Impact of State Marketing Board Operations on
Smallholder Behavior and Incomes: The Case of Kenya. International Development
Working Paper 119, Department of Agricultural, Food, and Resource Economics,
Michigan State University, East Lansing, Michigan.
McCown, R.L., B.A. Keating, M.E. Probert, and R.K. Jones. 1992. Strategies for sustainable
crop production in semi-arid Africa. Outlook Agric. 21:21–31.
McMillan, M.S., W. A. Masters, and H. Kazianga. 2011. Rural Demography, Public Services
and Land Rights in Africa: A Village-Level Analysis in Burkina Faso. NBER Working
Paper No. 17718 http://www.nber.org/papers/w17718 [accessed March 18, 2012].
Mellor, J. W. 1995. Agriculture on the Road to Industrialization. International Food Policy
Institute, Johns Hopkins University Press: Baltimore, MD.
Migot-Adholla, S., P. Hazell, B. Blarel, and F. Place. 1991. Indigenous land right systems in subSaharan Africa: A constraint on productivity? World Bank Economic Review, 5 (1): 155–
175.
Morduch, J. 2000. Sibling Rivalry in Africa. American Economic Review, 90 (2), pp. 405-409.
Moti J., B. Gebremedhin and D. Hoekstra. 2009. Smallholder commercialization: Processes,
determinants and impact. Discussion Paper No. 18. Improving Productivity and Market
217
 

Success (IPMS) of Ethiopian Farmers Project, ILRI (International Livestock Research
Institute), Nairobi, Kenya.
Mundlak, Y. 1978. On the Pooling of Time Series and Cross Section Data. Econometrica, 46,
69-85.
Murton, John. 1999. Population growth and poverty in Machakos, Kenya. The Geographical
Journal, March
Narain, Dharm. 1965. Impact of Price Movements on Areas under Selected Crops in India:
1900-1939, Cambridge: University Press.
Naude, W. 2010. The Determinants of Migration from Sub-Saharan African Countries. Journal
of African Economies, Vol. 19, number 3, pp. 330–356.
Nerlove, M. and I. Fornari. 1998. Quasi-rational expectations, an alternative to fully rational
expectations: an application to US beef cattle supply. Journal of Econometrics, 83(1/2):
129-161.
Nerlove, M. 1956. Estimates of the Elasticities of Supply Response in Sub-Saharan Africa: A
Critical Review. African Development Review, 2(2):83-99.
Norton-Griffiths, Mike. 2008. Revitalizing African agriculture: Secure property rights offers best
hope for poor farmers. PERC Reports, vol. 26 (3): 32-36.
Notestein, F. 1945. Population—the long view. In Food for the World, ed. TW Schultz, pp. 36–
57. Chicago: Univ. Chicago Press.
Omamo, S. W. 1998. Farm-to-Market Transaction Costs and Specialisation in Small-Scale
Agriculture: Explorations with a Non-separable Household Model. Journal of
Development Studies, 35(2): 152-163.
Ravallion, Martin. 1990. Rural welfare effects of food price changes under induced wage
responses: Theory and evidence for Bangladesh. Oxford Economic Papers, New Series,
42(3), 558–574.
Ricardo, D. 1821. On the Principles of Political Economy and Taxation (3rd. ed). In P. Sraffa,
ed. The Works and Correspondence of David Ricardo. vol. 1. Cambridge: Cambridge
University Press.
Robinson, Warren and Wayne Schutjer. 1984. Agricultural Development and Demographic
Change: A Generalization of the Boserup Model. Economic Development and Cultural
Change, Vol. 32, No. 2 (Jan., 1984), pp. 355-366.

218
 

Papke, L.E. and J.M. Wooldridge. 2008. Panel data methods for fractional response variables
with an application to test pass rates. Journal of Econometrics, 145: 121–133.
Pender, J. and D. Alemu. 2007. Determinants of smallholder commercialization of food crops:
Theory and evidence from Ethiopia. Discussion Paper No. 75. IFPRI (International Food
Policy Research Institute), Washington, DC, USA.
Pender, J. and B. Gebremedhin. 2006. Land Management, Crop Production, and Household
Income in the Highlands of Tigray, Northern Ethiopia: An Econometric Analysis. In
Pender, John, Frank Place and Simeon Ehui (ed), Strategies for Sustainable Land
Management in the East African Highlands, International Food Policy Research Institute,
Washington DC, 107-139.
Pender J., S. Ehui and F. Place. 2006. Conceptual framework and hypothesis. In: Pender J, Place
and Ehui S (eds), Strategies for sustainable land management in the East African
highlands. IFPRI (International Food Policy Research Institute), Washington, DC, USA.
Pender, J., E. Nkonya, P. Jagger, D. Sserunkuuma and H. Ssali. 2006. Strategies to Increase
Agricultural Productivity and Reduce Land Degradation in Uganda: An Econometric
Analysis. In Pender, John, Frank Place and Simeon Ehui (ed), Strategies for Sustainable
Land Management in the East African Highlands, International Food Policy Research
Institute, Washington DC, 165-189.
Pender, J., F. Place, and S. Ehui. 2006. Strategies for sustainable land management in the East
African highlands. Environment and Production Technology Division Discussion Paper
No. 41. International Food Policy Research Institute. Washington, D.C.
Pinckney, Thomas C. and P.K. Kimuyu. 1994. Land Tenure Reform in East Africa: Good, Bad
or Unimportant? Journal of African Economies, 3(1), 1–28.
Pingali, P. L., M. Hossain, R.V. Gerpacio. 1997. Asian Rice Bowls- The Returning Crisis? CAB
International. Wallingford.
Pingali, P.L. and M.W. Rosegrant. 1995. Agricultural commercialization and diversification:
Process and polices. Food Policy 20(3):171–185.
Pingali, P., Y. Bigot, and H. P. Binswanger. 1987. Agricultural mechanization and the evolution
of farming systems in sub-Saharan Africa. Baltimore, Md., U.S.A.: Johns Hopkins
University Press.
Place, F., and S. Migot-Adholla. 1998. The Economic Effects of Land Registration on
Smallholder Farms in Kenya: Evidence from Nyeri and Kakamega Districts. Land
Economics 74(3):360–73.

219
 

Shi, Y.J., T.T. Phipps, and D. Colyer. 1997. Agricultural Land Values Under Urbanizing
Influences. Land Economics, 73:90–100.
Shipton, P. and M. Goheen. 1992. Understanding African Land-Holding: Power, Wealth, and
Meaning. Journal of the International African Institute, 62(3): 307-325.
Singh, I., L. Squire and J. Strauss. 1986. A Survey of Agricultural Household Models: Recent
Findings and Policy Implications, The World Bank Economic Review, 1(1): 149-179.
Simon, Julian. 2003. The Great Breakthrough and Its Cause, 4th Edition, Ann Abhor, MI:
University of Michigan Press.
Smucker, T. 2002. Land tenure reform and changes in land-use and land management in semiarid Tharaka, Kenya. The Land Use Change, Impacts and Dynamics (LUCID) Project
Working Paper Series Number 11, International Livestock Research Institute, Nairobi,
Kenya.
Stone, Glenn D. 2001. Theory of the square chicken: advances in agricultural intensification
theory. Asia Pacific Viewpoint, Vol. 42(2/3):163-180.Tegene, A., W.E. Huffman, and
J.A. 1988. Miranowski, Dynamic Corn Supply Functions: A Model with Explicit
Optimization. American Journal of Agricultural Economics, 70: 103-11.
Strasberg, P.J., T.S. Jayne, T. Yamano, J. Nyoro, D. Karanja and J. Strauss. 1999. Effects of
agricultural commercialization on food crop input use and productivity in Kenya.
International Development Working Papers No. 71. Michigan, USA.
Suri, T. 2005. Selection and Comparative Advantage in Technology Adoption. Unpublished PhD
Dissertation, Department of Economics, Yale University / Tegemeo Institute.
Swanson, L. 1988. Agriculture and Community Change in the U.S. The Congressional Research
Reports. Boulder, CO: Westview Press.
Tiffen, M., M. Mortimore, and F. Gichuki. 1994. More people -- less erosion: Environmental
recovery in Kenya. London: John Wiley & Sons.
Timmer, C.P. 1997. Farmers and markets: The political economy of new paradigms. American
Journal of Agricultural Economics 79(2):621–627.Wooldridge, J. 2010. Econometric
Analysis of Cross-Section and Panel Data. 2nd Edition. Massachusetts: MIT Press.
Weerahewa, J., K.D. Meilke, R.J. Vyn and Z. Haq. 2008. The determinants of farmland values in
Canada. Canadian Agricultural Trade Policy Research Network (CATPRN) Working
Paper 2008-03, Canada. http://www.uoguelph.ca/catprn/PDF-WP/Working_Paper_20083_Weerahewa.pdf [accessed November 29, 2012].

220
 

Wehbe, M., H. Eakin, R. Seiler, M. Vinocur, C. Avila and C. Marutto. 2006. Local Perspectives
on Adaption to Climate Change: Lessons from Mexico and Argentina. AIACC Working
Paper No. 39. September 2006.
Wilson, M.F.J., B. O'Connell, C. Brown, J.C. Guinan and A.J. Grehan. 2007. Multiscale terrain
analysis of multibeam bathymetry data for habitat mapping on the continental slope.
Marine Geodesy 30: 3-35. Available from: http://srtm.csi.cgiar.org/.
Wiggins, S., J. Kirsten, and L. Llambi. 2010. The Future of Small Farms, World Development,
38: 1341-48.
Wooldridge, J. 2010. Econometric Analysis of Cross-Section and Panel Data. 2nd Edition.
Massachusetts: MIT Press.
World Bank. 2007. World Development Report 2008: Agriculture for Development. Washington
DC: World Bank.
World Bank. 2006. World Development Report 2006: Equity and Development. Washington,
D.C.: World Bank.
Zaal, Fred and Remco H. Oostendorp. 2000.Explaining a miracle: intensification and the
transition towards sustainable small-scale agriculture in dryland Machakos and Kitui
Districts, Kenya. World Development, 30:1271-1287.
Zhang, Y., X. Liao, B.J. Butler and J. Schelhas. 2009. The Increasing Importance of Small-Scale
Forestry: Evidence from Family Forest Ownership Patterns in the United States. SmallScale Forestry, 8: 1-14.
Zhao, M., F. A. Heinsch, R. R. Nemani, and S. W. Running. 2005. Improvements of the MODIS
terrestrial gross and net primary production global data set. Remote Sensing of
Environment. 95: 164.176. Dataset is available from:
ftp://firecenter.umt.edu/pub/MODIS/Mirror/MOD17A3.LATEST/ .
Zilverberg, C., U. Kreuter and R. Conner. 2010. Population Growth and Fertilizer Use:
Ecological and Economic Consequences in Santa Cruz del Quiche´, Guatemala. Society
and Natural Resources, 23:1–13.
 
 

221