IMPACT OF INPUT SUBSIDIES ON HOUSEHOLD FOOD AVAILABILITY

IN RURAL ZAMBIA: A GENDERED PERSPECTIVE

A Research Report Presented to the Department of Agricultural Economics and Extension

of the University of Zambia

By

BLAIR SYAKOBBOLA

In Partial Fulfilment of the Requirement for the Degree of Bachelor of Science

in Agricultural Economics

© BLAIR SYAKOBBOLA, 2018

 ACKNOWLEDGEMENT

I thank God for guiding and strengthening me throughout this journey full of turbulences. Let his
name be forever glorified.

First and foremost, I would like to express my profound gratitude and heartfelt appreciation to my
supervisors; Dr Kiwanuka-Lubinda, R., Dr Chewe Nkonde and Dr Mofya-Mukuka, R., without
whom this study would not have been possible, for their guidance, constructive suggestions and
professional comments at every stage of this thesis.

My special thanks goes to my mother Mrs. Charity Mutampuka for her support and ceaseless
prayers, am forever indebted. Further, I would like to thank my relatives and friends for their love
and support throughout this journey especially Chena T. Lesa, Simon D. Zulu, my sister, Grace
Sikalundu-Mutenguna and Mukuka Chilalika, too mention just a few.
TABLE OF CONTENTS
LIST OF TABLES ..................................................................................................................................... iii
LIST OF FIGURES ................................................................................................................................... iv
LIST OF ABBREVIATIONS .................................................................................................................... v
ABSTRACT ............................................................................................................................................... vii
CHAPTER ONE ......................................................................................................................................... 1
INTRODUCTION....................................................................................................................................... 1
1.1 Background ....................................................................................................................................... 1
1.2 Problem Statement............................................................................................................................ 3
1.3 Rationale ............................................................................................................................................ 4
1.4 Study objectives................................................................................................................................. 5
1.4.1 Specific objectives ...................................................................................................................... 5
1.5 Research questions ............................................................................................................................ 5
CHAPTER TWO ........................................................................................................................................ 6
LITERATURE REVIEW .......................................................................................................................... 6
2.1 Introduction ....................................................................................................................................... 6
2.2 Household food security by gender of household head.................................................................. 6
2.3 Determinants of participation in input support programmes from a gender perspective ......... 8
2.4 Gendered Impact of participation in ISPs on food security .......................................................... 9
2.5 Conceptual Framework .................................................................................................................. 12
CHAPTER THREE .................................................................................................................................. 14
METHODOLOGY ................................................................................................................................... 14
3.2 Data .................................................................................................................................................. 14
3.2 Variables used in the model ........................................................................................................... 14
3.3.1 Key outcome Variables ............................................................................................................ 14
3.3.2 Independent Variables ............................................................................................................. 15
3.4 Empirical approach ........................................................................................................................ 16
3.4.1 Measuring the number of MAHFP ........................................................................................ 17
3.4.2 Determinants of participation in FISP by gender of the household head ........................... 17
3.4.3 Gendered impact of participation in FISP on MAHFP ........................................................ 19
CHAPTER FOUR..................................................................................................................................... 21

i
RESULTS AND DISCUSSION ............................................................................................................... 21
4.1 Introduction ..................................................................................................................................... 21
4.2 MAHFP by gender of household head .......................................................................................... 21
4.3 Factors affecting smallholder farmers’ participation in FISP from a gender perspective ...... 24
4.4.1 Female Decision Makers .......................................................................................................... 30
4.4.2 Participation in FISP ............................................................................................................... 31
4.4.3 The Interaction of FISP participation and female decision makers .................................... 31
4.4.4 Social-demographic characteristics ........................................................................................ 32
4.4.5 Distance to market ................................................................................................................... 32
4.4.6 Agro-ecological zones ............................................................................................................... 32
REFERENCES .......................................................................................................................................... 36

ii
 LIST OF TABLES

Table 1. Characteristics of decision makers ................................................................................. 23

Table 2. Comparison of means for selected sample characteristics of participants in FISP by
gender of decision maker .............................................................................................................. 25
Table 3: Comparison of percentages for selected sample characteristics of participants in FISP
by gender of decision maker ......................................................................................................... 26
Table 4. Comparison of means or selected sample characteristics by participation in FISP ....... 27
Table 5: Comparison of percentages for selected sample characteristics by participation in FISP
....................................................................................................................................................... 28
Table 6 FIML estimates of ESR for total fertilizer applied on FISP and Non FISP users. .......... 29
Table 7: Impact results using POLS and CRE .............................................................................. 33

iii
LIST OF FIGURES

Figure 1: Conceptual framework .................................................................................................. 13

Figure 2: Percentage of gender of decision maker in crop production by survey year ............... 22
Figure 3: Comparison of the distribution of MAHFP in rural households by gender of decision
maker in crop production. ............................................................................................................. 24

iv
 LIST OF ABBREVIATIONS

CRE Correlated Random Effects

CSO Central Statistical Office

ESR Endogenous Switching Regression

FAO Food Agricultural Organisation

FIML Full information maximum likelihood

FISP Farmer Input Support Programme

FRA Food Research Agency

FSP Fertilizer Support Programme

GDM Gender of the Decision Maker

GDP Gross Domestic Product

GII Gender Inequality Index

GHI Global Hunger Index G

IAPRI Indaba Agricultural Policy Research Institute

IFPRI International Food Policy Research Institute

ISP Input Subsidy Programme

MAHFP Months of adequate Household Food Provisioning

MGCD Ministry of Gender and Child Development

MIHFP Months of inadequate household food provisioning

MoA Ministry of Agriculture

POLS Pooled Ordinary Least Square

RALS Rural Agricultural Livelihoods Survey

v
SSA Sub-Saharan Africa

UNDP United Nations Development Plan

ZDHS Zambia Demographic and Health Survey

vi
 ABSTRACT

IMPACT OF INPUT SUBSIDIES ON HOUSEHOLD FOOD AVAILABILTY

IN RURAL ZAMBIA: A GENDERED PERSPECTIVE

By: Blair Syakobbola Supervisor: Dr Kiwanuka-Lubinda, R

This study estimates the impact of participation in Fertilizer Input Support Programme (FISP) on
gendered household food availability measured by months of adequate household food
provisioning (MAHFP). Using two waves of nationally representative Rural Agricultural
Livelihoods Survey (RALS) of small and medium-scale farming households in Zambia collected
in 2012 and 2015, the study measures and characterizes MAHFP by gender of decision makers
(GDM) in crop production. The results indicate that most primary decision makers are male and
most households have 9 MAHFP. In determining factors affecting participation in FISP, the study
used pooled full information maximum likelihood method which allowed the use of two waves of
panel data and controlled for endogeneity. Findings showed that participation in FISP is not
determined by gender, but other factors such as education levels and distance to market. To
determine impact of FISP on MAHFP by GDM, unobserved heterogeneity was controlled for by
employing correlated random effect for linear regression model. Results show that food
availability increases by 20.9 percent when households participate in FISP. However, households
participating in FISP and have female primary decision makers can increase food availability by
at least 23.6 percent. The study concludes that, on average, households are food insecure for at
least a quarter of the year. Worse still, more household with female primary decision makers have
lower MAHFP compared to those with male primary decision makers. Furthermore, Food
insecurity exists even in households participating in FISP. As such, it was recommended that
Government should prioritize investments in the key drivers of agricultural growth such as
extension services, research and rural infrastructure that would yield significant impact on
household food security. Further, it recommended that agricultural policies should include
deliberate strategies to improve resource allocation for women to enable them participate equitably
in programmes like FISP. It is more beneficial to target households with female primary decision
makers to enhance household food security in rural Zambia.

Key words: Impact, FISP, MAHFP, Gender, Rural Zambia

vii
 CHAPTER ONE

INTRODUCTION
1.1 Background

Majority of farmers in sub-Saharan Africa (SSA) are smallholders; they mainly produce crops for
home consumption and sell very little for cash to meet other non-food needs. Furthermore, over
65 percent of the rural population in developing countries is poor and food insecure, therefore,
improvement of agricultural production is one of the main strategies to reduce rural poverty and
food insecurity (World Bank, 2007). The low use of improved farm inputs in crop production,
especially fertilizer and hybrid seeds are among other factors that seriously impede agricultural
growth (Morris et al., 2007). Most smallholder farmers in SSA are unable to finance the purchase
of improved farm inputs to produce enough food and cash crops to meet household income and
food security requirements (Druilhe, 2012). In order to promote the use of fertilizer and hybrid
seeds, subsidies are one of the most pervasive policy instruments used by most governments in
developing countries (World Bank, 2007).

A few decades ago, most governments in SSA embarked on agricultural market reforms following
the structural adjustment programs which saw the removal of input subsidies for fertilizer and
maize seeds. However, with assertions that past mistakes have been realized and can be remedied
(Ricker-Gilbert et al., 2013), fertilizer and seed subsidies are once again at the center of many SSA
government’s agricultural development and poverty reduction strategies (Mason et al., 2013).
Predictably, most SSA governments, Zambia inclusive, are providing input subsidies programmes
(ISP) to smallholder farmers.

In Zambia, large-scale input subsidies were re-introduced in the 2002/2003 farming season through
the establishment of the Fertiliser Support Programme (FSP). The FSP mainly provided a maize
hectare pack which included a 400kg of inorganic fertiliser and 20kg of hybrid maize seed. In
2009, FSP was reviewed and renamed the Farmer Input Support Programme (FISP) after wide
stakeholder consultation. Although the objectives of FISP remained the same as its predecessor’s,
some changes accompanied the name change. Most notably, the input pack size was cut in half to
200kg of fertiliser and 10kg of hybrid maize seed. As part of the government’s push for crop

1
diversification, there was also an expansion in the range of crops included in the programme such
as rice, sorghum, cotton, and groundnuts (Mason et al., 2013).

The roll out of the newly introduced commodities of the programme was not countrywide. For
example, the provision of cotton input subsidies was only in Central, Eastern and Southern
provinces of Zambia in 2013/14 agricultural season (MAL, 2013). These provinces received a total
of 1,564MT and 156.40MT of fertilizer and cotton seeds respectively targeting 15,640
beneficiaries. Eastern province received the largest share of inputs (62.08 %) whereas Southern
province received the least (27.7%). Another example is the rice component of the input subsidy
which was only allocated to five provinces: Luapula, Northern, Western, Muchinga, and North-
Western. The total number of targeted beneficiaries for this crop was 15,890 with the largest given
to Northern Province (4,760). Similarly, the sorghum input subsidy was allocated to five
provinces: Copperbelt, Southern, Lusaka, Eastern and Western. A total of 21,420 beneficiaries
were targeted and Copperbelt province had the largest share of beneficiaries (7,020). Groundnuts
input subsidy was given to four provinces including Eastern, Muchinga, Northern and
North/Western. Generally, the number of target beneficiaries of the input subsidy has increased
from the initial 120,000 in the 2002/03 farming season (MAL, 2012) to the current target of over
1,000,000 in the 2017/18 farming season (MoA, 2018).

In the 2015/16 farming season, changes were made to FISP with the introduction of the electronic
voucher system. The changes were made to allow; increased private sector participation, timely
access of inputs by farmers, improved beneficiary targeting and promotion of agricultural
diversification (MAL, 2015). Fundamentally, Zambia has continued to make changes that improve
its agricultural input subsidy policies in order to achieve the intended goal of improved food
security. However, it remains unknown whether these policies are having the intended impacts
particularly when one considers the gender differences in FISP participation and their effect on
food security.

2
1.2 Problem Statement

In Zambia, women play a critical role in agriculture by providing 52 percent of labour as compared
to 48 percent of their male counterparts (MGCD, 2015). Also, women play a critical role in
improving the welfare of their communities. They manage domestic chores such as collecting
water and firewood and also sell surplus produce from their harvest. Additionally, rural women
are also responsible for the care and education of their families (Maria del, 2016). Despite the
important contribution of women towards agriculture, women continue to be viewed far more as
carrying out reproductive roles (unpaid care work) than productive roles. Much to their downside,
women have difficulties of balancing their reproductive roles with productive roles. The low
participation in reproductive roles by men is also a major constraint on women. Moreover, most
women are not in a position to make production decisions in terms of crops in households. What
is worse is that, even in some households headed by females, there are males making decisions on
production (Mofya-Mukuka Sambo, 2018).

Of concern is that, evidence from the Ministry of Gender and Child Development indicates a
continued decrease in participation of women in agriculture (MGCD, 2015). Amongst other
factors, this reduction has been attributed to women’s limited access to resources such as land and
inputs. For instance, out of the 1,540,390 households involved in agriculture in Zambia, 662,566
are supported by FISP and female headed households account for 17.6 percent of participants in
FISP compared to 82.4 percent of male headed households (MGCD, 2015).This disparity in access
to inputs by gender could partly be attributed to the eligibility criteria used for selecting
beneficiaries from the programme. To qualify for FISP, households are required to contribute 50
percent of the cost of the inputs of which most female-headed households are unable to do due to
lack of resources (ibid). This gender disparity in access to inputs from FISP could have far reaching
consequences on household food security. Hence, underscoring the need to provide empirical
evidence on the gendered impact of participation in FISP on household food security in order to
find improved means of avoiding these consequences.

Existing studies on the reintroduced farm input subsidies in SSA have focused on their direct and
general equilibrium impact (Sibande et al., 2016). Direct impact studies include effects on: (i)
maize output (Chibwana et al., 2010; Holden and Lunduka, 2010; Ricker-Gilbert and Jayne, 2011);

3
(ii) input markets (Xu et al., 2009; Chibwana et al., 2010; Ricker-Gilbert, et al., 2011); (iii) land
allocation (Holden and Lunduka, 2010; Chibwana et al., 2012) and (iv) household welfare,
including food security (Dorward and Chirwa, 2011); income from crops production, livestock
and asset worth (Ricker-Gilbert and Jayne, 2011; 2012); school attendance, health, household
shocks and stress (Chirwa et al., 2013). Studies that concentrate on investigating general
equilibrium effects have focused on maize prices, Gross Domestic Product (GDP) and agricultural
sector growth (Chirwa et al., 2013, Ricker-Gilbert et al., 2013).

In spite of vast empirical research carried out on the impact of FISP, there is no evidence that
specifically addresses the gendered impacts of participation in ISPs on rural household food
security. Therefore, this study will address this knowledge gap.

1.3 Rationale

Gender inequalities undermine the sustainable and inclusive development of the agriculture sector.
This causes disparities in development outcomes between men and women, mostly due to the fact
that rural women are often constrained by unequal access to productive resources and services
(FAO, 2011). As a result of these limitations on women, there are huge social, economic and
environmental costs on rural development and society as a whole, especially with regards to
agricultural productivity (Hill 2011).

If FISP is going to attain its goal of reducing poverty among the vulnerable small-scale farmers, a
deliberately biased approach towards female-headed households will have to be adopted so as to
narrow the gap between the two sets of households. The move will effectively reduce poverty
among female-headed households and uplift the lives of women. In view of the many challenges
faced by women, despite the key role they play in improving agricultural productivity and the
overall social and economic welfare of their households (FAO 2011), it is imperative that this
study is carried out so as to devise better ways of handling these challenges. This study is in line
with the Gender and Equality Act of 2015 that states that, there shall be special measures taken to
help benefit rural and peri-urban women and recognize the significant role they play in the
economic survival of their families and hence ensure they directly benefit from security
programmes such as FISP and any other programmes that affect development. This study will
therefore provide a basis on which to start the implementation of these policies.

4
1.4 Study objectives

The aim of this research is to estimate the impact of participation in FISP on gendered household
food availability measured by months of adequate household food provisioning (MAHFP).

1.4.1 Specific objectives

The specific objectives of the study are:

i. To characterize and measure MAHFP across gender of decision makers on crop production
ii. To determine the factors that influence smallholder farmers’ participation in FISP by
gender of the decision maker.
iii. To estimate the impact of participation in FISP on MAHFP by gender of the decision maker
on crop production

1.5 Research questions

1. How many months in a year did the farmers have adequate food?
2. What are the factors that influence smallholder farmer’s participation in FISP
conditional on the gender of the decision maker with respect to crop production
decisions?
3. What impact does FISP participation have on MAHFP by gender of the
decision maker?

5
 CHAPTER TWO

LITERATURE REVIEW
2.1 Introduction

This section looks at studies that have been carried out regarding this topic and hence provides a
basis on which this research was undertaken. The chapter begins by looking at literature on
household food security by gender of the head. Next, review of the determinants of participation
in input subsidy programmes from a gender perspective is presented followed by a section looking
at the impacts of IPSs on food security with a focus on the MAHFP. The chapter concludes with
a conceptual framework that provides a gist on how inter-related factors lead to food security in
female headed households.

2.2 Household food security by gender of household head

Food security is defined as a situation “when all people, at all times, have physical and economic
access to sufficient, safe and nutritious food that meets their dietary needs and food preferences
for an active and healthy life” (World Food Summit, 1996). There are various pathways to
improving food and nutrition security some of which include, increasing agricultural production
and productivity. These pathways can lead to a number of outcomes: an increase in food
availability and a reduction in food prices; increase in incomes and therefore people’s purchasing
power, and; ultimately empowering women (World Bank, 2014).

Studies show that women play a key role in attaining food security through a range of activities
they carry out; from production on the family plot, to food preparation and distribution within the
household (Njuki et al., 2016). However, the relationship between enhancing yields and food
security through improving women’s access to inputs/technologies is not that simple, since it is
mediated by several factors. This is evidenced by the fact that, even when women and men have
equal access to inputs, that balance is not reflected in their agricultural productivity. This is mainly
because of gender norms, market failures, or institutional constraints that alter the effectiveness of
these resources for women (World Bank, 2014).

6
For instance, studies in most developing countries have shown that while men dominate in high
value cash crops, women concentrate more on food crops (Kamga et al., 2016). Also, the
cultivation of different crops and different varieties may vary by gender. In many places, local
varieties are considered women’s crops, while high yielding varieties are considered men’s crops
(Badstue et al., 2007). Other studies show that, women having secure property rights gain status
and have more influence in household decisions, including decisions on food production, nutrition,
and use of income (Smith et al., 2003; FAO, 2011; World Bank, 2011). However, in a recent study,
Mofya-Mukuka and Sambo (2018), found that in some female headed households, decision makers
are males. There is further evidence that suggests that, countries with the most severe hunger
problems also have the highest levels of gender inequality (IFPRI, 2009). In Zambia, the focus of
this study, the World Economic Forum (2015) ranks the country to be 116 out of 145 countries
worldwide in its gender gap index. While the average of undernourishment–measured using the
Global Hunger Index (GHI)–is around 11.3 percent and 23.8 percent for all countries in the world
and Africa respectively, the GHI for Zambia stands at 39 percent (IFPRI, 2016). This is among the
three highest rates of hunger in Africa and in the world. Ironically, Zambia has repeatedly produced
national surplus of food staples in recent years (Chapoto et al., 2015).

In Africa, estimates of the time contribution of women to agricultural activities are as high as 60
to 80 percent in some countries (FAO, 2011). Therefore, improvements in the status of women
both within and outside households is critical to ensuring better food security outcomes especially
in rural areas. According to the World Economic Forums Global Gender Gap report, productivity
(output per hectare) on women’s farms is significantly lower compared to men ranging from 13
percent in Uganda to 25 percent in Malawi. The implication of this is that, for example, males in
Uganda managed to produce 13 percent more food compared to women. In order to address this
gender gap, there is a surge by many research and development organizations to find better
solutions for improving women empowerment in agriculture (World Economic Forum, 2013).

7
2.3 Determinants of participation in input support programmes from a gender perspective

Gender inequalities negatively affect sustainable and inclusive development especially in

agriculture. This has caused great differences in development outcomes between men and women.
In most cases, rural women are often constrained by unequal access to productive resources and
services (FAO, 2011).

As a result of these constraints, women face a number of challenges in ISP participation. Mason
et al. (2013) intimate that most of the fertilizer channeled through Zambia’s FISP is not allocated
to the poorest households but households that cultivate more land (above 2 hectares); and are more
likely to fall above the poverty line. In addition, to qualify for FISP, households are required to
contribute 50 percent of the cost of inputs of which most female-headed households are unable to
due to lack of resources. This has further contributed to the decline in the proportion of female
headed household that are supported under FISP (MGCD, 2015). Yet many studies show that
targeting FISP towards poorer households as well as female-headed households would increase
the amounts of fertilizer available in household for application in the fields, a move likely to
increase crop yields other factors remaining the same (Ricker-Gilbert, et al,. (2011); Mason &
Jayne (2013)

The UNDP Gender Inequality Index (GII) reflects gender-based inequalities in three dimensions
namely reproductive health, empowerment, and economic activity. Zambia has a GII value of
0.611, ranking the country 133 out of 149 countries in the 2013 index assessment. This can be
attributed to lack of education and unequal access to opportunities between male and female
(ZDHS, 2015). According to the ZDHS 2013-14, 8.4 percent of women aged 15-49 years
compared to 3.7 percent of men of the same age group had never attended any level of formal
education. Furthermore male compared to their female counterparts were literate, 82.7 percent and
67.5 percent, respectively, and 48.8 percent of the total number of women aged 15-49 years were
currently employed compared to 72 percent of men in the same age category (MGCD, 2015). The
lack of education is a serious impediment to FISP participation by most women. This is because
with increased education comes many other opportunities such increased capability to make
income and increased knowledge in the importance and better ways of achieving improved

8
agricultural production.

Recently, issues of decision making in agricultural practices have been of concern among policy
makers. It is believed that women participate less in decision making of crop production. For
example, in Ethiopia, in a study to understand gender participation and decision making in farming
activities, Mulugeta and Amsalu (2014), found that most rural women did not have any decisional
role in any crop field related activities such as purchasing of chemicals, deciding what crops to
plant and purchasing any type of inputs despite being highly involved in critical farming activities
such as weeding, seed preparation and harvesting.

2.4 Gendered Impact of participation in ISPs on food security

According to the United Nations Population Information Network (1995), women’s empowerment
encompasses women having a sense of self-worth; their right to have and determine choices; their
right to have access to opportunities and resources; their right to have the power to control their
own lives, both within and outside the home; and their ability to influence the direction of social
change to create a more just social and economic order, nationally and internationally.

Women are an important part of the agricultural labor force, and agricultural and agriculture value
chains are equally important to women as a source of food and employment. Aggregate data shows
that women represent about 43 percent of the agricultural labor force globally and in developing
countries (FAO, 2011).

A study carried out in Malawi by Sibande et al. (2015) shows that there is a positive relationship
between FISP and the number of months of household food security. The study further establishes
that the percentage incremental effect is highest for poorest and most food insecure households.
Measured in levels, the effect is higher among the most food secure households. However, the
magnitude of the effects of subsidized fertilizer on food security were not large enough to eradicate
food insecurity among poor households in isolation. The researchers further argue that input
subsidy programs could improve food security if they targeted large food producers and not the
smallholder farmers. Hence, they conclude that ISPs are less useful if the objective is to reduce
poverty. However, this assertion may be hard to accept especially from the end of policy makers.
It is noteworthy to realize that redistribution of income and any other basic need to ordinary

9
citizens is difficult in developing countries (Ravallion, 2009). Thus, if ISPs targeted only large
producers, the welfare of smallholder farmers who are in fact the majority can be worsened off
(Mason et al. 2013). This can cause a further increase in poverty and food insecurity.

In Lesotho, Thaphelo and Li (2016) find that the use of the subsidized inputs results in increased
food production. However, very few households have access to the input support programmes due
to the requirements needed for one to access the support whereby most small-scale farmers fail to
meet.

Most studies on the impact of ISPs show that there is a positive relationship between the ISPs and
increased production. Even though this is the case, the gendered impact of the ISPs is not taken
into account. This is worrying considering that most of the rural population consists of women that
are mostly poor (Sibande et al. 2015).

Several other studies that determine the impact of FISP on smallholder farmers have been
undertaken in Malawi. For instance, Chibwana et al. (2010) investigated whether Malawi’s FISP
had influenced farmers to diversify or simplify their crop varieties and cropping patterns and
whether it resulted into specialization of crops. The study noted that FISP had reduced allocation
of land to legume crops which could have negative impacts on soil fertility. Further findings in
this study suggest that maize and tobacco, the crops that were included in the FISP package
received a larger proportion of land allocation compared to other crops. The study also found that
older household heads allocated more land to traditional maize than young households. On the
other hand, educated households allocated less land to traditional maize than uneducated heads.
The patterns were opposite in terms of land allocation for uneducated heads.

Further findings by Chibwana et al. (2010) suggest that even though larger farms used more
fertilizer for maize production. The intensity of fertilizer use was high in small farms providing
greater returns per unit of fertilizer applied. Female headed households used less fertilizer for
maize than their male counterparts. This could be attributed to the constraints that women face in
accessing resources.

10
Shively (2013), investigated the impact of increased application of fertilizers due to subsidy
programmes on maize yield in Malawi. The findings found a positive correlation between the
amount of fertilizer used and yield. However, at higher application rates of fertilizer, total yield
exhibits diminishing returns to fertilizer use.

Meanwhile in Gwembe district of Southern province, Zambia, Sianjase and Seshamani (2013),
carried out a research on the impact of FISP on maize yield. The study identifies three groups of
farmers in the area. Firstly, farmers that have small land and solely depend on FISP to produce,
secondly, farmers that depend on FISP but have a relatively larger size of land and thirdly, farmers
that do not necessarily depend on FISP and also have huge land. The study finds that the first group
of farmers was less responsive to input subsidy. The second group of farmers was most responsive
group to input subsidies and recorded greater yields, while the third group’s yield was not as high
as expected. Hence, they concluded that if the Zambian government wants to increase household
food security and reduce poverty among its rural population, the target group of farmers should be
those that are most responsive to subsidies. Yet, the incidences of poverty are high among the
smallholder farmers especially those cultivating less than 2 hectares of land (Mason et al, 2013).
Hence, it is not easy for policy makers to neglect very small farmers in allocation of fertilizer
subsidies despite their being less responsive to the input support program. Therefore, if policy
makers are to make this decision, they should first find alternative ways of taking care of this group
as they are the most vulnerable farmers in rural areas.

Maximilian et al. (2018), find that the subsidy programme has failed to substantially reduce
poverty and improve food security through reduced maize prices in Zambia. Although farm
incomes have moderately increased, the overall costs exceed the benefits. This failure is attributed
to poor targeting, diversion and leakages.

Given the above rich literature, understanding the link between FISP and months of adequate
household food provision focusing on gender is critical to gleaning yet another aspect that has
probably not been explicitly investigated in previous studies

11
2.5 Conceptual Framework

Food security is defined into four main dimensions; access, availability, utilization and stability
(FAO, 2008). Food availability encompasses the supply side of food security and is determined by
the physical levels of food production, stock and even net trade. Food access is concerned with
economical and physical access of food at household level. Food utilization refers to the way the
body utilizes various nutrients in the food, taking into account the issue of dietary diversity, intra-
household distribution of food and feeding practices. All of which result into sufficient energy and
nutrient intake by individuals and when combined with a good biological utilization of food
consumed, improve the nutritional status of individuals. Stability is the component that deals with
the consistent availability, accessibility and utilization of food. Although all the four components
of food security are simultaneously required, this study mainly focuses on food availability which
will be measured by MAHFP in this study.

The framework used in this study shows how FISP can lead to increased crop production. In most
rural set ups, gender has an influence on land accessibility and ownership (FAO, 2011). Land
access in turn has an effect on participation in FISP since part of the requirement for access to
FISP is access to land. There is enough evidence to suggest that FISP leads to increased crop
production and crop production can lead to increased food access and food availability which
determine the MAHFP (Sibande et al, 2015).

Increased women support through women empowerment and support programmes can also result
into improved welfare of women through increased resource ownership which can also enable
them acquire FISP and hence increase crop production. In addition, increased women
empowerment can result into increased knowledge for women which in turn can result into
management of resources thereby contributing to food availability at household levels and the
national community at large.

Income is another important consideration when acquiring FISP; 50 percent of the inputs must be
paid for by the farmers (MGCD, 2015). Income access is however dependent on so many factors,
such as the level of education, age, skills, employment and even crop production.

The inter-connections between the factors that lead to the MAHFP and ultimately food security is
the reason the study adopts this conceptual framework. The study however is a gendered impact

12
study whose overall aim is to measure the gendered impact participation in FISP on MAHFP, a
study that needs a lot of empirical studies as there has been very little works done to fully
understand this knowledge gap. Figure 1 illustrates the links between different factors that affect
the MAHFP and ultimately food security.

Figure 1: Conceptual framework

13
 CHAPTER THREE

METHODOLOGY
3.1 Introduction
A quantitative research approach was used for this study. The rest of the chapter describes the
research design beginning with explaining the data used. This is followed by highlighting the
empirical approach on which the study is grounded by detailing the analytical methods employed.

3.2 Data
This study used the nationally representative Rural Agricultural Livelihoods Survey (RALS) of
small and medium-scale farming households in Zambia collected in 2012 and 2015. The RALS is
a longitudinal survey conducted by the Indaba Agricultural Policy Research Institute (IAPRI) in
conjunction with the Central Statistical Office (CSO) of Zambia and the Ministry of Agriculture
(MoA). The RALS used the new sampling frame derived from the 2010 census. A total of 8,839
and 7,934 households were interviewed during the 2012 and 2015 surveys respectively. The survey
collected data on a number questions related to the following main themes: demographic
characteristics of household members; farmland and use; crop sales from own production; input
and credit acquisition; livestock ownership and marketing; off-farm income sources; food security
indicators, and; other themes such as kinship ties of the household head. Information on access to
inputs from various sources including the government’s FISP were also collected

3.2 Variables used in the model

Various pathways leading to food availability guided the variables to include in the model. The
variables used included both the outcome and the independent variables.

3.3.1 Key outcome Variables

In this study, the key outcome variable is MAHFP. MAHFP was used to measure household food
provisioning as a proxy measure of household food availability. In the RALS 2012 and 2015 data,
this variable was captured as months of inadequate household food provisioning (MIHFP). Stata
14 software was used to generate MAHFP by subtracting MIHFP from 12. According to Deitchler
et al. (2010), over time, the MAHFP indicator can capture changes in the household’s ability to
address vulnerability in such a way as to ensure that food is available above a minimum level year-
round. Measuring the MAHFP has the advantage of capturing the combined effects of a range of

14
interventions and strategies, such as improved agricultural production, storage, and interventions
that increase the household’s purchasing power.

3.3.2 Independent Variables

(a) The decision to produce a crop in a field

In order to ensure food availability in female headed households, women are required to have the
capacity to produce and be the ones to decide what crop to produce and from which field.
According to FSLF (2013), women empowerment can lead to household food availability, poverty
reduction and ultimately food security. In the Women Empowerment Agricultural Index, the
gender of the decision maker (GDM) on the decision of crop production is used as proxy measure
for women empowerment (IFPRI, 2012). To come up with this variable, RALS 2012 and 2015
data was used to identify decision making of production by gender at field level. The next step was
keeping households that had females making decision of production on any of the fields. The final
step was to collapse the files to household level in order to allow merging with the original file of
our data set. GDM is a dummy variable which indicates 1 if the decision maker is female and 0
otherwise.

(b) Participation in FISP

Participation in FISP is vital for increase in physical food production. FISP supplies farmers with
inputs needed for production such as maize seeds and fertilizer (MAL, 2015). There is evidence
suggesting that farmer participation in FISP results into increased production, increased income
and relatively large fields (Chibwana et al. 2010; Shively, 2013; Sianjase and Seshamani, 2013).
The variable FISP is a dummy variable indicating 1 if a farmer participated in FISP

(c) Social-demographic characteristics

Besides the gender of the decision maker (GDM) and participation in FISP (Fisp-KG), food
availability may also be increased by several social-demographic characteristics (Harris-Fry et al,
2015). Therefore, this study will control for several social-demographic characteristics such as
land size, land category, education levels of the household head, off farm income and the age of
the household head.

15
(d) Distance to markets in hours

Distance to markets can have an effect on food availability in that both inputs and outputs require
the existence of markets to enable production and selling of produce. A study by Omiti et al. (2009)
established that households located near urban centers sold more than those that were located in
rural areas because the former had reduced transaction costs which enabled participation in the
market. Distance to markets was captured in hours. As such this study will look at variables related
to market such hours to the nearest place with at least 500000 inhabitants.

(e) Agro-ecological zones

Zambia has varying climatic condition and is divided into 3 zones: I, II and III. These can have an
effect on crop production and reduce the availability of food at household level especially going
in the near future (Wineman and Crawford, 2017). According to the Zambia Agricultural Research
Institute, Region I receives less than 800mm of rainfall and covers Southern province and parts of
Western and Eastern provinces. Region II receives between 800mm and 1000mm. This region is
sub-divided into IIA and IIB. It covers the Central, Lusaka, Southern and Eastern fertile plateau of
the country and has naturally fertile soils. Region III receives between 1000mm and 1500mm of
rainfall. The region covers the Copperbelt, Luapula, Northern and Northern Western provinces of
the country. Following the importance of agro-ecological zones, this study will control for these
agro-ecological conditions.

3.4 Empirical approach

The empirical approach of the model outlines the methods that were employed to estimate the
impact of FISP on MAHFP by gender of household. The underlying model is presented followed
by the estimation strategy.

16
3.4.1 Measuring the number of MAHFP

The first objective which sought to know the MAHFP by gender of households was obtained using
descriptive statistics. The MAHFP was stated for each household. This was done using Stata 14, a
statistical software. A tabulate command was run for the variable MAHFP. This was followed by
the same command by gender of the decision makers. Tabulating the variable by gender of the
decision makers helped in understanding how MAHFP was distributed by households in terms of
gender in decision making regarding crop production.

3.4.2 Determinants of participation in FISP by gender of the household head

Previous studies on participation of FISP have used a two-stage instrumental variable regression
(see for example, Chibwana et al., 2010). Some studies have used the maximum likelihood
estimation. The use of instrumental variables has an advantage in that it counteracts issues with
measurement errors and also overcomes issues of omitted variables (Angrist, 2001).
Notwithstanding its advantges, this approach is unreliable especially when the intruments are weak
and in cases of more than one endogenous variable. It also does not yield consitent standard erros.
Thus, based on Lokshin and Sajaia (2004), this study made use of the pooled full information
maximum likelihood method (FIML) to simultaneously estimate the selection equation and the
two regime equations outlined below. This was done to correct the inefficiencies inherent in other
estimation methods like the two-step least square or maximum likelihood estimation. The FIML
yields consistent standard errors. This empirical strategy helped in identifying determinants of
FISP participation and assessed how they differed across gender of decision makers. The use of
pooled FIML in the analysis entailed estimating parameters with data from the two waves of the
panel taken as a cross-section.

The underlying model that describes the relationship between the dependent variable (quantities
of fertilizer received) and the two key independent variables (participation in the FISP and gender
of the household head) ceteris paribus was depicted as follows:

𝑦 = 𝑋 ′ 𝛽 + 𝛿𝐼 + 𝛼𝐺𝐷𝑀 + 𝑢 (1)

17
where 𝑦 is the quantity of fertiliser received (in kilogrammes), 𝑋 is a vector of exogenous
characteristics, 𝐼 is a dummy variable defined as 𝐼 = 1 if the household received subsidized
fertilizer from FISP and 0 otherwise, GDM is a dummy variable defined as GDM = 1 if the gender
of the decision maker in a household is female and 0 if male, and 𝑢 is the error term. The parameter
estimates are represented by 𝛽, 𝛿 𝑎𝑛𝑑 𝛼.

The regression model outlined above assumes that 𝐼 is uncorrelated with the error term (exogeneity
condition). This approach might lead to biased parameter estimates because of this strong
exogeneity assumption. The implication of this is that some unobserved characteristics that
influence the probability that a farmer participates in FISP could also influence the quantities of
fertilizer received by the farmer. To control for these methodological challenges, the study used
the endogenous switching regression (ESR) estimation strategy.

Using the ESR, the underlying model depicted in equation 1 now consists of two parts. The first
part is a selection equation, which models the decision to participate or not in the FISP using a
standard limited dependent variable method. The second part is the outcome equation estimated
for each group (FISP and non-FISP farmers). Thus, the model is specified as follows:

1 𝑖𝑓 𝐼𝑖𝑡∗ > 0
Selection equation 𝐼𝑖𝑡∗ = 𝑍𝑖𝑡 𝛾 + 𝛼𝐺𝐷𝑀 + 𝜀𝑖𝑡 , 𝐼𝑖 = { (2)
0 𝑖𝑓 𝐼𝑖𝑡∗ ≤ 0

Regime 1 (FISP) 𝑦1𝑖𝑡 = 𝑋1𝑖𝑡 𝛽1 + 𝛼𝐺𝐷𝑀 + 𝑢1𝑖𝑡 𝑖𝑓 𝐼𝑖𝑡 = 1 (3)

Regime 2 (Non-FISP) 𝑦2𝑖𝑡 = 𝑋2𝑖𝑡 𝛽2 + 𝛼𝐺𝐷𝑀 + 𝑢2𝑖𝑡 𝑖𝑓 𝐼𝑖𝑡 = 0 (4)

Where, in the selection equation, 𝐼 ∗ is a latent variable for choice of fertilizer source, 𝐼 is what is
observed (𝐼 = 1 if the household received subsidized fertilizer from FISP and 0 otherwise), 𝑍 is a
vector of observed farm and non-farm characteristics determining the decision to receive or not to
receive subsidized fertilizer, GDM is as defined above and 𝜀 is the error term. For the second part,
𝑦 represents the outcome variable (quantity of fertilizer received in kilograms), 𝑋 is a vector of
variables that capture household characteristics, market environment, district fixed effects and the
agricultural policy environment. Similarly, GDM is as defined previously

18
3.4.3 Gendered impact of participation in FISP on MAHFP

The third objective sought to measure the impact of participation in FISP on MAHFP conditional
on GDM and was estimated as follows;

𝑦𝑖𝑡 = 𝒙𝑖𝑡 𝜷 + 𝑐𝑖 + 𝑢𝑖𝑡 (5)

where for household 𝑖 and time period 𝑡, 𝑦𝑖𝑡 denotes each indicator measuring food security, 𝒙𝑖𝑡
is a 1 x K vector of covariates such as technology factors, ownership of capital assets, labour
supply, factors that affect the household’s production and environment, FISP_kg and GDM. The
term 𝑢𝑖𝑡 is the idiosyncratic error term while 𝑐𝑖 is the time invariant unobserved heterogeneity (e.g.
farmer ability) and 𝜷 is a K x 1 vector of parameter estimates.

To estimate equation 5, the two commonly used approaches in micro-econometrics are the random
and fixed effects frameworks. However, they both have shortcomings, which render them
inappropriate for the proposed application. In the random effects framework, the underlying
assumption is that the unobserved heterogeneity (𝑐𝑖 ) is uncorrelated with the observed explanatory
(𝒙𝑖𝑡 ) variables (Wooldridge, 2010). This strong assumption might not be entirely plausible. The
fixed effects framework, on the other hand, imposes no restrictions on the relationship between 𝑐𝑖
and 𝒙𝑖𝑡 .

To overcome these shortcomings, this study estimated the gendered impacts of participating in
FISP on MAHFP using the correlated random effects (CRE) that allows modeling the dependence
between 𝑐𝑖 and the𝒙𝑖𝑡 . The study made use of the Mundlak-Chamberlain device (CRE for linear
in parameters model) to model the aforementioned relationship between 𝑐𝑖 and the 𝒙𝑖𝑡 . Thus, the
estimation equation was as follows:

𝑦𝑖𝑡 = 𝛽0 + 𝑘𝑖𝑡 𝛽1 + 𝑙𝑖𝑡 𝛽2 + 𝑧𝑖𝑡 𝛽3 + 𝛽4 𝐹𝐼𝑆𝑃_𝑘𝑔𝑖𝑡 + 𝛽5 𝐺𝐷𝑀 + 𝛽6 (𝐹𝐼𝑆𝑃_𝑘𝑔 ∗ 𝐺𝐷𝑀)𝑖𝑡 + 𝑐𝑖 + 𝑢𝑖𝑡

(6)

Where all the variables are as previously defined. The main parameter of interest that this study
sought to estimate is 𝛽6, the coefficient on the interaction term between FISP and whether or not
the decision maker is female in equation (6). Hence, this study answered the question, how does

19
food availability change for households when the gender of the decision maker is female and is
participating in FISP, ceteris paribus?

20
 CHAPTER FOUR

RESULTS AND DISCUSSION

4.1 Introduction

This chapter presents results and discusses the findings of the study. Results will be organized in
accordance with the objectives with each finding followed by a plausible explanation.

4.2 MAHFP by gender of household head

Gender at household level was analyzed in terms of the decision maker in crop production (figure
2). This gave an accurate understanding of gender. It is thinkable that in a female headed
household, a man, perhaps, an older son or relative is making production decisions. In both survey
years 2012 and 2015 the majority of decision makers in crop production were male. In 2012, male
decision makers were 5,916 (66.9%), while the female decision makers were 2,924 (33.1%). Even
though the percentage of female decision makers slightly increased in 2015, male decision makers
were still more than female with 4,846(61.1%) being male decision makers and 3,088 (38.9%)
being females decision makers.

21
Figure 2: Percentage of Gender of Decision Maker in Crop Production by survey year

Percentage of GDM by survey year

80

70

60
percentage

50

40

30

20

10

0
2012 survey 2015 survey

Males Females

Source: CSO/MoA/IAPRI 2012/2015 and authors own calculation.

Further, characterisation of decision makers were analyzed in detail (Table 1). This helped in
understanding who exactly was making decisions in households. Male decision makers were
slightly younger (45.37) than female decision makers who were 49.86 years. Male decision makers
had more years of education (6.5) than their female counterparts (5.4).

The sample population comprised of 20.1% female headed households. Worryingly, only 54.2
percent of these female heads were decision makers in the household. The other 45.8 % of decision
makers in female headed households were male. However, in the male headed households,
comprising of 79.9% of households, 98.9 percent of decision makers were male with 1.1 percent
being female. It is startlingly surprising to find a high presence of male decision makers in female
headed households. This finding could entail the inability of most females to make solid and sound
decision on production of crops. Also, this shows that men have greater decision control in most
undertakings at household level than females. Hence, this may help explain why most women are
disadvantaged against in accessing resources. These findings are in agreement with findings of
Mofya-Mukuka and Sambo (2018). Lower levels of education and high age among female decision
makers could explain the presence of male decision makers in female headed households. This

22
 finding confirms a report from the Ministry of Gender and Child Development which indicates
higher illiteracy levels among women compared to men (MGCD, 2015).

Table 1. Characteristics of decision makers

Sub-population .

Variable description Population Male decision maker Female decision maker

Age of head (years) 46.98 45.37 49.86

Education level (years) 6.1 6.5 5.4

Female headed HH (%) 20.1 45.8 54.2

Male headed HH (%) 79.9 98.9 1.1

Source: CSO/MoA/IAPRI 2012/ 2015 and authors own calculation.

The study further compared the MAHFP by gender of the decision maker. Figure 2 shows the
distribution of MAHFP by gender. Most households with female decision makers 842

(34.37 %,) had 9 MAHFP, while the least had 2 (0.08%) MAHFP. The highest MAHFP was 11
months (0.78%, n= 19) while some households did not have MAHFP in any month (0.61%, n=
15). Similarly, most households with male decision makers 1,371 (35.38%) had 9 MAHFP with
the least having 2 (0.49%). The highest MAHFP was also 11 months (1.06%, n=41).

Of concern is that a good number of households had adequate food available for less than 8 months
and no household had MAHFP for the entire 12 months. Especially worrisome is the fact that some
households did not have MAHFP in any month. What this implies is that most households in rural
Zambia are poor, food insecure, and threatened with poor healthy which can cause a further
downward movement of their production. These findings are consistent with (Maximilian et al,
2018; World Bank, 2007; Sibande et al. 2015), that mention high levels of food insecurity in
countries like Zambia and SSA in a broader perspective.

Most households with lower MAHFP of 8 and below were those with female decision makers. The
opposite was true for MAHFP of 9 and above. This suggests that, households with male making
production decisions have more months of food availability than households where females are

23
making decisions. These results are not surprising considering the disparity in access of resources
by men and women (MGCD, 2015; FAO 2011). Another reason for this may be that tradition
sparingly affords women with opportunities to participate in, among other things, production of
high value crops, accessing land and markets (World Bank, 2014).

Figure 3: Comparison of the distribution of MAHFP in rural households by gender of

decision maker in crop production
40
35
30
Percentage

25
20
15
10
5
0
0 1 2 3 4 5 6 7 8 9 10 11
Number of Months

females males

Source: CSO/MoA/IAPRI 2012/2015 and authors own calculation

4.3 Factors affecting smallholder farmers’ participation in FISP from a gender perspective
Table 2 displays a comparison of sample characteristics of participation in FISP by gender of the
decision makers. This is was done using the t-test statistic for difference in means for continuous
variables. Categorical Variables made use of chi-square statistic for variable association.

As can be seen in table 2 , at 1% significance level, gender of the decision maker could have an
effect on: food availability in months, number of fields cultivated, fertilizer applied (KGs)
education levels, age of the head, access to markets and income. Also at 5% level land hold in size
is affected by gender. This agrees with findings that there is a difference in accessing resources as
a result of gender (MGCD, 2015; FAO 2011). Households with male decision makers had on
average more months of food availability than their female counterparts.

24
Table 2: Comparison of means for selected sample characteristics of participants in FISP
by gender of decision maker
Variable Overall Female Male decision
decision maker
maker

Food availability (Months) 8.31 8.16 8.41 ***

Number of fields cultivated 2.52 2.62 2.46 ***

Fertilizer applied (kgs) 275.14 255.16 286.3***

Education of the household head (Years) 6.1 5.4 6.45 ***

Age of the household head (Years) 46.98 49.86 45.37***

Tropical Livestock Unit 3.59 3.63 3.57

Full time adult equivalent 2.85 2.62 2.98 ***

Livestock value (ZMW) 1.07 1.38 8.97

Landholding units (Hectares) 4.20 3.99 4.32**

Lagged purchases by FRA 11508.26 12113.7 11169.7***

Distance to the nearest urban center (Hours) 13.45 13.07 13.66***

Significance: *** p<0.01, ** p<0.05, * p<0.1.

Source: CSO/MoA/IAPRI 2015 and authors own calculation.

25
Table 3: Comparison of percentages for selected sample characteristics of participants in
FISP by gender of decision maker
percentages

Variable Overall Female Male decision

decision maker
maker

Participation in FISP 39.66 39.63 39.71

Agro-ecological zones

Household domiciles in Zone IIA 43.37 43.32 43.46 ***

Household domiciles in Zone IIB 6.46 6.1 7.12 ***

Household domiciles in Zone III 39.74 40.62 38.26 ***

Significance: *** p<0.01, ** p<0.05, * p<0.1.

Source: CSO/MoA/IAPRI 2015 and authors own calculation.

The gender of decision maker is associated with agro-ecological zone at 1% significant level. This
could mean that household decision in crop production could be determined depending on the
region. Some crops do not do well in some regions and hence the significance of the variable agro-
ecological zone in terms of gender.

Similarly, table 4 and 5 show a comparison of sample characteristics in terms of participation in

FISP by gender. Table 4 shows that, at 1% percent level, there is a difference between participants
and non-participants in FISP and not participating in terms of food availability, number of fields
cultivated, quantity of fertilizer applied (Kgs), education of the head, age of the household head,
landholding and access to markets. This shows that, for instance, a high land hold size could make
one participate in FISP than a lower one. In most rural areas, land hold size is a symbol of wealth
and ownership of other assets. Ownership of resources play a great deal in farmer participation in

26
FISP. The more the number of fields the more fertilizer that will be needed for use in crop
production. The need for more fertilizer mainly is key driver for farmers to access FISP.

Table 4. Comparison of means or selected sample characteristics by participation in FISP

Variables Overall Participants Non-
participants

Means

Food availability in months 8.31 8.65 8.16 ***

Number of fields cultivated 2.52 2.94 2.25 ***

Kgs of fertilizer applied 275.13 499.66 127.5***

Education of the head 6.1 6.74 5.66 ***

Age of the head 46.9 48.03 46.29***

Tropical livestock Unit 3.59 4.94 2.71***

Full time adult equivalent 2.85 3.37 2.50***

Livestock value 1.07e 1.44 8.26

Landholding 4.20 5.31 3.47 ***

Lagged purchases by FRA 11508.2 12822. 10647.69***

Hours to nearest urban centre 13.46 12.84 13.87 ***

Significance: *** p<0.01, ** p<0.05, * p<0.1.

Source: CSO/MoA/IAPRI 2015 and authors own calculation.

27
Table 5: Comparison of percentages for selected sample characteristics by participation in
FISP

Variable Overall Participants Non-

participants

Percentages

Female Gender of decision maker 35.8 35.9 35.9

Agro-ecological zones

Household domiciles in Zone IIA 43.3 45.5 42.0 ***

Household domiciles in Zone IIB 6.5 2.5 9.1***

Household domiciles in Zone III 39.7 42.8 37.7***

Significance: *** p<0.01, ** p<0.05, * p<0.1.

Source: CSO/MoA/IAPRI 2015 and authors own calculation.

As shown above, difference in agro-ecological zones can determine whether one participates in
FISP or not. It is unprofitable to grow certain crops in some regions which can affect participation
in FISP.

4.3.1 Econometric results on factors that affect participation in FISP

To determine factors that affect participation in FISP, the study made use of FIML to estimate the
parameters. The results are shown in Table 6.

28
Table 6: FIML estimates of ESR for total fertilizer applied on FISP and non-FISP users

Variable Selection equation FISP participant Non-FISP

Fertilizer Kgs participants
Fertilizer Kgs
Tropical livestock unit .01*** 10.667*** 5.66***
Age of the head (years) .002 -.949 -.003
GDM (1= Female) -.05 -54.29* -6.67
Level of education (years) .02*** 11.03*** 3.969***
Full time adult equivalent .01 6.44* 5.98***
Land hold size (ha) .01** 11.18*** 3.56***
Hours to Nearest Urban -.01*** -4.29*** -2.007***

Number of fields cultivated .14*** .86 6.602***

If hh owns poultry (1=yes) .07 97.79*** 20.34*
Zone IIA .03 41.3 46.33***
Zone IIB -.35** 45.43 -7.12
Zone III .25*** 31.28 37.37***
Food availability in months .05*** .-6.55 .863
Lagged maize purchases by FRA -2.92* .00 .00086***
in MT
Kinship ties -.24***
Local -.07
Years of settlement .002
Member of any group 2.02***
Constant -2.63*** 356.48*** -36.05***
Sigma
Rho -.2709***
Rh1 -.137***
N 5841
Note: (1) Significance levels *** at 1%, ** at 5%, * at 10%.

(2) LR test of indep. Eqns: chi2 (2) = 51.08 Prob > chi2 = 0.0000

Source: Authors Econometrics Analysis using Stata software and data obtained from
CSO/MoA/IAPRI 2015.

29
The selection equation in Table 6 shows that, at 1% significance level, food availability in months,
education of the head, hours to the nearest place with 500000 inhabitants, total livestock units,
count of members, hectares cultivated, Zone III, being a member of a cooperative and kinship ties
determine participation in FISP. Zone IIB and land hold are significant at 5% level while lagged
maize purchases by Fra is significant at 10%. Unsurprisingly, distance to nearest place with at least
500000 inhabitants and zone II B show a negative effect of participation in FISP. This means that,
the further away a household is from a market center, the less likely they are to participate in
economic activities which can potentially exclude them from participation in FISP. In the case of
households domiciled in zone II B, it could be that crops do not do well in this region and fearing
losses, households in this region participate less in FISP.

Just as expected high levels of education, months of food availability, being a member of any
group or cooperative and total livestock unit indicate an increased participation in FISP. Education
levels and total livestock unit are associated with income earning which is needed to participate in
FISP (MGCD, 2015). Also, the programme FISP is naturally designed to target farmers belonging
to groups such as cooperatives.

The variable for Gender of the Female Decision Maker is negative but insignificant. This implies
that female decision makers participate less in the FISP programme. However, the result itself
indicate that gender does not affect participation in the programme.

4.4 Impact of FISP on the month of adequate food provisioning by gender

Table 7 shows results of regression when unobserved heterogeneity is controlled for and when not.
This was obtained using POLS with robust standard errors and CRE that controls for unobserved
heterogeneity.

4.4.1 Female decision makers

Taking into account the assumption that all covariates are uncorrelated with unobserved
heterogeneity, Ci in equation 5 using pooled ordinary least square (POLS) estimations. The study
finds that gender if female is the decision maker is negative and significant at the 1% level. The
indication is that, households with females making decisions to produce crops are likely to have
less food availability in months by 29%. With voluminous constraints that women face in terms
of opportunities, inter alia, in accessing resources, land, education and markets, though gloomy,

30
these results are not surprising. Information from RALS data by CSO/MoA/IAPRI 2015 show that
most female decision makers (57%) had land below 2ha on top of having less education and more
years than male decision makers. These results are in agreement with the findings by several
studies on the gender challenges faced by women (MGCD, 2015; Mason et al. (2013); World
Bank, 2014; The World Economic Forum (2015); FAO, 2011)

4.4.2 Participation in FISP

Similarly, bearing in mind the assumption that covariates are uncorrelated with unobserved
heterogeneity in POLS, participation in FISP was positive at the 1% level indicating that
participating in FISP is likely to increase food in months by 20.9%. Participation in FISP provides
farmers with hybrid seeds and fertilizer which is likely to improve crop production (MAL, 2015).
Therefore, participation in FISP can result into more months of adequate food at household level.
These results are consistent with findings by several authors such as Chibwana et al. (2010);
Shively, (2013); Sianjase and Seshamani, (2013); Ricker-Gilbert, Jayne and Chirwa (2011); and
Sibande et al. (2015).

4.4.3 The Interaction of FISP participation and female decision makers

Interestingly, when female decision makers participate in FISP, there is a positive and high
likelihood of an increase in MAHFP. Estimates on the interaction of participation in FISP and the
female decision maker show this positive impact at 1% level indicating an increase of food in
months by 23.7 percent without controlling for unobserved heterogeneity between the covariates
Ci. Likewise, when unobserved heterogeneity is controlled for, the impact of the interaction is
positive at 1 level indicating 23.6 percent increase of MAHFP. The positive impact of the
interaction term on food availability entails that, if women have access to resources for production,
they too can be productive and contribute in reducing food insecurity. Several studies indicate the
ability of women to be as productive given resources (Smith et al., 2003; FAO, 2011; World Bank,
2011; Njuki et al., 2016).

31
4.4.4 Social-demographic characteristics

Results of POLS show a negative impact of age of the household head on food availability at 1%
level indicating a reduction of food in month by 0.8%. Similarly, using CRE, age of the household
head is significant and negative at 1% level signifying a decrease in months of food availability by
1.8 %. Probably, this could be that household heads attach less importance to producing more
crops as they advance in age. Another reason could be that with old age comes diminishing returns
which decreases the level of crop production for older household heads. However, POLS results
show that land category of between 2 and 5 ha was significant at 1% whereas land category above
5 ha was significant at 5% level with both categories having positive impact on the food
availability.

4.4.5 Distance to market

Hours to the nearest place with 500,000 inhabitants was found to have a negative effect on the
MAHFP at 1% level using POLS indicating a 1.7% decrease in food availability. Equally, CRE
results show the same at 1% level. When a place is located far from a market, there is a less chance
of individuals accessing the market. Even if they did, there would be a cost attached.

4.4.6 Agro-ecological zones

Results indicate that agro-ecological zones have a significant effect on the MAHFP at 1% level.
Specifically, Zones IIa and III which receives good amounts of rainfall show a high likelihood of
increasing MAHFP. Crop production depends on good amounts of rainfall. However, these results
also show that high rainfall alone is not enough. Zone II B indicates a negative impact on MAHFP
despite receiving the same amount of rainfall with zone II A. Zone IIB has bad soil type which
makes production of crops problematic.

32
Table 7: Impact results using POLS and CRE
Variables POLS Standard error CRE Standard error
(Coefficients) (Coefficients)
Female GDM -0.290*** 0.059 -0.135 0. 092
If FISP is 1 0.209*** 0.053 -0.089 0 .098
FISP participants 0.237*** 0.901 0.236*** 0.091
and GDM of female
Household size 0.008 0.008 0.018 0.024
Education of 0.0002 0.002 0.005 0.016
household head
(years)
Age of Head (years) -0.008*** 0.002 -0.018*** 0.007
Land category 0.228*** 0.047 0.054 0.066
(2-5 ha)
Land category 0.212** 0.025 -0.131 0.131
(>5 ha)
Distance to urban -0.017*** 0.003 -0.017*** 0.003
place (hrs)
Zone II A 0.748*** 0.089 0.736*** 0.088
Zone II B -0.415*** 0.124 -0.402*** 0.124
Zone III 0.352*** 0.092 0.349*** 0.092
Constant 8.42*** 0.130 8.20*** 0.156
Observation 5,870 5,870
Note: Significance levels *** at 1%, ** at 5%, * at 10%.

Source: Author’s Econometrics Analysis using Stata software and data obtained from
CSO/MoA/IAPRI 2012/2015.

33
 CHAPTER FIVE
CONCLUSIONS AND RECOMMENDATIONS
5.1 Conclusions
In conclusion, the findings of this study suggest that food insecurity, in rural Zambia is eminent.
On average, most households do not have adequate food in at least a quarter of a year. Worse still,
more households with female primary decision makers have lower MAHFP compared to those
with male primary decision makers. The findings show that landholding size is larger for
households with male primary decision making compared to those with female primary decision
makers. Similarly, there is higher participation in FISP for households with male decision makers
compared to those with female primary decision makers. . The findings further reveal that food
insecurity exists even in households participating in FISP.

Additionally, the findings have showed that participation in FISP is not determined by gender,
rather it is affected by factors such as education levels, number of fields cultivated, tropical
livestock units, agro-ecological zones and distance to markets. Unsurprisingly, in rural Zambia,
women have lesser ownership of resources than their male counterparts which reduces their
participation in empowerment programmes such as FISP.

The study also finds that food availability increases by 20.9% when households participate in FISP.
However, households participating in FISP and have female primary decision makers can increase
food availability by at least 23.6 percent.

Furthermore, the study finds significant variations across the agro-ecological Zones. While crop
production has less impact on increasing food availability in region IIB producing crops in region
IIA and III can increase food availability.

Lastly, the study revealed that distance to market place in hours has a negative impact on MAHFP.

34
5.2 Recommendations

 There is need to prioritise investments in key drivers of agricultural growth that would
yield significant impact on food household food security. These include extension
services, research and rural infrastructure to enhance access to markets. Furthermore,
investments in research and extension would lead to improved yields and impact MAHFP
positively.
 Agricultural policies should include deliberate strategies to improve resource allocation
for women to enable them participate equitably in programmes like FISP.
 It is more beneficial to target households with female primary decision makers to enhance
household food security in rural Zambia. Even better, it is imperative to empower women
to participate in agriculture decision making.
 There is a need to have specific agricultural strategies for each agro-ecological region,
unlike the current blanket FISP distribution.
 Investment in rural road infrastructure is essential to reduce the distance travelled to the
market places by the rural households. While currently the Government has embarked on
massive road construction, this is mostly in the urban areas while most rural roads remain
impassable

35
 REFERENCES
Angrist , J.D & Krueger , A.B 2001, ‘instrumental Variables and search for identification: From
supply and demand to Natural Experiments’ Journal of economic perspectives, vol.15, no.4, pp.
69-85

Badstue L.B, Mauricio R, Bellon J.B, Alejandro R, Dagoberto F, & Xochit, J. 2007, “The
Dynamics of Farmers‟ Maize Seed Supply Practices in the Central Valleys of Oaxaca, Mexico’
World Development, vol.35, no. 9 pp. 1579–93.

Chapoto, A, Chisanga, B, Kuteya, A. & Kabwe, S. 2015,’ Bumper harvests a curse or blessing for
Zambia: Lessons from the 2014/15 maize marketing season’ IAPRI Working Paper No. 93.
Lusaka, Zambia: Indaba Agricultural Policy Research Institute.
Chibwana, C, Fisher, M, Shively, G, Jumbe, C. & Masters, W. 2013, ‘ Measuring the impacts of
Malawi’s farm input subsidy programme’ African Journal of Agriculture and Resource Economics
vol.9, no.2, pp. 132-147.

Chibwana, C, Fisher, M. & Shively, G. 2012, ‘Cropland Allocation effects of Agricultural Input
Subsidies in Malawi’ World Development, vol.40, no.1, pp. 124-133.

Chibwana, C, Fisher, M., & Shively, G. 2010, ‘Land allocation effects of agricultural input
subsidies in Malawi’ Purdue University, West Lafayette, IN

Deitchler, M, Ballard, T, Swindale, A. & Coates, J. 2010 ‘Validation of a measure of household

hunger for cross cultural use’ Food and Nutrition Technical Assistance (FANTA), USAID
Technical report.
FAO (Food and Agriculture Organization of the United Nations). 2011, ‘the State of Food and
Agriculture 2010–11: Women in Agriculture: Closing the Gender Gap for Development. Rome:
FAO. Viewed 21st july, 2018.http:// www.fao.org/docrep/013/i2050e/i2050e00.htm.
FAO (Food and Agriculture Organization of the United Nations). 2011. The State of Food and
Agriculture 2010–11: Women in Agriculture: Closing the Gender Gap for Development. Rome:
FAO. Viewed 21st july, 2018.http://www.fao.org/docrep/013/i2050e/i2050e00.htm.

Food and Agricultural Organisation (FAO). (2012). Gender and Nutrition, Issue Paper-Draft.
Rome: FAO.

36
Food and Nutrition Technical Assistance (FANTA). 2010, ‘Months of adequate household food
provisioning (MAHFP)” for measurement of household food accessing indicator guide. Viewed
25may,2018https://www.fantaproject.org/sites/default/files/resources/MAHFP_June_2010_ENG
LISH_v4. pdf

Government of the republic of Zambia, Ministry of gender and child development (MGCD) 2015,
gender status report. Viewed 25 June, 2018.

Harris-Fry,H, Azad, K, Kuddus, A, Shaha, S, Nahar,B, Hossen,M, Younes,L, Costello, A &

Forttrel,E. 2015. ‘Socio-economic determinants of household food security and women’s dietary
diversity in rural Bangladesh:a cross section study’ Journal of Healthy, pouplation and Nutrition
vol.33, no.2 https://doi.org/10.1186/s41043-015-0022-0

Hill, C. 2011, ‘enabling rural women’s economic empowerment: Institutions, opportunities and
participation Background paper for Expert Group Meeting, Commission on the Status of Women’
United Nations Entity for Gender Equality and the Empowerment of Women (UN Women), Accra,
Ghana, 20–23 September.

IFPRI. (2016). Healthy diets year round – A case study captured on film in Zambia. Viewed august
20, 2018 at http://a4nh.cgiar.org/2016/02/19/healthy-diets-year-rounda-case-study-captured-on-
film-in-zambia/.

Kamga, R.T, Tchouamo, I.R, Ronal Chendjou, R., Bidogeza, J. C and Afari-Sefa, V. 2016,’
Gender inequality in smallholder onion (allium cepa l.) production in the far north region of
Cameroon. Journal of Gender’ Agriculture and Food Security, vol.1, no.3 pp. 85-103.

Lokshin, M & Sajaia, Z. 2004, ‘Maximum likelihood estimation of endogenous switching

regression models’ Stata Journal, vol.4, no. 3, pp. 282-289,
http://EconPapers.repec.org/RePEc:tsj:stataj:v:4:y:2004:i:3:p:282-289.

Maria, d & Hildago,G. 2016. ‘THE ROLE OF WOMEN IN FOOD’ , FAO, (May), vol.4. no.3,
pp.282-289.

Mason, N. M, Jayne, T. S, & Mofya-Mukuka, R. 2013. ‘Zambia’s input subsidy programs’ ‘Agric.
Econ’ doi: 10.1111/agec.12077.

37
Mason, N. M., Jayne, T. S. & Myers, R. J. 2012, ‘Smallholder behavioral responses to marketing
board activities in a dual channel marketing system: The case of maize in Zambia’ Selected paper
prepared for presentation at the International Association of Agricultural Economics Triennial
Conference, Foz do Iguaçu, Brazil, 18-24 August, 2012. Accessed May 2018, available at
http://ageconsearch.umn.edu/bitstream/126927/2/MasonEtAl.pdf.

Mason, N, & Ricker-Gilbert. J. 2013, ‘Disrupting demand for commercial seed: Input subsidies in
Malawi and Zambia’ Unpublished manuscript.

Mason, N.M, & Tembo, S.T. 2015. Do input subsidy programs raise incomes and reduce poverty
among smallholder farm households? Evidence from Zambia. Working Paper 92. Indaba
Agricultural Policy Research Institute (IAPRI).
Maximilian, Z, Jonathan, M, Manfred, Z. 2018, ‘Effects of Fertilizer Subsidies in Zambia: Have
the Goals been Achieved?’University¨ at Hohenheim 25 January 2018

Mihret, M & Tadesse, A, 2014. Gender, Participation and Decision Making in Farming Activities:
the case Yilman Densa District, Amhara Region, Ethiopia.

Ministry of Agriculture and Livestock. (MALI). Farmer Input Support Programme (FSIP).
Electronic voucher implementation manual 2015/2016 agricultural season.

Ministry of Agriculture and Livestock. (MALI). Farmer Input Support Programme (FSIP).
Electronic voucher implementation manual 2017/2018 agricultural season.

Njuki, J, Parkins, J.R. and Kaler, A. 2016. ‘Gender, agriculture, and food security: where are we?,
Transforming gender and food security in the global south: Routledge studies in food society and
environment, New York.

Omiti, J.M, Otieno, D.J, Nyanamba T, & McCcullough E. 2009. ‘Factors influencing intensity of
market participation by smallholder farmers: A case Study of Rural and peri-urban areas of Kenya.
African journal of Agricultural and Resource Economics. Vol.3, no.1. pp. 57-82

Paula Bilinsky and Anne Swindale June 2010, Months of Adequate Household Food Provisioning
(MAHFP) for Measurement of Household Food Access: Indicator Guide, Food and Nutrition
Technical Assistance III Project (FANTA) FHI 360 1825 Connecticut Avenue, NW Washington,

38
DC 20009-5721 T 202-884-8000 F 202-884-8432 fantamail@fhi360.org www.fantaproject.org

Ravallion, M, 2009. ‘The Developing World’s Bulging (but Vunerable) “Middle Class”’
Developing Research Group, World Bank 1818 H street NW, Washington DC, 20433, USA

Rhoda Mofya-Mukuka & Jairos Sambo. 2018, ‘Household Dietary Diversity Impact of Women
Control over Income from Agriculture in Zambia’IAPRI,working paper 136. April 2018.

Ricker-Gilbert, J, Jayne, T.S. & Shirvely, G. 2013. ‘Addressing the “wicked problem” of input
subsidy programs in Africa’ Applied Economic Perspective and Policy vol.35, no.2. pp.322-340.

Ricker-Gilbert. J. & Jayne, T.S. 2011, ‘What are the enduring effects of fertilizer subsidies on
recipient households?’ Staff Paper 2011-09. Department of Agricultural Food and Resource
Economics, Michigan State University. East Lansing, MI. USA.

Shively, Gerald E. & Ricker-Gilbert, J. 2013, ‘Measuring the Impacts of Agricultural Input
Subsidies in Sub-Saharan Africa: Evidence from Malawi's Farm Input Subsidy Program’ Purdue
Policy Research Institute (PPRI) Policy Briefs: Vol.1. no.1, Article 4.Available at:
http://docs.lib.purdue.edu/gpripb/vol1/iss1/4

Sianjase, A & Seshamani, V 2013, ‘Impacts of Farmer Inputs Support Program on Beneficiaries
in Gwembe District of Zambia’ Journal of Environmental Issues and Agriculture in Developing,
vol.5, no.1, pp. 41-49.

Sibande,L, Bailey, A.& Davidova, S. 2015, ‘The impact of farm input subsidies on household
welfare in Malawi’ Paper presented at the International Conference of Agricultural Economists.
Milan, Italy.

Smith, L. C, Ramakrishnan, U., Ndiaye, A., Haddad, L. & Martorell. R. 2003, ‘The Importance of
women’s status for child nutrition in developing countries’ IFPRI Research Report 131.
Washington, DC: International Food Policy Research Institute (IFPRI).

Thapelo, JM, & Li, X. 2016. ‘Development evaluation of Lesotho agricultural input subsidy policy
based on rural households’ food security and access to inputs: Evidence from Mohale’s Hoek
District” Huazhong Agricultural University, Department of Agricultural Economics and
Management, No.1, Shizishan Street Hongshan District, Wuhan, Hubei Province, 430070 P.R.

39
China.

The Food Security Learning Framework. 2013, ‘The M&E Harmonization Group of Food Security
Partners’ July 2013.

Wineman, A & Crawford, EW 2017, ‘Climate change and crop production choice in Zambia: A
mathematical programming approach’, NJAS-Wageningen Journal of Life Sciences, vol. 81,
pp.19-31.

Wooldridge, J.M. 2010, Econometric Analysis of Cross Section and Panel Data, MIT Press Books,
the MIT Press, 2nd Edition, Vol.1.

World Bank, 2007. World Development Indicators 2007. World Development Indicators 2007.
Washington, DC; World Bank. http://open knowledge.worldbank.org/handle/10986/8150

World Bank. 2011, World Development report: Promoting gender equality and women’s
empowerment. Washington, DC: World Bank Group.

World Bank. 2014. Voice and Agency: Empowering women and girls for shared prosperity. Wa
gton, DC: The World Bank Group.

World Economic Forum. 2015. The Gender gap report 2015. Cologny, Switzerland: World
Economic Forum. Available at: https://www.weforum.org/reports/global-gender-gap-report-
2015/.

World Food Summit, 1996. Rome Declaration on World Food Security.

40