Journal of Rural Studies 32 (2013) 126e136

Contents lists available at SciVerse ScienceDirect

Journal of Rural Studies

journal homepage: www.elsevier.com/locate/jrurstud

Rural livelihood change? Household capital, community resources and

livelihood transition
Prem B. Bhandari*
Population Studies Center, University of Michigan, 426 Thompson Street, Ann Arbor, MI 48106-1248, USA

a b s t r a c t
Keywords:
Capital
Farm exit
Livelihood
Nepal
South Asia

Using the sustainable livelihoods approach, this study examines the extent to which household human,
natural and economic capital, socio-cultural background and physical resources contribute to livelihood
change of farm household to non-farm activities in a rural agrarian setting of Nepal. A number of studies
examine the inuence of various macro-level, particularly economic factors on farm exit in developed
countries. However, we know much less about micro-level household and community assets that
contribute to decisions on livelihood transition by farm households in developing countries. I use the
unique longitudinal panel data between 1996 and 2001 collected from 1180 farm households from a
rapidly changing rural agrarian setting of Nepal. The ndings reveal that the availability of household
labor, particularly children, access to cultivated land, and livestock ownership hinder decision to livelihood transition net of other factors known to inuence livelihood change. Moreover, proportion of nonfarm households in the community signicantly and positively inuenced livelihood transition of farm
households. These ndings provide important insights on livelihood transition in a rapidly changing poor
rural agrarian context.
2013 Elsevier Ltd. All rights reserved.

1. Introduction
This study uses the sustainable livelihood approach to examine
the extent to which the access to various capital inuence a
households livelihood transition from farming to non-farm activities also called farm exit in a poor rural agricultural setting of
Nepal. A number of studies explain farm exits in developed countries such as in the United States, Canada, Israel, Germany, Austria
and Finland (Bragg and Dalton, 2004; Foltz, 2004; Glauben et al.,
2006; Goetz and Debertin, 2001; Kimhi and Bollman, 1999;
Pietola et al., 2002; Stiglbauer and Weiss, 2000; Vare and Heshmati,
2004). Many of these studies focused on socioeconomic forces
inuencing farm exit such as government payments, off-farm
employment, land size, types of farm enterprises, land and livestock ownership, and returns from off-farm employment opportunities. Some of them also examined the inuence of demographic
factors such as farm operators age, their marital status, gender,
family size, and number of children (Glauben et al., 2006; Kimhi
and Bollman, 1999; Pietola et al., 2002; Stiglbauer and Weiss,
2000; Vare and Heshmati, 2004). Because a very small proportion
of the population of these countries is engaged in agriculture, these

* Tel.: 1 734 764 6349 (work); fax: 1 734 615 3557.

E-mail address: prembh@umich.edu.
0743-0167/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jrurstud.2013.05.001

studies are motivated by policies designed to retain farms (Bragg

and Dalton, 2004; Foltz, 2004; Goetz and Debertin, 2001).
Studies of farm exit or livelihood transition are almost entirely
absent in poor rural agrarian contexts of developing countries
including Nepal. This study contributes to the existing knowledge
gap by empirically examining the inuence of various livelihood
assets on livelihood transition of farm households to non-farm
activities in a poor rural agrarian context of Nepal. This investigation is important for several reasons. First, about three-quarters of
poor people in developing countries directly or indirectly depend
on subsistence agriculture for their livelihoods. The World Bank
(2008) recognizes that promotion of agriculture is important in
agriculture-based countries particularly those in Sub-Saharan Africa for achieving the Millennium Development Goal (MDG)
through reducing poverty and hunger. However, in agricultural
transforming countries such as those of South and East Asia, the
Middle East and North Africa, the World Bank suggests assisting
farmers to help move out of agriculture in addition to other alternatives such as shifting to high value agriculture and promoting
non-farm activities as important pathways out of poverty.
Second, shift of farm occupation by individuals and households
to non-farm activities referred to as farm exit or livelihood transition is increasing recently in Nepal. The Nepal Labor Force Survey
reported a signicant decline in the proportion of population
currently employed in agriculture from 76 percent in 1998 to 67

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

percent in 2008 (Central Bureau of Statistics, 1999, 2009).1 Moreover, within households, it is often not only one or two individuals
but all members who change from farming to non-farming occupations. For example, in the Western Chitwan Valley the setting for
this study, the 1996 Chitwan Valley Family Study (CVFS) reported
that about 7.5 percent of households left farming between 1996 and
2001 (Bhandari, 2006). While this rate of attrition might not seem
rapid, that it occurred in only ve years makes it signicant.
Moreover, much less is known about various factors that contribute
to livelihood transition by farm households in developing countries
including Nepal.
Third, understanding of this issue is also important in the
Nepalese context because increasing pressure of population in
agriculture has been considered one of the important problems
facing the country (Nepal Agriculture Perspective Plan, 1995; Ashby
and Pachico, 1987). It is believed that increased population pressure
on agricultural land has contributed to low agricultural productivity due to increased marginal land under cultivation (Chitrakar,
1990; Karan and Ishii, 1996). Therefore, lessening the pressure of
population in agriculture by diverting farm based individuals toward non-farm activities such as formal and informal sector jobs,
tourism, and business has been the policy agenda of the Nepalese
government (NPC, 1998; NPC, 2003). In addition, the World Bank
(2008) also recognizes that assisting farmers to help move out of
agriculture is one of the important pathways out of poverty. By
analyzing the unique longitudinal panel data of households at two
points in time, 1996 and 2001, this study affords to fulll the
existing knowledge gap by empirically examining the inuence of
various capital assets on livelihood transition of farm households to
non-farm activities.
2. Study setting
The Western Chitwan Valley situated in the southern plain of
central Nepal is the setting for this study. The Valley is surrounded by the Rapti River and the Royal Chitwan National Park
on the south, the Narayani River on the west and north, and
Barandabar forest on the east. The area covers part of the Bharatpur municipality and 12 Village Development Committees
surrounded by the Narayani River, the Mahendra Rajmarg (the
national Highway), and the Chitwan National Park. Narayanghat
is the largest market center in the District, is the main business
hub.
The household economy is primarily agriculturally based. Before
the 1950s, the area was inhabited primarily by Terai Janajati
(indigenous) groups such as the Tharu, Darai and Kumal. At that
time, the area was heavily infested with malaria-bearing mosquitoes. In 1956, the government initiated the Rapti Valley Development Project (RVDP) with aid from the United States Agency for
International Development (USAID) to initiate a rehabilitation
program in the Valley by eradicating malaria. The government also
provided land to migrants ranging from 4 bighas (1 bigha 0.68 ha
or 1 ha 1.5 bigha) to 100 bighas by clearing the dense forest
(Shrestha, 1990). Currently, the Valley is inhabited mostly by inmigrants, especially from the Hill and the high Hill as well as
other Terai districts including India (Blaikie et al., 2000; Guneratne,
1998). The Valley now is home to diverse ethnic communities that
range from Terai Janajati (indigenous) (e.g. Tharu, Kumal, Darai) to
high caste Hindu (e.g. Brahmin and Chhetri), Hill Janajati

1
The Nepal Labor Force Survey denes currently employed as e if a person did at
least one hours work in the previous seven days or if the person had a job
attachment (Central Bureau of Statistics, 1999, 2009). This is in line with the
standards of the International Labor Organization (ILO).

127

(indigenous) (e.g. Gurung, Magar and Tamang), Dalit (e.g. Kami,

Sunar, Damai and Sarki) and Newar.
Transformation in the Valley has resulted in a proliferation of
government services, businesses, and wage labor jobs in Narayanghat and Chitwan (Shivakoti et al., 1999). Various governmental and non-governmental organizations such as the District
Agricultural Development Section, Agricultural Statistics Substation, Agricultural Inputs Corporation, District Cooperative Section, Cooperative Union, Nepal Food Corporation, Nepal Bank
Limited and Agriculture Development Bank including the Institute
of Agriculture and Animal Science provide various services. A few
large industries such as Bottlers Nepal and many other small scale
industries have been established in and around the Valley. More
recently, many poultry production farms have been established
(Bhandari and Ghimire, 2013; Shrestha and Bhandari, 2000;
Shrestha et al., 1998/1999). Transportation and communication
networks such as roads, radio, television, and telephone facilities
are relatively well developed compared to other parts of the
country. These transformations have generated off-farm employment opportunities in trade, agribusiness, tourism, and industry
(Shivakoti and Pokharel, 1989).
3. The conceptual framework, empirical evidence and
hypothesis
This study uses the sustainable livelihoods approach, a tool
developed to improve understanding of livelihoods, particularly the
livelihoods of the poor people. I use this framework e the access to
capitals and capabilities (e.g. Bebbington, 1999; DFID, 1999;
Scoones, 1999) e in understanding a households livelihood transition from farming to non-farm activities. This sustainable livelihoods approach is relevant for several reasons. First, this approach
recognizes the importance of capabilities, assets and activities
required for a means of living. Second, this approach helps understand the links between individual or household assets and the
activities in which households engage with a given set of assets.
Third, this approach brings together various critical factors that
affect the vulnerability or strength of survival strategies (Allison
and Ellis, 2001; Carney, 2002; Ahmed et al., 2008). Specically, I
examine the inuence of the access to various livelihoods capital e
human capital (e.g. labor availability and skill), natural and nancial
capital (e.g. operational land holding and ownership of land and
livestock), socio-cultural context (e.g. caste/ethnicity), and physical
resources (e.g. the access to non-family community resources) on a
households livelihood transition from farming to non-farm activities in a poor subsistence agrarian setting of Nepal.
3.1. Human capital and livelihood transition (farm exit)
Human capital comprises of amount and quality of labor available, skills, knowledge and health that together enable individuals
or households to pursue different livelihood strategies to achieve
their livelihood objectives (DFID, 1999). In subsistence-based poor
rural agrarian societies, human capital such as the quality and
quantity of available labor is a building block for acquiring livelihood objectives and sustaining livelihood outcomes. A household
uses traditional labor-using inputs, for example, bullocks (and human power) for plowing farm land, farmyard manure for fertilizing
crop elds, and weeding and taking out diseased or insect infested
plants as long as family labor is available to carry out these activities. Their use depends upon the number of working-age family
members e men, women and children e available to the farm
household. The availability of working-age family members which
is directly used in household production determines the size of the
households farm labor force (Food and Agriculture Organization

128

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

[FAO], 1986; Low, 1986). Subsistence farming in developing countries employs a large number of mostly unskilled family labor, and
to a large extent, the labor needed for performing farm activities
comes from within the household (Bhandari et al., 1996e1997;
Bhandari, 2006; Cain, 1977; Chitrakar, 1990; Filmer and Pritchett,
1997; Karan and Ishii, 1996; Kumar and Hotchkiss, 1988;
Loughran and Pritchett, 1997; Mamdani, 1972). In the meantime,
simultaneous employment of all working-age unskilled family
members is difcult in off-farm sector jobs. Thus, family labor
characteristics such as the availability of working age men and
women, elderly, and children may have important implications in
farm continuation or discontinuation.
Age of the household head, another important demographic
variable, is thought to have an important effect on livelihood
transition (farm exit). Because younger individuals are more likely
to change occupations and they prefer non-farm work more than
older individuals (Mahesh, 2002; Moore, 1966; Ogena and De Jong,
1999). However, the existing literature from developed countries
suggests that the age of a farm operator positively contributes to
farm exit (Glauben et al., 2006; Kimhi and Bollman, 1999; Pietola
et al., 2002; Vare and Heshmati, 2004). This relationship holds
true in countries such as the United States, Canada, or Austria,
where retirement plans are available for farmers. But government
retirement plans for farmers do not exist in Nepal. Elderly people
continue farming as long as they can contribute to the farm.
Moreover, for elderly farmers there are few chances of off-farm
employment and change of livelihood may not be a viable option
for them.
Involvement of children in family farming is signicant in poor
rural agrarian settings of developing countries (Cain, 1977;
Chitrakar, 1990; Filmer and Pritchett, 1997; Karan and Ishii, 1996;
Kumar and Hotchkiss, 1988; Loughran and Pritchett, 1997;
Mamdani, 1972). Children provide support to their parents in a
variety of productive activities (both farm and off-farm work), as
well as enabling labor, for example, meal preparation and child care
(Cain, 1977; Kumar and Hotchkiss, 1988; Skouas, 1994). Children
share a major portion of the household work burden in Nepal
(National Planning Commission and United Nations Childrens
Fund/Nepal [UNICEF/Nepal], 1996). Boys under 14 years of age
perform activities such as farming, livestock grazing, and collecting
rewood and fodder. After age 14, they assume the full responsibility of adult males such as plowing, digging and chopping
logs. Similarly, girls, in addition to aforementioned jobs, help in
fetching water, cleaning, washing clothes, and caring for younger
children. Therefore, the number of children in a household means
there is a potential supply of child workers (Groothaert and Kanbur,
1995). Children may take over farm responsibilities from their
parents as successors (Glauben et al., 2006). Moreover, a household
with many children may be relatively risk averse and would not risk
leaving their current occupation unless the expected income from
an off-farm job is signicantly higher compared to their current
farm income. In Austria, Stiglbauer and Weiss (2000) found a strong
and positive effect of the size of family on farm succession and a
negative effect on farm exit. Similarly, Kimhi and Bollman (1999)
also reported that heads of larger families were less likely to exit
farming in Israel.
From a gender perspective, womens involvement in agriculture
is signicant in developing countries (Boserup, 1971; Patnaik and
Debi, 1991). For example, in India, women play a crucial role in
the conservation of land, water, ora and fauna and in overall
agricultural development (Prasad and Singh, 1992). Moreover,
some of the agricultural operations are gender specic (Acharya
and Bennet, 1981; Agarwal, 1992; Boserup, 1971, 1990; Prasad and
Singh, 1992; Rani and Malaviya, 1992; Sachs, 1996; Singh et al.,
1992). While land preparation for crop cultivation, irrigation, and

threshing of grains are predominantly performed by men, transplanting of rice, sowing, manure application, weeding, intercultural
operations, and harvesting are primarily done by women (Agarwal,
1992). Moreover, women spend much longer hours in farming than
their male counterparts (Acharya and Bennet, 1981; Jackson, 1995;
Kumar and Hotchkiss, 1988). This scenario is not an exception to
Nepal. Therefore, the gender composition of family labor may also
affect the decision to livelihood transition. Overall, it is hypothesized that a family unit with a larger number of working-age members
(men, women, and children) and/or with elderly people is expected to
continue farming (H1).
Education, another important dimension of human capital, is
one of the key determinants of structural change in agriculture.
According to Stiglbauer and Weiss (2000), education can have two
opposing effects on livelihood transition of farm households. Education increases access to information, which enhances farmers
ability to process information ultimately helping to increase their
income. The increased income from farming will encourage an
educated farmers likelihood of continuing to farm. Conversely, an
increase in education also increases skill and the opportunity for
employment outside agriculture. As the wages and returns from
agriculture are rather seasonal, and relatively less rewarding than
off-farm jobs, it is expected that educated individuals may be more
likely to leave farming (H2). This later explanation seems plausible
for Nepal because educated individuals have a tendency to leave
traditional agriculture and join the modern off-farm sector.
Although the authors did not nd a signicant effect of education
on farm exit as well as dairy farm exit, respectively in Austria and
the US (Stiglbauer and Weiss, 2000; Bragg and Dalton, 2004), in
India, a survey conducted by National Sample Survey Organization
reported that about one-half of the farmers wanted to quit farming.
It is believed that a job in off-farm sector is expected to have
relatively higher return than farming, and educated individuals
wanted to take up a job in off-farm sector rather than taking up
agriculture as a profession (Watts, 2009).
3.2. Natural and economic capital and livelihood transition (farm
exit)
Natural capital includes all natural resource stocks such as land,
ora and fauna, water, air and environmental services from which
livelihoods are derived. In rural agrarian societies, the access to
farm land and its ownership is crucial for sustaining livelihoods.
These resources are also the economic capital because the access to
land provides employment and income to farmers (De Janvry, 1981;
Findley, 1987). In addition, the ownership of land is an important
criterion in dening ones position in the socioepolitico-economic
class hierarchy (De Janvry, 1981; Findley, 1987; Blaikie et al., 2002;
Sudgen, 2009). The ownership of this most vital resource increases
control over other resources such as income earned from land,
political power, and access to other institutions, for example, banks.
In Nepal, the access to and ownership of land, in general, is associated with caste hierarchy with high caste farmers being the ones
with most access to and ownership of land (Sudgen, 2009). Evidence suggests that an increase in the access to operational land
reduces the tendency to close down farms (Glauben et al., 2006;
Goetz and Debertin, 2001; Kimhi and Bollman, 1999; Pietola et al.,
2002). Large farms provide higher incomes to farmers and therefore, increase farm survival (Kimhi and Bollman, 1999). Similarly,
Goetz and Debertin (2001) and Kimhi and Bollman (1999) found a
lower rate of farm exit among large land owners in the United
States and Canada.
Livestock keeping is another integral activity of farm households
in most rural agrarian settings. Moreover, crop-livestock mixed
farming is commonly adopted by the farmers as an important

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

livelihood strategy. Animals provide food, employment, and income to farm families. Nevertheless, livestock and crop production
are closely interlinked (Gurung, 1987). While animals provide
manure and draft power for crop production, crops provide food for
animals, thus increasing the bond between crop production and
livestock production. In Canada, Kimhi and Bollman (1999) found a
lower tendency to exit if farmers operated mixed farms with crops
and livestock. Moreover, animals are also used as economic capital.
Farm households may sell animals and earn incomes. Therefore, the
access to and ownership of natural/economic capital such as land and
livestock increases a households capability to continue farming and
thus, reduce the likelihood of farm exit (H3).
Moreover, large farmers have better access to economic/nancial capital such as incomes, savings and credit and have capability
to use these resources to strengthen their livelihood. They can
afford to purchase modern farm inputs such as chemical fertilizers,
pesticides and improved farm implements, and use them on their
farms. Although the use of modern farm inputs is sometimes
argued to be scale neutral, their use tends to favor those who
control the means of production other than labor (Lu, 1985;
Shrestha, 1990; Shrestha and Conway, 1985). Specically, technological advances tend to favor larger farms (Glauben et al., 2006; Lu,
1985; Shrestha, 1990). This implies that farmers who cannot use
modern inputs may receive lower prots from farming and can no
longer compete with households that use modern inputs (Buttel
et al., 1990; Shrestha, 1990). Therefore, the non-adopting households may be more likely to exit agriculture (H4).
3.3. Social capital and livelihood transition (farm exit)
Social capital refers to various social resources upon which
people draw in pursuit of their livelihood objectives (DFID, 1999:
Section 3.2). According to DFID, these resources are developed
through (i) networks and connections that increase peoples trust
and ability to work together and expand their access to wider institutions; (ii) membership of formalized groups which is often a
reection of adherence to mutually-agreed or commonly accepted
rules, norms and sanctions; and (iii) relationships of trust, reciprocity and exchanges that facilitate co-operation and may provide
the basis for informal safety nets. However, in the absence of these
direct measures, I consider the socio-cultural context e the casteethnic diversity existing in Nepal and the Valley. The casteethnicity entails important socio-cultural structure of Nepal
(NESAC, 1998; Bhandari et al., 2007; Bennet et al., 2008) and enables or hinders the access to assets and activities through social
relations (Allison and Ellis, 2001).
Nepal is characterized by various caste, sub-caste and ethnic
groups, an admixture of Indo-Aryan and Tibeto-Mongoloid origins.
The 2001 census reported 100 such caste/ethnic groups in Nepal
(Dahal, 2003). The hierarchical caste system is fundamental to the
Hindu religion. Upper caste Hindus (e.g. Brahmin and Chhetri) are
at the top of the social hierarchy and are presumed to be socioculturally, economically and politically advantaged as compared
with other caste/ethnic groups (Dahal, 2003; Levine, 1987; Bennet
et al., 2008). Historically, Dalits (e.g. Kami, Sunar, Damai, Sarki) are
untouchables and positioned at the bottom of the social hierarchy
and are among the disadvantaged groups in many respects. Ethnic
groups (e.g. Gurung, Tamang, Magar, Newar2 also called Janajati)
may also have hierarchy within their system, but the social hierarchy is not as distinct as in other caste groups. Although
discriminatory practices based on caste/ethnicity were legally
abolished in 1962, such practices are pervasive in Nepal.

Newar are also considered Janajati, yet have their own caste system.

129

The issue of socio-cultural disparity/discrimination in every

sector of society and economy has received much attention
recently. People are often discriminated along caste/ethnicity and
gender lines (Pradhan and Shrestha, 2005; Norwegian Refugee
Council/Global IDP Project, 2003; ADB, 2002; NESAC, 1998;
Lawati, 2001; Pandey et al., 2006; Bennet et al., 2008; World Bank/
DFID, 2006). The high caste Hindu, particularly Brahmin and
Chhetri, and Newar caste people are among the historically privileged groups and are considered the elites (Bennet et al., 2008;
World Bank/DFID, 2006). It is believed that these groups, particularly the high caste Hindu, have the most access to various economic and non-economic opportunities (Acharya and Bennet,
1981). The Norwegian Refugee Council/Global IDP Project (2003)
also reported a disproportionate distribution of wealth and power
in favor of higher castes, while lower castes (or Dalits) and minority
ethnic groups (Hill and Tarai ethnic groups) are disproportionately
affected by widespread poverty. These groups are relatively
disadvantaged in terms of educational achievement, income, life
expectancy and overall Human Development Index (NESAC, 1998;
Bennet et al., 2008).
The farming practices also vary among ethnic groups. In general,
the Terai Janajati group (Tharu, Darai, Kumal and Chepang) primarily follow traditional agricultural practices. Brahmin and
Chhetri households keep cattle, buffalo, sheep and goats, whereas
households of other ethnic groups keep these animals as well as
poultry and pigs. Dalit and other ethnic communities are more
likely to have marginal lands and use less advanced agricultural
techniques compared to Brahmin, Chhetri and Newar. The access to
and ownership of land, in general, is also associated with caste
hierarchy with high caste farmers being the ones with most access
to and ownership of land (Sudgen, 2009). For these reasons, it is
believed that livelihood transitions may vary by socio-cultural
context and households of other ethnic groups are expected to be
more likely to exit farming than Brahmin and Chhetri (H5).
3.4. Physical capital and livelihood transition (farm exit)
Physical capital, according to livelihoods approach, includes the
basic infrastructures and producer goods essential for supporting
livelihoods such as transportation, road, market, water supply,
schools, banks, cooperatives, employment centers, community
development programs, communication, health and sanitation and
many other non-family services (DFID, 1999; Bebbington, 1999;
Ahmed et al., 2008). A farm households access to these various
neighborhood services may inuence the decision to farm continuation or discontinuation.
Farmers located closer to urban centers and markets, nearer
transportation arteries, community services, including education
and health facilities, and off-farm employment possibilities have a
number of options open to them that are lacking in many other
communities. Although not directly related to livelihood transition,
Shivakoti et al. (1999) found that community characteristics were
strongly related to land use in this setting of Western Chitwan
Valley. Similarly, Ghimire and Axinn (2010) also found that proximity to various non-family services mentioned above were found
to be strongly related to changing expectations about family life.
They also found that living in agricultural settings inuenced family
formation behavior by increasing the speed of having child births.
These evidences lead us to believe that proximity to such services
may also have important implications on the decision to farm
continuation or exit. Therefore, it is hypothesized that households
nearby non-family services are more likely to exit farming (H6).
Similarly, geographic proximity of a household to the urban center
might inuence a households shift from farming to a non-farming
activity as off-farm employment opportunities are primarily

130

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

available in urban areas. Therefore, I hypothesize that households

close to the urban center are more likely to exit farming (H7). Similarly, the presence of non-farm households in the community may
also affect farm exit decision. Goetz and Debertin (2001) in their
study of the U.S. counties found that off-farm employment
encouraged farm exits in counties where the number of farm
proprietors declined. This implies that the proportion of non-farm
households may inuence farm exit decisions. Secondly, non-farm
households in the community exert pressure for off-farm services,
which may encourage farm households to engage in newly created
off-farm jobs. Hence, I argue that households living in communities
with large proportion of non-farm households are more likely to
exit farming (H8).

the area farthest from it. From each stratum, 10 settlements were
selected based on 1991 population census. Then, each settlement
was divided into small neighborhood clusters with xed
geographic boundary. Each neighborhood clusters consisted of 5e
15 households living inside the boundary. Then, a representative
group of neighborhood clusters were chosen using the systematic
sampling techniques making a total of 171 neighborhood clusters.
From each neighborhood, neighborhood histories of communitylevel changes over time such as bus services, schools, health services, markets, dairy, cooperatives and other community services
were collected. The information was collected from each community using in-depth interviews, key informant surveys, and other
secondary data sources.

4. Methods

4.2. Measures

4.1. Data sources

4.2.1. Measure of livelihood transition or farm exit

The 1996 household consumption and agriculture survey
conrmed the farming or non-farming status of a household by
asking: Does your household do any farming? Similarly, the 2001
household consumption and agriculture survey also conrmed the
farming status of each household that was surveyed in 1996 by
asking the same question. In both surveys, the validity of the
response on farming status was conrmed by examining the actual
size of land the household was cultivating during the survey period,
whether it is equal to or more than 10 dhurs. A households farming
status recorded in 1996 was compared to the farming status of the
same household in 2001. If a household was farming in 1996 and
was not farming in 2001, this change in status was considered as
livelihood transition from farming to non-farming or farm exit.
Thus, farm exit is measured as a dichotomous measure coded 1 if
a household reported a shift from farming to non-farming occupation and 0 otherwise.

Data for this study come from multiple surveys collected by the
Chitwan Valley Family Study (CVFS) from the Western Chitwan
Valley of south central Nepal. Specically, the 1996 household
census data, household consumption and agriculture survey data
1996 and repeated in 2001, and the 1996 neighborhood history data
are used in this study (for details, http://perl.psc.isr.umich.edu/).
A farming household is the unit of analysis. The information
comes from the households that were farming during the 1996
household consumption and agriculture survey. A farming household is dened as a household in which at least one member (not
necessarily the head, the reference person or the main income
earner) is operating a holding as dened by the FAO (1986, p. 144).
Specically, the survey considered a household as farming if the
household reported that it was engaged in the production of any
kind of crop in at least 10 dhurs (0.500 kattha 0.017 ha) of land
during the survey period. The survey also asked the actual size of
land under various crops during the survey year. The validity of the
response on farming status was conrmed by examining the actual
size of land the household was cultivating during the survey period,
whether it is equal to or more than 10 dhurs. The 2001 household
consumption and agriculture survey also dened farming in the
same way as the earlier 1996 survey and the same procedure was
applied to identify a farm household.
The 1996 CVFS household census data and the 1996 household
consumption and agriculture survey data were collected from 1805
households located in 171 neighborhood clusters. However, due to
resource limitations, the 2001 household consumption and agriculture survey data were restricted to 1523 households living in 151
neighborhoods in 1996. Despite the fact, households that were not
included in the 2001 interview did not differ signicantly from the
1996 households. The data were collected using a face-to-face
interview technique that administered a scientically designed
interviewer assisted structured schedule. Information on the entire
asset measures such as human, natural, economic and sociocultural context used in this study come from the 1996 survey.
The household farming status, the measure of livelihood transition
or farm exit, was obtained from the 1996 and 2001 surveys.
Neighborhood or community level information was also
collected in 1996 to construct measures of community characteristics or community physical resources. The neighborhood clusters
were the lowest level sampling units considered for CVFS. In brief,
prior to choosing samples of these neighborhoods, the study area of
the Western Chitwan Valley was rst divided into three different
strata based on the approximate distance from Narayanghat, the
urban center of Chitwan District, to select a representative sample
of neighborhoods (see Barber et al., 1997; for detail). Strata 1
included the area nearest to Narayanghat while strata 3 included

4.2.2. Measures of capital

Among human capital are family labor characteristics such as the
availability of working age males and females, elderly, children and
education. Availability of working age family members is the
number of working-age males and females 15e64 years of age
living in the household at the time of the 1996 survey. Similarly, the
presence of elderly is measured as the number of elderly individuals over 64 years of age. I also examined separately the effects
of the number of children below 6 years of age (non-working age),
and 6e14 years (working age). Since the household survey does not
identify a household head, the age (in years) of the oldest male is
used. If there was no male in the household, the age of the oldest
female was used. Similarly, education of the oldest male member in
a household is used as number of years of schooling. If there was no
male in a household, the education of the oldest female was used.
The access to and ownership of land, ownership of livestock and
farm inputs use are considered under natural and/or economic
capital. In the Chitwan Valley, two types of farm lands are available
e bari and khet. Bari is upland, usually un-irrigated, and generally
not suitable for rice cultivation. Khet is low lying land that can be
irrigated during the monsoon season and is suitable for planting
rice. The 1996 household consumption and agriculture survey obtained information about the ownership of bari and khet land
separately by asking Does your household own the land, is it
sharecropped, is it mortgaged, is it on contract to you, are you the
tenant of the land or are there some other arrangements? Based on
the responses, farm households were categorized as (i) full owners,
(ii) owner plus sharecroppers (part-owners), and (iii) sharecroppers. Similarly, the access to cultivated land is measured as the total
of bari and khet land cultivated by a farm household during the
survey year in the local unit, bigha and kattha (1 ha 1.5 bigha 30

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

kattha). Livestock ownership is measured as a dichotomy whether a

household has animals coded 1 or not coded 0. Use of biochemical (chemical fertilizers and pesticides) and mechanical (a
tractor and other improved farm implements) technologies is
coded as whether a farm household used (1) or not (0).
As commonly used in this setting (e.g. Axinn and Barber, 2001),
caste/ethnicity of each household is grouped as high caste Hindu
(e.g. Brahmin and Chhetri), Dalit or low caste Hindu (e.g. Kami,
Damai, Sunar, Sharki), Newar, Hill indigenous or Janajati (e.g.
Gurung, Magar and Tamang), and Terai indigenous or Janajati (e.g.
Tharu, Darai, and Kumal).
The access to community physical resources such as banks, cooperatives, bus services, schools, health services, and employment
opportunities was measured in time-to-walk in minutes to the
nearest service from the neighborhood. Since most of these services
are likely to be concentrated in one place, an index was constructed
by adding up the dichotomously constructed measures (less than
10 min 1 and more than 10 min 0) that ranged from 0 to 6;
0 implying no services were available within a 10-min-walk, and 6
implying all services in question were available within a 10-minwalk. Moreover, the presence of non-farm households in the
community or neighborhood is measured in percent. The study
area is divided into three different strata based on the approximate
distance from Narayanghat, the largest urban center. Households in
strata 1 are closest to the urban center while the households in
strata 3 are located farthest from Narayanghat.

131

farming (4.83 years) was slightly higher, but not statistically

different from those who continued farming (4.18 years).
The cultivated size of land is signicantly larger for households
that continued farming (26.18 kattha 0.87 ha) than those that left
farming (11.74 kattha 0.40 ha). While only 7.8 percent of the full
owners left farming, 11.4 percent of the part owners left farming.
Interestingly, only 1.6 percent of the share cropper households left
farming. Livestock ownership also varied by farming status. Those
who did not own any animals were more likely (23.4 percent) to
leave farming compared to those who owned any livestock (5.1
percent). Signicantly small proportions of the households that
used either bio-chemical or mechanical inputs in crop production
left farming. Although more proportions of households that
belonged to hill Janajati exit farming compared to other groups, the
bi-variate association between ethnicity and farm exit was not
statistically signicant.
Farm households in communities where a larger proportion of
the households were engaged in non-farm activities had a greater
tendency to exit. In addition, as expected, farm exit rates were
higher in strata 1 and 2 compared to strata 3. While slightly over 4
percent households living in areas farthest from the main market
center of Narayanghat (strata 3) changed their farming occupation,
about 8 percent of those living in the vicinity of main market center
(strata 1) changed their farming occupation to non-farm activities
by 2001.
5.2. Multivariate results

4.3. Data analysis

The analysis proceeds in two steps. First, descriptive data on all
the measures used in this study and collected in 1996 are compared
between households that continued farming and those that left
farming. Their bi-variate associations or differences were examined
using one-way ANOVA or Chi-square test as appropriate. Secondly,
because the outcome measure, farm exit, is a dichotomy, the binomial logistic regression is used as a multivariate tool to examine the
relationships between livelihood transition and the access to
various livelihood capital. Two models e a full model and a reduced
model e are constructed. In the full model, all sets of capital measures examined in this study are simultaneously included. The
reduced model included only those measures that statistically
signicantly contributed to the outcome measure e farm exit.
Collinearity diagnostics and tolerance statistics were also used to
diagnose potential multicollinearity problems (Menard, 1995). No
evidence of multicollinearity was found in the data. The results of
binomial logistic regression are presented as unstandardized logistic regression coefcients and odds ratios. The unstandardized
logistic regression coefcients are interpreted as the increase or
decrease in the logged odds of the dependent variable due to a oneunit change in the independent variable (Pampel, 2000). For
simplicity, the results are interpreted as odds ratios.
5. Results and discussion
5.1. Descriptive results
Table 1 provides comparative descriptive statistics of all the
measures in 1996 by household farming status in 2001. The results
indicate that the average size of household that continued farming
is signicantly larger than those that left farming in terms of the
number of working-age women (1.76 vs. 1.40) and men (1.70 vs.
1.35), number of working-age children (1.43 vs. 0.96), and the age of
the household head (42.24 years vs. 38.14 years). Mean differences
are statistically signicant except for elderly and children below 6
years of age. The level of education of the household head that left

Table 2 provides the results of multivariate analysis. Model 1

(full model) of Table 2 presents odds ratio estimates (in parenthesis) predicting the effects all the measures e human, natural and
economic capital, socio-cultural context and physical resources e
used in the study on livelihood transition or farm exit. Altogether
these measures explain over 26 percent (Pseudo R-square 26.2
percent) of the variation on the outcome measure e farm exit and
the model ts the data (model Chi-square 127.598, p < .000). The
results in Model 2 (reduced-model) provide the results of only
those measures that statistically signicantly contributed to the
outcome measure farm exit in Model 1. Statistically non-signicant
variables were sequentially removed from the analysis, beginning
with the weakest measure removed rst. Only ve measures,
number of working age males, number of working age children,
land size, livestock ownership and percent non-farming households in the community were found to be statistically signicant.
These ve measures explained 22.4 percent of the variation on the
outcome measure perhaps suggesting that these ve measures are
important determinants of livelihood transition in this setting of
Western Chitwan Valley.
5.2.1. Human capital and livelihood transition (farm exit)
Overall, the results in Model 1 (Table 2) suggest that a one
person increase in a household belonging to each category of the
life cycle, except the elderly people, decreases the odds of farm exit.
However, interestingly, only the effect of working-age children (6e
14 years) statistically signicantly contributed to farm exit. This
result indicates that a one person increase in working-age (6e14
years) children reduced the odds of farm exit by 26 percent (odds
ratio 0.744; p < .05) adjusting for all other measures of human,
natural and economic capital, socio-cultural context and physical
resources that are known to inuence farm exit. When the statistically non-signicant variables were sequentially removed from
the analysis, beginning with the weakest measure removed rst, in
the reduced model (Model 2), interestingly, the statistical signicance of the effects of working-age men appeared and the effect of
working-age children became much stronger. This result suggests

132

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

Table 1
Descriptive statistics: demographic, socioeconomic, and neighborhood characteristics by household farming status (N 1180).
Measuresa

Farming status
Left farming by 2001 (n 80)

Continued farming by 2001 (n 1100)

SD

Minemax

Mean or (%)

SD

Minemax

1.40***
1.35**
0.23
0.71
0.96***
38.14**
4.83

0.67
0.78
0.53
0.90
1.18
13.31
5.17

0e4
0e4
0e2
0e4
0e5
15e80
0e16

1.76
1.70
0.24
0.75
1.43
42.24
4.18

0.97
1.01
0.52
0.95
1.24
12.32
4.50

0e10
0e8
0e2
0e5
0e7
15e78
0e16

11.74***

12.59

1e67

26.18

23.62

1e200

7.8*
11.4***
1.6

e
e
e

0e1
0e1
0e1

92.2
88.6
98.4

e
e
e

0e1
0e1
0e1

5.1***
23.4

e
e

94.9
76.6

e
e

5.7***
13.0
5.7**
11.2

e
e

0e1
0e1

94.3
87.0
94.3
88.8

e
e
e

0e1
0e1

5.7
7.7
10.3
2.9
7.3

e
e
e
e
e

0e1
0e1
0e1
0e1
0e1

94.3
92.3
89.8
97.0
93.2

e
e
e
e
e

0e1
0e1
0e1
0e1
0e1

23.09***
2.78***
7.7*
9.4
4.4

24.87
1.74
e
e
e

0e91.67
0e6
0e1
0e1
0e1

10.30
2.19
92.3
90.6
95.6

12.69
1.44
e
e
e

0e89.47
0e6
0e1
0e1
0e1

Mean or (%)
Human capital
Family labor availability by type
Number of working-age females
Number of working-age males
Number of elderly persons (>64 years)
Number of children (<6 years)
Number of children (6e14 years)
Age of household head
Education of household head
Natural and economic capital
Size of cultivated land (kattha)
Land ownership
Full-owners
Part-owners
Sharecroppers
Livestock ownership:
Owned 1
Do not own 0
Technology use in agriculture
Bio-chemical: used any (1)
Did not use (0)
Mechanical: Used any (1)
Did not use (0)
Socio-cultural context e caste/ethnicity
High caste Hindu (Brahmin/Chhetri)
Dalit
Hill Janajati
Newar
Terai Janajati
Physical resources (neighborhood context)
Non-farm households in neighborhood (%)
Number of services within a 10-min walk
Proximity to urban center (Ref strata 1)
Strata 2 (between strata 1 and 3)
Strata 3 (farthest from the urban center)

e
e

e
e

1 ha 1.5 bigha 30 kattha.

a
Data collected in 1996.
b
One-way ANOVA F test (for categorical Chi-square test) ***p < .001; **p < .01; *p < .05.

that the availability of working-age children and men are the two
statistically important human capital measures in the decision to
farm exit.
There could be several reasons behind this result. Rosenzweig
(1977, p. 124) pointed out two important roles of children: as
durable commodities which yield psychic income and productive
laborers. in agricultural households. In rural Nepal, children
provide support to their parents in a variety of productive roles (for
example, farm and off-farm work), as well as enabling labor (for
example, meal preparation and child care) (Kumar and Hotchkiss,
1988). They share a major portion of the household work burden
(National Planning Commission and UNICEF/Nepal, 1996). For
instance, while boys under 14 perform activities such as farming,
livestock grazing, and collecting rewood and fodder, after 14 years
they take on the full responsibility of adult males such as plowing,
digging and chopping logs. Girls, in addition, perform activities
such as fetching water, cleaning, washing clothes, and caring for
younger children.
Another human capital measure, education, however, did not
statistically signicantly contribute to farm exit, although the direction of the effect was as expected (odds ratio 1.030; p > .05,
Model 1). This result corresponds to the ndings of Stiglbauer and
Weiss (2000) among farmers in Austria and among dairy farmers in
the US (Bragg and Dalton, 2004). It could be due to less variation in
the level of education: 44 percent of the household heads were not
educated. Among those who were educated, about 68 percent of

them had less than 10 years of schooling. I also examined the effect
of mean years of schooling of household members as used by Axinn
and Ghimire (2011) but the result was not statistically signicant
(results not shown).
5.2.2. Natural (and economic) capital and livelihood transition
(farm exit)
Among various measures of natural and economic capital
assessed in this study, only the cultivated size of land and livestock
ownership statistically signicantly contributed to farm exit. Net of
all other factors, a one kattha (30 kattha 1 ha) increase in cultivated land signicantly decreased the odds of farm exit by about 4
percent (odds ratio 0.962; p < .01, Model 1), illustrating the
importance of the access to cultivated land on livelihood transition
of these smallholder farmers. Stated differently, a 10 kattha increase
(one-third of a hectare) in cultivated land holding would decrease
the odds of exiting farming by 40 percent. When this effect was
further examined by land ownership categories, this result held
true among full land owners but not for sharecroppers and part
owners (results not shown).
Interestingly, ownership of land did not statistically signicantly
contribute to farm exit. Adjusting for all other measures of capital,
full land owners were not statistically different from sharecroppers
and part-owners in terms of occupation change. Surprisingly, net of
other factors, both part-owners and sharecroppers were found to
be less likely to exit farming compared to full land owners,

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

133

Table 2
Logistic regression models for predicting farm exit by household demographic, socioeconomic and neighborhood characteristics (N 1180).
Measures

Human capital
Family labor availability by type
Number of working-age females
Number of working-age males
Number of children (<6 years)
Number of working-age children (6e14 years)
Number of elderly persons (>64 years)
Age of the household head (years)
Age of the household head squared
Education of household head (years)
Natural and economic capital
Size of cultivated land (kattha)
Land ownership: (Ref Full owners)
Sharecroppers
Owners plus sharecroppers
Any livestock (yes 1)
Technology use in agriculture
Bio-chemical technology: used any (1)
Mechanical technology: used any (1)
Socio-cultural context
Ethnicity: (Ref high caste Hindu e Brahmin/Chhetri)
Dalit
Hill Janajati
Newar
Terai Janajati
Physical resources (neighborhood context)
Percent non-farm households
Number of services within a 10-min walk
Proximity to urban center (Ref strata 1)
Strata 2 (between strata 1 and 3)
Strata 3 (farthest from urban center)
Intercept
Model Chi-square
Degrees of freedom
2LL
Nagelkerke R-square (percent)
Percent correctly classied

Multivariate models
Model 1 e full model

Model 2 e reduced model

0.107 (0.898)
0.263 (0.769)
0.054 (0.948)
0.296 (0.744)*
0.105 (1.111)
0.061 (0.941)
0.001 (1.001)
0.030 (1.030)

e
0.348 (0.706)*
e
0.356 (0.700)**
e
e
e
e

0.038 (0.962)**

0.049 (0.953)***

0.148 (0.862)
0.774 (0.461)
0.720 (0.487)*

e
e
0.735 (0.480)*

0.392 (0.676)
0.334 (0.716)

e
e

0.035 (0.966)
0.141 (1.151)
1.237 (0.290)
0.010 (0.990)

e
e
e
e

0.043 (1.044)***
0.023 (0.977)

0.040 (1.041)***
e

0.600 (1.823)
0.194 (1.214)
0.581 (1.789)
127.598***
22
457.450
26.2
93.4

e
e
e
107.927***
5
477.122
22.4
93.3

Wald Chi-square *** p < .001; ** p < .01; * p < .05; 1 ha 1.5 bigha 30 kattha.
Figures in parenthesis are odds ratios.

although this effect was not statistically signicant. In addition,

when the analysis was disaggregated by land size categories, the
land ownership also did not play a signicant role on farm exit
(results not shown). I further compared sharecroppers alongside
marginal land owners as a single class with larger land owners. A
slightly stronger but still statistically not signicant relationship
was evident. This result suggests that the access to operational land
size was important in the decision to livelihood transition from
farming to non-farming activities in this setting.
Farm households that owned animals such as cattle, buffalo,
sheep and goat were over 51 percent less likely to exit farming
compared to those that did not own these livestock (odds
ratio 0.487; p < .05, Model 1). Households that used modern farm
technologies were assumed to enjoy the benets from their use and
therefore, were not expected to leave farming as frequently as those
who do not use these inputs. As expected, the use of bio-chemical
inputs (such as chemical fertilizers and pesticides) and mechanical
inputs (such as tractors, pumpsets, and farm implements) negatively inuenced livelihood transition net of all other factors. But
both of these effects were not statistically signicant. This could be
because the farming is still subsistence based and is not fully
commercialized.
5.2.3. Socio-cultural context and livelihood transition (farm exit)
Interestingly but unexpectedly, the socio-cultural background
of farm household measured by caste/ethnicity was not

statistically important for livelihood transition. Although small

differences were observed in the rate of farm exit among caste/
ethnicity groups at the bi-variate level, these differences were
not statistically signicant suggesting any evidence of caste/
ethnicity differential in livelihood transition in this setting. One
reason could be that Chitwan is considered quite different from
rest of the country in terms of its egalitarian nature and as a
popular migrant destination. Although not directly related to
livelihood transition, other studies in this setting and Nepal
provided mixed results of differences by caste/ethnicity. For
example, Bhandari et al. (2007) using the 2001 Nepal Demographic Health Survey Data found that Brahmin and Chhetri
(high caste Hindu) were not signicantly different from other
caste/ethnic groups such as Dalit, hill Janajati, Terai Janajati and
Newar in child immunization. Similarly, Axinn and Ghimire
(2011) reported that household-level vegetation consumption e
use of common land for grazing or collection of fodder did not
vary by caste/ethnicity in this setting of Western Chitwan Valley.
On the other hand, other scholars reported a caste/ethnicity
differential in human capital endowment such as fathers and
mothers work and social capital endowment such as migration
of parents (e.g. Massey et al., 2009); migration of individuals (e.g.
Massey et al., 2010) and childbearing (e.g. Ghimire and Axinn,
2010; Ghimire and Hoelter, 2007) to mention a few. Therefore,
further exploration is necessary to understand why farm exit
decision was not related to caste/ethnicity in this setting.

134

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

5.2.4. Physical resources and livelihood transition (farm exit)

The results of the effects of various community resources also
referred to as physical capital on farm exit indicate that the proportion of non-farm households in the community signicantly and
positively contributed to farm exit (odds ratio 1.044; p < .001;
Model 1) net of all other factors. A one percent increase in the
proportion of non-farm households in the community increased
the odds of livelihood transition from farming to non-farming activities by 4 percent. Interestingly, the effect of the number of nonfamily services available in a community, which was statistically
signicant and positive without controlling for the effect of proportion of non-farm households (results not shown), turned out to
be statistically non-signicant when the latter variable was
simultaneously included in the analysis. Similarly, rural or urban
location of a farm household was also not important in farm exit
decision. This could be because households living in communities
with relatively greater proportions of non-farm households demand additional services, which help create off-farm employment
opportunities, thus encouraging other households to leave farming.
6. Summary and conclusions
Farm households in Nepal are changing their livelihood strategies by shifting their farming occupation to non-farm activities also
called farm exit. Using the sustainable livelihood framework, this
study contributes to our understanding of why some households
leave farming and why others continue it by examining the access
to various assets or forms of capital known to inuence livelihood
transition.
The ndings revealed that various dimensions of human, natural, and economic capital and community resources inuence
decisions on livelihood transition in this poor rural agrarian setting
of Nepal. The availability of working-age family labor pool, particularly the presence of working-age males and more importantly the
presence of working-age children are found to be important human
capital that discouraged livelihood transition from farming to nonfarm activities. This nding is plausible because children are widely
used in carrying out farming and other household activities and
share a major portion of the household work burden (Filmer and
Pritchett, 1997; Kumar and Hotchkiss, 1988; Loughran and
Pritchett, 1997; National Planning Commission and UNICEF/Nepal,
1996). Moreover, children may take over farm responsibilities
from their parents as successors (Glauben et al., 2006). From a
policy perspective, this nding suggests that if farmers are
encouraged to invest household resources in human capital
development of children such as schooling, rather than using their
labor on farm and household activities, the effect of the availability
of working-age children on farm exit may be altered. However,
exactly why the presence of working-age children in a household
hinders farm households occupation change to non-farm activities
in this setting remains unclear. Information on childrens time
allocation in farming, schooling, and other activities, and household
income sources, and their distribution in various activities,
including child development (for example, health and schooling)
may be necessary to answer this question.
Access to cultivated land and keeping livestock also reduced the
likelihood of farm exit. Other way round, farm households with
small size of operational land holding and with no or fewer livestock were more likely to change their livelihood strategies to nonfarm occupations. These ndings have important policy relevance
in the Nepalese context where an overwhelming majority of
households are engaged in farming. Moreover, many farm households own and cultivate very small holdings that are marginal or
just enough to sustain livelihoods of their families. For example, in
2001, over one-half of farm households in Nepal reported to have

less than one hectare of land. In addition, the farming system is

crop-livestock mixed. In this situation, development of off-farm
employment opportunities suitable for small land holders as well
as those with no or few animals may encourage these farmers to
move out of farming. While employment of small farm holders in
the off-farm sector may result in higher incomes, their movement
out of farming will help relieve the pressure of population in
agriculture.
Interestingly but unexpectedly, the socio-cultural background of
farm households was not statistically important in the decision to
livelihoods transition. The farm households that belonged to hill
Janajati, Dalit, Newar and Terai Janajati were not statistically
signicantly different from those that belonged to the high caste
Hindu. This nding provides an important insight of no differences
in livelihoods transition based on caste/ethnicity. One of the reasons could be that Chitwan is considered signicantly different
from rest of the country in terms of its egalitarian nature, migrant
population, centrality of geographic location and economic development. Further investigation is necessary to understand the
mechanism why farm exit decision was not related to caste/
ethnicity in this setting.
The evidence indirectly suggests that increased access to physical resources such as non-family services may positively inuence
farm exit, the effect of which was mediated by the proportion of
non-farm households in the community. Development of community resources such as schools, health services, banks, cooperatives and bus service may provide off-farm employment
opportunities to individuals. This likely encourages households to
leave farming occupation thus increasing the proportion of nonfarm households in these communities. These non-farming
households may also demand additional off-farm services in the
community, which may further encourage other farm households
to exit due to an increase in off-farm employment opportunities.
While this is a plausible interpretation of this nding, conrmation
of this explanation requires data on the employment and occupations of those who left farming.
This study is not free from some methodological limitations.
One of the important limitations is the conceptualization of livelihood change e farm exit. This study denes livelihood transition in
a very simple way e whether a household that was farming in 1996
continued or discontinued farming in 2001. In reality, there are
complications. While households may say they are involved in
farming at the time of survey, their livelihood strategies may be
diverse and this study does not explore other livelihood strategies
such as part-off-farm employment (such as business or jobs or
wage labor) or migration of one or more members of the household
adopted by a farm household. Households may continue farming,
but farming may have played a very limited role in their overall
livelihood because sons or daughters may have professional income
or income from migration or from other off-farm sources. Some
farmers may farm land during the monsoon but leave fallow during
dry season when agriculture is more risky. Moreover, this research
also does not provide information on alternative livelihood strategies adopted by households that exited farming. Investigation is
also needed in order to create necessary services and facilities in
that direction. Further, it is also not clear whether the livelihood
outcomes of the households that changed from farming to nonfarming occupations are better or worse in off-farm sectors or
whether these households return to farm sector jobs.
In conclusion, this study provides evidence that households
livelihoods assets inuence their livelihoods transition. These
ndings reveal that the presence of working-age children
encourage continuation of farming as a livelihood strategy. Thus,
from a policy perspective, encouraging farm households to invest in
the human capital development of children may increase farm exits

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

in the long term. This study also provides evidence that the access
to natural and/or economic capital such as the size of cultivated
land and livestock keeping are two important factors that inhibit
livelihood transition from farming to non-farm activities. If Nepal
continues to pursue policies that encourage small farm holders to
leave farming, then appropriate economic policies must focus in
generating alternative employment opportunities in the non-farm
sector. Such policies should only be pursued if non-farm employment increases in sufcient numbers to absorb those who exit
farming. In the absence of employment growth in the non-farm
sector, such policies will likely be self-defeating and tend to undermine the agricultural sector, while further exacerbating urban
unemployment and congestion.

Acknowledgment
This research was supported by a number of grants from the
National Institute of Child Health and Human Development
(NICHD) (Grant #R01-HD032912, Grant #R01-HD033551, and
Grant #R01HD033551-13). I thank William G. Axinn (PI) for
providing access to the data. Sincere thanks are due to Shannon C.
Stokes and Leif Jensen, my mentors at the Pennsylvania State University for their guidance and valuable suggestions in my dissertation research. I thank Dirgha J. Ghimire for his continuous
encouragement. I would also like to thank the staff of the Institute
for Social and Environmental Research-Nepal for their contributions to the research reported here. I offer many thanks to three
anonymous reviewers who provided excellent feedback to improve
the quality of this manuscript. Last but not least, I owe a special
debt of gratitude to the respondents who continuously welcome to
their homes and share their invaluable experiences, opinions,
thoughts and have devoted countless hours responding to our
survey questionnaires. All errors and omissions remain the responsibility of the author.

References
Acharya, M., Bennet, L., 1981. The Status of Women in Nepal: the Rural Women of
Nepal. Tribhuvan University, Centre for Economic Development and Administration, Kathmandu, Nepal.
Agarwal, Bina, 1992. Impact of HYV rice technology on rural women. In: Punia, R.K.
(Ed.), 1992. Women in Agriculture, Their Status and Role, vol. 1. Northern Book
Center, New Delhi, pp. 233e260.
Ahmed, Nesar, Allison, E.H., Muir, J.F., 2008. Using the sustainable livelihoods
framework to identify constraints and opportunities to the development of
freshwater prawn farming in Southwest Bangladesh. Journal of the World
Aquaculture Society 39 (5), 598e611.
Allison, E.H., Ellis, F., 2001. The livelihoods approach and management of smallscale sheries. Marine Policy 25, 377e388.
Ashby, J.A., Pachico, D., 1987. Agricultural ecology of the mid-hills of Nepal. In:
Turner II, B.L., Brush, S.B. (Eds.), Comparative Farming Systems. The Guilford
Press, New York, pp. 195e222.
Asian Development Bank, 2002. Poverty Reduction in Nepal: Issues, Findings, and
Approaches. http://www.adb.org/Documents/Reports/Poverty_Reduction_NEP/
poverty_analysis.pdf (retrieved on September 20, 2005).
Axinn, William G., Barber, Jennifer S., 2001. Mass education and fertility transition.
American Sociological Review 66, 481e505.
Axinn, William G., Ghimire, Dirgha J., 2011. Social organization, population, and
land use. American Journal of Sociology 117 (1), 209e258.
Barber, J.S., Shivakoti, G., Axinn, W.G., Gajurel, K., 1997. Sampling strategies for rural
settings: a detailed example from the Chitwan Valley family study, Nepal. Nepal
Population Journal 6, 193e203.
Bebbington, A., 1999. Capitals and capabilities: a framework for analyzing
peasant viability, rural livelihoods and poverty. World Development 27 (12),
2021e2044.
Bennet, L., Dahal, D.R., Govindasamy, Pav, 2008. Caste, Ethnic and Regional Identity
in Nepal: Further Analysis of the 2006 Demographic and Health Surveys. Macro
International Inc., Calverton, Maryland, USA.
Bhandari, P.B., 2006. Technology Use in Agriculture and Occupational Mobility of
Farm Households in Nepal: Demographic and Socioeconomic Correlates (Unpublished doctoral dissertation). The Pennsylvania State University, Department of Agricultural Economics and Rural Sociology, University Park.

135

Bhandari, Prem, Ghimire, Dirgha, 2013. Rural agricultural change and fertility
transition in Nepal. Rural Sociology. http://dx.doi.org/10.1111/ruso.12007.
NIHMSID: NIHMS431748.
Bhandari, P., Chaudhary, P., Shrestha, S., Tiwari, P., 1996e1997. The Naike System:
hired labor mobilization in rice production in western Chitwan. Journal of the
Institute of Agriculture and Animal Science 17e18, 65e70.
Bhandari, Prem, Shrestha, Sundar S., Ghimire, Dirgha J., 2007. Sociocultural and
geographical disparities in child immunization in Nepal. Asia-Pacic Population
Journal 22 (2), 43e64.
Blaikie, P.M., Cameron, J., Seddon, D., 2000. The Struggle for Basic Needs in Nepal.
Adroit, Delhi, India.
Blaikie, P.M., Cameron, J., Seddon, D., 2002. Understanding 20 years of change in
west-central Nepal: continuity and change in lives and ideas. World Development 30 (7), 1255e1270.
Boserup, E., 1971. Womens Role in Economic Development. George Allen and
Unwin, London, England.
Boserup, E., 1990. Economic and Demographic Relationships in Development: Essays Selected and Introduced by T. Paul Schultz. The Johns Hopkins University
Press, Baltimore, London.
Bragg, L.A., Dalton, T.J., 2004. Factors affecting the decision to exit dairy farming: a
two-stage regression analysis. Journal of Dairy Science 87, 3092e3098.
Buttel, F.H., Larson, O.F., Gillespie Jr., G.W., 1990. Sociology of Agriculture. Greenwood Press, New York, NY.
Cain, Mead T., 1977. The economic activities of children in a village of Bangladesh.
Population and Development Review 3, 201e227.
Carney, D., 2002. Sustainable Livelihoods Approaches: Progress and Possibilities for
Changes. Department for International Development, London, UK.
Central Bureau of Statistics, 1999. Nepal Labour Force Survey 1998/99 Statistical
Report. National Planning Commission Secretariat, Kathmandu, Nepal.
Central Bureau of Statistics, 2009. Report on the Nepal Labor Force Survey 2008.
National Planning Commission Secretariat, Central Bureau of Statistics, Kathmandu, Nepal.
Chitrakar, P.L., 1990. Planning, Agriculture and Farmers: Strategies for Nepal. Ganesh
Devi Chitrakar, Kathmandu, Nepal.
Dahal, D.R., 2003. Social Composition of the Population of Caste/ethnicity and
Religion. In: Population Monograph of Nepal 2003. Central Bureau of Statistics,
Kathmandu, Nepal.
De Janvry, A., 1981. The Agrarian Question and Reformism in Latin America. Johns
Hopkins University Press, Baltimore, MD.
DFID, 1999. Sustainable Livelihoods Guidance Sheets. Department for International
Development, London, UK.
Filmer, D., Pritchett, L., 1997. Environmental Degradation and the Demand for
Children: Searching for the Vicious Circle (World Bank Policy Research Paper
No. 1623). The World Bank, Washington, DC.
Findley, S.E., 1987. Rural Development and Migration: a Study of Family Choices in
the Philippines. West View Press, Boulder, CO.
Foltz, J.D., 2004. Entry, exit, and farm size: assessing an experiment in dairy price
policy. American Journal of Agricultural Economics 86, 594e604.
Food and Agriculture Organization [FAO], 1986. Population and the Labor Force in
Rural Economies (FAO Economic and Social Development Paper No. 59). Food
and Agriculture Organization of the United Nations, Rome, Italy.
Ghimire, Dirgha J., Axinn, William G., 2010. Community context, land use, and rst
birth. Rural Sociology 75 (3), 478e513.
Ghimire, Dirgha J., Hoelter, Lynette F., 2007. Land use and rst birth timing in an
agricultural setting. Population and Environment 28 (6), 289e320.
Glauben, T., Tietje, H., Weiss, C., 2006. Agriculture on the move: exploring
regional differences in farm exit rates. Jahrbuch fur Regionalwissenschaft 26,
103e118.
Goetz, S.J., Debertin, D.L., 2001. Why farmers quit: a county-level analysis. American
Journal of Agricultural Economics 83, 1010e1023.
Groothaert, C., Kanbur, R., 1995. Child labour: an economic perspective. International Labour Review 134, 187e203.
Guneratne, A., 1998. Modernization, the state, and the construction of Tharu
identity in Nepal. The Journal of Asian Studies 57, 749e773.
Gurung, O.P., 1987. Inter-relationship Among Pasture, Animal Husbandry and
Agriculture: Case Study of Tara. In: National Resource Management Paper Series
No. 2. Winrock, Kathmandu, Nepal.
Jackson, C., 1995. Environmental reproduction and gender in the third world. In:
Morse, S., Stocking, M. (Eds.), People and Environment. UCL Press, London,
England.
Karan, Pradyumna, Ishii, Hiroshi, 1996. Nepal: a Himalayan Kingdom in Transition.
United Nations University Press, Tokyo, New York, Paris.
Kimhi, A., Bollman, R., 1999. Family farm dynamics in Canada and Israel: the case of
farm exits. Agricultural Economics 21, 69e79.
Kumar, S.K., Hotchkiss, D., 1988. Consequences of Deforestation for Womens Time
Allocation, Agricultural Production, and Nutrition in Hill Areas of Nepal
(Research Report No. 69). International Food Policy Research Institute, Washington, DC.
Lawati, M.L., 2001. Racial Discrimination toward the Indigenous People in Nepal. In:
A Non-government Report for the Third World Conference against Racism
(WCAR). Paper Presented at the National Conference on the NPC in Kathmandu,
April 26, 2001. http://www.mtnforum.org/resources/library/lawom01a.htm.
Retrived on September 11, 2005.
Levine, N.E., 1987. Caste, state, and ethnic boundaries in Nepal. The Journal of Asian
Studies 46 (1), 71e88.

136

P.B. Bhandari / Journal of Rural Studies 32 (2013) 126e136

Loughran, D., Pritchett, L., 1997. Environmental Scarcity, Resource Collection, and the
Demand for Children in Nepal (Research Report). World Bank, Washington, DC.
Low, A., 1986. Agricultural Development in Southern Africa: Farm-Household Economics and the Food Crisis. James Currey, London, England.
Lu, Yao-Chi, 1985. Impacts of technology and structural change on agricultural
economy, rural communities, and the environment. American Journal of Agricultural Economics 67, 1158e1163.
Mahesh, R., 2002. Labour Mobility in Rural Areas: a Village-level Study (Discussion
Paper No. 48). Center for Development Studies, Kerala Research Programme on
Local Level Development, Thiruvananthapuram, India.
Mamdani, Mahmood, 1972. The Myth of Population Control: Family, Caste, and Class
in an Indian Village. Monthly Review Press, New York and London.
Massey, Douglas S., Williams, Nathalie, Axinn, G. William, Ghimire, Dirgha, 2009.
Community services and out-migration. International Migration. http://
dx.doi.org/10.1111/j.1468-2435.2009.00581.x.
Massey, Douglas S., Axinn, William G., Ghimire, Dirgha J., 2010. Environmental
change and out-migration: evidence from Nepal. Population and Environment
32 (2e3), 109e136.
Menard, S., 1995. Applied Logistic Regression Analysis. Sage, Newbury Park, CA.
Moore, W.E., 1966. Changes in occupational structures. In: Smelser, N.J., Lipset, S.M.
(Eds.), Social Structure and Mobility in Economic Development. Aldine, Chicago, IL.
National Planning Commission, 1998. The Ninth Plan 1997e2002. His Majestys
Government of Nepal, National Planning Commission Secretariat, Kathmandu,
Nepal.
National Planning Commission, 2003. The Tenth Plan: Poverty Reduction Strategy
Paper 2002e2007. His Majestys Government of Nepal, National Planning
Commission, Kathmandu, Nepal.
National Planning Commission and United Nations Childrens Fund/Nepal [UNICEF/
Nepal], 1996. Children and Women in Nepal: a Situation Analysis. National
Planning Commission, Kathmandu, Nepal.
Nepal Agriculture Perspective Plan, 1995. Nepal Agriculture Perspective Plan.
Agricultural Projects Services Center, Kathmandu, Nepal.
NESAC, 1998. Nepal Human Development Report 1998. Nepal South Asia Centre,
Kathmandu, Nepal.
Norwegian Refugee Council/Global IDP Project, 2003. Prole of Internal Displacement: Nepal. Compilation of the Information Available in the Global IDP
Database of the Norwegian Refugee Council. Geneva, Switzerland.
Ogena, N.B., De Jong, G.F., 1999. Internal migration and occupational mobility in
Thailand. Asian and Pacic Migration Journal 8, 419e446.
Pampel, F.C., 2000. Logistic regression: a primer. Sage, Newbury Park, CA.
Pandey, Tulsi Ram, Mishra, Surendra, Chemjong, Dambar, Pokhrel, Sanjeev,
Rawal, Nabin, 2006. Forms and Patterns of Social Discrimination in Nepal e a
Report. UNESCO, Kathmandu, Nepal.
Patnaik, S.C., Debi, S., 1991. Womens economic contribution. In: Punia, R.K. (Ed.),
Women in Agriculture: Their Status and Role. Northern Book Center, New Delhi,
India.
Pietola, K., Vare, M., Lansink, A.O., 2002. Farmers Exit Decisions and Early Retirement Programs in Finland. Paper presented at the 10th EAAE Congress
Exploring Diversity in the European Agri-Food System. Zaragoza, Spain.
Pradhan, R., Shrestha, A., 2005. Ethnic and Caste Diversity: Implications for
Development. Working Paper Series No. 4. Nepal Resident Mission, Asian
Development Bank.

Prasad, C., Singh, R.P., 1992. Farm women: a precious resource. In: Punia, R.K. (Ed.),
Women in Agriculture: Education, Training and Development. Northern Book
Center, New Delhi, India.
Rani, Seema, Malaviya, A., 1992. Farm mechanization and women in paddy cultivation. In: Punia, R.K. (Ed.), Women in Agriculture: Education, Training and
Development. Northern Book Center, New Delhi, India.
Rosenzweig, M.R., 1977. The demand for children in farm households. Journal of
Political Economy 85, 123e146.
Sachs, Carolyn, 1996. Gendered Fields: Rural Women, Agriculture, and Environment.
Westview Press, Inc., Colorado.
Scoones, I., 1999. Sustainable Rural Livelihoods: a Framework for Analysis. IDS
Working Paper 72. Institute of Development Studies, Brighton, UK.
Shivakoti, G.P., Pokharel, B., 1989. Marketing of Major Crops in Chitwan: a Case
Study of Six Village Panchayats. In: Research Report Series No. 8. Winrock,
Kathmandu, Nepal.
Shivakoti, G.P., Axinn, W.G., Bhandari, P., Chhetri, N., 1999. The impact of community
context on land use in an agricultural society. Population and Environment 20,
191e213.
Shrestha, N.R., 1990. Landlessness and Migration in Nepal. Westview Press, Boulder,
CO.
Shrestha, S.S., Bhandari, P., 2000. Analysis of Protability, Price Behaviour, and
Market Integration of Poultry Industry in Chitwan (Research Report). Tribhuvan
University, Institute of Agriculture and Animal Science, Chitwan, Nepal.
Shrestha, N.R., Conway, D., 1985. Issues in population pressure, land resettlement and development. Studies in Comparative International Development
20, 55e82.
Shrestha, S.S., Bhandari, P., Bhattrai, G.R., 1998/1999. Economics of poultry production in Chitwan District of Nepal. Journal of the Institute of Agriculture and
Animal Science 19e20, 125e132.
Singh, Ram Iqwal, Singh, Daulat, Singh, R.K., 1992. Green revolution and rural
women. In: Punia, R.K. (Ed.), 1992. Women in Agriculture: Their Status and Role,
vol. 1. Northern Book Center, New Delhi, pp. 277e284.
Skouas, E., 1994. Market wages, family composition and the time allocation of
children in agricultural households. The Journal of Development Studies 30,
335e360.
Stiglbauer, A.M., Weiss, C.R., 2000. Family and non-family succession in the UpperAustrian farm sector. Cabiers dEconomie et Sociologie Rurales 54, 6e26.
Sudgen, Fraser, 2009. Neo-liberalism, markets and class structures on the
Nepali lowlands: the political economy of agrarian change. Geoforum 40,
634e644.
Vare, M., Heshmati, A., 2004. Perspectives on the Early Retirement Decisions of
Farming Couples (Discussion Paper No. 1342). Institute for the Study of Labor,
Bonn, Germany.
Watts, Himangshu, 2009. 40% of Farmers Will Quit Farming in India: Survey. One
World South Asia. http://southasia.oneworld.net/todaysheadlines/40-offarmers-will-quit-farming-in-india-survey#.UGsKylFmNZA. Retrieved on July
1, 2012.
World Bank, 2008. World Development Report 2008: Agriculture for Development.
World Bank Publications.
World Bank/DFID, 2006. Unequal Citizens: Gender, Caste and Ethnic Exclusion in
Nepal. A Co-publication of The World Bank and the Department for International Development, U.K., Kathmandu, Nepal.