1
CHAPTER I: INTRODUCTION
Explaining disparities in wealth, health and education between rich and the poor and
informing interventions that can bridge this gap is one of the fundamental issues in Social
Work. In all countries of the world, we find some groups who are rich and others who are
poor. These groups generally vary across ethnic and geographic communities. Why do
some groups or communities become poor while others become rich? What determines
the relationship between wealth, health and education? More importantly, how best
should society be organized so that all humans have the capacity to “live well” regardless
of who they are and where they live? These are the primary questions that motivate this
dissertation study. These are important questions to the applied social sciences and to
governments and organizations concerned with promoting human well-being.
Despite the global advancement in technology and economic growth, one group that
has remained poor throughout the world is indigenous peoples. The United Nation’s
recent report on the State of the World’s Indigenous Peoples, 2009 (UNPFII, 2009) warns
that poverty among indigenous peoples throughout the world is pervasive and persistent.
For example, the life expectancy of an indigenous child is 20 years shorter than that of
his or her non-indigenous counterpart in Australia and in Nepal; 13 years shorter in
Guatemala; 11 years shorter in New Zealand; 10 years shorter in Panama; and 6 years
shorter in Mexico. However, little is known about why indigenous peoples continue to be
poor while other groups become rich. Much less is known about what should be done to
bridge this gap.
One emerging theory on why some societies become rich while others remain poor is
the theory of institutional design (North, 1990). According to this theory, societies
2
become poor because their institutions – in particular, the rules of law--constrain the
economic behaviors of the citizens in those societies. According to this view, what
matters are the rules of the game in a society, as defined by prevailing explicit and
implicit laws and their ability to create appropriate incentives for desirable economic
behaviors. By extension, this theory implies that the laws of a society, such as a country’s
constitution, constrains the economic productivities of some groups (such as indigenous
peoples) while providing incentive structures for the others (the elites). This view is
strongly associated with North (1990) and Ostrom (1990), and consistent with works of
Sherraden (1991) and others.
This institutional framework was used as a theoretical guide for an in-depth
investigation of poverty among indigenous peoples in Nepal, where poverty is the norm
for most. In particular, this study examines Nepal’s first constitution of 1964 (Muluki
Ain) to determine the extent to which an institution is a source of socioeconomic disparity
between indigenous and non-indigenous peoples in Nepal.
This dissertation serves as the first empirical study to examine the socioeconomic
disparity between indigenous and non-indigenous peoples in Nepal using a nationally-
representative sample. The dependent variable in this study is asset-poverty, as measured
by wealth index (Rutstein & Johnson, 2004). The individual-level independent variables
are ethnicity/caste and productivity characteristics such as education, health, employment
and occupation. The community-level independent variables are geographic isolation,
development regions, and ecological regions. The institutional variable is the education
law that prohibits use of indigenous languages as a language of instruction in public
3
schools. The control variables are gender, age, marital status, household size and gender
of the head of household.
This study will contribute to two areas of social science enquiry: global poverty
discourses and institutional theory. The findings of this study will have important
implications on poverty reduction strategies and on institutions which can provide
incentives for self-governance of the indigenous communities. In particular, the study will
shed light on whether geo-ethnically targeted approaches are needed to reduce disparity
between indigenous and non-indigenous peoples in Nepal.
A. Statement of the problem
Poverty is a serious social problem in Nepal. Over 40% of the population of Nepal
lives under poverty (ILO, 2000). However, little is known about who these people are and
why they are poor. To date, no empirical studies have been conducted to determine
whether indigenous peoples in Nepal are at significantly higher risk of poverty than
nonindigenous peoples.
Since the 1960s, it is becoming increasingly clear that being an indigenous or ethnic
minority significantly increases an individual’s risk of poverty (Psacharopoulos &
Patrinos, 1994; Plant, 1998; Carino, 2009, Eversole, 2005). In the words of
Psacharopoulos & Patrinos (1994), there is a cost to “being indigenous.” However, these
studies have been conducted mostly in the industrialized countries (US, Australia,
Canada, and New Zealand) or in the Latin Americas where non-indigenous peoples are
White-Europeans. The extent to which socio-economic disparity exists between
indigenous and non-indigenous peoples in other developing countries (where
4
nonindigenous peoples are non-White Europeans), and the factors that contribute to such
disparity is currently unknown.
There is a reason to believe that indigenous peoples in Nepal may be more vulnerable
to poverty than non-indigenous peoples. Since the establishment of Nepal as a nationstate
in the 1770s, the settlers (the caste group) have dominated the political and economic life
of Nepal, including those of the indigenous peoples (Aadibasi Janajati). The first
constitution of Nepal (Muluki Ain 1854) brought about great divisions in Nepali society.
This constitution served as a basis for exclusion of the indigenous peoples in governance,
politics, and in the economies of the country. In particular, the constitution prohibited use
of indigenous languages as a language of instruction in public schools. While caste
peoples are allowed to study in their own mother-tongue (Khas language), the indigenous
peoples are prohibited from studying in their mother-tongues. School text books are
written only in Khas language. The primary purpose of schooling has been to assimilate
indigenous peoples into the culture of caste peoples. The extent to which the prevailing
institution constrains the indigenous peoples’ ability to accumulate human capital--and its
subsequent effect on their poverty--is currently unknown.
Nepal is a multilingual and multi-ethnic country, with two distinct racial groups of
people: the caste group (Ariyan of Indian origin, the settlers) and the indigenous group
(Mongoloid, known as Adibasi Janajati). The caste group consists of three caste
hierarchies—high-caste, mid-caste, and low-caste. The indigenous group consists of over
60 distinct ethnic or linguistic groups. About 70% of the people in Nepal consider
themselves as indigenous (detail on Nepal is provided in Chapter IV: Context). Research
is needed to disaggregate the national poverty estimate into individual ethnic groups
5
(indigenous peoples) and caste groups (non-indigenous peoples) so that a precise estimate
can be made about the determinants of poverty for each of the groups.
The prevailing public perception in Nepal is that, within a caste group, low-caste
peoples will have lower socio-economic status than high-caste peoples due to
castediscrimination in the Hindu caste system. However, within the indigenous group,
there is no reason to believe why some ethnic groups will have different socioeconomic
status than others (except for Newar which has Hindu caste system). Ethnic groups of
Nepal are culturally diverse but socially non-hierarchical. To the extent that there is a
significant difference in wealth (poverty) between various ethnic groups within
indigenous peoples, it will be important to understand why.
To date, there have been very few studies that explicitly looked at the poverty among
indigenous peoples vis-à-vis non-indigenous populations. Much of this existing research
on indigenous poverty, however, is descriptive, and some of it is inductive (mostly from
anthropology), but there is much less deductive analytical work. Very little of the research
is applied. In addition, none of the existing research was conducted in Nepal, the area of
concern in this dissertation (Lama, 2010).
Of the previous poverty research on indigenous peoples, most have focused either on
a small indigenous group or on a single geographic community. These studies are
scattered here and there; and they often gloss over the disparity that may exist between
ethnicities within the indigenous group. To my best knowledge, no studies have
systematically looked at indigenous poverty using a nationally-representative sample or
using poverty indicators that are reflective of the indigenous peoples’ well-being (Lama,
2010).
6
Previous studies on indigenous poverty in other countries and communities have
found that indigenous peoples typically have a low level of education (schooling) and
many health problems. Most are largely employed, and many work as farmers or
seasonal laborers (Psacharopoulos & Patrinos, 1994; Eversole, 2005; Humapage, 2005).
These individual productivity characteristics (education, health, employment, occupation)
were associated with poverty among indigenous peoples (Psacharopoulos & Patrinos,
1994). Previous studies have also noted that indigenous peoples live largely in isolated
geographic areas, and the poverty map closely coincides with the geographic territories of
indigenous people (Plant, 1998). These studies, however, do not explain why indigenous
peoples have low human capital and poor health status or live in isolated geographic areas
in the first place.
In recent years, there has been emerging evidence that suggests that geography or
‘where you live’ plays a significant role in determining an individual’s access to quality
healthcare (Wennberg, 1970; Raghavan et al, 2010) and quality education (Wilson, 1990;
Garner & Raudenbush, 1991). Geography is the key determinant of climate and of natural
resource endowments, and it can also play a fundamental role in the disease burden
(Rodrik & Subramanian, 2003) and infra-structure development. Geography can
influence agricultural productivity and the quality of human resources (Diamond, 1997;
Sachs, 2001). Since indigenous peoples live in isolated or poor geographic communities
(Plant, 1998), they are less likely to have access to quality education and quality
healthcare services. The lack of access to quality education and quality healthcare
services is likely to result in poor educational and health outcomes. Poor education and
poor health, combined, are likely to put indigenous peoples at high risk of poverty.
7
However, to date no empirical studies have looked at the effect of geography on the
relationship between education, health and poverty. Furthermore, why geographic
communities of indigenous peoples are poor or isolated is currently unknown.
Further research is needed to better understand the geographic and institutional
contexts in which indigenous peoples live and how these contexts influence the risk of
poverty. This dissertation study proposes an in-depth investigation of the socioeconomic
status of the indigenous peoples in Nepal, where poverty is an established part of life for
the indigenous peoples. Each year, many indigenous peoples in Nepal face deaths due to
poverty- induced problems such as malnutrition and tuberculosis.
The theory of institutional design (1990) predicts that societies/communities become
poor due to institutional structure, such as constitution or laws, which are designed by
elites of the society to further their own best interests. This dissertation research used the
theory of institutional design as a theoretical guide to determine the extent to which
indigenous peoples of Nepal are poor due to the institutional structure of Nepal. In
particular, this study examined if the geographic territories of indigenous peoples are
systematically isolated (made poor) by the prevailing institutions of Nepal, and whether
this isolation is driving the observed socioeconomic disparity between indigenous and
non-indigenous peoples. The study also examined if the low level of educational
attainment among indigenous peoples is a result of the prevailing education laws, and
whether the low level of educational attainment is contributing to their risk of poverty.
The findings of this study will shed light on our understanding of why it matters ‘who
you are’ and ‘where you live’ with regard to a person’s capacity to ‘live well’.
8
A1. Research Objectives
The specific objectives of this dissertation study are as follows:
1. To determine which ethnic/caste groups in Nepal are at the highest risk of poverty.
2. To determine the extent to which poverty is driven by individual productivity
characteristics (education, health, employment and occupation).
3. To determine the extent to which poverty is driven by geographic characteristics.
4. To examine the extent to which disparity in education, health and wealth (poverty)
between indigenous and caste groups are driven by prevailing institutions of
Nepal.
A2. Background and significance of the research
Over 370 million peoples across the world consider themselves as indigenous peoples
(UNPFII, 2007). They represent over 5,000 of the estimated 7,000 distinct culture and
language groups in the world and live in more than 90 countries across the globe
(UNPFII, 2007). Despite the vastly varied geographic and cultural contexts in which they
live, they all share one common problem--poverty (Eversole, 2005).
Indigenous peoples throughout the world suffer a disproportionately higher risk of
poverty than non-indigenous peoples (Eversole, 2005). Although indigenous peoples
represent about 5% of the global population, they comprise over 30% of the world’s 900
million extremely poor and 15% of all poor worldwide (State of the World’s Indigenous
9
Peoples, 2009). Over 72% of indigenous peoples are extremely poor (make less than $1 a
day) and almost 99% of indigenous peoples are classified as poor (make less than $2 a
day). Global poverty, therefore, is largely a de facto poverty of the indigenous peoples.
It is an understatement to say that poverty is a serious social problem. Poverty has
serious consequences on human health and well-being. Poverty has been shown to cause
general health problems (Pytell, 2007), mental health problems (BMA, 2006), conflicts
(Justino, 2008), crime (Hsieh & Pugh, 1993), suicide (Chuanc & Huang, 1997), and poor
educational outcomes (Brooks-Gunn et al., 2000). However, little is understood about
what causes poverty itself or how to overcome it. Much less is known about the
determinants of poverty among indigenous peoples and the interventions that can help
overcome it.
Global poverty literature has largely ignored the indigenous peoples. Much of the
global poverty literature is focused on the economic structures of the society; but it gives
little attention to this question: Who are the poor people in a society? Indigenous peoples
are rarely the subject of academic discussion in global poverty discourses.
Poverty has traditionally been the primary subject of economists. However, economic
studies conceptualize poverty purely as an economic problem and give little attention to
the local institutions and social contexts that shape the socioeconomic behavior of the
local people. Economic studies, in general, assume homogeneity across all demographic
and social groups. A typical economic analysis of poverty often focuses on the efficiency
of the economic system of a country rather than on the question of whether such a system
is efficiently endangering the vulnerable population into further risk of poverty.
10
Even in the poorest countries of the world, not all people are equally poor-- some
individuals or groups (the elites) in poor countries are, in fact, as well-off as those in rich
countries. Rich individuals or groups (the elites) in these societies may, sometimes, be the
reasons for poverty of the commons due to exploitative labor relationships or other
factors such as a caste system or slavery. Unfortunately, such group differences and local
contexts have rarely been the focus of economic analysis of global poverty –largely
because economists do not deal with the question of social justice. In a typical economic
analysis, poverty is often romanticized, but poor people are largely ignored, or even
dehumanized (for examples: Collier, 2007; Sachs, 2005; Easterly, 2006). Such analyses
see poverty as something that needs to be “attacked” or to be “fought a war against.” Poor
people are often portrayed in a negative light (e.g. Collier, 2007) and frequently treated as
less than human. Such misguided analyses do not capture the nature of society that often
hosts the determinants of poverty. Research is needed to understand the social
determinants of poverty among indigenous peoples and to inform intervention that can
alleviate it.
In recent years, there has been increasing number of interventions purported to be
addressing poverty in developing countries. These works are often spearheaded by
nonprofits or NGOs (Non-governmental Organizations) which market their products to
the poor under the banner of micro-credit, micro-finance, or micro-enterprises. These
approaches are conducted on a trial-and-error basis without any theoretical foundations
that are empirically valid. To date, no countries have seen substantial economic growth
and development or poverty reduction as a result of these approaches (Morduch, 1998),
although these approaches have been implemented for the last 40 years since the work of
11
Muhammad Yunus in the early 1970s. For example, Bangladesh, where micro-credit has
become established as a model of economic development, remains one of the poorest
countries in the world, where the poverty rate is 49.85% (UNDP, 2008).
There are reasons to believe that poverty is rooted in the institutional structure of the
society, rather than based purely on the economic behavior of the poor or economic
structure of a society (North, 1990). Poverty can no longer be analyzed in isolation of the
institutional context in which the poor people live, at least in the case of indigenous
peoples. Multidimensional approaches are needed to study poverty (North, 1990). This
study was undertaken to investigate the endogenous relationship between poverty, health
and education in the context of geographic communities and institutions in which the
indigenous peoples live. The findings will shed light on the knowledge gap in our
understanding of the nature and determinants of poverty among indigenous peoples.
This research will contribute to two areas of social science inquiry: global poverty
discourses and institutional theory. This research is the first study to examine the
socioeconomic disparity between indigenous and non-indigenous peoples in Nepal using
a nationally-representative sample. Understanding how institutional and community
contexts influence the relationship between poverty, health and education will contribute
to efforts to understand the determinants of poverty and will inform interventions that
improve the living conditions of those who experience poverty, poor health and poor
education.
12
A3. Defining poverty or “Living well”
Poverty has been defined in various ways; and considerable disagreement exists
among the scholars over its definition and measures (Psacharopoulos & Patrinos, 1994).
The conventional income or consumption-based definition and measures of poverty have
been critiqued as being limited and narrowly focused (Sherraden, 1991; Sen, 1999;
Iceland, 2005; Blank, 2008; Rutstein & Johnson, 2004). Furthermore, critics argue that
the non-indigenous concept of poverty is misleading and reflects the hedonistic consumer
culture of the market-economy rather than the true well-being of the people (Carino,
2009).
There is a growing consensus among scholars that any measures of indigenous
peoples’ social and economic status must necessarily start from indigenous peoples’ own
definitions and indicators of poverty (Eversole, 2005; Carino, 2009). These scholars argue
that the definition of poverty should be comprehensive and should encompass not only
economic but also health and social dimensions. However, to date, no formal definition of
indigenous poverty exists in the literature.
One approach has been to utilize the indigenous concept of “living well” as an
alternative to poverty (Carino, 2009). According to this conceptualization, poverty may
be thought of as a lack of capacity to live well. This concept is thought to reflect the
values of indigenous peoples, who believe that the purpose of any socioeconomic
development policy or program should be to promote “living well” or “living a good life”
(Eversole, 2005; Carino, 2009). Intrinsic in this definition is the idea that “well-being” is
a multi-dimensional quality of living. In this conceptualization, at least three basic needs
are necessary --wealth, health and knowledge (wisdom/education). These three needs are
13
thought of as interdependent and serve as balancing forces to each other. As an
aggregate, they are a necessary condition for living well. This concept is represented by a
graph in Figure 1.
Figure 1. Conceptual model of well-being
This study utilized Wealth Index developed by Rutstein & Johnson (2004) as a
measure of wealth (poverty). The Wealth Index is a composite of household assets and
services consistent with the indigenous conceptualization of wealth. The Wealth Index
provides a relative measure of wealth (poverty) and has been widely used in other studies
(Rutstein & Johnson, 2004). (See the method section of this document for more details on
Wealth Index). One critique of the Wealth Index is that it is skewed to urban areas.
Well-being
Geo-Society
Wealth
Health
Education
14
Another critique is that poverty based on wealth may produce a different poverty rate
than poverty based on income (Rutstein & Johnson, 2004). Since the primary focus of
this study is on the determinants of wealth, rather than on the measurement methods of
wealth, this document focuses on the relationship between wealth and its predictors. The
construction of the Wealth Index merits a separate chapter and is beyond the scope of this
paper (Please see Rutstein & Johnson, 2004, for details on construction of this index).
15
CHAPTER II: BACKGROUND
B. Defining Indigenous Peoples
According to Webster’s Dictionary, the term “indigenous” is rooted in the Latin
indigenous, meaning “Having originated or occurring naturally in a particular region or
environment.” The term indigenous is synonymous to native, innate or inborn---to the
land. The origin of the concept of “indigenous peoples” as a group is traced back to
colonization when the colonizers or settlers used the concept to differentiate themselves
from the native people who were already living on the land.
Currently, there is no formal universal definition of indigenous peoples. The
general understanding among the scholars of indigenous peoples is that such a universal
definition is neither necessary nor sufficient to describe the scope and complexity of the
diversity that exists within indigenous peoples as a group (Eversole, 2005; Carino, 2009).
The current working definition used by the United Nations Permanent Forum on
Indigenous Issues is that indigenous communities, peoples and nations are as follows:
…those which having a historical continuity with pre-invasion and pre-colonial
societies that developed on their territories, consider themselves distinct from other
sectors of societies now prevailing in those territories, or parts of them. They form at
present non-dominant sectors of society and are determined to preserve, develop,
and transmit to future generations their ancestral territories, and their ethnic identity,
as the basis of their continued existence as peoples, in accordance with their own
cultural patterns, social institutions and legal systems. (UNPFII/2004/WS.1/3, p2).
16
The key to this definition is self-identification:
An indigenous person is one who belongs to these indigenous populations through
self-identification as indigenous (group consciousness) and is recognized and accepted
by these populations as one of its members (acceptance by the group). This definition
preserves for these communities the sovereign right and power to decide who belongs
to them, without external interference (UNPFII/2004/WS.1/3, p2).
Common characteristics of indigenous peoples include being original inhabitants
of a land later colonized by others, and forming distinct, non-dominant sectors of society,
with unique ethnic identities and cultural systems. Indigenous characteristics also include
strong ties to land and territory; experiences or threats from their ancestral territory; the
experience of living under outside, culturally-foreign governance and institutional
structures; and the threat of assimilation into dominant sectors of society and loss of
distinct identity (McNeish & Eversole, 2005).
Indigenous people may include, but are not limited to, Aborigines or First Nation
of Australia, New Zealand, and North America; the hill tribes, ethnic minorities, ethnic
nationalities, original inhabitants, scheduled tribes and other indigenous groups of Asia
and the subcontinent; the indigenous campesinos (peasants) or indios (Indians) of Latin
America; the indigenous peoples of Russia and Scandinavia; and even to some extent the
tribal peoples or ethnic groups of Africa. Each category in turn contains great diversity,
comprising many groups and sub-groups, distinguished by language or lineage or
geographical areas (McNeish & Eversole, 2005, p. 6).
Using the term “indigenous peoples” rather than “indigenous people” recognizes
this diversity. Unlike indigenous populations, the term indigenous peoples recognizes that
17
a shared identity, as a people, exists within each distinct group. Adding an “s” represents
an effort to acknowledge the vast diversity contained within this umbrella term. It is an
effort to avoid the danger of oversimplification, of indicating a stereotypical
“indigenousness.” As noted by McNeish and Eversole, “When we speak of indigenous
peoples, we recognize that we are dealing with no clearly defined group. Rather, we are
placing under a single conceptual umbrella many different peoples” (McNeish &
Eversole, 2005, p.6).
B1. Indigenous peoples in the world
An estimated 40 million indigenous peoples, speaking over four hundred different
languages, live in Latin America and comprise nearly 10 percent of the total Latin
American population (Partridge & Uquillas 1996, cited in Eversole, 2005, p.30). These
people include the descendants of complex civilizations such as the Maya, Aztec, and
Inca, as well as tribes of the forests and lowland plains, peoples such as the Yanomamo,
Xavante, Miskito, and Guarani (Eversole, 2005). The largest indigenous peoples are
found in Bolivia, Peru, Ecudor, Guatemala and Mexico (Gonzalez, 1994).
An estimated 70 percent of the world’s indigenous peoples live in Asia (IFAD
2000/2001). The ‘indigenous peoples’, a category that first came to existence as a
reaction to the legacy of Western European colonialism has become problematic in this
part of the world because many governments refuse to recognize the distinction advanced
by dissident ethnic groups between indigenous and non-indigenous populations (Barnes
et al. 1995 p.2, quoted in Eversole, 2005: 31). As a result, the indigenous peoples of Asia
do not have the same well-defined, long-standing and recognized status as indigenous
18
peoples in recently colonized areas such as the Americas, Australia or New Zealand
(McCaskill and Rutherford, 2005). The indigenous peoples are often defined as prior
rather than original inhabitants (Eversole, 2005). For example, many people of the
Chittgong Hill Tracts in Bangladesh are not the original inhabitants of that region – only
the Kuki peoples can make that claim- but they all pre-date recent efforts by the
Bangladesh army to colonize the area through violent attacks on villages (Eversole,
2005).
Many ethnic groups in Africa pre-date the arrival of European colonizers yet do
not identify themselves as indigenous peoples. Other terms such as “tribes” or “ethnic
groups” are generally preferred (Eversole, 2005). In Africa, the indigenous peoples are
generally pastoralists or hunter-gatherers, such as the Pygmies, Hadzabe, Maasai and
Tuareg, (ILO 1999 p.3).
B2. Indigenous Peoples versus Minorities
Not all indigenous peoples are population minorities. In many countries, such as
Nepal and Bolivia, indigenous peoples are the population majority. Indigenous peoples
are also not necessarily a minority in terms of socioeconomic status. For example, in
Nepal, Newar and Thakali, indigenous groups have achieved their economic status that is
par with the non-indigenous groups. These economic achievements have been made,
however, at the cost of their linguistic and cultural identity (Bhattachan & Webster, 2005).
Except for a few groups, most of the indigenous peoples live in extreme poverty and are
political, socio-cultural and religious minorities in the countries where they live (State of
the World’s Indigenous Peoples, 2009; Eversole, 2005; Pscharopoulos & Patrinos, 1994;
19
etc.). What differentiates indigenous peoples from other minority groups is their historical
significance as the “native” or “original people of the land” as opposed to the settler who
migrated later. Of course, this distinction is a relative one because, historically, every
category of people has migrated from one place to another, perhaps originally from
Africa, including the indigenous peoples. The implied meaning here is prior rather than
original. The point of origin of indigenous peoples as a group is the colonization or the
establishment of current nation-states.
B3. Nature and extent of poverty among indigenous peoples
Indigenous peoples experience poverty at various levels of society. At an
individual level, indigenous peoples experience abject poverty. At a community level,
they experience neighborhood poverty--no roads, piped water, hospitals, communication
technologies, or higher educational organizations in their communities. At a national
level, countries themselves are poor (except the U.S., Australia, and New Zealand). At a
group level, they experience relative poverty and inequality –indigenous peoples are at
higher risk of poverty than their counterparts, both in developed and developing
countries.
Poverty is pervasive among indigenous peoples (Psacharopoulos & Patrinos,
1994, Kelly, 1988; Stephen & Wearne, 1984; del Aguila, 1987). In the United States, the
reservation-based indigenous peoples typically have the lowest income and housing
(Cornell, 2005). The poverty rate among the Native Americans and Alaska Natives is
23.2%, compared to only 12.5% of the general population (US Census Bureau, 2000).
20
The percent of Native Americans who live in crowded households (more than one person
per room) is 18%--three times higher than the percent nationwide. The percent of Native
American and Alaska Native homes that lack safe and adequate water supply and/or waste
disposal facilities is 13 times higher than the homes for the U.S. general population (Indian
Health Service, 2009). About 18 percent of all Native American households live in crowded
households (more than one person per room), compared to 6 percent nationwide. Thirteen
percent of Native American and Alaska Native homes lack safe and adequate water supply
and/or waste disposal facilities (Indian Health Service, 2009).
In Canada, especially in cities, over 60 percent of indigenous children live below
the poverty line. In Winnipeg, 80 percent of inner-city indigenous households reported
incomes below the poverty line (a much higher percentage than for poor non-indigenous
families). Similarly, indigenous homes are 90 times more likely to be without piped water
than non-indigenous homes. Indigenous homes are generally overcrowded, and one
reserve in four has a substandard water or sewage system. About 55 percent live in
communities where half of the houses are inadequate or sub-standard, manifested in
deteriorated units, toxic mold, lack of heating and insulation, and leaking pipes (Carino,
2009).
In Australia, indigenous peoples overall have lower incomes than the
nonindigenous population (Eversole, 2005). Indigenous households are half as likely to
own their own homes – 34 percent of indigenous peoples owned their own home,
compared to 69 percent of the non-indigenous population (Carino, 2009). Over a quarter
of the indigenous peoples was reported living in overcrowded conditions. The situation is
worse in rural and remote communities where people frequently do not have access to
21
adequate food, water and housing and have poor access to basic services and
infrastructure
(Altman et al, 2008; Carino, 2009). In 2001, 46 percent of the Australian indigenous
communities had no connection to a town water supply (Bolstridge, 2008). In New
Zealand, Maori as a group has a lower level of income and housing relative to non-Maori
(Humapage, 2005).
The economic situation of indigenous peoples in Latin America is not any better
(Pscchapropoulos and Patrinos, 1994). In Paraguay, poverty is 7.9 times higher among the
indigenous peoples, compared to the rest of the population (Plant, 1998). In Panama,
poverty rates for indigenous peoples are 5.9 times higher, in Mexico 3.3 times higher, and
in Guatemala 2.8 times higher than for non-indigenous peoples (ECLAC, 2007 p.152).
Poor, in Latin America, is synonymous with being indigenous; in addition, virtually all
the indigenous peoples living in municipalities where more than 90% of the peoples are
indigenous are extremely poor (Plant, 1998).
Similarly, Africa does not seem to offer any better situation for indigenous
peoples. In South Africa, the Nama and San people constitute some of the poorest of the
poor, stigmatized as a rural under-class fit only for menial labor (Eversole, 2005). The
Batwa in Rwanda, Burundi, Uganda and Eastern Democratic Republic of Congo have no
access to forests, have little or no land, and are desperately poor. Most of the Pygmy
indigenous peoples suffer hardship and work as servants on farms that do not belong to
them, or practice small-scale, informal mining activities; some must resort to begging
(Carino, 2009). Less than two percent of Batwa peoples have sufficient land to cultivate,
very few own livestock, and most are either squatters or tenants on other people’s land
22
(Mugarura & Ndemeye, 2003).
Although over 70% of the world’s indigenous peoples live in Asia (IFAD
2000/2001), the exact poverty status of many of the indigenous peoples in this region is
currently unknown. The reason is that statistics on the poverty status of indigenous
peoples are not readily available because few countries collect data disaggregated by
ethnicity (Carino, 2009 p.29). Of the few studies that have been done, these suggest
similar conditions in Asia. For example, in China, the “lack of fuels for fire, insufficient
clothing and shoes, several months’ shortage of grain each year, and extreme scarcity of
animal protein are common conditions” among the indigenous peoples (Tapp 1995: 215);
and in Taiwan, the country’s so-called economic miracle has left the indigenous peoples
with lower average incomes than the general populations (Eversole, 2005).
Poverty among indigenous peoples is not only pervasive, but also persistent (Hall
& Patrino, 2005). In the 1980s, poverty rates among indigenous peoples were 60% in
Peru, over 70% in Bolivia, 80% in Ecuador 80%, and as high as 90% in Guatemala and
Mexico (The World Bank, 2007; Carino, 2009). Twenty years later, with only Guatemala
the exception, the poverty rates remained the same in all of the countries (The World
Bank, 2007; Carino, 2009). Similarly, in Vietnam, poverty rates in regions where
indigenous peoples are concentrated remained high in the 1990s--73 percent in the
northern highlands and 91 percent in the central highlands--despite the fact that the
poverty rates for the country as a whole decreased from 58 to 37 percent (ILO, n.d.;
Eversole, 2005 p.32).
CHAPTER III: THEORY AND LITERATURE REVIEW
23
H. Theories of indigenous poverty
The review of the literature indicates that, to date, there are no specific theories of
indigenous poverty. The existing economic theories of poverty are based on industrialized
economies and are designed to explain individual economic behavior of the poor. Since
indigenous peoples experience poverty at multiple levels (individually, as a family, and as
a whole community) and since their livelihoods are based on subsistent economies, the
existing theories have little relevance to indigenous peoples.
Poor indigenous peoples may be viewed as a subset of the world’s poor peoples. Poor
people have been the central theme in both political and religious discourses throughout
history. About 2600 years ago, Buddha (563BCE -483 BCE) saw that poor people could
not get the opportunity for self-actualization because they were not able to fulfill their
basic material needs. He saw that the poor were often those who were at the bottom of the
hierarchy of the Hindu caste system. He saw the caste system as the most inhumane and
unjust system and those who took advantage of the poor as lower forms of life.
Years later, Socrates (469 BC–399 BC), Plato (428 BC -348 BC) and Aristotle (348
BC -322 BC) also advocated for social justice, primarily for the poor. About another 300
years later, Jesus (0 -30 AD est.) also addressed the inhumane treatment of the poor. They
all believed that the unjust socio-political system was the cause of sufferings, and the
victims were always the poor.
It is only in recent centuries that the condition of poor people have been
conceptualized in more abstract form as poverty, and discussed in academic discourses. In
particular, the writings of Adam Smith (1723 -1790), Thomas Malthus (1766 -1834) and
Karl Marx (1818 -1883) appear to have set the stage for poverty discourse. Smith saw in
24
each individual the potential to overcome poverty and suffering through hard work and
intelligence. He saw free-market as the necessary political condition for individuals to be
able to exercise economic behaviors and maximize profits of their labor. Marx, on the
other hand, saw the very political structure proposed by Smith as a both necessary and
sufficient condition that brews poverty. He argued that free-market economic structure
allows the owner of the means of production, the upper-class, to exploit the poor, the
lower-class, who form the pool of labor or means of production. He advocated for a
classless society, or regulated market, as opposed to a free market so that the exploitation
of the poor may be minimized or even eradicated. Malthus, on the other hand, was
primarily worried about the scarcity of resources and saw that poor and uneducated
people were “digging their own graves” by overpopulating and over-utilizing scarce
resources. His analysis gave rise to a popular analogy known as “tragedy of the
commons” – the idea that poor people don’t understand the consequences of their own
actions. This thesis was widely used (or abused) by early crusaders and religious
missionaries who described indigenous peoples as savage and primitive, thus
rationalizing the invasion and colonization of the indigenous nations. The colonization of
sovereign indigenous nations is thought to be the beginning of sufferings and poverty
among indigenous peoples.
The thinking of Thomas Malthus not only contributed to justification of colonization
at the time, but also continues to influence the current thinking of both Smithian and
Marxian economists who converge on the idea of “scarce resources” and the
“competition of interests” over those resources. Indeed, several theories have been
proposed to explain poverty based on these philosophical traditions. These theories “can
25
be simplistically lumped into two groups--theories that focus on individual behaviors and
theories that focus on social structures” (Sherraden, 1991 p. 35).
The individual theories of poverty (Schultz, 1963; Becker, 1964; etc.) claim that
determinants of poverty are found in the individual characteristics of the poor themselves
and not in the structural characteristics of the society. The basic premise of individual
explanation is the assumption that each individual human is a rational being who seeks to
maximize his or her own interests over those of the others. The responsibility for poverty
lies on the individual decision or choice about his or her own behaviors. Those who fail to
make right choices become poor.
Structural theories of poverty (Burton, 1992; Doeringer & Piore 1971; Blau,
Ferver & Winkler, 1998; Rank, 1994; Sherraden 1991; North, 1990; etc.), on the other
hand, claim that the major determinants of poverty are found “not in the characteristics of
the poor themselves, but in the structural elements of the larger society” (Burton, 1992, p.
149). The basic premise of structural explanation is that society does not treat its
members equally and fairly, and that there’s no “level playing field” for all members of
the society. Individuals who start off with better socio-economic conditions have more
choices and take advantage of those who start off with lower socio-economic status and
who have fewer choices. Due to this comparative advantage, the rich will always be
richer and those who are poor will always remain poor regardless of their individual
productivity (i.e. hard work). Structural theorists contend that changes in the structure of
a society are necessary to change the socio-economic status of the poor.
These existing theories, however, are fraught with sectarian ideologies of the
theorists and are often at odds with each other. These economic theories of poverty are
26
largely based on the assumption of market economies of the industrialized countries. For
one thing, these theories assume availability of markets and cash income in all
economies. Even the most progressive theories, such as Asset Theory (Sherraden, 1991),
assumes that poor people have cash incomes, and that what is lacking is the institutional
incentive structures (e.g. financial inclusion) to facilitate saving these cash incomes for
future use.
These assumptions are, however, at odds with the empirical reality of the
indigenous peoples. Except for a few groups in the industrialized countries, indigenous
peoples largely live in subsistent economies of developing countries in which cash
incomes are scarce. Indigenous peoples are largely self-employed farmers, and a few are
cattle grazers or hunter-gatherers. Exchange of goods and services in these economies is
often transacted through a bartering system rather than through cash. Where cash is used,
it is of minimal amount. Except for the industrialized countries (US, Australia, New
Zealand, and Europe), the majority of the indigenous peoples lives in countries or
communities that are themselves poor or under-developed. Theorizing indigenous poverty
requires not only explaining individual economic behaviors at present, but it also requires
explaining the factors that shape the evolution of the local contexts in which indigenous
peoples live, and how they became who they are now. Given this background, the
prevailing cash income-based market theories of poverty seem to have little relevance to
indigenous poverty.
The theory of institutional design (North, 1990) appears to capture both the
evolution of the macro-structures of a society and how these macro-structures shape
micro (individual) behaviors of the people. In the absence of specific theories of
27
indigenous poverty, this study used institutional design as an overall theoretical
framework. The human capital theories were used to guide the assessment of the
relationships between education, health and poverty. The asset theory (Sherraden, 1990)
was used as a guide to measure asset poverty among indigenous peoples.
C1. Theory of Institutional Design
The theory of institutional design was proposed by Douglas North (1990).
According to this theory, institutions are “the rule of the game in a society, or more
formally, is the humanly devised constraints that shape human interaction” (North, 1990
p.3). The major role of the institutions in a society is to reduce uncertainty by establishing
a stable (but not necessarily efficient) structure to human interaction (North, 1990: 6).
These institutions could be regulatory--formal rules such as a country’s laws and
constitutions, or they could be normative--informal rules such as codes of conduct and
social norms that embody shared understandings of acceptable behavior. These
institutions serve as the basis for rewarding and punishing individual acts of conformity
or deviance; violators of these rules are punished (Lesorogol, 2003).
According to North (1990), institutions are different from organizations.
Organizations are the “players” of the game. They include such entities as governments,
firms, universities, non-profits, clubs, and teams. Organizations are created to take
advantage of the opportunities provided by the institutions. As these organizations pursue
their objectives, they act as agents of institutional change. These organizations are the
crucial instruments of rule-enforcement for the stability of the institutions. They impose
the values, taste, culture, religion and language of the elites over other members of the
28
society in order to ensure that the established institutions serve their best interests and that
they remain stable over time.
There are several theories of institutional emergence and change. These theories
may be grouped into two schools of thought: evolutionary explanation (Knight, 1995 etc.)
and design explanation (North, 1990). According to evolutionary explanation, institutions
emerge naturally from local processes and change through natural selection processes.
The stronger institutions survive and the weaker ones die out.
According to the theory of institutional design, however, institutions do not
naturally emerge from nowhere, but rather they are deliberately designed. They are
designed and controlled by the few elites to further their own best interests. These
institutions are changed when they no longer serve the interest of the elites, whereupon
new rules are created.
Institutional theories are largely concerned with the emergences and changes in
institutions, and some are concerned with how such institutions affect performance of the
economies or societies. In particular, theory of institutional design is concerned with the
question of how best to design and organize human societies such that the growth and
development of these economies may be maximized. The primary assumption of this
theory is that countries/economies (and by extension indigenous economies) become poor
because they lack basic institutional structures for economic growth. In particular, poor
societies have high transaction costs and lack security over property rights.
Since the origin of indigenous poverty began with the origin of modern
nationstates, and since these nation-states are largely designed and controlled by the
nonindigenous peoples, the design theory of institution seems more relevant than the
29
evolutionary theory of institution. According to the theory of institutional design,
indigenous poverty may be thought of as the result of institutional structure of the society,
rather than due to lack of individual productivity characteristics. By extension,
indigenous peoples are thought to be poor because they lack “self-governance,” that is,
sovereignty over their territories.
Although the theory of institutional design is largely concerned with the question
“Why do some countries become rich, while others remain poor?” (North, 1990), by
extension this theory has important implications for the question “Why do some groups
within a country become poor, while others remain rich?” This is the central question of
this dissertation study. Institutions matter because our present and future choices are
shaped by the past; and our future is connected to the past through society’s institutions.
The assumptions of institutional theory with regard to indigenous poverty may be
specified as follows: Indigenous peoples are poor : (1) The constitution of the country
constrains indigenous peoples from self-governing their own territories, which makes
their territories poor and isolated; (2) The education laws of the country constrains
indigenous peoples using native language as a language of instruction in public schools-
this leads to low human capital attainment among indigenous peoples; (3) The health laws
of the country constrains poor people’s access to quality healthcare; (4) The labor laws or
other informal institutions related to occupation (e.g. caste system) constrains indigenous
peoples’ ability to move to better paying jobs; and (5) Laws governing financial system
excludes indigenous peoples and provide little incentive for economic growth.
The institutions of primary interest in this study are both regulatory or formal
institutions such as a country’s constitution, and informal institutions such as a caste
30
system. Although indigenous peoples in Nepal do not have castes, the caste system
(differential treatment of peoples based on their ethnicity and occupation) was imposed
on the indigenous peoples through national law ‘Muluki Ain 1854’ of Nepal (Hofer,
2004). The national law “Muluki Ain 1854” was constructed based on the principles of
the caste system and have been the de facto instrument of government control over its
citizens. While indigenous and non-indigenous peoples may live under a different set of
informal institutions (such as cultural or religious norms), they are both assumed to live
under the same formal institutions (the country’s laws or constitutions).
C2. Human Capital Theory
The human capital theory was originally conceptualized by Theodore Schultze in
the 1960s, and later followed by Gary Becker’s 1964 monograph “Human Capital”,
which has ever since served as a benchmark of the subject (Blaug, 1976, p. 827). The
human capital theory contends that people invest themselves in diverse ways, not for the
sake of present enjoyments, but for the sake of future returns. According to this theory, all
purchases of health, education, job search, information retrieval, migration, and inservice
training may be regarded more as investment than consumption, regardless of whether
purchases were made by individuals on their own behalf or society on behalf of its
members (Blaug, 1976, p.829).
According to this theory, schooling contributes to individual productivity which,
in turn, leads to higher individual earnings. The earning advantage of the more educated
relative to the less educated is subject to the laws of supply and demand-- as the number
of the more educated increase, their earnings advantage declines and the minimum
31
qualifications for given jobs rise in line with increased relative supplies (Schultz 1961;
Mincer 1974; Becker 1975; Psacharopoulos & Patrinos, 1994; p.46). The lack of human
capital – i.e. training, education, experiences, skills etc. – would mean less competition in
the labor market which then would lead to poverty (Rank, 1994, p. 26-27). According to
human capital theory, indigenous poverty may be thought of as a result of low human
capital attainment among indigenous peoples.
Psacharopoulos & Patrinos (1994) tested the human capital hypothesis among
indigenous peoples in Latin America. They found that indigenous peoples on average
receive less schooling compared to their non-indigenous counterparts, and the schooling
was positively correlated with earnings, supporting the prediction of the human capital
theory.
However, this study also found substantial earning differentials between
indigenous and non-indigenous peoples even after equalizing the human capital and other
productive characteristics. That is, even if indigenous peoples were endowed with equal
human capital, they would still earn only 50 percent of the non-indigenous earnings
(Pscharopoulos & Patrinos, 1994, p.xxi). The other 50% income disparity between
indigenous and non-indigenous peoples remains unexplained. Furthermore, this study
does not enlighten as to why indigenous peoples lack human capital in the first place.
C3. Asset Theory
The welfare theory of asset was conceptualized and developed by Michael
Sherraden (1991). This theory emerged as a reaction (or an alternative) to the prevailing
income-based welfare policy approach to poverty in the U.S. This theory contends that a
32
welfare system that relies on income-transfer may provide a safety-net for the poor but
does not help the poor become rich. What poor people lack, according to this theory, is
not income but asset. What is needed for the poor is an institutional incentive structure
(such as financial inclusion) that will facilitate accumulation of assets and access to
financial systems that encourage the poor to save their assets for the future growth.
Saving is thought of as the necessary condition to lift the poor from their poverty. The
assumptions of this theory are consistent with that of the theory of institutional design in
that they both emphasize the institutional structures, such as incentive structures.
Since the publication of this seminal work, asset has become an established
approach to measuring poverty. Asset is a more reliable indicator of wealth (poverty) than
income because when people lose income, they still rely on their assets. Assets may
include savings, bonds, houses, lands or any other movable and immovable possessions
of capital value. According to this theory, poor people lack assets because they lack
institutional incentive structures that facilitate their saving behaviors. Emerging empirical
studies have found support for asset theory in the industrialized countries (Loke &
Sherraden, 2009; Han et al. 2007).
Asset is more relevant to indigenous poverty than income or consumption.
However, although asset-based measure is thought to be innovative in the US where
income is widely used, the asset-based measure of poverty, in fact, is the oldest method in
the world. Its origin dates back to the history of humankind when hunter-gatherers began
to accumulate materials for future use. In Biblical times, wealth was measured by the size
of the house, number of livestock owned, amount of land holdings, and possession of
jewelries and other valuable metals and minerals. Those who owned less were considered
33
relatively poor. Those who owned the least or did not own anything were the poorest in
society. They were the slaves and indentured or bonded laborers. Still today, in many
villages of developing countries--especially among indigenous communities where
subsistence agriculture is the primary means of livelihood--the land, house and livestock
are the commonly used indicators of wealth. Where exchange of goods and services takes
place, it often takes the form of bartering. Most of the farmers are self-employed and
selfreliant. This study will use asset theory as guidance to measuring poverty among the
indigenous peoples.
D. Empirical Evidence
The review of the literature indicates that there have been very few experimental
studies on indigenous poverty. Empirical studies on socioeconomic status of indigenous
peoples began only recently in the 1990s. The movement towards empirical studies of
indigenous poverty was motivated by the studies of racial disparities between Blacks and
Whites in the United States in the 1960s (Psacharopoulos & Patrinos, 1994). The first
empirical study of indigenous poverty was conducted by Psacharopoulos & Patrinos
(1994) in Latin America. Since then, there have been very few other studies in Latin
America (Plant, 1998; Hall & Patrinos, 2005), the US (Cornell, 2005), Australia and New
Zealand (Altman et al, 2008). The findings from these studies are summarized here.
D1. Geographic isolation and indigenous poverty
Previous studies have noted that indigenous peoples live in different geographies
than non-indigenous peoples, and the indigenous territories are largely rural or isolated
34
(Plant, 1998; Hall & Patrinos, 2005). The isolation of the geographies can have a direct
effect on the level of poverty of their inhabitants. These areas are less likely to have
infrastructure development (roads, electricity, piped water, irrigation system), social
services (education opportunities, healthcare services), and economic opportunities
(markets, industries). Living in these poor or isolated communities is likely to increase
the risk of poverty compared to those who live elsewhere. For example, in Mexico,
Panadiges (1994) found that indigenous communities have significantly less access to
public services. In geographic areas where the majority of the populations were
indigenous, only 16.1% had piped water (compared to 62.5%), 48.9% had electricity
(compared to 92.9%) and only 2.4% had telephone services (compared to 22.2%). Living
in areas where 50% or more of the population is indigenous increases the probability of a
household being poor by 24.5% (Panagides, 1994). In almost all Latin American
countries, the poverty map coincides with indigenous peoples’ territories (Hall & Patrino,
2005). Across the world, indigenous women tend to live in more impoverished
municipalities (SanchezPerez et al. 2005).
Geography is thought to effect indigenous poverty in two ways. First, different
geographic areas are endowed with different levels of resources. Second, different
geographies are likely to have different institutional arrangements. The variation in
poverty outcomes across geographies is likely to be either due to variation in endowed
resources or due to variations in institutions. Some studies suggest that differential
poverty rates across geographies are likely to be due to differences in institution rather
than due to differences in endowed resources (Cornell, 2002). The studies at Harvard
Project of Native American Poverty found that, independent of the endowed resources,
35
the Indian Nations which were governed by the natives themselves (self-governance), and
which had higher sovereignty over their territories were economically better off compared
to those Indian Nations which did not have complete sovereignty, and which were merely
acting as administrators of the U.S. Federal Government. This finding suggests that the
‘self-governance’ or ‘sovereignty’ of the indigenous territories is an important predictor
of indigenous well-being.
Geography can also have an indirect effect on poverty through health. In recent
years, it has become increasingly clear that geography plays an important role in
determining an individual’s access to needed healthcare services (Raghavan, Lama et al;
2010). In the words of Wennberg and colleagues (1973; 1982, 1987), “Where you live
matters,” at least with regard to receiving quality healthcare services. In geographically
isolated areas, little health care services and opportunities are available to indigenous
peoples. As a result, indigenous peoples are likely to suffer from disproportionately high
rates of death due to preventable causes such as childbirth, diarrhea and others.
The lack of infrastructure and public services in indigenous areas constrains
indigenous peoples’ ability to advance their economic well-being in multiple ways. Due
to the lack of educational opportunities in their isolated areas, the people have little
chance to develop human capital. As they have very little human capital endowment, their
ability to move out of poverty is much slimmer than non-indigenous peoples even if they
migrate to urban areas.
Due to geographic isolation, fewer or no opportunities are available for
indigenous peoples to advance their economic well-being (Psacharopoulos & Patrinos,
1994). Their traditional livelihoods are hindered by the decline in the quality of the land
36
due to soil erosion and deforestation, which in turn leads to decline in food production.
Indigenous farmers lack modern irrigation systems and modern tools for agricultural
production. They rely on natural rain for irrigation, so the harvests are not guaranteed.
They are vulnerable to natural disaster such as famine, and their food supplies are not
secure. Food shortages are further exacerbated by the population growth, which shrinks
the amount of land per capita. Food shortages lead to high malnutrition, which makes the
indigenous peoples vulnerable to health problems and early death.
In isolated geographic areas, indigenous peoples have little opportunity to
network with people who control resources and opportunities outside their own
communities. This lack of social capital minimizes their ability to emerge from poverty.
Another effect of geographic isolation on poverty is seen through the housing condition
of indigenous peoples. In indigenous communities in Mexico, a larger percentage of
homes were built with low quality materials such as wood (21%) compared to
nonindigenous communities (6%). Compared to 71% of the homes in non-indigenous
communities, only 29% homes in indigenous communities were constructed with
concrete and brick (Panadiges, 1994 p.136). The lower quality housing has a direct effect
on the health status of the people residing in those houses. As the indigenous peoples
suffer from health problems for which there are no health services in their communities,
they suffer loss of economic productivity and disproportionately higher rates of mortality,
spiraling them or their families into deeper poverty.
While the previous studies have made observations on the relationship between
geographic isolation and indigenous poverty, they offer little explanation as to why the
indigenous peoples are geographically isolated in the first place. From the perspective of
37
the theory of institutional design, the observed geographic isolations of indigenous
peoples may be thought of as the direct result of institutions, which deny indigenous
peoples’ ability to self-determine, or self-govern, their own territories. Such constraints
have detrimental effects on self-development of the indigenous communities. It appears
that by isolating the indigenous peoples geographically, the elites effectively disfranchise
them from forming collective action or revolts. Elites then reap the benefits of excess
cheap labors of the indigenous peoples.
.
D2. Education system and indigenous poverty
Many studies have reported that lack of human capital is an important determinant of
poverty among indigenous peoples (Champagne, 2007; UNICEF, 2003; Steele, 1994;
Panadiges, 1994; Macisaac, 1994 etc.). Empirical studies in Latin America found that
education is the most critical determinant of poverty among indigenous peoples in Bolivia
(Wood & Patrinos, 1994), Guatemala (Steele, 1994), Mexico (Panadiges, 1994) and Peru
(Macisaac, 1994). In Mexico, for example, an average of 6.5 years of education would
decrease an individual’s probability of being poor by 22.5 percent (Panagides,
1994).
Illiteracy is the highest among indigenous peoples throughout the world. Among the
H’mong of Vietnam, 83% of men and 97% of women are illiterate (UNICEF, 2003). In
some indigenous communities of Australia, 93% of the populations are illiterate. In
Ecuador, illiteracy among the indigenous peoples in 2001 was 28% compared to national
rate of 13% (Carino, 2009; p.132). In Venezuela, the illiteracy rate among indigenous
peoples is 32%, five times higher than the non-indigenous rate (6.4%) (ECLAC/CEPAL,
38
2006 p.177 cited in Carino, 2009 p. 132). In Guatemala, 53.5% of indigenous young
people aged 15-19 have not completed primary education as compared to the 32.2%
nonindigenous youth (Carino, 2009). In Nepal, 30 percent of the indigenous people have
literacy rates far below the national average (UNDP, 2004 p.63). Indigenous women tend
to have even lower educational attainment (Sanchez-Perez et al. 2005).
In the U.S. fewer indigenous children graduate from high school and far fewer go to
colleges and universities. According to Assembly of First Nations (2009), about 70% of
First Nations students on Reservations will never complete high school. Graduation rates
for the on-reserve populations range from 28.9% to 32%t annually. Only about 27% of
First Nations populations between 15 and 44 years of age hold a post-secondary
certificate diploma or degree, compared to 46% of the Canadian population within the
same age groups (Assembly of First Nations, 2009). Dropout rates from primary schools
are significantly higher among Native American students compared to their
nonindigenous counterparts. Only 7.6 percent of Native Americans have a bachelor’s
degree compared to 15.5% of the total population (Tsai & Alanis, 2004).
While the previous studies shed some light on the educational status of indigenous
peoples and its relationship to poverty, these studies fall short of explaining why
indigenous peoples lack human capital in the first place. There are reasons to believe that
low human capital attainments among indigenous peoples are likely to be due to the
nature and quality of the education system rather than due to lack of individual
motivation for personal growth. In many countries of the world, where non-indigenous
peoples are the elites of the society, it seems that education systems are designed to
domesticate the indigenous peoples into the culture of the elites, rather than to enlighten
39
their citizens. In these societies, the elites of the society use organizations such as schools
and universities as change agents to impose their language, culture and religion on the
indigenous peoples under the pretext of assimilation or integration. In such societies, the
language of instruction in public schools is generally the language of the non-indigenous
elites; and in many countries such as Nepal, the languages of indigenous peoples are
legally prohibited as a language of instruction in public schools. The teachers in public
schools are generally non-indigenous elites who neither understand the culture or
language of the indigenous peoples, nor do they seem to genuinely care about the
progress of the indigenous communities. Teachers often act as masters of the indigenous
pupils rather than as a guide to bettering their future.
Furthermore, since indigenous peoples live in remote isolated areas, their
communities generally do not have access to educational opportunities to begin with;
where schools are available, they are either of low quality, taught in a language that is
different from their own, or simply too expensive. Faced with a double burden of foreign
languages and culture, most indigenous children give up schooling altogether or fail to
compete against their non-indigenous counterparts. As a result, the high illiteracy rate
(UNICEF, 2003), high dropout rate, and lower graduation rate (Assembly of First
Nations, 2009) among the indigenous peoples seem only natural, reflecting the
underlying educational system of the society. By depriving indigenous peoples of the
human capital that is necessary to achieve elite status, the non-indigenous elites
effectively ensure continuation of the status quo.
40
D3. Health system and indigenous poverty
Several studies have documented that indigenous peoples suffer
disproportionately high rates of health problems. The high prevalence of health problems
among indigenous peoples are thought to have direct consequences on their poverty. The
high cost of health care and loss of economic productivity due to illness and death of the
family member is likely to prevent the indigenous peoples from overcoming poverty. An
empirical study in Bolivia found that being healthy lowers the probability of being poor
by 5.3 percent (Wood & Patrinos, 1994).
Indigenous peoples experience a disproportionately lower level of life expectancy
(State of the World’s Indigenous Peoples, 2010; Sanchez-Perez et al. 2005; Eversole,
2005; Humapage, 2005). In Australia, an indigenous child can expect to die 20 years
earlier than his non-native compatriot (Cooke, Mitrou, Lawrence, Guimod & Beavan,
2007). The life expectancy gap is also 20 years in Nepal, while in Guatemala it is 13
years, in New Zealand it is 11, in Panama, it is 10 years and in Mexico, it is 6 years (State
of the World’s Indigenous Peoples, 2009). In the U.S., a Native American’s life
expectancy is on average 2.4 years lower than that of the general population. In Canada,
life expectancy was 8.1 years less for male and 5.5 years less for female Canadian
Indians than for general Canadian populations. The low life expectancy rate means fewer
years in economic productivity and higher risk of poverty for the families and the
communities.
Indigenous peoples experience disproportionately high levels of death due to
unnatural causes such as tuberculosis, diabetes, alcoholism, and suicide (State of the
World’s Indigenous Peoples, 2010). Compared to the general population, Native
41
Americans and Alaska Natives have a 600 percent higher death rate due to tuberculosis
(Indian Health Services, 2006). In Canada, the Inuit TB rate is over 150 times higher.
Among indigenous Americans, death rates due to diabetes are 189 percent higher than
non-indigenous peoples (Indian Health Services, 2006). Worldwide, more than 50 percent
of the indigenous adults suffer from Type-2 diabetes (State of the World’s Indigenous
Peoples, 2009). Indigenous Americans have a 510 percent higher death rate due to
alcoholism, 229 percent higher rate due to motor vehicle accidents, 152 percent higher
death rate due to unintentional injuries, 61 percent homicide rate, and 62 percent higher
suicide rate (Indian Health Services, 2006). Death before age 75 due to suicide or
unintentional injury among indigenous Canadians is four and half times higher than the
general population (Health Canada, 2007). Suicide rates, particularly among indigenous
youth, are considerably higher in many countries, for example up to 11 times the national
average for the Inuit in Canada (State of the World’s Indigenous Peoples, 2009). High
level of deaths among indigenous peoples means loss of potential labor forces, and higher
risk of poverty.
Indigenous women across the globe suffer from higher infant mortality rates
(State of the World’s Indigenous Peoples, 2009; Sanchez-Perez et al. 2005; Eversole,
2005; Humapage, 2005). Indigenous children are more likely to die before one year of
age than others in their countries (Damman, 2005). The child mortality rate in Latin
America is 70 percent higher among the indigenous peoples (ECLAC, 2007 p.191,
quoted in Carino, 2009). In Canada, Inuit children are 2.2 times more likely to die before
one year of age compared to children in the general population; Metis and other Canadian
Indian children are 1.9 times more likely to die than their counterparts. A similar trend
42
exists for the U.S. (Eversole, 2005). A high child mortality rate means loss of return for
investment in child rearing. The loss of children puts pressure on parents to produce more
children (to increase the survival rates), which increases the cost and reduces the
economic activities of the parents, again leading them further into poverty.
In addition to low life expectancy, high death rate and child mortality rates,
indigenous peoples generally have lower health status. Reservation-based indigenous
groups have some of the lowest health status in the US (Cornell, 2005). In New Zealand,
Maori as a group continue to demonstrate lower levels of health relative to non-Maori
(Humapage, 2005). Indigenous adults in Australia are twice as likely as non-indigenous
adults to report their health as fair or poor, are twice as likely to report a high level of
psychological stress, and are twice as likely to be hospitalized (Cooke, Mitrou, Lawrence,
Guimod & Beavan, 2007). In parts of Ecuador, indigenous peoples have a 30 times
greater risk of throat cancer than the national average (State of the World’s Indigenous
Peoples, 2009). The lower health status and higher disease prevalence increases health
care costs in terms of time and money, perpetuating poverty.
While the high rate of health problems among the indigenous peoples is likely to
prevent the indigenous peoples from coming out of their poverty, it is also possible that
health problems themselves are the consequence of poverty. The relationship between
poverty and health problems is likely to be endogenous. For example, for many
indigenous children, poverty is an a priori condition. In many cases, indigenous children
are born in an impoverished families, and not that they become poor as a result of their
health status or other individual economic behaviors. Being born in poverty leads them to
poor health due to lack of adequate nutrition and needed health care. The poor health and
43
cost of healthcare, in turn, are likely to lead them into further poverty as they are not able
to invest their time and resources in human capital development and economic
productivity, which would help them exit poverty in later life.
In this study, health is viewed as independent variable because population health
is more likely to depend on the healthcare system of a society, rather than solely on the
socioeconomic status of individuals. Poor countries are less likely to provide quality
healthcare services to their citizens. Lack of access to quality healthcare is likely to be the
reasons for low health status among indigenous peoples. Furthermore, due to ethnic
mismatch between the healthcare providers and the consumers, indigenous peoples are
less likely to utilize healthcare services where they are available. Since indigenous
peoples have low human capital, they are less likely to be the healthcare providers. The
healthcare providers are generally the non-indigenous peoples, who speak different
language and practice different cultures. In most poor countries, the health care markets
are designed and controlled by non-indigenous elites who are motivated by profits than
services to the people. The high-cost of healthcare; compounded by linguistic and cultural
barriers are likely to discourage indigenous peoples from seeking health services
provided by non-indigenous peoples.
D4. Discrimination (inequality) and indigenous poverty
In the words of Psacharopoulos & Patrinos (1994), there is a “cost of being
indigenous.” Discrimination against indigenous peoples exists in many forms and in
44
many spheres of their livelihood. Studies have extensively documented discrimination
against indigenous peoples in land tenure, income, employment, housing, and education
(Psacharopoulos & Patrinos,1994; Wood & Patrinos, 1994; ILO, 2007; Taylor & Kalt,
2005; Altman, Biddle & Hunter, 2008; Eversole, 2005; Carino, 2009; Freeman & Fox,
2005 etc.). Social exclusion is the primary form of discrimination. In many countries,
indigenous peoples are excluded from participating in public affairs, civic engagement,
educational organizations and governance. Where indigenous peoples are allowed to
participate, they often experience differential treatment from the system.
Psacharopoulos & Patrinos (1994) have documented that in Latin America,
indigenous peoples are widely discriminated against in the labor market. Indigenous
workers in Latin America make on average about half of what non-indigenous workers
earn. For example, in Guatemala, Mexico and Peru, indigenous peoples earn only 50% of
the earnings of the non-indigenous peoples even after equalizing the human capital and
other productivity characteristics. In Bolivia, the earning differential was 28 percent and
the probability of being poor for indigenous peoples is 16% greater than probability for
non-indigenous counterparts (Wood & Patrinos, 1994). Overall, about 25-50 percent of
the income gap in Latin America is “due to discrimination and non-observable
characteristics, such as quality of schooling” (ILO, 2007 p.27).
Similarly in the U.S., the average income of Native Americans is less than half the
average for the general population of the U.S. (Taylor & Kalt, 2005). The
reservationbased indigenous peoples have among the lowest income (Cornell, 2005). In
Australia, indigenous peoples overall have lower incomes than the non-indigenous
population (Eversole, 2005). The median indigenous income in Australia is just over half
45
of the nonindigenous income (Altman, Biddle & Hunter, 2008). In New Zealand, Maori
as a group has a lower level of income relative to non-Maori (Humapage, 2005). In
Taiwan, the indigenous peoples have much lower average incomes than the general
population
(Eversole, 2005).
Discrimination also exists in housing. The reservation-based indigenous peoples
have among the poorest housing conditions in the U.S. (Cornell, 2005). The percent of
Native Americans who live in crowded households (18%) is three times higher than the
percent nationwide. The percent of Native American and Alaska Native homes that lack
safe and adequate water supply and/or waste disposal facilities is 13 times higher than the
homes for the U.S. general population (Indian Health Service, 2009). In New Zealand,
Maori as a group has lower level housing relative to non-Maori (Humapage, 2005).
In Canada, the Royal Commission on Aboriginal Peoples (RCAP) reported that
houses occupied by indigenous people are twice as likely to be in need of major repairs as
compared to house of other Canadians. The indigenous homes are 90 times more likely to
be without piped water than non-indigenous homes (Carino, 2009). In Australia,
indigenous households are half as likely to own their own homes – 34 percent of
indigenous peoples owned their own home, compared to 69 percent of the nonindigenous
population (Altman, Biddle & Hunter, 2008).
Studies have also documented discrimination in employment. While the total
unemployment rate in the U.S. declined from 6.5 to 5.9 percent between 1994 and 2003,
during the same period, it increased from 11.7 to 15.1 percent among American Indians
and Alaska Natives (Freeman & Fox, 2005 p.122). In Canada, Aboriginal people have
46
poor access to jobs. In 2005, the unemployment rate of Canada’s western provinces of
Manitoba, British Columbia, Alberta and Saskatchewan was as high as 13.6 percent
among indigenous people, compared to only 5.3 percent among the non-indigenous
population (Statistics Canada, 2005). In Australia, the indigenous unemployment rate was
15.6 percent in 2006, over three times higher than the non-indigenous rate. A study in
Latin America found that, among the indigenous households in Bolivia, living in a
household where the household head is unemployed (not working but looking for work)
was the most substantial factor contributing to the probability of being poor (Wood &
Patrinos, 1994).
While unemployment rate has been found to associate with poverty among the
indigenous peoples in the cash-based market economy of both developed and developing
countries, it is not clear whether this association remains valid in less developed market
economies. Published reports have documented that in some countries, indigenous
peoples work long hours, and their employment rate is higher than non-indigenous
peoples, yet their income is lower and poverty rate is higher than the non-indigenous
peoples. The likely reason is thought to be discrimination in labor market, which
systematically puts them at lower paying, manual labor work.
The source of discrimination (inequality) is thought to be inherent institutional
design, such as labor laws, which reserve better paying jobs for the elites. For example, in
Nepal many of the indigenous peoples were legally barred from running for government
offices and joining national armies until the 1990s. Low caste people were structurally
prevented from changing their occupations, such as from being a cobbler to a teacher.
Such constraints in occupational change prevented their socioeconomic mobility. Still
47
today, the majority of government bureaucrats, technocrats, aristocrats and political
leaders are non-indigenous peoples who design market structure to their own advantage.
It is in the interest of the social elites to design institutions in ways that benefit them the
most and to ensure the stability of such an institution over time. Since the social elites are
non-indigenous, they develop rules of engagement (e.g. bribery, nepotism as acceptable
norms) which treat indigenous peoples differently from others. These rules are enforced
through various organizations, firms and markets. The unequal outcome is generally felt
as discrimination by the indigenous peoples.
48
CHAPTER IV: CONTEXT - NEPAL
Nepal is a landlocked country, located between China on the north and India on
the south, east and west. Nepal is one of only a few countries in the world that has never
been colonized by Europeans. Nepal was established as a nation-state in the 1770s (AD).
Prior to the 1770s, the native peoples of the Himalayas governed independent kingdoms
of their own. The evidences of their advanced civilizations are well-reflected in the
cultural monuments such as monasteries, temples, and Stupa (Buddhist shrines). Still
today, these monuments stand tall across Himalayan regions, including Kathmandu valley
(Yambu). For example, the Boudha Stupa (Jhyarung Khasyor Chyorten) and
Swayambu Stupa (Phapa Singun Chyorten) are thought to have been built around 300
BC. These native Buddhist monuments still impose the unmistakable identity of the
advanced civilizations enjoyed by indigenous peoples in the Himalayas prior to the
arrival of the Malla (1200s) and the Khas people (1770s), both of which were Hindus.
Still today, Nepal is well-known because of such native heritages as the Boudha Stupa,
Swayambu Stupa, the Buddha Dharma, the Namo Buddha, the Mt. Everest (Jhyomo
Longma), the Gurkhas, and the Sherpas.
Currently Nepal is considered one of the poorest countries in Asia, with a poverty
rate estimated to be more than 31% in 2004 (World Bank). Much of this poverty is
thought to constitute indigenous peoples.
The population of Nepal was estimated to be 28.5 million in 2009 (CIA the World
Fact Book). The population of Nepal is broadly classified into two groups--indigenous
peoples, or Adibasi Janajai (Mongoloid) and the Khas people or settlers (Aryan). This
distinction is based on race/ethnicity and the history of Nepal. The historical point of
49
convergence of these two groups under one political system is the establishment of Nepal
as a nation-state or one kingdom in the 1760s. The peoples who had been living on the
land and who were dominant groups prior to the establishment of Nepal as a nation-state
are the indigenous peoples (Adibasi Janajati or Mongoloid). Those who came later as
settlers are the non-indigenous group (Hindu-Aryan). These non-indigenous groups are
also known as the Khas people (or peoples of the caste system). Even Prithvi Narayan
Shah, who is credited for unifying Nepal as a nation-state (and destroying the indigenous
nationalities), recognized these differences when he proclaimed Nepal as a garden for
Chaar Jaat (four Hindu castes) and Chattis Varna (36 non-Hindu indigenous nationalities).
Prithvi Narayan Shah, however, was ignorant of the existence of many of the native
peoples of the Himalayas which he invaded. Today, the number of indigenous groups in
Nepal is thought to be over 120.
Review of the literature on indigenous peoples in Nepal indicates that there is no
authoritiative data on the indigenous population in Nepal. Until the 1990s, Nepali
government banned collecting demographic and ethnicity data in the census. The reason,
it is believed, is that the government was afraid that the census data would reveal
indigenous peoples as the population majority. This information, if published, was
considered a threat to the government—it would potentially question, or even overthrow,
the two-and-half century old Bahun/Chetri regime in Nepal.
The government of Nepal allowed collection of demographic and ethnographic
information in its census only in recent years, the 1990s. However, there has been a
growing concern about the validity of the government-collected data. The indigenous
peoples have disputed the census data claiming that the Nepali government manipulates
50
the data to show Bahun/Chetri (non-indigneous groups) as the majority population of
Nepal. The indigenous peoples claim that their true population number has been
undercounted, while Bahun/Chetri population numbers have been exaggerated. They
claim that the Nepali government does this to justify the continuation of the Bahun/Chetri
hegemony in the country.
Without valid census data, it is difficult to estimate the exact proportion of the
indigenous population in Nepal. One report suggests that indigenous peoples in Nepal
represent over 72% of the population (Leslie et al, 2010).
Much of the literature on Nepal’s history and population has been written by the
non-indigenous scholars, particularly Brahmin and Chetries, and a few Newars. Since the
Brahmin/Chetries began to hold political dominance in Nepal in the 1800s, a new history
of Nepal has been written. The new history often glorifies the Khas peoples, particularly
Bahun and Chetries. The historical documents of Nepal reflect the views of the
Bahun/Chetries and their biases against indigenous peoples. The views of the indigenous
peoples and the history seen and experienced by them have been absent in the national
literature and poltical discourses of Nepal. The written history of the indigenous peoples
prior to the invastion by Bahun/Chertries, and thereafter, has been lost. However, the
stories of the indigenous peoples as preserved through oral traditions and songs continue
to prevail and provide a solid base for social science to examine the facts, which were
systematically excluded in the written history of Nepal.
The sections that follow briefly provide the historical processes that have shaped
Nepal to its current state. These historical processes reflect the views that are shared
among the indigenous peoples.
51
E1. Indigenous Peoples of Nepal (Adibasi Janajati or Mongoloid Group)
Indigenous peoples of Nepal are the native or original inhabitants of the
Himalayas and the southern plains known as Terai. Indigenous peoples as a group
constitute about 72% of the total population of Nepal (Leslie et al, 2010). This estimate,
however, is likely to be undercounted because census data is generally tempered [by the
government] to project large Hindu or caste population (Lawoti, 2001). Indigenous
peoples speak more than 60 languages and practice diverse religions and cultures
including Buddhism, Shamanism (Bon practice), and other local religions. The
indigenous peoples look distinct in their physical features due to their Mongoloid racial
origin. (Note: some Khas peoples especially K.C.- Kasheko Chetri or fallen Chetries--
may resemble Mongoloids due to interracial marriage with Mongoloids). According to
the Nepal Federation of Indigenous Nationalities (NEFIN), as of 2007, the government of
Nepal recognized 59 nationalities as the indigenous peoples. The major indigenous
peoples of Nepal include Magar, Tharu, Tamang, Newar, Gurung, Rai, Limbu, Sherpa and
Thakali.
Due to the process of “Sanskritization”
1
and proselytizing,
2
some of the
indigenous groups are presently divided into Buddhists and Hindus. Newars, for example,
1
According to Indian anthropologist Srinivas, “Sanskritization” is a process whereby a less powerful group
adopts the religious and cultural attributes of the more powerful group in order to achieve upward mobility
or a higher degree of acceptance.
2
Gopal Gurung, the author of The Hidden Facts in Nepalese Politics (1998), contends that many Buddists
were converted to Hinduism forcefully rather than voluntarily.
52
were originally Buddhists until the Malla came to Kathmandu Valley. However, during
the Malla regime (1200 AD – 1770AD), some of the Newars were converted into
Hinduism. The Malla (wrestler in Sanskrit) settlement in Kathmandu valley began in the
12th century
3
. The Mallas are believed to have migrated from north Bengal and the
southeast region of Nepal (Bhojpur area)4 during the Mogul invasion of India. The Mallas
were the ones who began to call the people of Kathmandu valley “Newar” (napa: the
citizen of nepa). As the Mallas began to settle in Kathmandu valley, they intermarried
with the local Newars in their effort to co-exist. It was through intermarriage with the
Newars that Mallas gained political prominence among the Newars. The first Malla to be
the king of Newar was Ari Malla. It was also through intermarriage that Mallas
propagated Hinduism to Newars, building Hindu idols (statues) on the periphery of
Buddhist temples. This is thought to be the beginning of the mixing of Hindu-Buddhist
idols in religious temples of Kathmandu valley. However, it was Jayashtiti Malla (r.1382-
1395) who is credited for structuring Newar society by implementing the caste system.
With the regime of Jayasthiti Malla, the Hindu Newars began to practice the caste system.
Many of the Hindu temples in Kathmandu valley were built after the Jayasthiti Malla
regime. As a result, unlike Hindu temples in India, the Hindu temples in Nepal look like
Buddhist monasteries because they were built with local Buddhist architecture -the
Pagoda style that is prevalent among Buddhists in much of Asia, China, Korea, and
3
Andrea Matles Savada ed. Nepal: A Country Study. Washington: GPO for the Library of Congress, 1991. 4
Lama (2011). Newari Buddhist’s Account of Malla Regime in Kathmadu Valley. The story of how Mallas
infiltrated Newar community, became the king of Newars, and sidelined those who resisted
Hinduproselytizing is a common story among the Buddhist Newar. Their story is based on their own
collective memories and lived experiences rather than written history.
53
Japan. It is a common belief among Buddhist Newars that, during the Jayasthiti Malla
regime, Buddha’s idols in the Buddhist temples began to be replaced with Hindu-god
idols. The Hindu-god idols took the center-stage in many of the Pagoda style Buddhist
temples, while Buddha-idols were pushed to the periphery. The major Hindu Newars
include Shrestha, Joshi, Pradhan, and Baidya. The major Buddhist Newar include
Shakya, Bajracharya, Tamrakar, Ranjitkar, and others.
Prior to the establishment of Nepal as a nation-state in the 1760s (AD), the
indigenous peoples of Nepal governed independent kingdoms (principalities) of their
own. However, currently the indigenous peoples are the political and economic minority
of Nepal. As a result of the establishment of Nepal as a nation-state, indigenous peoples
lost sovereignty over their territories. Their native system of governance and inherent
cultural institutions were destroyed by the government (Bhattachan, 2008; Lawoti, 2001)
and their lands and resources on them were confiscated by people who ran the
government. During the Rana regime (1846 -1950 AD), most of the indigenous groups
(and some Khas groups) were legally barred from attending schools. In particular,
Tamang indigenous peoples were barred from holding public offices or government jobs
including military and police; they were also prohibited from joining the British Gurkha
Army and leaving the country to seek a better future elsewhere. Tamang men were used
for free labor as servants and porters. Tamang women were used as concubines in the
Rana palaces and then sold to India for prostitution. Trafficking of women in Nepal
originated during the Rana regime but did not end with it. Even today, Tamang women
continue to be the victims of this slave trade (KC et al, 2001).
54
Tamangs are thought to be the first peoples to settle in the Kathmandu (Yambu)
valley.
4
Kathmandu valley was once thought to be a giant lake. As the lake began to
drain, Tamangs began to settle on the hills surrounding Kathmandu valley (since the top
of the hills, and not the bottom pit, were the areas where the water drained out first). Still
today, the majority of the peoples living on these hills are Tamangs--evidence that is often
used as a support to this claim. The current political movement to create Tamang
Autonomous Region: Tamsaling (Tam = language, Sa = land, ling =territory: land of the
Tamang peoples) is based on this historical claim. In Tamang language, a king is called
Ghleh. Still today, the descendents of Tamang kings are called Ghle (mispronounced in
Nepali language as Ghale). Newars (also known as Jyahphu; Jya =work, phu =farming),
who were mostly farmers and traders, migrated later and began to settle near the river
bank as the lake drained completely and provided fertile land for farming.
During the Rana regime, some of the caste groups, including the Shahs, were also
oppressed. However, the higher-caste groups overturned the Rana regime in the 1950s
and have enjoyed the privilege status since then. The lower-caste groups, especially the
untouchables or occupational castes, and some Chetri caste remain marginalized.
Indigenous peoples are still not allowed to educate their children in their native
languages in public schools. One exception is the Newar, who are the natives of
Kathmandu valley as mentioned earlier, and some of whom are Hindus. Due to the
geographic and social proximity with the caste rulers of Nepal, some Newars have
enjoyed relatively better socioeconomic status than other indigenous groups. However,
4
Lama (2011) Collective Memory of the Tamang Nation. Common belief among the Tamang peoples
about their place in the Tamang territories. There is no written history of Tamang people because the
Tamangs lost their written scripts when they lost their kingdom.
55
these economic achievements have been at the cost of their linguistic and cultural identity
(Bhattachan & Webster, 2005). The other indigenous groups continue to be excluded from
participating in political process. The executive, legislative and judiciary branches of the
government are controlled by non-indigenous or caste peoples. To date, not a single non-
caste indigenous person has ever become the prime-minister of the country despite the
fact that they constitute population majority. Up until 1999, the Newar, Brahmin, and
Chetri castes jointly held more than 81.7% of the leadership position in executive,
judiciary and legislative branches of the governance of Nepal (Lawoti, 2001). Almost all
of the political parties of Nepal are controlled by the caste groups. The elite castes use
these political organizations to control the affair of indigenous peoples, and to constrain
indigenous peoples’ ability to form their own political parties.
E2. Caste peoples of Nepal (Khas or Aryan group)
The caste, or Khas, peoples of Nepal are the settlers who migrated to Nepal in the
1500s as refugees from the low land, what is currently known as India (Note: Prior to the
1950s, there was no country called India). They are believed to have initially migrated to
the western hills of Nepal--the Parbat district. Hence, they are also known as Parbatia.
Their migration towards the hills is thought to have been driven by the Mogul (Muslim)
invasion of India. Prior to the 1760s, no caste people were found on the east of Gorkha,
including Kathmandu valley. Until the 1950s, caste people were assumed to be less than
7% of the total population of Nepal and mostly concentrated in the far western part of
Nepal. However, their population grew exponentially, and today, they constitute about
32% of the total population (Lama et al, 2010). They are spread throughout the country,
56
except on the mountain regions of Nepal. In some parts of the country, including major
parts of Kathmandu, they have displaced or overpopulated the native peoples. Today,
Brahmin and Chetri, combined, represent 38% of the population of Kathmandu compared
to only 31.8% Newar, and less than 5% Tamang (Subedi, 2010). Prior to the
establishment of Nepal as a nation-state in the 1760s, Newar and Tamang (the indigenous
groups) were the only two native peoples in Kathmandu. Newars were largely
concentrated on the downstream river banks of the Kathmandu valley, while Tamangs
surround the upstream and the hills around Kathmandu valley.
The Khas people are a monolithic group (one language, one religion, one culture,
one race) who practice Hinduism and caste system, (although in recent years, some caste
groups, especially the lower castes have converted to Christianity, Muslim or other
religions in response to caste discrimination). Khas group are divided into four caste
categories- Brahmin, Chetri, Baisya and Suddra. Although the amended version of
Nepalese law, Muluki Ain 1963, prohibits caste-discrimination, it is perceived to be
widely practiced in Nepal. The low or occupational castes are discriminated against in
jobs, education, and other areas of social life (Bhattachan, Sunar, Bhattachan, 2007). On
the top of the caste hierarchy is the Brahmin or the priest group, and at the bottom of the
hierarchy is Sudra or untouchables. They all speak one language, the Khas language (In
recent years, due to one language policy of the government, some indigenous groups,
particularly Newar and those who are young and educated in Nepali schools, speak only
Khas language. Many of the indigenous languages are at the verge of extinct due to the
imposition of Khas language as the official language of the government. Everyone must
learn Khas language in order to advance education, or work for the government.
57
Indigenous peoples are often belittled if they can’t speak Khas language). The Khas
people are a homogenous group in their physical features and cultural practices, and one
cannot identify their caste from their look alone. The easier way to identify their caste is
by their last names or by asking them directly. Common last names of Brahmins and
Chetries include Sharma, Upadhaya, Pandey, Shah, Rana, Thapa, Paudel, Pohkhrel,
Upreti, Panth, KC, among others. Common last names of lower caste people include
Kami, Biswokarma, Damai, Nepali, and Sarki, among others. Figure 2 presents the
distinction between native peoples and caste peoples of Nepal.
Figure 2. Indi genous Peoples and Caste People of Nepal
Native of the Himalayas (Nepal):
Migrants from the lowland (India):
Indigenous Peoples
(Aadibasi Janajati: Mongoloid)
Caste People
(Khas: Ariyan)
58
E3. Brief History of Nepal and its Institutions
Nepal became a nation-state in 1768 AD. However, the history of the movement
to create a Nepali state may be traced further back to the arrival of caste peoples to the
Himalayan nations. The caste peoples are thought to have initially transitioned from
refugees to political prominence in Nepal after they took over a princely nation of Magars
in the Lig Lik Kot, the current Gorkha region of Nepal (the midwestern part of Nepal).
Prior to the taking over of their kingdoms by Khas peoples, Magars had a tradition of
changing kings every year. Each year, during a festival, a new king would be selected
Language
Khas
:
Religion
Hindu
:
Bahun
Chetri
Sudra (Dalit)
Baishya
Language
Each group has distinct languages
:
Religion
Buddhist and local religions
:
Taman
g
Ma
g
a
r
Limbu
Thakali
Sher
p
a
N
ewa
r
Tharu
Gurun
g
Rai
Others
(
50+
)
Malla
Effect
59
based on the overall qualities of that person. One such quality was a physical attribute--
the ability to win a marathon race to the top of a hill. The winner of the race would
become king for a year. Only Magars could qualify for the race. However, in the year
1559, when the Magars were celebrating their festival to coronate their new king to the
throne, one Drabya Shah, a Chetri caste member, pleaded to participate in the marathon.
It is said that Drabay Shah himself was not a physically robust man, but he had the
backing of Brahmins who conspired trickery for him. Initially he was denied participation
because he was a foreign refugee. However, as the Magars were drunk during the
festivities, Drabay Shah took the advantage and participated in the marathon anyway.
While the Magars followed the designated route, Drabya Shah deceptively took a short
cut and won. In Magar tradition, Drabya Shah was crowned king of Magar Nation
(Magaranti) for one year (Encyclopedia Britanica). However, Drabya Shah soon declared
hereditary monarchy, and from that year on, no new king was ever again chosen. Drabya
Shah would be king for life and his son would succeed him. Most of the Magars were
killed or subjugated, their nations dissolved, and their lands and properties confiscated.
Magars became slaves in their own nation.
Drabya Shah, then, declared his stolen nation as Gorkha Kingdom (named after
his ancestral land, Gorakhpur in India, and his patron saint, Gorakh Nath). His great-
grandson, Prithvi Narayan Shah (1723 -1775), is credited with expanding the Gorkha
Kingdom in 1768 to establish what is presently a Nepali state. (When Prithvi Narayan
Shah conquered Kathmandu Valley in 1768, it is said that he ordered his commanders to
bring him twelve baskets of noses, twelve baskets of tongues, and twelve baskets of ears
of the subjugated Newars). For the next 240 years until 2006, the feudal Shah regime was
60
marked by brutal murder, dictatorship, killings, stealing, nepotism and family feuds. The
latest saga emerged recently with reports that in June 2001, the prince supposedly
murdered the whole royal family except his uncle, Gyanendra Shah, and his family.
Gyanendra Shah became king in 2001 after the royal massacre but was deposed in 2006
(see Stiller, 1973; Whelpton, 2005; Shah, 1992 etc. for detailed history of Nepal).
Since the establishment of Nepal as a nation-state in the late 1760s A.D, mostly
Brahmins and some Chetries have become the de facto political elites of Nepal. The kings
of Nepal belonged to the Chetri caste, and they were considered by Hindus as the re-
incarnation of their Hindu god, the Bishnu, even though most Hindus in Nepal practice
Shivaism.
5
The Shah kings were the de facto kings of the Hindus. The Hindu kings tried
to impose the caste system on the indigenous peoples, claiming to be rulers of all Nepali
citizens when in reality the indigenous peoples were marginalized and treated as
secondclass citizens. The attempt by the final king, Gyanendra Shah, to continue as an
invincible Hindu-god king failed, resulting in the abolition of the feudal Hindu monarchy
altogether in 2006. At present, Nepal is a democratic republic. As of February, 2011,
Nepal is in the process of writing a new constitution which will determine the structure
and future direction of Nepal as a nation-state.
During the 240 years of a feudal monarchy and centralized government system,
the property rights of indigenous peoples were under constant threat. Often the fertile
lands of the indigenous peoples were confiscated and handed over to the Khas people
under the system called Birta or under the pretext of land reform.
5
Shiva is the god of destroyer among the three gods in Hinduism- Brahma, Bishnu and Shiva. Brahma is
the god of creation, and Bishnu is the god of protection. Most Hindus of Nepal and India are Shivayats, and
their guru is Sankara Acharya, who lives in India.
61
E3a. Institutions of Nepal
The first formal (written) institution of Nepal, after it became a nation-state or one
kingdom in the current form, was the National Civil Code called Muluki Ain of 1854. The
Muluki Ain, hereafter MA, was promulgated by Rana
6
ruler Junga Bahadur Kunwar
(Rana). Although MA was not exactly a constitution in the modern sense of the term, it
was the primary law of the land by which all peoples of Nepal were judged (Hofer,
2004).
The MA was designed by Brahmins and Chetries who were the de facto political
elites at the time, and are still today. The MA became a source of division among peoples
of Nepal, dividing them into five hierarchical groups (based on Hofer, 2004, p.9):
1. Those who wear the holy cord: not supposed to drink alcohol. Brahmins and
Chetries.
2. Non-enslavable: those for whom culture permits alcoholic drinks (indigenous
group)
3. Enslavable: those for whom culture permits alcoholic drinks (indigenous group)
4. Impure but touchable: low-caste Hindu Newar, Muslims and Europeans
(Brahmins/Chetries do not eat food or drink water from these groups but can be
touched)
5. Untouchable: lower-caste Hindus--blacksmiths, tailors and musicians, cobblers,
fisherman, etc. As a group, they are currently called Dalit or occupational caste.
6
Ranas are a clan of Chetri caste. Ranas ruled Nepal for 104 years (1846 -1950) under the nominal Shah
Kings.
62
(Brahmin/Chetries do not eat food or drink from these groups and cannot be touched).
It is important to note that MA classifies Europeans as one of the lowest group--
impure. It is believed that prior to the 1950s when Nepal was under Rana rule, anyone
who went outside of Nepal and who came in contact with Europeans had to go through
three days of cleansing or purification process before they could enter Nepal. This absurd
classification of people into arbitrary hierarchy clearly reflects the Brahmin/Chetri
prejudice against other races and religions at the time. At best, the first constitution of
Nepal, the Muluki Ain of 1854, may be described as a racist manifesto designed to
dehumanize the indigenous peoples of Nepal. Although this manifesto has been amended
several times since the 1950s, it has made a lasting impact on the psychological and
economic well-being of the indigenous people of Nepal. To date, no empirical studies
have investigated the extent to which people of Nepal are cognizant of this document and
the extent to which it has impacted their socioeconomic well-being.
Among many of the prejudices in this document, one includes prohibition of killing of
Brahmin and cows. The killing of non-Brahmins by Brahmins or Chetries, however, is
permissible, often without consequences.
The MA was amended in 1963 by king Mahendra. Under the amended MA, hereafter
new MA, killing of cows in Nepal is still illegal, with sentence of 12 years or even
lifetime in prison. Prior to promulgation of MA, killing of cows was a common practice
among the native peoples of Nepal. Because of this law, many indigenous peoples are
thought to be in prison for killing cows.
63
Although the new MA was supposed to reflect the aspirations of the people, it
inherited much of the prejudices of the old MA. Among others, the new MA prohibited
indigenous peoples from using their native language as a language of instruction in public
schools. In addition, the new MA continued to declare Khas language as the official
national language of Nepal. This language is currently known as Nepali language. The
government appropriated national resources for Brahmin/Chetries to teach Nepali or
Sanskrit languages in public schools and universities, while it continued to ban the
languages of the native peoples into the 1990s.
In response to the popular democratic movements of the 1990s, the constitution of
Nepal was amended again – more substantially this time. The revisions allow indigenous
peoples to use their mother tongues as a language of instruction in public schools, up to
5th grade. However, the government does not provide funding for the indigenous language
education, while it continues to fund Khas (Nepali) language education.
Prior to the invasion of the Himalayan indigenous kingdoms by the Khas peoples, the
indigenous peoples taught their children in their native languages. The MA, and its
subsequent documents, effectively cut off the system of inter-generational
knowledgetransfer among the indigenous peoples. As a result, many of the indigenous
languages and cultures are on the verge of extinction. Likewise, the MA also prohibits
women from inheriting ancestral property, effectively making women vulnerable to
poverty and abuse. It appears that MA was designed primarily to control economic
mobility of the indigenous peoples, women and lower-caste non-indigenous peoples. The
MA classification of people appears to be the de facto classification of occupations and
division of labor by their ethnicity and castes. The MA reserves the professional jobs
64
(government offices, rulers, army, teachers, etc.) to the first group (Brahmin & Chetries),
and the menial jobs to the last group--the untouchables. The philosophical foundation of
MA may have been derived from Hindu texts such as Manusmriti, and the idea for
population division seems to have been borrowed from the Hindu caste system that the
Brahmins and Chetries brought with them to Nepal, and which they still practice. Under
the Hindu caste system, change of occupation is considered sin. For example, if a cobbler
becomes a teacher or a doctor, it would be considered a sin. Similarly, if a blacksmith
(Kami) or tailor (Damai) becomes a priest or ruler, they have committed sin. The
positions of priest, ruler, teacher, lawyer or doctor are largely controlled by the Brahmins
and Chetries, and it is virtually impossible for the lower-caste Hindus to move to these
positions of power.
For the indigenous peoples who are outside the Hindu-caste system (mostly
Buddhists or other non-Hindu religions) and who are largely farmers and traders, the
access to education is limited. Much of the public educational institutions and educational
curriculum in Nepal have been controlled by Brahmins and Chetries. Curriculum is
taught in Khas language, which only Brahmins/Chetries or Khas peoples can understand.
This provides a distinct advantage to Khas language speakers and puts indigenous
peoples at a disadvantage in educational attainment. It is not clear, however, to what
extent the MA constrains the indigenous people’s ability to accumulate human capital and
move to positions of power.
Recently, since the democratic movements of the 1990s, there has been surge of
private boarding schools in which the medium of instruction is English. Both Khas
language speakers and native language speakers are equally treated in terms of language.
65
However, only a few people can afford to send their children to private boarding schools.
It is worth noting that it is not illegal to study religious texts in Newars, Tibetans,
Tamangs or other native languages in religious institutions such as temples, monastery or
private schools. However it is illegal to do so in public schools. There are no public
schools in which teachers are paid to teach, for example, Buddhist religious texts. On the
other hand, although Sanskrit is a religious text, it is compulsory in public schools and is
fully funded by the government.
As of March 2012, Nepal is in the process of rewriting Nepal’s constitution
altogether. The indigenous peoples continue their struggle against the inherent injustices
in the constitution. They demand a provision in the constitution for the autonomy of their
territories and self-governance over those territories. Whether or not the aspirations of the
indigenous peoples will be reflected in the new constitutions remains to be seen.
E3b. Nepal Geography
At present, Nepal is divided into 14 administrative zones. These 14 zones are
further divided into 75 districts. These political boundaries were designed and created in
the 1960s to disenfranchise indigenous peoples from forming geo-political units among
themselves. During the three decades (1960 – 1990) of the autocratic “partyless”
Panchayat system, political parties were prohibited from carrying out any forms of
political activities. For the purpose of development priorities, Nepal was divided into five
developmental regions: East, Central, West, Mid-west and Far-west regions. These
66
regions were further divided into sub-regions totally 13. The administrative zones and
districts are clustered into one of these sub-regions. These sub-regions serve as the
geographic units of Nepal National Planning Commission, a centralized government
agency which decides the development priorities of these geographic regions. These
subregions provide logical units of geographic analysis.
Ecologically, Nepal has great physically diversity ranging from the plain region
(about 300 meters above sea level) to the highest point on Earth, Mount Everest (8,848
meters above sea level). Based on this ecological diversity, Nepal is divided into three
ecological regions: the Mountain Region, the Hill Region, and the Terai Region (the
plain) (Figure 2). These three parallel each other, from east to west, as continuous
ecological belts, occasionally bisected by the country’s river system. The Terai Region
(plain) is the most fertile for agricultural productivity and is considered the bread basket
of Nepal. The Hill Region is not fertile for agricultural productivity and has limited
economic potential, except for several valleys and river basins. The Mountain Region is
characterized by severe climate and rugged topographic conditions. Although not fertile
for agricultural productivity, it has economic potential for tourism.
These ecological regions and the political administration of these regions have
implications on the economic status of the people who live there. After the establishment
of Nepal as a nation-state, most of the fertile land of the Hill Region (valleys and river
basins) and the Terai Region were confiscated from the native peoples and transferred to
the caste peoples under the system called Birta
7
or under other pretexts such as land
7
Birta is a land grant system by which government officers were allowed to confiscate lands of the
indigenous peoples as a reward for working in remote areas.
67
reform. The native indigenous peoples were pushed further up on the hill and mountain
regions.
Figure 2a. Ecological and Administrative Map of Nepal
Source: United Nation, Nepal Information Platform (http://un.org.np/node/10274)
CHAPTER V: KNOWLEDGE GAP AND CURRENT STUDY
F. Knowledge gap
The review of the literature indicates that while much has been done on global
poverty, little has been done to understand the indigenous poverty throughout the world.
To date, there have been very few studies that explicitly look at poverty among
indigenous peoples vis-à-vis non-indigenous populations. Much of this existing research
on indigenous poverty, however, is descriptive, and some of it is inductive (mostly from
anthropology), but there is little deductive analytical work produced that is applied, that
is, done to test a theory or intervention..
68
Of the previous poverty research on indigenous peoples, most have focused either
on one small indigenous group or on a single geographic community. These studies are
scattered here and there, and they often gloss over the variations that may exist within
various indigenous groups. No known studies have systematically looked at indigenous
poverty using a nationally representative sample or using poverty indicators that are
reflective of the indigenous peoples’ well-being. The extent to which poverty is driven by
the isolation of indigenous territories and the factors that contribute to this isolation are
currently unknown.
Theories of indigenous poverty are underdeveloped; in fact, to date no known
specific theory of indigenous poverty exists. Poverty research on indigenous communities
is fraught with shortcomings. The prevailing theories of poverty are based on the
marketeconomy of industrialized countries. The effort to alleviate poverty has been
hampered by the fact that researchers in this area have tended to take a sectarian approach
to their efforts. There is, therefore, a need for theorizing indigenous poverty based on
indigenous indicators of poverty.
Previous studies on indigenous poverty have used indicators, such as income
(Psacharopoulos & Patrinos, 1994) or consumption (World Bank), that are not reflective
of indigenous well-being (Carino, 2009). The income or consumption-based measures of
poverty have been widely criticized as being limited and narrowly focused (Sherraden,
1991; Sen, 1999; Iceland, 2005; Blank, 2008; Rutstein & Johnson, 2004) or even
misleading for subsistent economies (Carino, 2009). The income or consumption
measure reflects the hedonistic consumer culture of the market-economy rather than the
true well-being of the people (Carino, 2009). In recent years, there is a growing
69
consensus among scholars that any measure of indigenous peoples’ social and economic
status must necessarily start from their own definitions and indicators of poverty
(Eversole, 2005; Carino, 2009). However, there is currently no knowledge of such a
measure.
Much of the empirical work on institutions has been on how institutions emerge
and change (e.g. Ostrom, 1990; Knight, 1992; North, 1990), but little empirical work has
been done to understand how institutions determine the socioeconomic status of various
groups. The pre-1990 constitution of Nepal did not allow indigenous peoples to study in
their native languages (Hofer, 2004). Article 18(2) of the new Constitution of 1990 does
not sanction native language instructions in public schools beyond primary level. The
state does not financially support native language instruction even at the primary level.
On the other hand, the government spends millions of rupees [national currency of Nepal]
for the Sanskrit pathsalas [schools] and the Sanskrit University whose beneficiaries are
male Brahmins (Lawoti, 2001). In addition, by imposing compulsory Sanskrit throughout
Nepal, the state is systematically imposing Hindu values and norms on all communities
of Nepal. The effect of this constraint on the human capital development, and
subsequently on the health and economic well-being of indigenous peoples, is currently
unknown.
F1. Current study
In an effort to begin to address these problems and fill in the knowledge gap, this
study investigates the poverty among indigenous people in Nepal using a nationally
representative sample and indicators of poverty that are native to the indigenous peoples.
70
This study uses the Wealth Index developed by Rutstein & Johnson (2004) as a measure
of poverty. The wealth index is constructed from using asset indicators that are reflective
of indigenous people’s socioeconomic well-being. Use of assets is a well-established
approach to studying poverty (Sherraden, 1991). The asset indicators include land,
livestock, housings, and other household items that are native to subsistent economies.
In the current literature on poverty, there have been two approaches to the
problem. In one approach, poverty has been used as an independent variable. Such an
approach attempts to establish the effect of poverty on other variables such as health,
education or other outcomes. The second approach uses poverty as a dependent variable
and attempts to identify the determinants of poverty, that is, to explain what causes
poverty. This study uses the second approach.
Most studies have used level of education as a predictor of poverty; however, it is
equally plausible that poverty is a predictor of education. Similarly, health has generally
been used as an outcome variable, but in this study, it is hypothesized that health may
actually be a determinant of poverty. In reality, however, the relationship between
poverty, health and education is likely to be endogenous.
This study will test the structural relationship between poverty, health and
education to identify the causal link. It will then test whether this relationship changes by
ethnicity and geography. For some ethnic groups, the relationship between poverty and
education may not be the same as for other ethnic groups. The return on investment in
education for indigenous groups is likely to be different than return on investment for
non-indigenous peoples. For example, there is a wide-spread perception that, due to the
caste system, indigenous peoples are less likely to find jobs in Nepal even if they are
71
equally qualified. However, the empirical validity of these perceptions has not established
to date.
F2. Conceptual model, research questions and hypotheses
Drawing from the theories and literature, Figure 3 presents a conceptual model of
indigenous well-being. In this model, well-being (i.e. human capacity to “live well”- an
indigenous definition) is conceptualized as an inter-related relationship of health,
knowledge and wealth. The level of capacity of an individual and/or family is
conceptualized as a function of the community and the institutions which govern their
communities; in turn, the individual and/or family’s capaicity provides feedback that
affects the institution.
Figure3. Conceptual model of indigenous well-being (capacity)
72
F3. Model explanation:
Indigenous peoples are thought to be poor (have low wealth) because (1) they
lack education (human capital); (2) they have poor health; (3) they live in geographically
isolated areas; and (4) they are treated unequally (discriminated against) by the society in
which they live. They lack education and health (human capital endowment) because
their communities are geographically isolated, and because they are discriminated against
by the society. Geographic isolation and discrimination, in turn, are thought to occur due
to institutions. The institutions are assumed to be designed by, and in the interest of, the
non-indigenous peoples (or Bahun/Chetries, in the case of Nepal). It is hypothesized that,
if these institutions remain unchanged, the indigenous peoples will remain poor even if
they have the same level of human capital or productivity characteristics as the
73
nonindigenous peoples. However, if the institutions are changed such that geo-integration
and equality are maximized to an optimal level, it will improve the education, health and
ultimately the wealth of the indigenous peoples. When the indigenous peoples attain
higher levels of wealth, education and health (i.e. capacity to “live well”), they will, in
turn, be able to design new institutions that best serve the well-being of the people. In an
aggregate, the increased wealth of the indigenous peoples will increase the wealth of the
nation.
The conceptual model may be specified as follows: A.
Institutional Level (Macro):
1. Good institutions (such as constitutions or laws that treat citizens equally and
allow for self-governance of and a multilingual education system) lead to
greater geographic integration (i.e. uniform development of all geographic
communities, eliminating indigenous isolation).
2. Good institutions lead to equality among all caste/ethnic groups (less
discrimination or social exclusion of certain ethnic/caste groups). B.
Community/Group Level (Meso):
1. Geographic integration (less geo-isolation) of communities leads to less disparity
in education.
2. Geographic integration of communities leads to less disparity in health.
3. Geographic integration of communities leads to less disparity in wealth.
4. Equality in treatment of various ethnic/caste groups generates equality in health
among all the ethnic/caste groups.
74
5. Equality in treatment of various ethnic/caste groups generates equality in
education among all the ethnic/caste groups.
6. Equality in treatment of various ethnic/caste groups generates equality in wealth
among all the ethnic/caste groups. C. Individual/Family Level (Micro):
1. More education leads to better health.
2. More education leads to more wealth and vice versa (through changes in
occupational status).
3. Better health leads to more wealth and vice versa.
4. The relationship between education and wealth (poverty) is moderated by
geography.
5. The relationship between health and wealth (poverty) is moderated by
geography.
D. Feedback Loop:
More capacity (more education, better health and more wealth) will, in turn, lead to
advancement of better institutions, and the loop will continue.
Testing all the relationships specified in the conceptual model is beyond the scope of this
study. This dissertation study tests some of the hypothesized relationships which are
described in the section that follows.
F4. Specific Research Questions and Hypotheses
1. Are some ethnic/caste groups in Nepal at significantly higher risk of poverty than
others?
75
H1: Compared to Brahmins (the de facto political elites), all the other caste/ethnic
groups will be at higher risk of poverty.
H1a: The risk of poverty for indigenous peoples as a group will be significantly different
from caste people as a group.
H1b: Within a caste group, people of lower caste will be at significantly higher risk of
poverty than people of higher caste.
H1c: Within an indigenous group, the risk of poverty for some ethnic groups (e.g. Newar,
Gurung) will be significantly higher than for the other ethnic groups (Tharu, Magar,
Tamang).
H1d: Some ethnic groups are as well off as high-caste groups (i.e. the risk of poverty for
some ethnic groups like Newar and Gurun is not significantly different than for people of
higher caste, Brahmin).
H1e: Some ethnic groups are as poor as lower-caste groups (i.e. the risk of poverty for
some ethnic groups such as Tharu, Tamang, and Magar are likely to be as high as for
lower-caste groups).
Literature indicates that there is a cost of being indigenous, that is, indigenous peoples
are at higher risk of poverty than non-indigenous peoples (Psacharopoulos & Patrnos,
1994). However, previous research used to support this claim was conducted in the
countries where non-indigenous peoples were mostly White-European. The extent to
which such relationship exists in other developing countries where non-indigenous
peoples are non-White European is currently unknown. This dissertation study tests the
validity of previous findings under different socio-economic and political contexts.
76
Furthermore, the theory of institutional design (North, 1990) contends that institutions are
deliberately created and designed to serve the interest of the elites of the society. The de
facto political elites of Nepal are largely the Brahmins. This study tests the extent to
which Brahmins are socioeconomically better off than other groups. Previous
studies have documented between-group differences concerning indigenous and non-
indigenous peoples. These studies, however, assumed homogeneity among the indigenous
groups and glossed over the variations that may exist within different indigenous groups.
Since indigenous peoples vary in their cultural, linguistic and geographic contexts, there
is reason to believe that some indigenous groups within a country may be at higher risk of
poverty than other indigenous groups. This study expands the understanding of whether
and how different indigenous groups experience the risk of poverty.
There is a general perception that caste discrimination is widely practiced among
caste peoples in Nepal. The lower-caste (Dalits) are perceived to be socioeconomically
worse off than the rest of the population. This study tests the extent to which there is an
empirical validity to that claim.
The constitution of Nepal (Muluki Ain) categorized indigenous peoples of various
ethnic groups into different social hierarchies solely based on their ethnicities. As a result,
some groups such as Newars, who are largely Hindus, are more likely to be better off than
others.
2. To what extent do the individual productivity characteristics (education, health,
employment, and occupation) determine the risk of poverty and, conversely, how
does poverty influence these characteristics?
H2a: The higher the education, the lower the risk of poverty and vice versa.
77
H2b: The higher the health problems, higher the risk of poverty and vice versa. H2c:
Those who are employed are at lower risk of poverty than those who not not
employed,and vice versa.
H2d: Farmers are at higher risk of poverty than non-farmer.
Previous studies on poverty among indigenous peoples have documented that
indigenous peoples have low education and high health problems (Corina, 2009;
Psacharopoulos & Patrnos, 1994). They also note that, for the most part, indigenous
peoples are employed, working in agriculture. This study expands our understanding of
the relationship between individual productivity characteristics and poverty in Nepal.
An individual’s occupation in Nepal is generally specified by his or her ethnic and
caste identity. Within the indigenous peoples, occupation is a personal choice and largely
determined by one’s level of education and training (except Hindu Newars). Indigenous
peoples work as monks (priests), traders, farmers and other professionals. However,
within the caste peoples, the occupation is generally determined by their castes according
to Hindu religion. The low-caste (Dalits) are supposed to be entertainers, tailors, metal
workers (blacksmiths), cobblers, butchers and other menial jobs such as cleaning or
janitorial works. The mid-caste (Chetri) is supposed to be in the military. The so-called
high-caste (Brahmin) is supposed to be priests. The Chetries and Brahmins, who are not
in the military or the priesthood, are supposed to be farmers (Baishya). The change of
occupation is forbidden in the Hindu caste system--it is considered a sin. Such religious
indoctrination and internalized oppressions discourage the mobility of the low-caste
peoples to better jobs. This study sheds light on the effect of occupational segregation on
poverty in Nepal.
78
3. To what extent do the geographic characteristics of the community determine the
risk of poverty, and vice versa?
H3: Those living in isolated geographic areas are at higher risk of poverty than those
living in non-isolated geographic areas.
This hypothesis is informed by previous literature that suggests that geography is a
key determinant of poverty because different geographies are endowed with different
levels of resources (Diamond, 1997; Sachs, 2001; Rodrik & Subramanian, 2003).
Individuals living in geographies with well-endowed natural resources (such as fertile
land, availability of water, oil, and minerals) are likely to be economically better off than
those living in resource-deprived geographies. More importantly, individuals living in
geographies with poor human-made resources (such as a lack of infrastructure with no
access to roads, electricity, irrigation systems, hospitals, schools or colleges) are likely to
be at higher risk of poverty. Indigenous territories are less likely to have good
infrastructure development due to lack of self-governance (Cornell, 2002). The
constitution (Muluki Ain) of Nepal does not allow self-governance among indigenous
peoples. This constraints the ability of indigenous peoples to self-determine the
development of their communities. The infrastructure development of communities is
generally determined by government, rather than by individual poor who live in those
communities, especially if the government is centralized, like in the case of Nepal.
There is a concern that the relationship between geography and poverty may be
endogenous-- that poverty may cause geographic isolation as much as geographic
isolation may cause poverty. It is possible that poor people move (from urban) to isolated
79
geographies because that is where they can afford to live. In the case of indigenous
peoples, however, it is unlikely that they move to isolated geographies because they are
poor. Historically indigenous peoples have been pushed to the remote/isolated places not
because they were poor, but because they were rich compared to the settlers. It was a
political process. The establishment of Nepal as a nation-state and its constitution (Muliki
Ain) were instrumental in pushing the indigenous peoples to the periphery of
development processes. It is more likely that people become poor because they live in
isolated geographies, rather than that their geographies are isolated because they are poor.
. This study expands our understanding of the relationship between indigenous peoples,
geographic isolation and poverty.
4. To what extent do the differences in individual level characteristics and
geographic characteristics explain the differences in the risk of poverty between
various ethnic/caste groups?
H4: The observed differences in the risk of poverty between various ethnic/caste
groups will disappear when the individual level characteristics and geographic
characteristics are controlled. In other words, if all the ethnic/caste groups were
equal in their individual characteristics (education, health status, employment status,
and occupation) and geographic characteristics, there would not be ethnic/caste
disparity in poverty (wealth).
The Hypothesis 4 is informed by the theory of institutional design (North, 1990).
If the country’s constitution (Muluki Ain) or prevailing social system treats its
citizens equally, the data should support the Hypothesis 4. If, however, the data
80
rejects the Hypothesis 4, it will indicate that the country’s constitution (Muluki Ain)
or social system treats its citizens differentially based on their ethnicity/caste. The
extent to which the ethnic/caste disparity in poverty (wealth) remains unchanged even
when the individual and structural variables (geography) are controlled for, it will
indicate that some ethnic/caste groups will remain poor even if they have same
education, health status, employment, occupation; and live in the similar geographic
areas. This study expands our understandings of the various sources of ethnic/caste
disparity in poverty (wealth); and informs appropriate level of interventions:
individual/family, community or institution.
CHAPTER VI: METHOD
This section presents the research design, data and samples, measures, and
analytical techniques that were used to answer the research questions and test the
hypotheses described in the previous section.
G. Research design
This study utilized a cross-sectional survey design. Nationally representative data
was collected for major indigenous and caste groups of Nepal for one time point. In
particular, this study identified twelve of the largest groups: nine ethnic groups and three
caste groups. The nine ethnic groups, also known as Mongoloids, are collectively called
indigenous peoples. They include Sherpa, Tamang, Magar, Gurung, Rai, Limbu, Thakali,
Newar and Tharu. The three caste groups, also known as Hindu-Aryan or Khas people,
are collectively called non-indigenous people. They include Brahmin, Chetri and Dalit
81
(also known as low-caste or occupational caste).The probability of being poor for each
group was estimated. Brahmin was treated as a reference group.
G1. Data and sample
Secondary data was obtained from the Demographic and Health Survey (DHS
Measures), a global database covering over 80 countries. The data is collected and
managed by Macro International in collaboration with in-country research partners.
According to the DHS Measures (http://www.measuredhs.com), data is collected every
five years for most of the countries. DHS is a public use data available upon written
request. The data has two parts: restricted and unrestricted. The restricted data contains
Geographic Information System (GIS) data and information on HIV/AIDs. For this study,
both restricted (GIS) and unrestricted data for Nepal was obtained.
The Nepal DHS covers a nationally representative sample. The Nepal DHS was
designed to provide current and reliable estimates for the whole country, both urban and
rural, and covers 13 domains obtained by cross-classifying the three ecological zones
(Mountain, Hill and Terai) and five development regions (East, Central, West, Mid-west,
and Far-west). Data includes indicators on key socioeconomic, health and demographic
characteristics of the national population. The Nepal Demographic and Health Survey
was conducted under the aegis of Nepal Ministry of Health and Population (NMOHP)
and implemented by New Era, a local research agency. The field data was collected
between February 2006 and August 2006 by 72 interviewers.
The Nepal DHS collected data on households (N = 8,707), women (N = 10, 793) and
men (N = 4,397) of age 15 -49 years. The sampling design was selected in two stages
82
using stratified and clustered sampling methods. In the first stage, 260 primary sampling
units (PSUs)--82 urban, 178 rural--were selected from the 2001 Population Census
sample frame. In the second stage, systematic sampling of 30 households per PSU in
urban areas and 36 households in rural areas were selected in all regions. Oversampling
was done in urban areas necessitating the weighting of the total sample. Technical details
on data collection methods have been discussed elsewhere (See 2006 NDHS Introduction
and Methodology, n.d.).
In addition to population information, the Nepal DHS 2006 survey also collected data
on location (geography) using a Geographic Positioning System (GPS). The GPS data
includes longitude/latitude coordinates for 260 clusters or primary sampling units (PSU)
in which the sample population resides. The 260 clusters represent the geographic
characteristics of the population in three ecological zones and five development regions.
Each cluster contains a sample size ranging from 18 to 99 people (women sample),
representing the characteristics of the people who reside in that cluster (community). The
GPS data allows for Geographic Information System (GIS) analyses of the sample
population.
This study utilized household, women, and men samples of the Nepal DHS. The age
group of 15-49 is a good fit for economic analysis since people in general are
economically productive during this age range. The study used data for the year 2006
(NDHS 2006), the latest year for which data is available. The study utilized GIS data for
geocoding and analyses of spatial relationships between geography, ethnicity and poverty.
83
DHS data structure
The Nepal DHS data contains 8,707 households. However, the household sample
does not include information on ethnicity or caste of the household. To identify the
ethnicity or caste of the head of household in the household data, the three datasets
(household, men and women) were merged. A unique seven-digit identifier variable was
created, combining the household numbers and cluster number in each of the datasets for
merging. In total, the ethnicity and caste of 7659 households were identified.
To correctly identify the ethnicity or caste of the head of household, first the
ethnicity and caste of each of the household members were identified in both women and
men samples. However, some households (e.g. in women sample) included as many as 30
members, and not all members of the same households were of the same ethnicity or
caste. Furthermore, the respondents were not always the head of the household.
More information was needed to correctly identify the ethnicity or caste of the
households in the household sample. Two additional variables in men and women were
analyzed: relationship structure and residency status. The relationship structure variable
contains information on the relationship of a household member to the head of that
household. The relationship variable identifies whether the household member is a
spouse, child, parent, grand-parent, grand-child, in-law, niece, nephew, “other relative” or
“unrelated” to the head of the household. The residency status variable identifies whether
a member is a usual resident of the household or whether he or she is a visitor.
The ethnicity or caste of the head of household was identified by the ethnicity or
caste of the closest relatives in the relationship structure. For example, if the member was
a spouse of the head of household, the ethnicity or caste of the spouse was assigned to the
84
head of household. If the member was a child of the head of household, his or her caste
was assigned to the head of household over the ethnicity or caste of another member such
as niece or in-laws. However, if the member was not a usual resident of the household or
if the member was “other relative” or “unrelated,” that person was excluded from the
analysis. There were three reasons for this exclusion. First, the unrelated member who
resided in the same household is most likely to be a domestic servant or other workers
(Rustien and Johnson, 2004). So the ethnicity and caste of this member may or may not
be the ethnicity or caste of the head of the household. Second, ethnicity or caste of the
“other relative” who was counted as a member of the household but who was not a usual
resident of the household may or may not be the ethnicity or caste of the head of the
household.
Third, it is unlikely that the household members who were identified as
“unrelated,” “other relative,” or “visitor” are the owners or recipients of the wealth of
their household. For example, although a domestic servant may be a member of a wealthy
household, he or she does not own--and is unlikely to inherit or enjoy--the wealth of the
household. Since this study is designed to understand the poverty/wealth status of the
individuals or household members based on the wealth status of the household, the
inclusion of these individuals in the analysis was thought to present the risk of incorrect
estimates.
Geographic Information System Data
The Geographic Positioning System (GPS) data includes geo-coordinates
85
(longitude/latitude) of geographic clusters or communities. However, this dataset does not
include attributes of the respondents who reside in those geographic clusters. To identify
the geographic location of the respondents, the cluster level GPS data needed to be
merged with the respondent level survey data. The merged data would allow for the
analysis of the relationship between the geography and respondents’ poverty status.
From the survey dataset, cluster-level aggregate data was generated for each of the
PSUs. The aggregated data identified distribution of sample population by poverty and
ethnicity or caste for each of the geographic units. The newly-generated aggregate data
was then merged with GPS data. The GPS data was used to geocode the geographic
clusters and associated respondent attributes using ArcView 10.
In addition to GPS data, another set of GIS base map data was requested of the
Government of Nepal. However, due to the high cost of base map layers (1 layer = N. Rs.
1000 = $14; and 100 layers = $1400), this plan was aborted. Alternatively, district level
and village development committee level GIS data for Nepal was obtained (Courtesy of
Dr. Keshav Bhattarai, professor/interim chair of Geography, University of Central
Missouri). This data identifies the administrative boundaries and was essential to test the
hypotheses on geographic distribution of ethnicity, caste and poverty. This data allows for
the cross-examination of the survey data with the native territories of the indigenous
peoples of Nepal.
Finally, administrative data was collected during the months of June 1 – July 30,
2011; and ethnographic data of an indigenous group was collected by conducting
fieldvisits to the indigenous territories in the High Himalayas during the months of
Ocboter –
86
December, 2011.
Household and Men Samples
The household and men samples do not include information on health. However,
health is an important variable in testing the hypotheses in this study. One approach was
to use only women samples, which includes all the variables proposed in the hypotheses.
However, as cautioned by Prof. David Gillespie (member of the dissertation committee),
exclusion of men and women samples in the analyses would run the risk of bias in favor
of the proposed hypotheses. In an effort to avoid the risk of bias, this study conducted
separate analyses on women, men and household samples. Detailed analyses were
conducted on the women sample as it contains the largest sample size among the three
datasets. The findings were then compared.
G2. Measures
The dependent variable in this study is wealth (or poverty). Based on the research
questions and theoretical framework, three levels of independent variables were used. At
the individual level, ethnicity/caste and productivity characteristics such as education,
health status, employment status and occupation were measured. At the community level,
structural or geographic characteristics (i.e. the degree to which a geographic community
is isolated and the geographic region of resident) are measured. At the institutional level,
laws governing the education system were measured by examining whether or not the
language of instruction in school is a mother tongue. In addition, demographic
87
characteristics such as age, gender, marital status, household size, and the gender of the
head of household were used as control variables.
Poverty, or wealth, is measured using asset indicators. Since indigenous peoples
live largely in subsistent economies of developing countries where cash incomes are
scarce, asset-based measures capture the material well-being of indigenous peoples better
than income or consumption-based measures (Rutstein & Johnso, 2006). Although
cashincome and wage labor market is increasing, indigenous peoples in Nepal are largely
selfemployed farmers who consider land, livestock, houses, and common pool resources
as their valuable assets. Exchange of goods and services in these economies is often
transacted through bartering rather than through cash. Where cash is used, it is of minimal
amount.
This study utilized the asset-based Wealth Index (NDHS 2006 Wealth Index)
developed by Rutstein and Johnson (2006). The Wealth Index is a composite of various
wealth indicators created by using a Principle Component Analysis (PCA) (Rutstein &
Johnson, 2006). The index was constructed using over 30 household assets including
land, houses, livestock, and ownership of household items ranging from a television to a
bicycle or car, as well as dwelling characteristics, such as source of drinking water,
electricity, sanitation facilities and type of material used for flooring, roofing, and walls.
Table 1 presents the list of assets commonly used for constructing wealth index.
Table 1. Typical Assets and Services Indicators used in DHS Wealth Index
1 Land
Housing:
Transportation items:
2 Floor Type
22 Car/Truck
88
3 Wall Type
23 Motorbike
4 Roof Type
24 Tempo
5 Kitchen Type
25 Animal cart
6 Toilet Type
26 Bicycle
Livestock:
Electronic
items:
7 Cow
27
Computer
8 Horse/donkey
28
Refrigerator
9 Goat
29
Phone
10 Sheep
30
Mobile phone
11 Chicken
31
TV
12 Duck
32
Radio
13 Pig
33
Fan
14 Yak
34
Clock
15 Buffalo
35
Bank account
Household items:
36
Pipe water
16 Sofa
37
Electricity
17 Cupboard
18 Chair
19 Table
20 Dhiki
21 Bed
Source: DHS Nepal 2006.
The Wealth Index is a relative measure of wealth rather than an absolute value of
wealth. Values range from 0 (no wealth) to 100 (great wealth). The Wealth Index is
precoded into five quintiles: the bottom 20% ( poorest), 21%-40% (poor); 41% -60%
(middle); 61%-80% (rich); and top 81% -100% ( richest). For this analysis, 40% was used
89
as the cut-off point to define poverty line. The reason for this is that Nepal’s national
poverty rate was estimated around 42% in 2000 (ILO, 2001). Those who fall under the
bottom 40% on wealth index are conceptualized as poor. If a person belongs to the
bottom 40%, he/she was coded as 1 (poor); otherwise, he/she was coded 0 (non-poor).
The probability of being poor is the probability that a person belongs to the bottom 40%
on the asset-based wealth distribution. This is a relative measure of poverty or wealth.
In an absolute term, however, the poor people (those who fall below the bottom
40% of the wealth index) in this measure broadly represent those who live in rudimentary
housing conditions such as mud, sand or dung floors; cane/palm/trunk, mud or sand
walls; and thatch/straw or ceramic tile roof. Also included are those who have no
television, car/truck, motorcycle, computer or refrigerator. Table 2 presents the selected
housing and other characteristics of the poor households.
Table 2. Selected Characteristics of the Poor Households
Indicator
Characteristics
Poor % (Weighted)
Housing Condition
Mud, sand or dung floor
99.31%
Cane/palm/trunk, mud,
sand, bamboo with mud,
stone with mud or wood
planks/shingles wall
97.07%
Thatch/straw, ceramic tiles
or metal roof
96.22%
90
Has no toilet facility, use
bush/field, pit-latrine with
or without slab
97.56%
Modernization/convenience
Television
0.19%
Refrigerator
0%
Car/Truck
0%
Motorcycle/scooter
0%
Computer
0.03%
Ethnicity/Caste identifies the caste and ethnicity of the sample population. This
variable was used to classify the sample population into indigenous and non-indigenous
categories. The variable was used to estimate the degree of inequality in wealth among
indigenous and non-indigenous groups. There are over 60 ethnic groups that are
collectively known as indigenous peoples. Of the 60 ethnic groups, this study identified
nine of the largest groups based on their self-reported ethnicity. Each of the indigenous
groups was then separately dummy-coded into distinct groups. This was done because
although they are all indigenous peoples, they each are considered unique in many ways
(e.g. they each have distinct language, culture and geographic territories). Tamang (the
inhabitants of the Mountain), for example, is in no way the same as Tharu (the inhabitants
of the flatland Terai) although they both are indigenous peoples. As a result, they each are
thought to suffer different socioeconomic disadvantages. Treating them as one group in
the analyses will gloss over the inter-group variation within indigenous peoples. The
largest indigenous groups identified in this study include Sherpa Tamang, Magar, Gurung,
Rai-Limbu, Thakali, Newar and Tharu. The ‘Other’ category consists of small groups
representing over 50 ethnic groups. Sherpa and Thakali each had a small sample size
91
(n<100), and they were included in the “other’ category for the analyses. Each of the
ethnic groups in the “other” category has too small a sample size for separate analyses.
Similarly, there are three caste groups that are collectively described as
nonindigenous peoples. They were identified based on their self-reported castes. The
caste people were dummy-coded into three caste groups: Brahmin (high caste), Chetri
(middle caste including Baishya), and Dalit (the low-caste on the caste hierarchy, also
known as occupational or Sudra caste). Although Dalit and Brahmin belong to the same
Khas group or non-indigenous people, they are thought to be distinct in many
characteristics including socioeconomics. Dalits are known as the oppressed group, or
victims of the caste system, while Brahmins are known as the oppressors, or the
perpetrators of the caste system. Previous studies have reported huge disparities between
these three castes (Bhattachan, Sunar & Bhattachan, 2007). The inclusion of Brahmin,
Chetri, and Dalit into one group is likely to gloss-over the inter-group variation among
the non-indigenous peoples. For this reason, they were coded separately and treated as
distinct groups.
Education measures the level of “human capital” endowment. It is generally
measured in two ways: (1) the number of years of schooling; and (2) the level of
educational attainment. Since drop-out and failing rates in Nepal are very high, the
number of years in schooling is less meaningful in measuring human capital endowment.
(For example, a person who fails 10 times in high school is in school for 20 years, but
may not necessarily have two times more education than his counterpart who only has 10
years of schooling). This study uses educational attainment as a measure of human capital
endowment. Typically, Nepal’s education system divides educational attainment into four
92
levels: no schooling; primary level (up to 5ht grade); secondary level (6 -10th grade but
not SLC); and SLC and above (SLC stands for School Leaving Certificate, which is given
to a student who successfully graduates from the 10th grade. If a student fails SLC, he/she
is not allowed to go for college, and his or her academic future ends here. For this reason,
it is commonly known as an Iron Gate. In 2006, for example, over 62% of the students
did not pass this test (The Government of Nepal, Ministry of Education, Office of the
Controller of Examinations, http://www.soce.gov.np/glance.php). The failing rate is
generally much higher for students in public schools compared to private schools).
Educational attainment is a categorical variable. Since there are only four categories
(levels) in this variable, and since the distance between each level is not symmetrical, this
variable cannot be treated as ordinal level or continuous level data. So it is dummy-coded
for each category, creating four variables to be used in multivariate analyses. Since the
survey has a large sample size, increasing the number of variables may not significantly
constrain degrees of freedom in multivariate analyses. “SLC and above” was treated as a
reference group.
Health, in this study, is measured by Body Mass Index (BMI) of the individual,
presence of anemia, child birth, and child death. BMI uses height and weight to measure
the thinness (malnourishment) or obesity status. BMI is defined as weight in kilograms
divided by height squared in meters (kg/m2) and adjusted for altitude. A cutoff point of
18.5 is used to define thinness or acute undernutrition, and a BMI of 25 or above usually
indicates overweight or obesity. According to the World Health Organization, if more
than 20% of a country’s population has BMI less than 18.5, the county is considered to be
in a serious public health disaster (Ministry of Health, 1998). This study uses 18.5 as the
93
cut-off point to assess the health status of a person. A respondent with BMI less than 18.5
was coded 1 (poor health); otherwise he or she was coded 0 (not in poor health). If the
person had anemia, it was coded 1; no anemia was coded 0. If a woman had given birth,
it was coded 1; otherwise a 0 was given. If the woman had ever had a child die, it was
coded 1 or else a 0.
Employment, or Occupation, measures the productivity characteristics of a person.
In this study, the employment variable was dummy-coded into four categories: farmer
(works on his/her own farm or works on another’s farm), labor (skilled or unskilled),
professional (technical, managers, clerical, sales, services), and “not-working.”
Professionals were treated as a reference category.
Geography is thought to capture two constructs: the extent to which communities
are endowed with resources and the extent to which communities differ in intrinsic
institutions (local laws, norms, cultures). Traditionally, studies have used urban vs. rural
differences in their geographic analyses. However, such analyses gloss over the huge
variations that may exist within urban or rural areas.
This study utilizes the Geo-Positioning System (GPS) to identify the degree of
geographic isolation of the small communities (neighborhoods) in which the sample
populations reside. The DHS 2006 survey collected GPS data from 260 geographic
communities representing the geographic diversity of Nepal, both in terms of natural
resources endowment and levels of development priorities of the Nepal government (i.e.
the five development regions and 13 sub-development regions). In this study, the 260
communities are classified into four categories of geographic isolation based on the
degree of their infra-structure development: (1) developed (capital city); (2) moderately
94
developed (small city); (3) less developed (town); and (4) isolated (not developed). The
degree of geographic isolation measures the degree of resource deprivation (community
poverty), such as lack of markets, roads, electricity, telecommunication, hospitals,
college/universities and so on. The 13 sub-regions measure the degree of isolation in
terms of government priorities and natural resource endowments.
Institutions are formal and informal rules or laws of the country. This study
measures the formal rule that governs the education system in Nepal and the informal
rule, the caste system. The national law of Nepal, Muluki Ain 1965, prohibits the use of
indigenous language as a language of instruction in public schools. Only the Nepali
language is permitted as a language of instruction in public schools. If the mother-tongue
of a respondent was Nepali, it was coded as 1. If the mother-tongue was an indigenous
language, it was coded as 0. The caste system, the informal social norm, was measured by
the caste of the person.
Since the DHS survey did not ask questions on the language of instruction in
public schools, it was not clear, at the individual level, whether or not the respondent
received education in his or her mother tongue. However, since the official language of
instruction in public schools in Nepal is Khas language, all the Khas or non-indigenous
people (Brahmin, Chetri and Dalit) were assumed to have learned in their mother tongue,
and all the indigenous peoples (including Newar) were assumed to have been prevented
from receiving education in their mother tongue.
One concern with this coding scheme is that, with increasing Khasinization of
Nepal’s population, many indigenous groups may have lost their mother tongue.
Although not their original language, Khas may be the only language they can speak.
95
This coding scheme will put them into a non-Khas speaking category.
Another concern is that some indigenous groups, particularly Newar, may have
been educated in their mother tongue. The majority of the Newar live in Kathmandu
valley, which when established as the capital of the nation, provided economic and other
advantages for them over other groups. As a result, they are thought to have sustained the
education system in their own mother tongue while other indigenous groups could not
due to increasing Khasinization of the education system of the country. The extent to
which Newar were taught in their native language is currently unknown, however. The
coding scheme in this study assumes that Newar as a group did not have education in
their own mother tongue.
Finally, there is also a concern that indigenous children who go to private
boarding schools may not be linguistically disadvantaged than the Khas children who also
go to private schools because the language of instruction in these schools is generally
English. Since the DHS survey does not ask about the types of schools they attended, the
language of instruction cannot be identified. The coding system assumes that the primary
language of instruction in all the schools is Khas language.
The institution variable, therefore, is only a proxy measure of institution
governing the language of instruction in public schools in Nepal. The coding scheme is
likely to underestimate the proportion of indigenous peoples whose mother tongue is the
language of instruction in schools. Please note that this coding scheme is identical to the
coding scheme applied to measure Khas people as a non-indigenous category and
therefore measures the same underlying construct.
96
Table 2a below presents the summary description of the variables that will be
used in this study.
Table 2a. List of variables, measures and codes
Variable/Construct
Measures/Instrument
Code
Dependent variables:
Poverty
Asset-based measure of
poverty/wealth.
Measures whether or not
a person is relatively
poor. DHS Wealth Index
was used.
1 = Poor (Bottom 40% of Wealth
Index)
0 = Non Poor (Top 60% of Wealth
Index)
Individual level predictors:
Ethnicity
Identifies the
ethnicity/caste of a
person.
Indigenous Group:
1 = Rai & Limbu, 0 = Not Rai &
Limbu
1 = Magar, 0 = Not Magar
1 = Tharu, 0 = Not Tharu
1 = Tamang, 0 = Not Tamang
1 = Newar, 0 = Not Newar
1 = Gurung, 0 = Not Gurung
1 = Sherpa, 0 = not Sherpa
1 = Thakali, 0 =, Thakali
1 = Other indigenous groups, 0 =
Not “Other indigenous groups’
Non-Indigenous Group:
1 = Brahmin, 0 = Not Brahmin
1 = Chetri, 0 = Not Chetri
1 = Dalits, 0 = Not Dalits
Education
Level of school
attainment. It is a proxy
measure of ‘human
capital’ endowment.
1 = No Education,
0 = Not ‘No Education’
1 = Primary education
0 = No primary education
1 = Up to secondary education,
0 = Not ‘Up to secondary
97
education’
1 = More than secondary education
0 = Not ‘More than secondary
education’
Health
Measures the health
status of a person;
Measures whether or not
the respondent ever have
child births and child
death
1 = Poor health (BMI =<.18.5)
0 = Good health (BMI>18.5)
1 = child births
0 = no child births
1= child deaths
0= no child deaths
Employment
/Occupation
Whether employed or
not and types of
occupation
1 = Farmer, 0 = Not farmer
1 = Laborer, 0 = Not laborer 1
= Professionals, 0 = Not
professionals
1 = Not working, 0 = Not “not
working”
Community level (structural) predictors:
Geographic Isolation
Measures the degree to
which communities are
isolated or poor.
1 = isolated (country side)
0 = not isolated (not countryside)
1 = developed (capital city)
0 = not developed (not capital city)
1 = moderately developed (small
city)
0 = not ‘moderately developed
(not small city)
1 = less developed (town)
0 = not ‘less developed’ (not town)
Geographic regions
Measures the degree to
which geographies vary
by development
priorities of the
government
1= Eastern Mountain
2 = Central Mountain
3 = Western Mountain
4 = Eastern Hill
5 = Central Hill ®
6= Western Hill
7 = Mid-western Hill
8 = Far-western Hill
9 = Eastern Terai
10 = Central Terai
11 = Western Terai
13 = Mid-western Terai
14 =Far-western Terai
Institutional (structural) predictors:
98
Institution
Measures whether the
mother tongue is Khas
language (the official
1 = Mother tongue is Khas language
0 = Mother tongue is not Khas
language
language)
Demographics:
Age
Measured in years
Continuous variable
Gender
Identifies whether the
person is male or female
1 = female
0 = male
Marital status
Identifies whether the
person is married or not
1 = married
0 = not-married
Household size
Number of people in a
household
1 = Large household>6 members 0
= Not large household =<6
rmembers
Gender of Head of
household
Identifies the gender of
the head of household
1 = female
0 = male
G3. Analytical techniques
First, univariate analyses were conducted to understand the sample characteristics
of the study population, and to assess the overall data distribution of each of the variables
in the study. Descriptive statistics were produced to describe each of the variables. For
categorical variables, weighted percents and unweighted frequency distributions were
provided (Table 2). For continuous variables, weighted means and standard deviation
were provided. Furthermore, for the continuous variables, assumptions of normality of
distribution were checked, skewness and data outliers were identified. Where the
normality assumption was violated, the data was transformed using an appropriate
method (such as recoding, log, square root or other types of transformation method).
Missing data was checked and verified. For the data missing at random, missing values
were imputed using an appropriate method (such as Multiple Imputation). For any data
that were not missing at random, the values were excluded from the analyses.
99
Second, bivariate analyses were conducted between the dependent variable
(poverty) and each of the independent variables to determine the strength and direction of
the relationship at bi-variate level. For the categorical variables, Chi-Square tests were
used. For example, Chi-square tests between ethnicity/caste (categorical variable) and
poverty (categorical variable) indicated whether poverty was significantly associated with
ethnicity/caste. In addition, it also produced poverty rates for each of the ethnic/caste
groups.
Similarly, Chi-square tests between education and poverty produced poverty rates
for each of the educational groups and indicated whether education was significantly
associated with poverty. Chi-square tests between health and poverty indicated whether
health status (good health vs. poor health) was significantly associated with poverty.
ChiSquare tests between occupation and poverty produced poverty rates for each of the
occupational types, and indicated whether occupational type was significantly associated
with poverty. Furthermore, Chi-square tests between geography and poverty indicated
which geographic characteristics were significantly associated with poverty. Poverty rates
for each of the geographic sub-regions were produced. Chi-square tests between
demographic characteristics (categorical) and poverty indicated which demographic
characteristics were significantly associated with poverty. The independent or control
variables that were not significantly associated with dependent variables at the bi-variate
level were excluded from further analyses (i.e. multivariate analyses).
Furthermore, to gain a better understanding of the characteristics of the
ethnic/caste groups and their association with poverty, bi-variate analyses were conducted
between ethnicity/caste and each the variables in the study. For example, bivariate
100
analyses between ethnicity/caste and education produced statistics on the educational
status of each of the ethnic/caste groups. Similarly, bivariate analyses between
ethnicity/caste and health status provided health characteristics of each of the ethnic/caste
groups. Bi-variate analyses between ethnicity/caste and occupation provided occupational
characteristics of each of the ethnic/caste groups. In addition, Bivariate analyses between
ethnicity/caste and geography provided geographic characteristics (e.g. geographic
clustering) of each of the ethnic/caste groups. Bivariate analyses between ethnicity/caste
and demographics provided demographic characteristics of each of the ethnic/caste
groups.
Finally multivariate regressions were conducted to test if the observed
characteristics of the ethnic/caste groups were associated with their poverty. Since the
dependent variable, poverty, was binary, binomial multivariate logistic regressions were
conducted.
In particular, to answer the research questions and test each of the hypotheses in
this study, the following analytical techniques were used:
1. To answer Research Question #1 (Are some ethnic/caste groups in Nepal at
significantly higher risk of poverty than others?) and test the hypotheses H1 – H1e,
bivariate analyses were conducted. Since both the dependent variable (poverty) and
independent variable (ethnicity/caste) were categorical, bivariate Chi-square tests were
conducted. The Chi-square tests produced poverty rates for each of the ethnic/caste
groups and tested the extent to which ethnicity/caste and poverty were associated at
bivariate levels (i.e. without controlling for other variables in the study). The result of the
Chi-Square tests showed which ethnic/caste groups were at the highest risk of poverty
101
and which were at the lowest risk of poverty as indicated by the proportion of poor within
each of the groups. The ethnic/caste groups were then rank-ordered based on their level
of poverty.
Multivariate logistic regressions were conducted to test the relative contribution of
each of the independent variables to the risk of poverty for each of the ethnic/caste groups
compared to the reference group. The dependent variable, poverty, was regressed on each
of the independent variables in the study. Odds ratios and confidence limits were
observed. The results of the multivariate regressions showed which ethnic/caste groups
were at higher risk of poverty, and whether or not the risk of poverty was significantly
different for each of the ethnic/caste groups compared to the reference group.
2. To answer Research Question # 2 (Do some ethnic/caste groups in Nepal have
significantly lower education and health than other groups?) and test the hypotheses H2a
–H2f, Chi-square tests were conducted. The Chi-square tests between ethnicity/caste and
educational status (categorical variable) indicated which ethnic/caste groups have less
education and whether or not the ethnic/caste groups are significantly different from each
other in their education. Similarly, the Chi-Square tests between ethnicity/caste and
health status indicated which groups have a lower health status and whether or not the
ethnic/caste groups are significantly different from each other in health status.
3. Finally, to answer Research Questions #3- #5 (3. To what extent do the individual
productivity characteristics {education, health, employment, and occupation} determine
the risk of poverty? 4. To what extent do the geographic characteristics of the community
102
determine the risk of poverty? 5. To what extent do the differences in individual level
characteristics and geographic characteristics explain the differences in the risk of
poverty between various ethnic/caste groups?) and to test each of the hypotheses
associated with these questions, step-wise multilevel (pooled method) multivariate
logistic regressions were conducted. This model may be expressed by the following
equation:
P(Y =1|ethnicity) = β1 demographics + β2 productivity characteristics + β3 geography +
β4 Institution +e; where Y = poverty.
In this technique, three models were tested. In the first model (M1), odds of
poverty (odds ratios) were determined for each of the ethnic/caste groups by controlling
for only demographic characteristics. This test investigated the extent to which the
observed inequality in poverty (wealth) between ethnic/caste groups was driven by mean
level differences in their demographic characteristics (Research Question #1). Fixed
effect of each of the demographic variables on poverty was tested.
In the second model (M2), individual productivity variables (education, health,
employment, and occupation) were added to the regression equation; and changes in the
odds ratios for each of the ethnic/caste groups were noted. This test determined the extent
to which observed inequality in poverty (wealth) between ethnic/caste groups were driven
by the mean level differences in productivity characteristics of the peoples
(random effects), controlling for the demographic characteristics (Research Question #3).
The fixed effects of each of the individual productivity characteristics on poverty were
noted as indicated by their odds ratios.
103
And finally, in the third model (M3), geographic variables were added to the
regression equation. The changes in the odds ratios for each of the ethnic/caste groups
were observed. This test determined the extent to which observed inequality in poverty
(wealth) between various ethnic/caste groups are a function of structural variables
(geographic isolation), controlling for the demographic and individual productivity
characteristics (Research Questions #4 and 5). Fixed effects of each of the geographic
variables on poverty were noted.
If the odds of being poor remained significantly different for each of the
ethnic/caste groups in Model 3 (that is, even after controlling for demographics,
individual productivity, geographic characteristics), then some unobserved forces are
assumed to be driving the inequality in poverty (wealth) between these ethnic/caste
groups. These unobserved forces are assumed to be institutions, such as a constitution,
that treat individuals differentially based on their ethnicity or caste. One proxy measure of
such institutions is the laws governing use of language in public offices including
schools.
To test the extent to which language (a proxy measure of institution) is associated
with the observed differential risks of poverty across ethnic/caste groups; language was
added to the regression equation (M4). The changes in the odds ratios for each of the
ethnic/caste groups were observed.
Furthermore, to better understand the relationship between ethnicity/caste and
poverty, this study further analyzed spatial relationships using the Geographic
Information System (GIS). GIS displayed the geographic distribution of ethnicity/caste
and poverty and visually displayed spatial relationships which were otherwise not
104
captured in the traditional statistical methods. For example, GIS provided information to
determine if the relationship between poverty and ethnicity is the same in place “A” as
and in place “B”. GIS also helped to determine if the relationship between poverty and
health vary by the geographic location in which the residents live. This data helped to
clarify the probability of an indigenous person being poor if he/she lives in Kathmandu
(capital city) vs. Sindhupalchok (small village) compared to non-indigenous person.
To rule out the bias due to sampling, the multivariate regression procedures were
repeated for men and household samples. The results were compared.
SAS version 9.2 was used to manage the data, merging women, men and
household datasets, generate and recode new variables, run univariate, bivariate and
multivariate analyses. ArcView 10 was used to geocode GPS data, perform Geographic
Information System (GIS) analyses and produce maps.
105
CHAPTER VII: RESULTS
This section provides the findings of the study. First, descriptive results from the
univariate analyses are presented (Table 3). Descriptive statistics include characteristics
of the overall sample population. Next, results of the bivariate analyses are presented
(Tables 4 – 11). Bivariate analyses provide an early exploration of the hypothesized
relationships between the variables of interest in this study. The results of the multivariate
regression analyses are presented next (Tables 12 –16). The multivariate results provide
detailed findings on each of the hypotheses tested in this study. Analyses are then reported
from the Geographic Information System (Figures 4 –14). Finally, the multivariate
regression results from the women samples are compared to those from household and
men samples (Tables 18 –19). Additionally, analyses of Tamang and
Brahmin sub-samples and other ethnic groups are provided (Tables 17, and 21-27).
1. Univariate Analyses: Characteristics of the Sample Population (Women sample,
N = 9836)
Poverty
Table 3 presents the description of the characteristics of the women sample (N =
9836). Overall, 38.34% of the sample population were estimated to be poor (i.e. fall
below the bottom 38.34% on wealth distribution). The percent distributions are weighted
to represent the national population. The probability of being poor is the probability that
an individual falls below 38.34% on wealth distribution.
106
Ethnicity and Caste
Indigenous peoples as a group represented 56.35% of the women sample. Among
the indigenous peoples, 11.14% were Tharu, 6.59% were Magar, 5.53% were Tamang,
4.44% were Rai and Limbu (hereafter Rai-Limbu), 2.94% Gurung, less than 1% Sherpa,
less than 1% Thakali, and 20.55% other indigenous groups (representing more than 50
indigenous groups). Since the sample size for Sherpa and Thakali women was very small,
they were included in the “other indigenous group” category in subsequent analyses.
The caste people as a group represented 43.66% of the women sample. Of this,
12.89% were Brahmin, 19.07% were Chetri and 11.7% were Dalit. (also known as
lowcastes, occupational castes or untouchables within Hindu caste system).
Compared to national estimates, the indigenous group appears to be
underrepresented in the sample. Some estimates suggest that indigenous peoples in Nepal
constitute over 70% of the national population (Leslie et al, 2010).
Education
More than half (54.07%) of the women in this study reported that they did not
have any education. 17.65% reported having only primary level education (up to 5th
grade), 19.85% reported having only secondary education (up to 10th grade but no School
Leaving Certificate). Only 8.43% reported having post-secondary education (SLC and
above).
Health
Almost a quarter (23.79%) of the sample population was underweight (BMI
=<18.5); and more than one third (35.57%) have suffered from anemia. The majority of
the women (74.1%) had given births to at least one child, with 21.21% of the women
107
experiencing the death of at least one child. On average a woman gave birth to 2.56
children (SD = 2.32), and the average child deaths per women was 0.33 (SD = 0.76).
Some women had up to eight child deaths.
According to the World Health Organization, if more than 20% of a country’s
population has a BMI less than 18.5, the county is considered to be in a serious public
health disaster. The findings indicate that Nepal clearly has a serious public health
problem.
Occupation
The majority of the women (70.66%) in the study sample self-identified as
farmers, 8.9% reported to be professionals (technical, managers, clerical, sales, services),
3.72% self-identified as laborers (skilled or unskilled), and 17.52% reported as
“notworking.”
Geographic Isolation
The majority of the women in the sample (82.92%) were from the geographically
isolated areas (under-developed, rural areas with few roads, limited electricity, piped
water, etc.). Only 5.08% were from the developed areas (the capital city), 7.16% were
from moderately developed areas (small cities), and 4.83% were from less developed
areas (small towns).
Development Region and Ecological Zone
In regards to development regions, 22.12% of the respondents in the women
sample were from the Eastern Development Region; 33.23% were from Central
Development Region; 22.53% were from Western Development Region; 13.42% were
from the Mid-Western Development region; and 11.64% were from the Far-Western
108
Development Region. In terms of ecological zone, half of the respondents in the women
sample (49.55%) were from the Terai, 43.21% were from the Hill, and 7.25% were from
the Mountain area.
To better understand the geographic distribution of the data, the sample was
further disaggregated to sub-development regions by cross-classifying the development
regions and ecological zones. This was necessary because it is thought that variations
exist within a development region and within an ecological zone. For example, within the
Central Development region, characteristics of the people living in Central Tarai are less
likely to be similar to those living in Central Hill or Central Mountain. The inhabitants of
Central Tarai are likely to be Tharus, whereas the inhabitants of Central Hill are likely to
be Tamangs and Newars, and that of the Central Mountain are likely to be Tamangs and
Sherpas.
Furthermore, the ecological variation within a development region is likely to
have differential consequences on the livelihood of the people who reside in those
ecological zones. People living in flatland fertile Central Terai, for example, are likely to
benefit from agricultural productivity, whereas such opportunities are less available for
those living in the dry and rugged terrain of the Central Hill or Central Mountain.
Similarly, within an ecological zone (e.g. the Hill ecological zone), the condition of
people living in one development region (e.g. Central Hill) are less likely to be similar to
those living in another region of the same ecological zone (e.g. the Far-western Hill). The
Central Hill is considered most developed whereas Far-western Hill is considered the
least developed. The variation in the level of development within an ecological zone is
likely to have different consequences on the livelihood of those who reside in different
109
development regions within an ecological zone. Analyses of the development region or
ecological zone alone are likely to gloss over the variations that may exist within a
development region or an ecological zone. Analyses at the sub-development regions seem
imperative to better understand the geographic effects.
When the data was further disaggregated to the sub-development regions, the
largest sample came from Central-Hill (16.06%), followed by Central Terai (15.22%),
Eastern-Terai (14.46%) and Western-Hill (11.84%). The rest of the sub-development
regions each have less than 10% of the sample population. Eastern-Mountain and
Central-Mountain each represent less than 2% of the sample population (Table, 3, Map
4).
Language Policy (Mother Tongue)
The mother tongues of all the indigenous peoples in Nepal are regarded as
nonofficial languages. In the women sample, they represent 56.34%. Please note that the
distribution of this variable is the same as the distribution of indigenous peoples as a
group.
Khas language is the official language of Nepal, and it is the mother tongue of the
caste groups, namely Brahmin, Chetri, and Dalit. The de facto language of instruction in
public schools, in government offices, and in pubic media is Khas language. The
language policy variable is thought to capture institutions that constrain the indigenous
peoples’ ability to accumulate human capital, to access public information, and to
advance indigenous language, culture and education.
Distance to health facility: Two-fifths (40.6%) of the women reported that distance to a
health facility is a major problem.
110
Access to Piped Drinking Water: Only 13.76% of the women reported having piped
drinking water to their dwellings or yard.
Access to Electricity: Almost half the women (48.72%) did not have access to electricity.
Geographic Isolation, Development Region, Language Policy, Distance to health facility,
access to piped drinking water and access to electricity are proxy measures of institutions.
Since distance to a health facility, access to piped drinking water and access to electricity
were included in the construction of Wealth Index, they were excluded in the regression
analyses.
Demographics
Over three quarters of the women studied (76.8%) were married (or living with a
partner); 21.31% of the women lived in households headed by a female; 34.61% of the
women lived in large households (i.e. households with seven or more members. The
average household size was 6.06, SD = 3.04). The average age of the women was 29.15
years (SD = 9.84).
-----------------------------------------------------------------------------------------------------
Insert table 3 about here
-----------------------------------------------------------------------------------------------------
2. Characteristics of the Indigenous Peoples and Caste People (Women sample,
111
N = 9836)
Ethnicity, Caste and Poverty
Table 4 presents the distribution of poverty by ethnicity and caste. A huge disparity in
wealth (poverty) exists across ethnic and caste groups. Over half of the Tamang and Dalit
women in the sample were impoverished compared to only one-fifth of the Newar and
Brahmin women.
Among the indigenous women, the poverty rate (the proportion of women who fall
below bottom 40% on the Wealth Index) was highest among Tamang (51.41%). The rate
of poverty was 48.99% for Magar, 44.67% for Tharu, 37.81% for Rai-Limbu, 25.76% for
Gurung, 20.62% for Newar and 33.5% for those classified as “other ethnic group.”
Among the caste women, the poverty rate was highest among Dalit (52.44%). Poverty
rate was 43.64% for Chetri, and 19.36% for Brahmin.
Overall, the poverty rate was highest among Dalit (52.44%) and Tamang (51.41%),
and lowest among Brahmin (19.36%) and Newar (20.62%). The largest disparity was
within the caste group. Poverty among Dalit (52.44%) was more than two and half times
the poverty among Brahmin (19.36%). Disparity within the indigenous groups was
slightly less. Poverty among Tamang (51.41%) was slightly less than two and half times
the poverty among Newar (20.62%). The differences in poverty rates across ethnic and
caste groups were statistically significant (χ2=495.66, p<.0001) -------------------------------
---------------------------------------------------------------------- Insert table 4 about here
-----------------------------------------------------------------------------------------------------
112
Ethnicity, Caste and Education
Table 5 presents the distribution of education by ethnicity and caste. Huge
disparity was observed in education across ethnic and caste groups. A large proportion of
indigenous women had no education. Very few indigenous women had post-secondary
education. Similarly,very few Dalit women had post-secondary education. In contrast,
among Brahmin the proportion of women with no education was relatively small but the
proportion of women with post-secondary education was relatively large.
Overall, post-secondary education rates varied from 1.26% (Dalit) to 26.34%
(Brahmin). The proportion of women with no education was highest among the “other
ethnic group” (69.91%) and lowest among Brahmin (27.53%). Within the indigenous
group, post-secondary education rate was lowest among Tharu (1.99%) and highest
among Newar (18.65). Within the caste group, post-secondary rate was lowest among
Dalit (1.26%) and highest among Brahmin (26.34%). The disparity in education was
greater within the caste group than within the indigenous group or between indigenous
and caste groups. The caste group appears to be more heterogeneous in education than the
indigenous group.
Among the indigenous women, 65.41% of the Tharu had no education, 13.3% had
only primary education, 19.3% had secondary education, and 1.99% had post-secondary
education. Among the Tamang women, 63.05% had no education, 17.39% had only
primary education, 13.77% had secondary education, and 5.79% had post-secondary
education. Among the Magar, 55.07% had no education, 23.41% had only primary
education, 17.75% had secondary education, and 3.77% had post-secondary education.
Among Rai-Limbu , 36.18% had no education, 23.92% had only primary education,
113
29.6% had secondary education, and 10.3% post-secondary education. Among Gurung
women, 34.95% had no education, 24.88% had only primary education, 30.79% had
secondary education, and 9.38% had post-secondary education. Among the Newar,
33.62% had no education, 21.17% had only primary education, 26.56% had secondary
education, and 18.65% had post-secondary education. Among the “other ethnic group”,
69.91% had no education, 15.1% had only primary education, 10.9% secondary
education, and only 4.09% had post-secondary education.
Among the caste women, a large proportion of Dalit women, 67.69%, had no
education, 20.5% had only primary education, 10.54% had secondary education and only
1.26% had post-secondary education. Among Chetri, 48.26% had no education, 18.2%
had only primary education, 24.79% had secondary education, and 8.75% had
postsecondary education. Among Brahmin, 27.53% had no education, 14.38% only
primary education, 31.75% had secondary education, and 26.34% had post-secondary
education.
Among the indigenous groups, Newar appears to be an outlier in the distribution
of education. Among the caste group, Dalit appears to be an outlier. In terms of education,
Newars appear to be more similar to Brahmins than to their indigenous cousins, and
Dalits appear to be more similar to Tharu than to their caste cousins. The inclusion of
Newars in indigenous category and Dalits in caste category may produce biased estimates
in multivariate analyses. -----------------------------------------------------------------------------
------------------------ Insert table 5 about here
-----------------------------------------------------------------------------------------------------
114
Ethnicity, Caste and Health
Table 6 presents the distribution of health by ethnicity and caste.
Body Mass Index (Underweight): Huge variation was observed in BMI (Body
Mass Index) across indigenous and caste groups. The underweight (BMI=<18.5) rates
varied from 6.16% (Gurung) to 34.61% (Other ethnic group). Over one-third of the Tharu
women (34.36%) were underweight (BMI=<18.5). Underweight rate among Tamang was
11.15%, among Newar was 10.71%, among Magar was 8.82%, among Rai-Limbu was
7.59%, among Gurung was 6.16%, and among “other ethnic group” was 34.64%.
Similarly, among the caste people, 32.74% of Dalit women were underweight, 19.98% of
Chetri women were underweight, and 21.65% of Brahmin women were underweight.
Anemia: Significant variation was observed in anemia levels across indigenous
and caste groups. The anemia rates varied from 15.49% (Rai-Limbu) to 72.21% (Tharu).
Of those in the sample, almost three-quarters of the Tharu women (72.21%) were anemic.
Almost one-third of Tamang women (30.86%) were anemic. Simlarly, 16.78% Newar,
24.26% Magar, 15.49% Rai-Limbu, 22.69% Gurung, and 41.66% of “other ethnic
group” were anemic. Among the caste group, over one-third of Dalit women (35.05%),
32.43% Brahmin and 26.09% Chetri were anemic.
Child births: Little variation was observed in births among women of indigenous
and caste groups. Among the indigenous women, 73.03% Tamang, 71.7% Newar, 71.66%
Magar, 70.81% Tharu, 69.1% Gurung, 65.68% Rai-Limbu, and 77.98% “other ethnic
group” women had given birth to at least one child. Among the caste women, 81.51%
Dalit, 73.43% Chetri, and 71.42% Brahmin women had given birth to at least one child.
Overall, the proportion of women who had given birth was highest among the Dalit
115
(81.51%) and the lowest among Rai-Limbu (65.68%).
Child deaths: Variations was also observed in child deaths among women of
indigenous and caste groups. Among the indigenous women, 23.58% Tharu, 23.35%
Magar, 19.3% Tamang, 17.8% Gurung, 16.78% Rai-Limbu, 13.62% Newar, and 25.54%
“other ethnic group” women experienced the death of at least one child. Among the caste
women, 27.59% Dalit, 20.2% Chetri, and 12.32% Brahmin had at least one child die.
As high as 4.1% of Tamang, 3.44% Tharu and 3.21% Magar women had three or
more child deaths compared to only 1.57% Brahmin.
-----------------------------------------------------------------------------------------------------
Insert table 6 about here
-----------------------------------------------------------------------------------------------------
Ethnicity, Caste and Occupation
Table 7 presents the distribution of occupation by ethnicity and caste. Types of
occupation vary across indigenous and caste groups. The majority of the indigenous and
caste women sampled were farmers. Overall, the highest proportion of women
professionals was among the Newar (22.34%), followed by Brahmin (14.01%) and
Gurung (13.02%). Tharu women had the least proportion of professionals (2.4%) but
highest proportion of farmers (84.79%). Very few of the women worked as laborers
(skilled and unskilled). Newar had the highest proportion of women who worked as
laborers (12.63%), followed by Rai-Limbu (5.82%). Over 20% of Brahmin and nearly
116
30% of “other ethnic group” women were not working compared to only 8.97% Magar
women who were not working.
Among the Tharu women, 84.79% were farmers, 2.4% were professionals, 2.04%
were laborers, 10.77% were not working. Among the Magar women, 82.67% were
farmer, 4.98% were professionals, 3.38% were laborer, 8.97% were not working. Among
the Tamang women, 76.1% were farmer, 7.93% were professional, 3.91% were laborer,
12.07% were not working. Among the Gurung women, 69.41% were farmer, 13.02%
were professional, 1.54% were laborer, and 16.03% were not working. Among the
RaiLimbu women, 64.47% were farmer, 12.66% professional, 5.82% were laborer, and
17.05% were not working. Among the Newar, 46.34% were farmer, 22.34% were
professionals, 12.63% were laborer, and 18.69% were not working. Among the women in
‘other ethnic group’, 58.79% were farmer, 5.93% were professionals, 5.3% were laborer,
29.98% were not working.
Among the Dalit caste women, 79.45% were farmer, 7.93% were professionals,
2.14% were laborers, and 10.47% were not working. Among the Chetri caste women,
75.13% were farmers, 4.69% were professionals, 3.7% were laborers, and 16.49% were
not working. Among the Brahmin caste women, 63.26% were farmer,s 14.01% were
professionals, 1.74% were laborers, and 20.99% were not working. --------------------------
--------------------------------------------------------------------------- Insert table 7 about here
-----------------------------------------------------------------------------------------------------
117
Ethnicity, Caste and Geographic Isolation
Table 8 presents the distribution of ethnicity and caste by geography. A large
proportion of women of both indigenous and caste groups live in the countryside
(geographically isolated areas). Overall, the largest proportion of Tharu women live in the
country (94.7%), followed by Magar (90.53%). The largest proportions of women who
live in the capital city, small cities or towns were Newar (46.82%), followed by Brahmins
(27.43%). All other groups have less than 20% who live in the capital city, small cities or
towns.
Among tharu women, less than one percent live in the capital city, 2.92% live
small cities, 2.06% live in towns, and 94.7% live in country-side. Among Magar women,
2.43% reside in the capital city, 4.41% in small cities, 2.62% in towns, and 90.53% in
country-side. Among Tamang women, 8.16% reside in the capital city, 4.79% in small
cities, 2.76% in towns and 84.29% in country-side. Among Gurung women, 8% reside in
the capital city, 11.91% in small cities, 3.62% in towns, and 76.47% in country-side.
Among Rai-Limbu women, 4.13% reside in the capital city, 7.13% in small cities, 6.93%
in towns and 81.815 in country-side. Among Newar women, closse to a quarter (23.33%)
reside in the capital city, 15.85% in small cities, 7.64% in towns and 53.18% in
countryside. Among the ‘other ethnic group’ women, 3.72% live in the capital city, 5.64%
live in small cities, 4.5% live in towns, and 86.13% live in country-side.
Among the Dalit caste women, less than one percent live in the capital city, 5.81%
in small cities, 5.32 in towns and 87.895 in country-side. Among the Chatri caste women,
7.05% live in the capital city, 6.58% in small cities, 4.82% in towns, and 81.54% in
118
country-side. Among the Brahmin caste women, 5.7% live in the capital city, 13.785 in
small towns, 7.94% in towns and 72.57% in country-side.
-----------------------------------------------------------------------------------------------------
Insert table 8 about here
-----------------------------------------------------------------------------------------------------
Ethnicity, Caste, Development Regions and Ecological Zones
Tables 8.1. to 8.3. present the distribution of ethnicity and caste by development
regions and ecological zones. A clear pattern of geographic clustering of ethnic and caste
peoples was observed. Most of the Tharu in this study came from Terai region (98.32%),
and they were largely concentrated in the Far-Western Development region of Terai
(50.73%). Virtually no Tharus were found in the Mountain region and in most of the Hill
region.
Most of the Magars in this study came from the Hill (65.38%) and Terai (30.43%)
regions. They were largely concentrated in the Western (32.63%), Mid-western (16.29%)
and Central (13.58%) development regions of the Hill areas, and Western (20.12%) and
Mid-western (4.79%) development regions of the Terai.
The majority of the Tamang in this study came from the Central Development
Region (79.96%) of the Mountain, Hill and Terai ecological zones. Tamang were largely
concentrated in the Central-Hill (61.25%) and Central-Terai (15.18%). Just over seven
percent live in the Eastern-Mountain and Central- Mountain, 6.27% live in the
EasternHill, 6.75% live in Eastern-Terai. Very few Tamang live in the Western, Mid-
western or Far-western Development regions of the country. (Note: Tamang from Central-
119
Mountain appear to be under-represented while Tamang from Central-Terai appear to be
overrepresented in the sample. Tamang, along with Sherpa, are known as the people of
the mountain).
Almost all of the Rai-Limbu in this study came from the Eastern Development
Regions (91.1%). They were concentrated in the Eastern-Mountain (13.15%), Eastern-
Hill (41.4%) and Eastern-Terai (36.55%). Virtually no Rai-Limbu were found in the
Central and Western Mountain, and Western, Mid-western, and Far-western regions of the
Hill and Terai.
Newar were found in almost all sub-development regions of the country, and they
were largely concentrated in the Central-Hill (53.87%), Western-Hill (14.07%) and
Eastern-Terai (8.55%).
Many of the Guung in this study came from the Western and Eastern
Development Regions (93.99%). They were largely concentrated in the Western-Hill
(68.17%), Central-Hill (11.9%), Western-Mountain (5.95%) and Eastern-Terai (5.76%).
Unlike the indigenous peoples, the caste peoples were scattered all over the
geographic sub-regions and do not appear to form geo-caste enclaves. A relatively higher
proportion of Brahmins were found in Western-Hill (20.41%), Central-Hill (16.77%) and
Eastern-Terai (12.15%), but no more than 21% of Brahmins were concentrated in any
single sub-region. A relatively higher proportion of Chetri were found in the Central-Hill
(16.74%), Mid-western-Hill (16.86%) and Western-Mountain (15.59%) but no more than
17% of Chetri were concentrated in a single sub-region. Similarly, a relatively higher
proportion of Dalit were found in Central-Terai (21.54%) and Western-Hill (18.49%), but
no more than 22% of Dalits were concentrated in a single sub-region.
120
The geographic pattern of population distribution appears similar between caste
peoples and the Newar indigenous group. Since caste peoples are migrants who do not
have a particular place of origin in Nepal, it was not surprising to see that they were
spread out throughout the country. However, it was surprising to see that Newar were also
spread throughout the country. Newar are thought to be the original inhabitants of
Kathmandu valley. Newar may have been spread out to other regions of the country
through trade or as government employees. Newars are known as people of trade, and
also constitute a sizable proportion of government employees.
-----------------------------------------------------------------------------------------------------
Insert tables 8.1 – 8.3 about here
-----------------------------------------------------------------------------------------------------
Ethnicity, Caste and Institutions
Table 9 presents the distribution of institutional characteristics by ethnicity and caste.
Language Policy: None of the indigenous peoples (0%) was assumed to have used
official language (Khas language) as their mother-tongue. In contrast, all the caste people
(100%) were assumed to have used official language (Khas language) as their mother-
tongue.
Distance to health facility: Disproportionately higher proportion of indigenous
peoples reported that distance to health facility is a big problem. Distance to health
facility was a big problem for 58.83% of the Magars, 54.44% Tamang, 43.34% Gurung,
121
41.39% Tharu, 39.56% Rai-Limbu, and 35.4% Newar. However, distance to health
facility was a big problem for only 26.28% of the Brahmins. It was a big problem for
46.09% Dalit and 40.1% Chetri.
Access to piped drinking water: Less than one percent of Tharu have access to
piped drinking water. Among the indigenous groups, 13.15% Magar, 15.79% Tamang,
18.19% Rai-Limbu, 27.78% Gurung, and 45.05% Newar have access drinking water.
Among the caste groups, only 6.04% Dalits have access to piped drinking water
compared to 23.24% Brahmin and 14.8% Chetri. Overall, access to piped drinking water
was highest among Newar (45.05) and the lowest among Tharu (0.41%).
Access to electricity: Among the indigenous peoples, 33.6% of the Tharus,
37.22% Tamang, 43.3% Magar, 58.35% Rai-Limbu, 71.05% Newar and 74.69% Gurung
have access to electricity. Among the caste peoples, 34.8% Dalit, 50.98% Chetri and
75.39% Brahmin have access to electricity. Overall, access to electricity was lowest
among Tharu (33.6%), and highest among Brahmins (75.39%).
-----------------------------------------------------------------------------------------------------
Insert table 9 about here
-----------------------------------------------------------------------------------------------------
Ethnicity, Caste and Demographic Characteristics
Table 10 presents the distribution of demographic characteristics by ethnicity and caste.
Marital status: Among the indigenous peoples, the proportions of women married
or living together was 74.1% for Tharu, 77% for Magar, 73.43% for Tamang, 67,13%
for Rai-Limbu, 73% Newar and 68.43% Gurung. Among the caste peoples, the
122
proportions of women married or living together was 74.3% for Brahmin, 76.87% for
Chetri, and 83.72% for Dalit.
Female headed household: Among the indigenous peoples, 6.94% Tharu,
25.997% Magar, 22.15% Tamang, 32.44% Rai-Limbu, 18.64% Newar and 24.77%
Gurung lived in households headed by female.
Among the caste people, 26.07% Brahmin, 25.79% Chetri and 22.72% Dalit lived in
households headed by female. Overall, highest proportion of Rai-Limbu households were
headed by female whereas lowest proportion of Tharu households were headed by
female.
Household size: Overall, the proportion of women who lived in large households was
largest among Tharu (54.73%), and smallest among Brahmin (25.43%). Among the
indigenous peoples, 54.73% Tharu, 37% Magar, 31.54% Tamang, 28.74% Rai-Limbu,
25.95% Newar and 33.84% Gurung lived in large households (i.e. households with seven
or more members). Among the caste peoples, 25.43% Brahmin, 26.71% Chetri and
31.14% Dalits lived in larag households (i.e. households with seven or more members). -
---------------------------------------------------------------------------------------------------- Insert
table 10 about here
-----------------------------------------------------------------------------------------------------
3. Bivariate Analyses: Characteristics of the Poor
Table 11 presents the bivariate association between the dependent variable, poverty, and
each of the independent variables in the study (women sample, N = 98.36).
123
Education and Poverty
Poor women were largely those with no or little education. Poverty rate was as high
as 49.03% among the women with no education, 38.32% among the women with only
primary education, and 23.45% among the women with secondary education. Poverty rate
was lowest among the women with post-secondary education (6.48%). Education was
significantly associated with poverty (χ2= 790.673, p<.001)
Health and Poverty
Poor women were largely underweight (BMI=<18.5), had anemia, had given births to
at least one child, and had at least one child deaths. Poverty rate among underweight
women was 45.81% compared to only 36.25% among ‘not underweight’ women. This
difference was statistically significant (χ2= 68.108, p<.0001). Poverty rate among women
with anemia was 39.95% compared to 37.91% among women with no anemia (χ2= 3.88,
p<.05). Poverty rate among women who gave births to one or more children was 39.06%
compared to only 34.74% among women who had not given births (χ2= 14.399, p<.001).
Poverty rate among women who had one or more child deaths was 48.76% compared to
only 35.1% among women who did not have any child deaths (χ2= 118.031, p<.001).
Occupation and Poverty
Poverty rates varied from 6.52% to 49.7% across occupations. Poor women were
largely farmers. Poverty rate among farmers was as high as 49.7% compared to only
15.07% among laborers (domestic, skilled and unskilled manuals), 6.52% among
professionals (technical, managers, clerical, sales, services), and 12.95% among ‘not
working’ women (χ2= 1265.86, p<.001).
124
Geography and Poverty
Geographic Isolation: Huge variation in poverty was observed across geographic
places of residents. Poverty rate varied from 1.05% (capital city, not isolated areas) to
44.95% (country-side, isolated areas). Poor women largely live in country-side (isolated
places). Poverty rate among women who reside in country-side was 44.95% compared to
only 17.22% for those who live in towns, 4.47% for those who live in small cities, and
1.05% for those who live in the capital city (χ2= 870.36, p<.001).
Development Regions and Ecological Zones: Across the Development Regions,
poverty rate varied from 27.28% (Central Development Region) to 60.05% (Far-western
Development Region). Poverty rate was highest among women who live in the
Farwestern Development Region (60.05%) and lowest among women who live in the
Central Development Region (27.28%). The second highest poverty rate was among the
women who live in the Mid-Western Development Region (54.92%). The poverty rate for
Western Development Region was 30.34%, and Eastern Development Region was
31.85%. Across the Ecological Zones, poverty rate varied from 31.09% (Terai) to
69.93% (Mountain). Poverty rate was highest among women in the Mountain (69.93%)
and lowest among the women in Terai (31.09%). Poverty among the women in the Hill
was 41.31%.
However, when the poverty rate was disaggregated at the sub-regional level, greatest
disparity was observed within a Development Region, rather than between Development
Regions or Ecological Zones. Western Development Region appears to have the greatest
disparity. Although average poverty rate in the Western Development Region (30.34%) is
less than overall population average (38.48%), poverty rate within this region varies from
125
18.46% (Western-Terai) to 82.11% (Western-Mountain). Overall, poverty rate was
highest among women who live in the Western-Mountain (82.11%) and lowest among
women who lived in the Western-Terai (18.46%) - more than four-fold differences. The
greatest disparity was between ecological zones within a development region rather than
between development regions within an ecological zone.
Within the Eastern Development Region, poverty rates varied from 20%
(EasternTerai) to 62.25% (Eastern-Mountain). Poverty rate for Eastern-Hill was 51.84%.
Within the Central Development Region, poverty rates varied from 25.31% (Central-Hill)
to
45.3% (Central-Mountain). Poverty rate for Central-Terai was 36.52%. Within the
Western Development Region, poverty rates varied from 18.46% (Western-Terai) to
82.11% (Western-Mountain). Poverty rate for Western-Hill was 36.99%. Within the
Midwestern Development Region, poverty rates varied from 35.17% (Midwestern-Terai)
to 65.59% (Midwestern-Hill). Within the Far-western Development Region, poverty rates
varied from 52.51% (Far-western-Terai) to 72.73% (Far-western-Hill).
Similar disparities were also observed across Ecological Zones. Within the Mountain
Ecological zone, poverty rates ranged from 45.3% (Central –Mountain) to 82.11%
(Western-Mountain). Poverty rate for Eastern-Mountain was 62.25%. Within the Hill
Ecological zone, poverty rates ranged from 25.31% (Central- Hill) to 72.73%
(Farwestern-Hill). Poverty rate for Eastern-Hill was 51.84%, Western-Hill was 36.99%
and
126
Midwestern-Hill was 65.59%.Within the Terai Ecological zone, poverty rate ranged from
18.46% (Western-Terai ) to 52.51% (Farwestern-Terai). Poverty rate for Central-Terai
was 36.52%, for Eastern-Terai was 20%, and for Midwestern-Terai was 35.17% (χ2=
1221.64, p<.001).
Institutions and Poverty
Official language is a mother-tonuge: Poverty rate among indigenous women for
whom the official language is not their mother-tongue was slightly higher (39.7%) than
for Khas group for whom the language is a mother tongue (37.15%) (χ2= 6.702.042,
p<.01).
Distance to health facility: Poverty rate among women who reported that distance
to health facility is a big problem was almost two times higher, 53.14%, than those who
reported that distance to health facility is not a problem, 28.43%, (χ2= 609.042, p<.001).
Piped Drinking Water: Poverty rate among those who did not have access to
piped drinking water to their dwelling or yard was more than seven times higher
(43.59%) than those who had access to piped drinking water, 6.44%, (χ2= 676.69,
p<.001).
Electricity: Poverty rate among those with no access to electricity was almost nine
times higher, 69.24%, compared to those with access to electricity, 9.25%.(χ2=
3716.059, p<.001).
Demography and Poverty
Poverty rate among married women were slightly higher (39.33%) than not
married women (35.65%). Poverty rate among women who lived in female-headed
households was slightly higher (41.97%) than women who lived in male-headed
127
households. Poverty rate among women who lived in large households (seven or more
members) was similar (38.58%) to those who lived in smaller households (six or fewer
members) (38.27%).
Since the institutional characteristics- distance to health facility, piped drinking
water, and electricity were suspected of being used in the construction of wealth index,
they were excluded in the multivariate analyses.
-----------------------------------------------------------------------------------------------------
Insert table 11 about here
-----------------------------------------------------------------------------------------------------
4. Multivariate Logistic Regressions Analyses (Women sample, N = 9836)
To test the hypothesized relationships between the specified independent variables
and the dependent variable, a series of logistic regression models were constructed. First,
a model with only ethnicity/caste and demographic variables were conducted (Model 1).
This model tested the relationship between ethnicity/caste and poverty controlling for
demographic variables (Research question #1). Next, individual productivity
characteristic were added to the regression equation (Model 2). This model tested the
relationship between individual productivity characteristics and poverty (Research
128
question #2), conditional upon ethnicity/caste and demographic characteristics. Next,
geographic variables were added to the regression equation (Model 3). This model tested
the relationship between geography and poverty controlling for ethnicity/caste,
demographics and individual productivity characteristics (Research question #3). Finally,
a language policy variable was added to the regression model (Model 4). This model
tested the relationship between language policy and poverty controlling for all other
variables in the model.
The step-wise regression model allowed for the investigation of two types of
relationship between independent and dependent variable. First, the fixed-effect model
examined the relative contribution of each of the independent variables to the dependent
variable (Research questions #2 and #3). Second, the random-effect model examined the
extent to which the observed differences in poverty (wealth) across ethnicity/caste were
driven by the modeled variables (Research question #4). The final or full model
examined the relationship between each of the independent variable and the dependent
variable controlling for all other variables in the model.
Tables 12 –15 present the results of the multivariate logistic regressions. The results
show huge differences in the probability of being poor across ethnic and caste groups.
Demographics (household size and gender of head of household), individual productivity
characteristics (education, health, and occupation), geography, and language policy were
significant predictors of poverty (χ2= 3577.4279, p<.001). These characteristics, however,
did not account for the observed disparity in wealth (poverty) across ethnic and caste
groups (Table 16). Detail results of each of the regression models follow.
129
Model 1: To test the extent to which probability of poverty vary across ethnic and caste
groups; and to test the extent to which the observed variation was driven by the variation
in demographic characteristics; Model 1 was run only controlling for demographics
(marital status, household size, and gender of the household head).
Table 12 presents the results of the Model 1. Ethnicity/caste and demographic
characteristics were significant predictors of poverty (χ2= 590.57, p<.001). The results
show huge variation in probability of being poor across ethnic and caste groups.
Controlling for demographics, Tamang women were 300% more likely to be poor
(odds ratio =4.03, p<.0001) than Brahmin women. Rai-Limbu women were 260% more
likely to be poor (odds ratio =3.60, p<.0001) than Brahmin women. Magar women were
207% more likely to be poor (odds ratio =3.07, p<.0001) than Brahmin women. Tharu
women were 173% more likely to be poor (odds ratio =2.73, p<.0001) than Brahmin.
“Other ethnic group’ women were 88% more likely to be poor (odds ratio =1.88,
p<.0001) than Brahmin. Odds of poverty for Newar and Gurung women were not
significantly different from Brahmin women (p>.05).
Similalry, controlling for demographics, Dalit women were 353% more likely to be
poor (odds ratio =4.53, p<.001) than Brahmin women, and Chetri women were 196%
more likely to be poor (odds ratio = 296, p<.0001) than Brahmin women.
Compare to Brahmins, the risk of poverty were significantly higher for most of the
indigenous groups (except Newar and Gurung) and all the lower caste groups, even after
controlling for demographic characteristics (marital status, household size, and gender of
the household head). The result indicated that variation in poverty across ethnic and caste
groups were not solely due to mean level differences in demographic characteristics.
130
-----------------------------------------------------------------------------------------------------
Insert table 12 about here
-----------------------------------------------------------------------------------------------------
Model 2: To test the extent to which the observed variation in poverty across ethnic and
caste groups was driven by the variation in individual productivity characteristics,
education, health and occupation variables were added to the regression equation in
Model 2. The variation in probability of being poor across ethnic and caste groups were
noted.
Table 13 presents the results of the Model 2. Individual productivity characteristics
(education, health, occupation) were significant predictors of poverty (χ2= 2328.450,
p<.0001). When the individual productivity characteristics (education, health, occupation)
were added to the regression model, the odds ratios of poverty decreased slightly. For
example, the odds of poverty for Tamang decreased from 4.03 to 2.81 when individual
productivity characteristics were added to the model. Similarly, the odds of poverty for
decreased form 3.07 to 1.93 for Magar, from 2.73 to 1.5 for Tharu, from 4.53 to 2.89 for
Dalit, and so on (Please see Table 17). The results indicated that, to some degree, the
observed variation in poverty was related to variation in individual productivity
characteristics. However, compared to Brahmin, the odds of being poor remained
significantly higher for all the ethnic and caste groups (except Newar and
Gurung), even after controlling for the individual productivity characteristics.
131
Controling for demographics and individual productivity characteristics (education,
health and occupation), Tamang women were 181% more likely to be poor than Brahmin
women, Rai-Limbu women were 261% more likely to be poor than Brahmin women,
Magar women were 93% more likely to be poor than Brahmin women, Tharu women
were 50% more likely to be poor than Brahmin. “Other ethnic group’ women were 42%
more likely to be poor than Brahmin. Odds of poverty for Newar and Gurung women
were not significantly different from Brahmin women (p>.05). Similarly, Dalit women
were 189% more likely to be poor than Brahmin women, and Chetri women were 114%
more likely to be poor than Brahmin women.
The result indicated that variation in poverty across ethnic and caste groups were not
solely due to mean level differences in individual level productivity characteristics of the
women.
-----------------------------------------------------------------------------------------------------
Insert table 13 about here
-----------------------------------------------------------------------------------------------------
Model 3: To test the extent to which the observed variation in poverty across ethnic and
caste groups were driven by geography, geographic characteristics (also proxy measures
of institution) were added to the regression equation in Model 3. The variation in
probability of being poor across ethnic and caste groups were noted.
Table 14 presents the result of the Model 3. Geography was a significant predictor
of poverty (χ2= 3577.4279, p<.0001). When the geographic variables were added to the
regression model, the odds ratios of poverty increased for almost all the groups, including
132
Newar which was not significant in the previous model. Odds ratios remained about the
same for Rai-Limbu, and odds ratios remained not statistically significantly different for
Gurung.
Controlling for demographics, individual productivity characteristics (education,
health and occupation) and geography, Tamang women were 291% more likely to be poor
than Brahmin women. Rai-Limbu women were 259% more likely to be poor than
Brahmin women. Magar women were 185% more likely to be poor than Brahmin
women. Tharu women were 152% more likely to be poor than Brahmin. Newar women
were 54% more likely to be poor than Brahmin women. And “other ethnic group’ women
were 281% more likely to be poor than Brahmin. Odds of poverty for Gurung women
were not significantly different from Brahmin women (p>.05). Similarly, Dalit women
were over 300% more likely to be poor than Brahmin women, and Chetri women were
65% more likely to be poor than Brahmin women. ----------------------------------------------
-------------------------------------------------------
Insert table 14 about here
-----------------------------------------------------------------------------------------------------
Model 4: To test the extent to which the observed variation in poverty across ethnic and
caste groups were driven by language policy, a binary language variable (another proxy
measure of institution) was added to the regression equation in the Model 4. The variation
in probability of being poor across ethnic and caste groups were noted.
Table 15 presents the results of the Model 4. Language was a significant predictor
of poverty. When the binary language variable was added to the regression model, the
133
odds ratios further increased for all the indigenous groups. However, the odds ratios
decreased for caste groups. Controlling for demographics, individual productivity
characteristics (education, health and occupation), geography, and language, Tamang
women were over 700% more likely to be poor than Brahmin women, Rai-Limbu women
were 622% more likely to be poor than Brahmin women, Magar women were 492% more
likely to be poor than Brahmin women, Tharu women were 426% more likely to be poor
than Brahmin, Newar women were 216% more likely to be poor than Brahmin women.
Gurung women were 120% more likely to be poor than Brahmin women; and “other
ethnic group’ women were 694% more likely to be poor than Brahmin.
However, when demographics, individual productivity characteristics (education,
health and occupation), geography and language were controlled for, Chetri women were
only 5% more likely to be poor than Brahmin women, and Dalit women were over 285%
more likely to be poor than Brahmin women. ----------------------------------------------------
-------------------------------------------------
Insert table 15 about here
-----------------------------------------------------------------------------------------------------
4a. Effect of demographic characteristics on poverty
Controlling for all other variables in the model, gender of the head of household
and household size were significant predictors of poverty. Female-headed households
were 26% likely to be poor than male-headed households, and large households (i.e.
larger than average household size. The average household size was six) were 13% less
likely to be poor than small household (i.e. household with less than seven members).
134
Marital status was no longer significant when all the variables in the model was
controlled for.
4b.Effect of individual productivity characteristics on poverty
Productivity characteristics- education, health and occupation- were statistically
significantly associated with poverty (Table 16, Model 4). Controlling for all other
variables in the model, education was statistically significantly associated with poverty
women sample. Compared to women with post-secondary education (SLC and above),
the odds of being poor was 6.68 times higher for women with no education, 4.35 times
higher for women with only primary education, and 2.25 times higher for women with
secondary education (but not SLC).
Underweight women (BM=<18.5) were 36% more likely to be poor than women
who were not underweight. Anemic women were 12% more likely to be poor than
women who were not anemic. Women who had child births were 26% less likely to be
poor than women had had no child births. Women who had child deaths were not
statistically significantly different from women who did not have child deaths.
Compared to professional women (technical, managers, clerical, sales, services), the odds
of being poor was 4.2 time higher for farmer and 2.1 times higher for labor (skilled and
unskilled manual). Women who were not working were not significantly different from
women who were professionals.
135
4c.Effect of geography on poverty
Controlling for all other variables in the model, geography was statistically
significantly associated with poverty (Table 16, Model 4). Compared to women who live
in capital city, the odds of being poor were 24.4 times higher for women who live in
isolate areas (e.g. country-side), 9.8 times for women who live in less developed areas
(e.g. towns), and 5.6 times higher for women who live in moderately developed areas
(e.g. small cities).
Compared to women who live in the Western-Terai region of the country, the odds
of being poor was 9.7 times higher for those who live in Eastern-Mountain, 4.6 times
higher for those who live Central-Mountain, 13.8 times higher for women who live
Western-Mountain, 8.6 times higher for women who live in Eastern-Hill, 4.1 times higher
for women who live in Central-Hill, 4.03 times higher for women who live in
WesternHill, 17.7 times higher for women who live in Mid-Western Hill, 18.9 times
higher for women who live in Farwestern-Hill, 2.2 times higher for women who live in
CentralTerai, 3.6 times higher for women who live in Midwestern-Terai, and 5.5 times
higher for women who live in Farwestern-Terai. Women who live in Eastern-Terai were
not statistically significantly different from women who lived in Western-Terai.
4d.Effect of institution on poverty
Controlling for all other variables in the model, language policy (a proxy measure
of institution) was statistically significantly associated with poverty (Table 16, Model 4).
Compared to women whose mother-tongue is Khas language, an official national
136
language of Nepal (Brahmin, Chetri and Dalit), the odds of being poor was 2.2 times
higher for women whose mother tongue was Khas language.
Research Question 1: Are some ethnic/caste groups in Nepal at significantly higher risk
of poverty than others?
Bivariate Chi-square tests and Multivariate logistic regressions results show that,
among the women sample, some ethnic/caste groups in Nepal were at significantly higher
risk of poverty than others. Table 11 presents the results of the bi-variate Chi-square tests,
and Table 12 presents the results of the multivariate logistic regressions controlling for
demographic characteristics.
Bivariate Chi-Square analyses (Table 11) indicated a statistically significant
association between ethnicity/caste and poverty (χ2= 495.6613, p<.001). Among the
indigenous groups, Tamang women had the highest poverty rate (51.51%) and Newar had
the lowest poverty rate (20.62%). Among the non-indigenous or caste groups, Dalit
women had the highest poverty rate (52.44%) and Brahmin had the lowest poverty rate
(19.36%). Overall, Dalit had the highest poverty rate and Brahmin had the lowest poverty
rate.
The results of the multivariate regressions (Table ) indicate that some ethnic/caste
groups in Nepal were at significantly higher risk of poverty than others. In particular,
Tamang, Rai-Limbu, Magar, Tharu, and ‘other’ indigenous groups, and Dalit and Chetri
caste groups were at significantly higher risk of poverty than Brahmin. Newar and
Gurung were not significantly at higher risk of poverty than Brahmin.
137
H1a: Compared to Brahmins (the de facto political elites), all other ethnic/caste groups
are at higher risk of poverty.
Among the women sample, this hypothesis was partially supported by the data.
Multivariate logistic regression results shows that, controlling for the demographic
characteristics (marital status, gender of the head of household, and household size), the
odds of being poor were significantly higher for all the indigenous groups (except Newar
and Gurung) and all the caste groups compared to Brahmin (χ2= 590.56, p<.0001) (Table
12). Among the indigenous groups, the odds of being poor were 4.03 times higher for
Tamang (the highest among all the indigenous groups), 3.6 times higher for Rai-Limbu,
3.07 times higher for Magar, 2.7 times higher for Tharu, and 1.88 times higher for the
‘Other ethnic group’ compared to Brahmin. The risk of poverty for Newar and Gurung
were not statistically significantly higher than Brahmin as hypothesized.
Among the caste groups, the odds of being poor was 4.53 times higher for Dalit
(highest among the all the groups), and 2.96 times higher for Chetri compared to
Brahmin.
When all other variables were controlled for, the risk of poverty increased
significantly for all the ethnic groups.
H1b: Within a caste group, people of lower caste will be at significantly higher risk of
poverty than people of higher caste.
This hypothesis was supported by the results of the logistic regressions. Multivariate
logistic regression results reveal that, controlling for the demographic characteristics, the
138
odds of being poor was 4.53 times higher for Dalit (the lowest caste group), and 2.96
times higher for Chetri (the lower caste group) compared to Brahmin (Table 12).
H1c: Within an indigenous group, the risk of poverty for some ethnic groups (e.g. Newar,
Gurung) will be significantly higher than for the other ethnic groups (Tharu, Magar,
Tamang).
This hypothesis was supported by the results of the logistic regressions. Multivariate
logistic regression results reveal that, controlling for the demographic characteristics, the
risk of poverty was higher for Tamang (odds ratio = 4.03, p<.0001), Rai-Limbu (odds
ratio = 3.6, p<.0001), Magar (odds ratio = 3.1, p<.0001), and Tharu (odds ratio = 2.7,
p<.0001) than for Newar (odds ratio =1.12, p =0.38) and Gurung (odds ratio =1.07, p
=.06).
H1d: Between ethnic and caste groups: Some ethnic groups are as well off as high caste
group (i.e. the risk of poverty for some ethic groups, e.g. Newar and Gurung, are not
significantly different than for people of higher caste, Brahmin).
This hypothesis was supported by the results of the logistic regressions. Newar (odds
ratio =1.12, p =0.38) and Gurung (odds ratio =1.07, p =.06) indigenous groups were not
statistically significantly different from Brahmin caste group. Newar and Gurung appear
to be as well off as Brahmin.
H1e: Between ethnic and caste groups: Some ethnic groups are as poor as lower caste
groups (i.e. The risk of poverty for some of the ethnic groups such as Tharu, Tamang,
139
Magar, are likely to be as high as that for lower-caste groups?).
This hypothesis was supported by the results of the logistic regressions. The risk of
poverty for Tamang (odds ratio = 4.03, p<.0001), Rai-Limbu (odds ratio = 3.6, p<.0001)
and Magar (odds ratio = 3.1, p<.0001) indigenous groups were almost as high as that for
low-caste Dalit (odds ratio =4.53, p<.0001). The risk of poverty for Tharu (odds ratio =
2.7, p<.0001) indigenous group was as high as that for Chetri, a caste lower than Brahmin
(odds ratio =2.96, p<.0001).
H1e: The risk of poverty for indigenous peoples as a group will be significantly different
from caste people as a group.
This hypothesis was supported by the data. Multivariate logistic regressions results
showed that, controlling for the demographic characteristics, the odds of being poor was
statistically significantly different for indigenous peoples as a group compared to caste
people as a group (χ2=49.19, p<.0001). Indigenous peoples as a group was about 9% less
likely to be poor than caste peoples as a group (odds ratio = 0.914, p =.03).
However, when Dalit were excluded from caste group, indigenous people as a
group were 10% more likely to be poor than caste peoples as a group (odds ratio = 1.10, p
=.025). Furthermore, if Newar were excluded from the indigenous group, indigenous
peoples as a group were 20% more likely to be poor than non-indigenous peoples as a
group (odds ratio =1.20, p<.0001). Finally, if we only look in urban areas (and exclude
140
Newar and Dalit), indigenous peoples as a group were 29% more likely to be poor than
caste groups (odds ratio = 1.292, p =.04).
This hypothesis was supported by the data. However, the direction of the
relationship changed when Newar from the indigenous group and Dalit from caste group
were excluded. Stronger differences were observed in the urban areas. Newar appear to
be similar to caste group while Dalit appear to be similar to indigenous group in terms of
their poverty status.
Research Question 2. To what extent do the individual productivity characteristics
(education, health, employment, and occupation) determine the risk of poverty?
The results of the multivariate logistic regression show that individual
productivity characteristics are a significant predictor of poverty (Table 16, Model 4)).
H2a: Higher the education, lower the risk of poverty.
This hypothesis was supported by the data. Compared to women with
postsecondary education (SLC and above), women with no education was 568% more
likely to be poor, women with only primary education was 335% more likely to be poor,
and women with secondary education (but not SLC) was 125% more likely to be poor,
controlling for all other variables in the model (Table 16, Model 4)
H2b: Higher the health problems, higher the risk of poverty.
This hypothesis was partially supported by the data. Underweight (BMI=<18.5)
women were 36% more likely to be poor than women who were not underweight
141
(BMI>18.5), women with anemia were 12% more likely poor than women without
anemia, women who had child births were 26% less likely to be poor than women who
did not have child births. Women who had child deaths were not significantly different
from women did not have child deaths (Table 16, Model 4).
H2c: Employed are at lower risk of poverty than non-employee.
H2d: Farmers are at higher risk of poverty than non-farmers.
Since employment and occupation variables were integrated, these two
hypotheses were tested at once. The data partially support the employment and
occupation hypotheses. Farmers and laborers were at significantly higher risk of poverty
than professionals (technical, managers, clerical, sales, and services). Being a farmer
increases the risk of being poor by 321% compared to professionals; and being a laborer
(skilled & unskilled) increases the risk of being poor by 108% compared to professionals.
However, not-working women were not significantly at higher risk of poverty than
women who were working as professionals (Table 16, Model 4).
Research Question 3. To what extent do the geographic characteristics of the community
determine the risk of poverty?
The results of the multivariate regressions show that geography is a significant
predictor of poverty (Table 16).
H3: Those living in isolated geographic areas are at higher risk of poverty than those
living in non-isolated geographic areas.
142
This hypothesis was supported by the data. Women who live in isolated or
underdeveloped areas (country-side) were 2339% higher times more likely to be poor
than women who live in the developed areas (capital cities). Similarly women who
live in less developed (towns) were 880% more likely and women who live in
moderately developed areas (small cities) were 461% more likely to be poor than
women who reside in the developed area (capital city) (Table 16, Model 4).
Furthermore, among the development regions and ecological zones, women who
live in the Far-western hills were at the highest risk of poverty (1786% higher)
compared to women who live in the Western Terai. Women who live in the
Midwestern hill were 1673% more likely, and women who live Western-mountain
were 1284% more likely to be poor than women won live in Western Terai. Overall,
women who live in Eastern-development region, Western-development region,
Midwestern-development region, and Far-western-development region were at higher
risk of poverty than women who live in Central-development region. Women who
live in Hill ecological zone and Mountain ecological zone were at higher risk of
poverty than women who live in Terai-ecological zone (Table 16, Model 4). However,
women who live in Far-western Terai were about as likely as those who live in
Central-hill or
Central-Mountain to be poor.
Research Question 4. To what extent do the differences in demographics, individual
productivity characteristics and geographic characteristics explain the differences in the
risk of poverty between various ethnic/caste groups?
143
H4: The observed differences in the risk of poverty between various ethnic/caste
groups will disappear when the individual level productivity characteristics and
geographic characteristics are controlled for.
This hypothesis was not supported by the data. The observed differences in the risk of
poverty between Brahmin and each of the ethnic and caste groups remained significantly
high even after controlling for the individual level productivity characteristics and
geographic characteristics (Table 16, Model 4).
When the individual productivity characteristics (education, health and
employment/occupation) were controlled for, the observed differences in the risk of
poverty between Brahmin and each of the ethnic and caste groups reduced to a certain
degree (Tabel 16, Model 2). However, the difference still remained significant. The
results indicate that the observed differences in the risk of poverty (i.e. wealth disparity)
were not solely due to the mean level differences in the individual productivity
characteristics. In other words, indigenous groups and lower caste groups would remain
relatively poorer than Brahmin even if they had the same level of education, health or
employment/occupation as Brahmin. Increasing access to education, health or
employment/occupation does not appear to bridge the observed disparity in wealth that
exists between Brahmin and indigenous peoples and lower caster groups. Some other
factors appear to be driving the observed wealth disparity.
One plausible factor is thought to be geography. Indigenous peoples are assumed
to live in isolated areas and such isolation is thought to drive their poverty. When
geography was controlled for in the multivariate regression (in addition to demographics
144
and individual productivity characteristics), the disparity in wealth further exacerbated
(Table 16, Model 3).
The multivariate regression results indicate that the observed wealth disparity
between Brahmin and indigenous groups and lower caste groups were not driven by the
differences in the geographic communities in which they live. In fact, geography appears
to be a mitigating factor for wealth disparity. In other words, if Brahmin were to live in
the indigenous territories, the indigenous peoples would have been much worse off. The
geographic clustering of indigenous peoples appears to serve as a buffer zone against
their poverty. For example, Newar, which was not significantly different in the previous
model (Model 2, Table 16) becomes significant when geography was controlled for
(Model 3, Table 16). The results indicate that if Brahmin and Newar were to live in the
same geographic territories, the wealth disparity between Brahmin and Newar would
have been significantly greater.
However, the effect of geography on the wealth relationship between Brahmin and
Chetri was different. When geography was controlled for in the multivariate regressions,
the disparity between Brahmin and Chetri decreased by almost half (from 113% to 64%).
The results indicate that observed wealth disparity between Brahmin and Chetri was, to a
certain degree, driven by differences in geography.
When the language variable (a proxy measure of institution) was controlled for,
the disparity between Brahmin and all the indigenous groups were further exacerbated
(Table 16, Model 4). Gurung which was not significantly different in the previous models
now became significant. The results indicate that language may be a protective factor
against poverty for indigenous peoples. In other words, if the indigenous peoples spoke
145
the same language as the Brahmin, the wealth disparity between Brahmin and indigenous
peoples would have been much greater.
However, for Khas group, controlling for language significantly decreased the
wealth disparity between Brahmin, Chetri, Dalit. Language, therefore, appears to be a
factor that drives wealth disparity within the Khas group. In other words, if Chetri and
Dalit were to speak different languages than Brahmin, the wealth disparity between them
would have been less.
-----------------------------------------------------------------------------------------------------
Insert table 16 about here
-----------------------------------------------------------------------------------------------------
4e. Analyses of subsamples: Tamang and Brahmin
Table 17 presents the results of multivariate regressions conducted separately for
Tamang and Brahmin. The results show differential effects of the independent variables
on dependent variable conditional upon ethnicity or catse. For Tamang, there was no
significant relationship between most of the independent variables and the dependent
variable. For Brahmin, however, almost all the independent variables were significantly
associated with the dependent variable. For example, for Tamang, there was no
significant difference between post-secondary education and less than post-secondary
education; there was no significant relationship between occupation and poverty, orr
geographic isolation and poverty. For Brahmin, on the other hand, there was a significant
difference between post-secondary and less than post-secondary education. Brahmins
146
with post-secondary education was significantly less likely to be poor compared to
Brahmins with less than post-secondary education. Similarly, health, occupation, and
geography were significant predictors of poverty for Brahmin.
-----------------------------------------------------------------------------------------------------
Insert table 17 about here
-----------------------------------------------------------------------------------------------------
5. Results of the Geographic Information Systems (GIS) Analyses
Figures 4 –14 presents the results of spatial analyses of poverty using Geographic
Information System (GIS). Figure 4 displayes the map of Nepal with geographic
distribution of sample population by Development Regions and Ecological Zones. The
map shows that the sample population was evenly distributed across the five
Development Regions. However, in terms of Ecological Zones, they were concentrated
mostly in Terai and Hill zones. ---------------------------------------------------------------------
--------------------------------
Insert figure 4 about here
-----------------------------------------------------------------------------------------------------
Figures 5 displays the geographic distribution of indigenous population. The map
shows a clear geographic clustering of indigenous peoples by their ethnicity.
-----------------------------------------------------------------------------------------------------
147
Insert figure 5 about here
-----------------------------------------------------------------------------------------------------
Figure 6 displays the geographic distribution of caste population. The map shows that
caste population is spread out throught the country, and does not show geographic
clustering.
-----------------------------------------------------------------------------------------------------
Insert figure 6 about here
-----------------------------------------------------------------------------------------------------
Figure 7 displays the geographic distribution of poor as percent of the total population.
The bar chart indicates the proportion of individuals who are poor- taller the bar, higher
the proportion of poor. The map shows that there is a geographic concentration of poverty
in Nepal. Higher concentration of poverty is seen in the Mountain areas of the Eastern,
Central and Far-western Development regions, and lower concentration of poverty in the
areas fo the Eastern Development Region and Western Development region.
-----------------------------------------------------------------------------------------------------
Insert figure 7 about here
-----------------------------------------------------------------------------------------------------
148
Figure 8 displays the geographic distribution of poverty by geographic isolation. A clear
pattern is seen between poverty and geo-isolation. Poor people are largely concentrated in
isolated communities.
-----------------------------------------------------------------------------------------------------
Insert figure 8 about here
-----------------------------------------------------------------------------------------------------
Figure 9 displays the geographic distribution of poverty and the indigenous peoples. The
map shows that poverty is highly concentrated in areas where the indigenous peoples
live.
-----------------------------------------------------------------------------------------------------
Insert figure 9 about here
-----------------------------------------------------------------------------------------------------
Figure 10 displays the geographic distribution of poverty by caste groups. High poverty is
also seen in areas in which caste peoples live.
-----------------------------------------------------------------------------------------------------
Insert figure 10 about here
-----------------------------------------------------------------------------------------------------
Figure 11 displays the geograhic distribution of poverty by caste group (Brahmin/Chetri),
but without the lower-caste (Dalit). Notice that much of the poverty seen in figure 10
149
disappears when Dalit was excluded from caste group. The findingins indicate that much
of the poverty seen among caste people as a group was driven by poverty among the
Dalits. The findings also confirm the heterogeneity between Dalit and Brahmin/Chetri
that is independent of geography.
-----------------------------------------------------------------------------------------------------
Insert figure 11 about here
-----------------------------------------------------------------------------------------------------
Figure 12 displays the geographic distribution of poverty for Brahmin caste. The map
shows that proportion of Brahmins who are poor is very small in majority of the areas in
which they live. Very small proportion of Brahmins appears to be poor in the Central
Region and Eastern Development region. Far-western region appears to have higher
proportion of Brahmins who are poor.
-----------------------------------------------------------------------------------------------------
Insert figure 12 about here
-----------------------------------------------------------------------------------------------------
Figure 13 displays the geographic distribution of poverty for Tamang indigenous group.
Tamangs appear to concentrate around the Kathmandu valley. Proportion of poor appears
to be high among Tamangs who live around the capital city. Comparison of figure 12 and
150
figure 13 shows that the proportion of poor among Tamang is much higher than
proportion of poor among Brahmin (figure 12) although they live in the same region. The
finding indicates that the differences in poverty between Tamang and Brahmin are not
soley driven by the differences in geography.
-----------------------------------------------------------------------------------------------------
Insert figure 13 about here
-----------------------------------------------------------------------------------------------------
Figure 14 displays geographic distribution of poverty for Dalit. The map shows high
proportion of poor among Dalits in almost all places in which they live. --------------------
---------------------------------------------------------------------------------
Insert figure 14 about here
-----------------------------------------------------------------------------------------------------
The results from GIS analyses confirm and complement the findings from the
bivariate and multivariate regressions. The poverty maps show a clear geographic pattern
between poverty, geographic isolation, and ethnicity. Poverty appears to be concentrated
in areas that are geographically isolated. However, these isolated geographies with high
poverty appear to be the places where indigenous peoples are concentrated- i.e. poverty
map coincides with the territories of the indigenous peoples. Even in the geographically
isolated areas, indigenous groups and Dalits were more likely to be poor than Brahimns.
151
In all geographic areas, poverty was higher among indigenous peoples and Dalits than
among Brahmins.
6. Comparing the Multivariate Regression Results of Women Samples with Men and
Household Samples
Table 18 presents the distribution of poverty by ethnicity and caste among
Women, Men and Household Samples. The findings show comparable poverty rates
across these three samples.
Table 19 presents the results of multivariate logistic regressions predicting poverty
among Women, Men and Household Samples. The findings indicate that the findings
from women sample were consistent across findings from men and household samples.
Men sample (N = 4045):
Controlling for all other variables in the model, education was statistically
significantly associated with poverty among men sample. Compared to men with
postsecondary education (SLC and above), the odds of being poor was 5.23 times higher
for men with no education, 3.92 times higher for men with only primary education, and
2.46 times higher for men with secondary education (but not SLC).
Men who had child deaths were 29% more likely to be poor than men who had
had no child deaths.
Compared to men who work in professional jobs (technical, managers, clerical,
sales, services), the odds of being poor was 3.2 higher for men who work as farmers, 2.3
times higher for men who work as laborers (skilled and unskilled manual labors), and 1.6
times higher for men who were ‘not working’.
152
Controlling for all other variables in the model, geography was statistically
significantly associated with poverty among men sample. Compared to men who live in
capital city, the odds of being poor were 48.06 times higher for men who live in
countryside, 14.11 times for men who live in towns, and 6.74 times higher for men who
live in small cities.
Compared to men who live in the Western-Terai region of the country, the odds of
being poor was 8.99 times higher for those who live in Eastern-Mountain, 4.07 times
higher for those who live Central-Mountain, 11.75 times higher for those who live
Western-Mountain, 7.26 times higher for those who live in Eastern-Hill, 2.88 times
higher for those who live in Central-Hill, 3.52 times higher for those who live in
Western-Hill, 16.48 times higher for those who live in Mid-Western Hill, 18.78 times
higher for those who live in Farwestern-Hill, 2.09 times higher for those who live in
Central-Terai, 2.43 times higher for those who live in Midwestern-Terai, and 7.75 times
higher for those who live in Farwestern-Terai. Men who live in Eastern-Terai were not
statistically significantly different from men who live in Western-Terai.
Household sample (N = 7659):
Controlling for all other variables in the model, education was statistically
significantly associated with poverty among household sample. Compared to households
in which the head of the households had post-secondary education (SLC and above),
households in which the head of household had no education were 363% more likely to
be poor, households in which the head of household had only primary education were
153
232% more likely to be poor and households in which head of the household had
secondary education (but not SLC) were 78% more likely to be poor.
Households headed by farmers were 282% more likely to be poor than households
headed by non-farmer households (technical, managers, clerical, sales, services, skilled
and unskilled manual laborers.
Controlling for all other variables in the model, geography was statistically
significantly associated with poverty among household sample. Compared to household
in the capital city, the odds of being poor were 27.01 times higher for households in
country-side, 11.15 times for households in towns, and 6.71 times higher for households
in small cities.
Compared to household in the Western-Terai region of the country, the odds of
being poor was 7.43 times higher for households in Eastern-Mountain, 3.97 times higher
for households in Central-Mountain, 15.53 times higher for households in
WesternMountain, 6.68 times higher for households in Eastern-Hill, 3.88 times higher
for households in Central-Hill, 3.01 times higher for households in Western-Hill, 16.88
times higher for households in Mid-Western Hill, 17.75 times higher for households in
Farwestern-Hill, 1.99 times higher for households in Central-Terai, 3.21 times higher for
households in Midwestern-Terai, and 5.34 times higher for households in
FarwesternTerai. Households in Eastern-Terai were not statistically significantly
different from households in Western-Terai.
7. Analyses of Disaggregated Data by Geographic Sub-region and
154
Ethnicity/Caste
To rule out the possibility of bias due to small sample size for some geographic
subregions and ethnic groups, further multivariate analyses were conducted separately for
each of the sub-region and ethnic/caste groups for which data met the assumptions of
multivariate regression, including sample size. Tables 21-27 present the results of
multivariate regressions on these disaggregated data. The findings from this method were
consistent with the findings from the pooled method. However, the sub-regional analyses
revealed that among the indigenous peoples, Magar were at the highest risk of poverty in
the Central-Hill (Table 22) and in the Western-Hill (Table 23) regions; Tharu were at the
highest risk of poverty in the Eastern-Terai region (Table 24), but they were less likely to
be poor than Brahmin in the Central-Terai (Table 25); and again, Magar were at the
highest risk of poverty in the Western-Terai (Table 26) region.
The findings from multivariate regressions on each of the ethnic/caste sub-sample
further confirmed that education has differential effects on poverty for indigenous groups
compared to caste group, particularly Brahmin. There was no significant difference
between secondary and post-secondary education for any of the indigenous groups, but
difference was significant for Brahmin.
Summary of the findings
1. Tamang people are at the highest risk of poverty among the indigenous peoples
Among the indigenous peoples, Tamangs were at the highest risk of poverty.
Tamang women were over 700% more likely to be poor than Brahmin women,
controlling for all other variables in the model (Table 16, M4). Magar, Rai-Limbu and
155
Tharu were also significantly at higher risk of poverty than Brahmin. The risk of poverty
for Newar and Gurung was lowest among the indigenous peoples.
2. Dalit caste is at the highest risk of poverty among the caste peoples
Among the caste/Khas people, Dalit was at the highest risk of poverty. Dalit
women were 285% more likely to be poor than Brahmin women (Talbe 16, M4),
controlling for all other variables in the model. Chetri caste was also significantly at
higher risk of poverty than Brahmin. Brahmin caste was at the lowest risk of poverty
among all the groups.
3. Indigenous peoples have low human capital attainment and are trapped in
lowpaying occupation.
Overall, indigenous peoples had significantly lower level of education and were
employed in low-paying occupations (farming and labors), and lived in geographically
isolated areas than Brahmin (Tables 5 -7). The low level of education and low-paying
occupations were significantly associate with their poverty.
4. Indigenous peoples are geographically isolated
Indigenous peoples were geographically clustered, and lived in isolated areas. The
geographic isolation of indigenous peoples was significantly associated with their
poverty.
5. Demographics, individual productivity characteristics and geography are not the
sole drivers of indigenous poverty.
156
Significant differences wealth (or poverty) remain between Brahmin caste and the
indigenous peoples even after controlling for demographic, individual characteristics and
geographic variables (Table 16). The findings suggest that observed wealth disparity
between Brahmin caste and the indigenous peoples was not solely due to the mean level
differences in demographic characteristics, individual level productivity characteristics or
geography. Some other factors appear to be driving the disparity. These ‘other’ factors are
thought to be underlying institution.
6. Differential Return for Investment in Human Capital Investment
The analyses of the sub-samples for Tamang and Brahmin reveal that that there is a
differential return for investment in human capital development (education, health,
occupation) between the indigenous peoples (Tamang) and caste people (Brahmin) (Table
17). For example, if you are a Tamang, there is no significant difference in the risk of
poverty between having post-secondary education (SLC and above) and having only
secondary or primary education. However, if you are a Brahmin, the risk of poverty
decreases significantly if you have post-secondary education compared to having only
primary or secondary education. Similarly, if you are a Tamang, there is no significant
difference between underweight (BMI =<18.5) or not underweight. However, if you are
Brahmin, the risk of poverty increases significantly if you are underweight (BMI=<18.5).
Likewise, if you are a Tamang, there is no significant difference between being a farmer
or other professionals. However, if you are a Brahmin, the risk of poverty increases
significantly if you are a farmer. Furthermore, if you are a Tamang, the risk of poverty
does not vary by geography. However, if you are a Brahmin, the risk of poverty changes
157
significantly by geography. Brahmins living in the Mountains and Hills of the Mid-West
and Far-West Nepal are significantly at higher risk of poverty than Brahmins living in
Western Terai. The findings were consistent for other indigenous groups (Table 27).
CHAPTER VIII: DISCUSSION
Overview
The purpose of this study was to investigate the determinants of poverty among
the indigenous peoples of Nepal. The study used nationally representative samples of
women, men and households from Nepal Demographic and Health Survey (DHS 2006).
Analyses involved conducting maximum likelihood estimates of logistic regression
models and Geographic Information System (GIS) to test the hypothesized relationships
derived from extant literature and theories. The findings from this study reveal that
Tamang, Magar, Rai-Limbu, and Tharu indigenous peoples, and Dalit caste are
significantly at higher risk of poverty than Brahmin. The differences in the risk of poverty
were not solely driven by the differences in individual level productivity characteristics
and geography. Other factors appear to be driving the risk of poverty. Examination of the
first constitution of Nepal 1854, Muluki Ain 1854, reveal that the groups that were
designated as ‘lower’ in the constitution were the groups that were at higher risk of
poverty. The findings appear to suggest that the sources of wealth disparity (poverty) in
Nepal may be institutional, such as a constitution. The findings provide some support to
the theory of institutional design (North, 1990) that suggests that institutions, such as a
country’s constitution or laws, are the factors that drive some groups to become poor
while others to become rich.
158
The experiments conducted in this study provided support for all but one
hypothesis tested. The hypotheses related to ethnic and caste differences in poverty
(wealth) were supported by the data. The hypothesized relationship between individual
productivity characteristics, geography and poverty were also supported by the data.
However, the hypothesis that “differences in individual level productivity characteristics
and geographic characteristics account for the differences in wealth (poverty)” was not
supported by the data. The findings indicate that the ethnic/caste disparity in wealth
(poverty) in Nepal is not solely driven by the mean level differences in individual level
productivity characteristics or geography. Although individual productivity characteristics
(such as education, health and occupation) and geographies were important predictors of
poverty, poverty appears to be ultimately a function of underlying institutions,
particularly the constitution of Nepal, Muluki Ain 1854. The constitution appears to create
a caste-system like social structure, which, on one hand, isolate indigenous territories,
and on the other, prevent indigenous peoples from attaining human capital endowment
(education, health), thereby subsequently impoverishing the indigenous peoples, --to the
same extent as the lower-castes.
The findings of this study were consistent with the previous studies that suggest
that there is a cost of being indigenous (Pscharopoulos & Patrnos, 1994; Carino, 2009;
Eversole, 2005; etc.). However, the findings of this study further expand the current
understanding of indigenous poverty by demonstrating that the cost was primarily driven
by institutional factors rather than deficiencies of the individual peoples who are poor.
159
Institutions and Poverty
As the findings of this study reveal, indigenous peoples appear to suffer from
poverty as much as, or even more than, the lower-caste groups. How did the indigenous
peoples, who are the original inhabitants and rightful owners of the land, become so poor
in their own land, while the migrants, the caste people, became rich?
Poverty of the indigenous peoples and their countries in the Americas, Africa,
Australia, and India has largely been attributed to the colonization of their territories by
the Europeans (Psacharopoulos & Patrinos, 1994; Eversole, 2005; Carino, 2009 etc.).
However, since Nepal was never colonized by the Europeans, colonization cannot be a
plausible explanation to poverty of the indigenous peoples and of Nepal. An alternative
explanation to the poverty of the indigenous peoples of Nepal, and their country as a
whole, is warranted.
One explanation is offered by Bista (1991), who argues that fatalism and nepotism
among the Brahmin-Chetri castes were the primary causes of underdevelopment of
Nepal. In his seminal book, Fatalism and Development: Nepal’s Struggle for
Modernization, Bista argues that Brahmin/Chetri’s world view is largely shaped by Hindu
fatalism (not by rationality or science) and therefore, as long as Hindu Brahmin/Chetries
remain as political elites of the country, Nepal will never develop into a modern state.
Furthermore, he argues that since Brahmin/Chetries do not view Nepal as their own—
having migrated from India--they do not have a genuine interest in developing the
country into a modern society that will benefit non-Brahmin/Chetries. Consequently,
Bista advocates for indigenous leadership in governing the country as a solution to
Nepal’s underdevelopment.
160
While Bista’s assessment is considered candid by many accounts, critiques argue
that Bista, being a Brahmin himself, is simply trying to warn his fellow Brahmins against
their excess atrocities (Macfarlene, n.d.). Unlike other Brahmins, Bista saw the growing
indigenous movements against the Brahmin/Chetri dominance as an imminent threat to
the elites, including himself, and therefore wanted to devise a way to pacify the
indigenous peoples and their movements. But other Brahmins did not see the far-sighted
vision of Bista (Bista is currently missing and suspected to have been killed by other
Brahmins who were angered by his writings). Nevertheless, Bista’s writing considerably
influenced political discourses and movements in Nepal. In particular, the Communist
Party of Nepal, Maoists, was born as a brain-child of Bista’s book, and this party has
successfully brought in some of the indigenous peoples under Brahmin/Chetri control,
albeit under the pretext of the communist movement. It is important to note here that the
top three leaders of the Maoist party--Prachanda, Baidyae and Bhattarai--are all
Brahmins. Prachanda is a nom de guerre of Pushpa Kamal Dahal. It is alleged that
Prachanda uses his nom de guerre to avoid being identified as Brahmin.
The critically missing part in Bista’s analysis, however, is that he overlooks the
role of institutional structure under which the current Nepali state is built and
overemphasizes the role of the caste system. He analyzes the Hindu caste-system as if it
is a cultural universal when in reality, indigenous peoples of Nepal do not belong to the
castesystem. Another critical error in Bista’s work is that he fails to make the indigenous
peoples and their poverty the central focus of his work. Without understanding the
indigenous peoples and their problems, analysis of Nepal’s poverty seems incomplete, or
even misleading, since the majority of the Nepal’s population are indigenous peoples.
161
What appears to be the real problem of Nepal are its basic institutions, such as the
constitution. Although Muluki Ain 1854 mirrors the Hindu-caste system, by no means are
they the same thing. It appears that Brahmins/Chetries capitalize on both the caste-system
and the constitution to their own economic advantage. There is no evidence to support the
assertion that Brahmin/Chetries are naïve fatalists, as Bista seems to suggest. Rather they
appear to be rational actors who overlook humanity in the interest of economic benefit to
the few caste-based clan. The Muluki Ain 1854, designed by Brahmin/Chetries, appears to
be serving this purpose, at least until now. This may be one reason why although the
constitution of Nepal, has been amended twice (in 1965 and 1990), the consequences are
not realized. As predicted by North (1990), the constitution of Nepal appears to have been
changed to keep the status quo of the elites. By continuing to design and redesign the
constitution in ways that isolate indigenous territories and prevent or discourage
indigenous peoples from obtaining human capital, Brahmin/Chetries seem to achieve
their desired effect--keep the status quo and use indigenous men as cheap labor (e.g.
porters, servants in Brahmin/Chetri-owned businesses) and indigenous women as sex
objects for human trafficking. Such a system is what appears to be driving indigenous
peoples, and the country as a whole, into poverty.
Geographic Isolation and Poverty
The fixed effect model of the logistic regression revealed that geography was
positively associated with poverty (Table 16, M4). Geographic isolation significantly
increased the risk of poverty for all peoples, including the caste people. Since higher
proportions of indigenous peoples were geographically isolated than Brahmin, much of
their poverty appears to be driven by the geographic isolation of their communities. This
162
finding was further supported by the Geographic Information System (GIS) analyses
(Figures 4 - 14).
The random effect model revealed that the disparity between Brahmin and
indigenous peoples further widened when geographic variables were added to the
regression equation (Table 16, M4). The findings suggest that the disparity between
Brahmin and indigenous peoples was not driven the by the differences in geography
alone. In fact, if Brahmins and indigenous peoples were to live in the same geographic
territories (villages), the inequality between Brahmin and indigenous peoples would have
been much greater (Table 16, M4). Geography, therefore, appears to have protective
effects on indigenous peoples. Geographic clustering of the indigenous peoples appears to
mitigate the inequality between them and the caste people.
However, when the geographic variables were added to the regression model, the
disparity between Chetri and Brahmin were significantly reduced (Table 16, M4). The
findings suggest that much of the disparity between Chetri and Brahmin appear to be
driven by the differences in their geographies. If Brahmin and Chetri were to live in the
same geographic areas, the inequality between them would have been much smaller. The
remaining differences between Chetri and Brahmin are likely to be due to caste
discrimination.
The reasons for a disproportionately high level of geographic isolation among the
indigenous peoples are thought to be historical and political. Historically, indigenous
territories have been the targets of the Khas invasion. After the Khas invasion, much of
the fertile lands were expropriated to the Khas/caste peoples by the government, and
indigenous peoples were forced to move further into the hinterlands. Prior to the 1950s,
163
very few villages had any modern infrastructure development. In the 1960s, Nepal was
re-structured into 75 districts and 14 Zones. Each of the 75 districts was administered by
a CDO (Chief District Officer), appointed by the central government. The primary job of
a CDO is to maintain law and order in the district and to distribute development funds to
the villages in the district (Lama, 2011)
8
. Since the CDOs of all these districts are
generally Brahmin or Chetri (see Table 20), in effect, Brahmin/Chetris have become the
de facto rulers of all districts, including those in which indigenous peoples form the
majority of the population. Given the rampant nepotism, corruption and overt caste
favoritism among the Khas/caste bureaucrats (e.g. Bista, 1991), it is not unlikely that
these CDOs will favor villages of their own castes over those that belong to the
indigenous peoples. The asymmetric distribution of state funds to Bramhin/Chetri
villages, perhaps at the expense of indigenous villages, is likely to be the key factor that
drives indigenous villages into isolation.
Differential Return on Investment in Human Capital
The findings suggest that indigenous peoples are likely to be poor independent of
their education, health, and occupational status, and regardless of where they live. In
other words, indigenous peoples are poor not only because they have low education,
health, or occupation status, and live in a particular geography but rather that the system
systematically make them poor independent of these characteristics. Because of this,
indigenous peoples may have been discouraged from pursuing a better education, health
care, a new occupation, and even mobility to less-isolated geographies. Constitutional
8
Tara Lama, an indigenous journalist, is a research assistant based in Kathmandu. He assisted in collecting
archival data on CDOs of Nepal. He interviewed government officials on the role and responsibilities of the
CDOs.
164
prohibition of multilingual study in higher education is also thought to have further
prevented indigenous peoples from attaining the human capital necessary to move to a
higher-paying occupation.
For the Brahmins, on the other hand, it is clear that better education, better health,
better occupation, and better geographies provide better success in reducing poverty. To
the extent Brahmins are poor, the reason, it seems, is due to lack of individual motivation
for better education, health or occupation. This finding provide support for the human
capital theory (Schultz, 1960; etc.). Another reason, it seems, is geography. Individuals
living in isolated geographies are more likely to be poor than those living in not isolated
geographies, independent of caste or ethnicity (Table 16, M4).
Sources of Indigenous Poverty are Different from the Caste Poverty
The hypothesis- “H4: The observed differences in the risk of poverty between
various ethnic/caste groups will disappear when the individual level productivity
characteristics and geographic characteristics are controlled for’ – was not supported by
the data (Table 16, M4). This finding suggests that the sources of poverty among the
indigenous peoples are not the same as those of the caste people. The findings suggest
that even if indigenous peoples have the same level of education, health and occupational
status as the Brahmins, they are still likely to remain significantly at higher risk of
poverty than the Brahmin. It seems that the individual level variables (education, health,
occupation) may mitigate the risk of poverty for a certain degree, but they do not appear
to bridge the poverty gap that exists between Brahmin caste and indigenous peoples.
165
The determinants of poverty among the caste peoples appear to be, largely,
individual for the higher caste (Brahmin and Chetri), and caste-discrimination for the
lower-caste, Dalit. As stated in the previous section, the individual determents may
include lack of motivation for self-development, such as education, health or occupation
(Table 16). The findings indicate that theories that focus on individual characteristics,
such as human capital theory, appear to apply more to higher-caste groups.
Indigenous language appears to be a protective factor against inequality
When the indigenous language was added to the regression model, the disparity
between Brahmin and indigenous peoples was further exacerbated (Table 16, M4). This
suggests that the disparity between Brahmin and indigenous peoples is not driven by the
differences in their mother tongues. If the indigenous peoples were to speak the same
language (Khas) as the Brahmin, the inequality between Brahmin and the indigenous
peoples would have been much greater.
Furthermore, when the geographic and language variables were added to the
regression model, Newar, who were not significant in the previous models (Table 16, M1
–M3) became significant (Table 16, M4). The findings suggest that the disparity between
Newar and Brahmin would have increased if they both were to live in the same
geographic areas or speak the same language. Geography and language appear to drive
the relative economic advantage enjoyed by Newar. Findings suggest that Newars might
be relatively better off than other indigenous peoples only because of their geographic
proximity to the capital city--the most developed area and the center of the nation’s
economic activities.
166
Explaining poverty among the indigenous peoples of Nepal
There are several historical reasons why the indigenous groups may have been
“excluded” and how that exclusion might have contributed to their poverty. First, the
indigenous groups once held their own independent nations and polities prior to the
unification of Nepal. The territories of many of these groups were never conquered by the
Khas invaders. Rather, treaties were signed between the Nepal state and these nations.
Many local kings collected taxes in their territories until very recently, 2006
9
. The king of
Mustang, the Northwest district of the Himalayan regions of Nepal, for example, still
maintains some degree of sovereign rule over Mustang. Within the Kathmandu valley,
Chinia Lama collected taxes from the residents living around the Boudha area well into
the 1980s. Due to these historical connections to their land and culture, many of the
indigenous groups have recently come together as a social force and have begun
demanding autonomy of their territories. Some of the notable demands are autonomy of
the Tamangsaling (territory of the Tamang), Limbhuwan-Khumbhuwan (territories of the
Kiraties), Magaranti (territories of the Magars), Tharuwat (territories of the Tharus), and
Newa mandal (territories of the Newars). The findings of this study provide empirical
validity to the historical grievances of these groups. On the other hand, the emergence of
these groups provide a rationale for conducting historical analyses of their grievances as
reflected in the findings of this study.
9
In 2006, the 240 years of Khas monarchy as Hindu Kingdom was abolished, and Nepal
was established as Democratic Republic.
167
After Nepal was established as a nation-state in the 1770s, the indigenous peoples
lost their land and its resources, these lands then being expropriated to the Khas invader.
The Khas migrations depopulated many of the indigenous peoples in their territories.
More importantly, indigenous cultural institutions were destroyed. For example, killing of
a cow for meat consumption was an acceptable practice in the pre-Khas era. But after the
Hindunization of the country, killing a cow became illegal. Those who continued
practicing their culture were imprisoned or enslaved. Institutional constraints like these
may have stifled their income from trade and other economic activities. The destruction
of indigenous cultural institutions may have had a particularly devastating effect since it
prevented the transfer of traditional knowledge and wisdom to the next generation. The
lack of protection of property rights, combined with the destruction of the native culture
and constraints on the people’s traditional livelihood, may have contributed to indigenous
poverty.
Another historical reason for the exclusion of the indigenous peoples was the
opposition of the people to Hindunization of their land. The indigenous groups, such as
Tamang, Magar, Rai-Limbu, and Tharu were opponents of the Hindunization
(Khasanization) of their territories and annexation of their territories into the Nepali state.
These groups, in particular Tamang, were sidelined by the first constitution of Nepal,
Muluki Ain 1854, because of their fierce opposition to the Nepali state. Since the
indigenous peoples (except Hindu Newar) did not belong to the caste hierarchy, they
were arbitrarily designated as “lower” in the constitution. The Muluki Ain 1854
designated all non-caste peoples, including White Europeans, as “lower.” This
designation appears to have been designed primarily to exclude the indigenous peoples
168
from any socioeconomic and political processes of the Nepali state. The exclusion no
doubt limited their economic opportunities and mobility, and as a result they may have
become poor.
It appears that Tamang were the prime victims of the Muluki Ain 1854.
Historically, Tamangs were treated differently from the other groups by the Nepali state.
Tamangs are considered the protector of the Buddha Dharma in Nepal. Since Tamangs
were the Lama (priests) among the Buddhists, as were Brahmins among the Hindu, they
were least likely to be culturally subjugated by the Hindus. Since Tamang were not easily
tamed by proselytization, they were punished economically. They were prohibited from
working in any public offices (education sector, governments, army, policy etc.). Tamang
men were used as un-paid laborers to build palaces and roads, or as porters to transport
goods and services. Even until the 1950s (during the regime Juddha Samser Rana),
Tamang were forced to work as free laborers, while other laborers were paid their day
wage. Tamang women were used as concubines in the Khas palaces and later sold to
brothels in India. Tamang children were even prohibited from attending schools. The
Tamang community still suffers from the human trafficking problems perpetrated by the
Khas system. It wasn’t until the 1960s, when international organizations such as United
States Agency for International Development (USAID) provided funds to Nepal
government on the condition that education be accessible to all, that Tamang children
began to receive educational opportunities. These historical injustices, combined with
current structural barriers against them, appear to be plausible reasons why Tamangs are
at the highest risk of poverty.
169
The Gurkha Army may be responsible for the relative improvement in the
socioeconomic status observed among the Gurung, Magar and Rai-Limbu individuals.
During the unification process, the individuals who converted
10
to Hinduism and fought
against their own peoples and territories in favor of the Khas invaders were given special
status as Gurkha soldiers. Members of these groups were allowed to join the British-East
India Company as soldiers. Even after the British-East India Company was dissolved,
these groups continued to serve in the British Army. These Gurkha soldiers were exposed
to economic opportunities outside Nepal and were earning cash income in return for their
service. The relatively higher purchasing power of their cash income and other human
capital gained from exposure to the outside world may have helped them overcome their
poverty. The relatively better socioeconomic status, however, seems to have been
achieved at the cost of their own cultural and linguistic identity.
Unlike other groups, Newars likely enjoy a higher socioeconomic status due to
their close geographic proximity to the capital city. Newars live mainly in the urban areas
where access to education and economic opportunities are available. Some Newars,
especially those that are Hindu, were designated as “higher” in the Muluki Ain 1854.
These Newars do not experience as much discrimination as their indigenous cousins. If,
however, Newars lived in geographically isolated areas, their socioeconomic status may
not have been any better than other indigenous groups (Table 16, M4).
10
Unlike what …. () calls Sankritization- a process by which one converts to Hinduism in exchange for
economic benefits, the conversion is equally likely to have occurred due to forced proselytization.
170
Explaining poverty among the caste peoples of Nepal
Unlike the poverty among the indigenous peoples, the poverty among the Dalit is
thought to be primarily a function of caste discrimination. Dalits have historically been
treated as untouchable (i.e. Dalit cannot come in physical contact with higher castes-
Chetri or Brahmin without punishment). Historically, they have been excluded from
participating in any social and economic affairs of the state. They were trapped into
lowpaying occupations, such as tailoring, metal work (blacksmith), and entertaining.
Dalits are not just discouraged from changing their occupations, but according to the
Hindu religion, change is considered a sin. This internalized oppression may have
discouraged them from seeking alternative professions that was more profitable.
Furthermore, their work was not adequately compensated, and often the high-caste
Brahmin would demand labor and financial donations in return for their religious
services. Due to this system, many of the Dalits, who were not able to pay, became
bonded-labor – an intergenerational slavery system in which high caste Brahmin/Chetri
landlords force lowcaste Dalits into indentured labor – especially in the Western part of
Nepal where Khas/caste people are concentrated.
The findings on the socioeconomic status of Chetri were surprising. Since the
royal families of Nepal (Shah and Rana) belong to the Chetri caste, it was expected that
Chetries as a group would have a higher socioeconomic status than was observed in this
study. Shahs and Ranas (royal families) are by no means economically less well-off than
Brahmins. One plausible reason for this is that huge variation exists within Chetri.
Chetries who consider themselves as pure-Chetri (Shah, Rana, Thapa etc.) are thought to
be better off than Chetries who are degraded as non-pure. Non-pure Chetries may include
171
those who were given Chetri status for their service to Brahmin or King (e.g. foreign
diplomats, or public servants who have abandoned their original caste or ethnicity ) and
those fallen from their original caste- Khaseko Chetri (KC) due to inter-caste or
castethnic marriages. Pure Chetri vs. non-pure chetries were not separated in this
analysis. The relatively lower status of Chetri may have appeared due to the pooled
method used in this analysis. Further research on sub-group analyses are required to
determine the extent to which the observed socioeconomic status is the outcome of the
pooled method employed in this study.
Another plausible reason for the unexpected socioeconomic results of Chetri is
that a large number of Chetri live in the geographically isolated Western and Far-western
region of Nepal. Due to difficult topography and distance from the capital city,
infrastructure development is difficult to achieve. This geographic isolation may
contribute to their poverty.
A second but less plausible reason for lower socioeconomic status is the concern
regarding fidelity of the Chetri-caste data. It is suspected that lower-caste groups, such as
Dalit, may self-report as higher-caste such as Chetri. Since Cherti and Dalit physically
look alike, and since revealing one’s caste may invite humiliation if he/she is low-caste,
there is a slight chance that low-caste Dalit may have self-reported as Chetri. This may
have skewed the true distribution of Chetri-caste data. To rule out this possibility
completely, an identical survey with a nationally representative sample is needed. Clearly,
this is beyond the scope of this study. One approach to overcome this problem is to
conduct sub-analyses of Chetri data, identifying individual surnames and triangulating
this information with Geographic Information System (GIS) data. Since the primary
172
interest of this study is indigenous peoples and not caste peoples, this approach was not
implemented in this study. Researchers interested in Chetri caste should further explore
this.
Implication
One contribution of this study is to expand society’s understanding of why it
matters “who you are” and “where you live” regarding socioeconomic well-being.
Previous studies have pointed to race/ethnicity (Psacharaopoulos & Patrinos, 1994) and
geography (Sachs, 2005) as a source of variation in socioeconomic status across
populations. Others emphasize disparity in individual level productivity characteristics as
the source of variation in socioeconomic status. This study goes further to investigate
another source of variation--the institutions--and attempts to answer why race/ethnicity or
geography matters in the socioeconomic well-being of the population. The findings of
this study provide important implications for theory, method and practice.
The prevailing theories of poverty assume that all groups of people become poor
for the same reason. These theories do not take into account the unique historical
experiences of different groups of people. For example, human capital theory assumes
that Black and White Americans become poor for the same reason-- lack of human
capital. The findings of this study, however, suggest that the relationship between human
capital variables and poverty changes under different institutional conditions. While a
strong relationship was found between education and poverty for the Brahmin sample, no
significant relationship was found between education and poverty for the Tamang sample.
Similarly, while a strong relationship was found between occupation and poverty for the
173
Brahmin sample, no relationship was found between occupation and poverty for the
Tamang sample. Furthermore, while a strong relationship was found between geography
and poverty for Brahmin, no significant relationship was found between geography and
poverty for Tamang. The findings suggest that Tamang people become poor for entirely
different reasons than Brahmins. The findings imply that various groups of people
become poor for entirely different reasons, perhaps due to their unique historical
experiences.
Consistent with the findings of Psacharopolous and Patrinos (1994) in the
Americas, the findings of this study point to racial discrimination as a determinant of
poverty among the indigenous peoples. However, the findings of this study further
enlighten by pointing out that the source of discrimination is rooted in the society’s basic
institutions, such as a constitution--at least in the case of Nepal. These institutions appear
to create an unequal playing field at the onset, which then leads to inequality in treatment
(discrimination). Inequality, in turn, appears to drive the poverty among the peoples.
Implication for Method
This study employed multilevel modeling (pooled method) to examine the poverty
status of each of the indigenous groups and caste groups for which data was available.
The findings revealed a huge within-group variation among the indigenous peoples and
among the caste people. In fact, the within-group variation among the caste people
(Brahmin and Dalit) was much greater than the between-group variation between
indigenous group (Tamang) and caste group (Dalit). The findings suggest that Dalit and
Brahmin, although both belong to non-indigenous or caste-groups, are in no way similar
174
to each other, and therefore cannot be grouped as one class of people. Similarly, although
Tamang and Newar are both indigenous peoples, the huge economic disparity between
these two groups makes it illogical to treat them as one class of people. The findings
suggest that any analysis of indigenous peoples must consider inter-ethnic differences
within the indigenous group. Studies that classify population into binary indigenous vs.
non-indigenous categories gloss over the variation within these categories and are likely
to risk wrong estimates about the true relationship.
In addition to multilevel modeling, this study conducted geographic analyses at
sub-regional level by cross-classifying the ecological zones and development regions.
The results show a huge within-region variation in poverty (Western region: 18.46% to
82.11%). The within-region variation was much greater than the between regions
variation (27.28% to 60.05%). The findings suggest that regional level analyses that
ignore the within-region variation are likely to risk a wrong estimate of geographic
effects on poverty. Sub-regional level analyses appear to provide more precise effects of
geography than region or zone level.
This study controlled for ethnicity/caste and geography simultaneously in the
regression model. The findings revealed that indigenous peoples were invariably
clustered into certain geographies, whereas caste people were scattered throughout the
country. Due to geographic clustering of ethnicity, the effect of ethnicity on poverty is
often not distinguished from the effect of geography on poverty. This approach helps
overcome this problem.
Finally, this study used Geographic Information System (GIS) to perform spatial
relationships between ethnicity/caste and poverty. The maps produced from these
175
analyses reveal that the poverty map coincides with the territories of the indigenous
peoples. The findings suggest that GIS analyses help determine relationships which are
often not possible to see in quantitative or qualitative analyses alone.
Implication for Practice
Much of the programs and policies purported to help the poor focus on changing
the behaviors of the poor. These policies and programs often assume that providing
knowledge or skills to the poor will solve their poverty. Social workers, in particular,
have a reputation of being “poverty pimps”, a notion that social workers actually live off
of poor people under the guise of helping them. While this label may be applied to other
professions as well, such as doctors (disease pimp), lawyers (criminal pimp) and so on,
the underlying concerns appear to have some validity-- social workers do little to bring
about change in the social structure that makes people poor in the first place. Helping a
poor person who needs healthcare or shelter is one thing, but helping a person such that
he/she does not have to be poor in the first place is entirely another. Such an approach
requires changing the conditions rather than the individual’s behavior. Unfortunately,
even those programs purported to focus on structure do not go beyond policy structure.
Social policy, at best, brings about change in distributional structure, such as income
distribution (E.g. Medicaid system in the U.S.). However, in poor countries, where the
countries themselves are bankrupted, leaving little to be distributed, the change in
distributional structure does little to improve the condition of the poor. On the other hand,
micro- interventions such as micro- financing, micro-credit and so on do little to the
economy of scale. These approaches are often driven by traditional wisdom such as
176
‘teaching how to fish is better than giving a fish’, and are not informed by theories or
evidences that are empirically valid.
The findings of this study reveal that, first, different groups of people become
poor for entirely different reasons. Second, the lack of knowledge (education) or skill is
not the primary determinants of poverty, at least among the indigenous peoples. In
addition, under the current institutional condition, even if poor people have education or
other skills, groups such as indigenous peoples and lower-caste are still likely to remain
poor.
According to the findings of this study, change needs to come in the form of
freedom from the institutional constraints imposed on the poor by the political elites of
the society. Designing a society’s institution, such as a constitution, seems to be a starting
point, particularly in the case of Nepal. Such institutions should be unbiased and should
provide incentive structures such that each person living in that society can advance
his/her well-being to the fullest potential according to his/her own culture. A constitution
that allows for self-governance among indigenous peoples (Cornel, 2002; 2005), protects
indigenous property rights (Sened, 1997), promotes an indigenous education system (i.e.
system which advances indigenous language, culture and technology) and indigenous
health systems (i.e. system which advances indigenous health knowledge and medical
technology) may be the most powerful tool to help the poor attain well-being. Social
workers have important roles to play in creating such institutions and in effecting the
desired changes in the society, such that all humans, independent of who they are or
where they live, gain the ability to live a good life.
177
Implication for the development of indigenous peoples:
1. Global and national institutions, which constrain indigenous peoples’ capacity to
develop their own communities, must be eliminated.
2. Indigenous peoples should (be allowed to) design institutions to serve their best
interests.
3. Indigenous peoples should (be allowed to) develop their communities in their own
ways.
4. Indigenous peoples should (be allowed to) govern their own territories
(selfgovernance).
Limitation
The findings of this study were tempered by a number of limitations. First, this
study employed cross-sectional design. The study did not directly test the theories that
guided this research. The findings do not establish a causal relationship between the
independent and dependent variables.
Another recognized limitation is that this study utilized secondary data sources.
The scope of this research required a large data set representative of all indigenous
peoples and ethnic groups. Collecting a primary dataset on a national scale was beyond
the scope of this dissertation study. Therefore, DHS data presented the best among the
available data sources that could answer many of the research questions of this study. The
limitation of using secondary data in general, and the DHS dataset in particular, is that
178
there is little control over what has been collected. There is little control over survey
design, population coverage, and the types of questions asked.
The use of the asset-based Wealth Index as a measure of poverty is also a
limitation for this study. Wealth index is a multidimensional construct. It is a latent
construct rather than a direct measure of absolute wealth. Wealth Index does not include
income but instead includes measures of individual and household welfare such as land,
houses, livestock and other household items including access to electricity, piped-water,
and quality of housing. In this method, the probability of being poor is the probability that
a person or household belongs to the bottom 40% on the asset-based wealth/welfare
distribution. Thus this index does not directly measure poverty and does not tell how poor
an individual is in absolute terms. In other words, it is a relative measure of poverty. The
measurement of wealth index could be improved by decomposing it into its individual
dimensions and then reconstructing it with theoretically meaningful dimensions.
However, since the purpose of this study was to understand the determinants of the
overall socioeconomic well-being of the people rather than develop a measurement
model, the development of a new measurement method of Wealth Index was beyond the
scope of this study.
An additional limitation of this study is that it does not adequately capture health,
geographic isolation, and institutional variables. Health status variables were not
available in the DHS dataset, particularly for male samples. Inclusion of health variables,
such as tuberculosis, cancer, or maternal mortality would have strengthened the study.
Likewise, the geographic isolation variable does not capture all dimensions of isolation.
For example, some villages may be isolated, but isolation may also provide economic
179
values to its residents due to tourist attraction and revenue. Due to lack of data, this study
did not decompose the geographic isolation variables into those that attract tourists and
those that do not. Therefore the extent of positive value of geographic isolation to
economic well-being is not accounted for in this study. Similarly, measurement of
institutional variables is weak. Institution was partly captured by language and ethnicity
and partly through geographic isolation and development regions. Factors such as rights
to own property, security of contracts, institutions on financial inclusion (banking),
distance to Health Post, and resource allocation to different development regions are
important dimensions of institutions and their inclusion would have strengthened the
study. Collection of these variables would require tremendous amounts of time and
resources, which were beyond the scope of this study.
Finally, there is a concern of bias in the measurement of caste in DHS data,
another study limitation. Due to caste discrimination, there is a possibility that the
individuals of lower-castes, such as Dalit, may self-report as higher caste, such as Chetri
or Brahmin. Diagnosing this problem is challenging since there is no apparent racial or
cultural differences between Dalit, Brahmin and Chetri. They look similar (they look
Khas or Indian) and speak the same language (Khas language). Fortunately, the DHS data
identifies sub-castes or surnames for some of the caste groups. Since Dalit, Chetri and
Brahmin do not typically live in the same communities (tole), triangulating this
information with the geographic information system helps minimize this concern.
Furthermore, if this concern were valid, we would see under-sampling of the Dalit. In the
DHS women sample, 11.7% (weighted) self-identified as Dalit. This figure is comparable
to the national estimate. To completely rule out the possibility of bias, however, another
180
national survey, representative of all caste and ethnic groups, is needed. Clearly, this is
beyond the scope of this study. Furthermore, while the possibility of this bias is a
concern, to date no published reports have documented any inconsistencies in DHS data,
questioning the reliability of its measures. The bias is not a concern for measuring
indigenous groups.
Despite these limitations, this study represents the first- ever analysis of poverty
among the indigenous peoples of Nepal using a nationally representative sample of
women, men and households. Future researches should be mindful of the limitations
identified in this study and incorporate the suggestions for improvement.
Agenda for Continued Research
The findings of this study provide room for more questions rather than offer a
conclusive theory about the causes and consequences of poverty among the indigenous
peoples in general, and Nepal in particular. To begin with, this study was able to test only
a small part of the larger theoretical framework in which poverty was conceptualized as
cyclical. While this study lays a foundation for empirical studies of poverty among the
indigenous peoples in Nepal, it raises many questions: Why are the Tamang, who live in
the surrounding hills of the capital city, even poorer than those (Chetri) who live in the far
western region? Why are the Tamang villages that are so close to the capital city isolated?
Why are Tharu who live in Terai, the most fertile land of Nepal, as poor as Tamang, who
live in the Mountain? How have Brahmin and Chetries become so successful in
disfranchising the indigenous peoples and controlling the government of Nepal for over
181
240 years? Understanding these political processes has far-reaching implications--more
than simply understanding how a country, such as Nepal, becomes poor. This study
provides only anecdotal explanations to these questions.
What we know about poverty from empirical studies seems to provide only a
myopic view of this phenomenon. There is a need to understand poverty in the larger
political, institutional, and geographic contexts. As pointed out in the limitation section,
this study, too, was tempered by a number of limitations. There are theoretical and
methodological problems that need to be addressed in the study of indigenous peoples.
For example, the Western conceptualization of poverty, as something that needs to be
attacked or fought against, clearly seems to be irrelevant to the condition of indigenous
peoples. From the anecdotal evidences, it appears that indigenous peoples have been
made impoverished rather than that they became poor through their actions. They appear
to be systematically prevented from advancing their own interests rather than being left
behind in evolutionary processes. However, to confirm the validity of these claims, there
is a need to study indigenous poverty over time and across space.
As indigenous peoples and their culture are becoming more important as a source
of knowledge for various academic disciplines, including Anthropology and Social
Sciences, there is a greater needs to understand and urgently promote indigenous
wellbeing. In-depth studies of each unique group and large scale surveys to understand
the underlying factors that affect each group are imperative. Longitudinal surveys are
needed to document the changes in socioeconomic conditions over time. There is a need
to replicate this study in other countries with different institutional contexts such that
182
knowledge generated from this study can be validated and the findings may be
generalized.
CHAPTER IX: CONCLUSION
This study began with a question: Why do some people (or groups) become poor
while others become rich? This question was informed by the previous works,
particularly of North (1990), who asked a similar question: Why do some countries
become rich while others become poor? The rationale for pursuing the research question
was that, to better understand why some countries become poor, it is necessary to
understand who the poor people are in those countries and why they are poor. Drawing on
the theory of institutional design (North, 1990) as a guide, and Nepal, a poor country as a
case, this study investigated a specific research question: Who are the poor of Nepal, and
why are they poor?
In Nepal, there is a reason to believe that the lower-caste groups (Dalits) are likely
to be at higher risk of poverty than a higher-caste group (e.g. Brahmin). However, there is
no reason to believe that indigenous peoples of Nepal should be at higher risk of poverty
than the caste group (Brahmin). Indigenous peoples do not belong to the caste system and
are themselves the high priest group (e.g. Lama, Bonpo, Dhami, Jhankri etc.). Indigenous
peoples are, in fact, expected to be economically better off than the caste people since
they are the original inhabitants of the land and the rightful owners of the land and the
resources on them. The extent to which indigenous peoples in Nepal are poor, we wish to
know why.
183
Poverty of the indigenous peoples and their countries in the Americas, Africa,
Australia, and India has largely been attributed to the colonization of their territories by
the Europeans (Psacharopoulos & Patrinos, 1994; Eversole, 2005; Carino, 2009 etc.).
However, since Nepal was never colonized by the Europeans, colonization could not be a
plausible explanation for poverty of the indigenous peoples and of Nepal. An alternative
explanation to the poverty of the indigenous peoples of Nepal and their country as a whole
is warranted.
The findings of this study reveal that, like in other parts of the world, indigenous
peoples in Nepal are at a significantly higher risk of poverty than non-indigenous people
(Brahmin). In fact, some of the indigenous groups, particularly Tamang, Magar, Tharu,
and Rai-Limbu, are at as high a risk for poverty as the lower-caste groups (Dalits).
The findings suggest that poverty among the indigenous peoples in Nepal is not
solely driven by the lack of individual productivity characteristics or geography alone,
although they were significantly associated with poverty. There is a differential return on
investment in human capital – return on investment in education is less for indigenous
peoples than for Brahmins. Geographic isolation is also a significant risk factor of
poverty for indigenous peoples. However, beyond the effect of individual characteristics
and geographic isolation, indigenous peoples appear to be impoverished primarily due to
discrimination. Discrimination against the indigenous peoples in Nepal appears to be
institutionalized and practiced at multiple levels. At an individual level, indigenous
peoples appear to be discriminated against in education, health and occupation.
Indigenous peoples were less likely to receive post-secondary education and healthcare
services and more likely to work in low paying jobs such as farmers, porters and laborers.
184
In schools, indigenous children were treated poorly by their teachers. In healthcare
service agencies and hospitals, indigenous peoples experience humiliating treatments--
they are asked to pay or show money before services are delivered or denied services if
they can not pay. A majority of the teachers and healthcare professionals are
nonindigenous people (Brahmins and Chetries). At a community level, indigenous
villages are less developed (e.g. lack of electricity, roads, schools, universities, hospitals,
irrigation) than Brahmin villages. Indigenous villages appear to be resource-deprived and
left out of the development processes. At a national level, indigenous peoples seem
systematically excluded in all forms of governance. Even if indigenous peoples have the
same level of education as the Brahmins, indigenous peoples seem less likely to find
government jobs or be able to move to a position of power. Furthermore, when
indigenous peoples try to develop business entrepreneurship, they experience humiliation
and discouragement at every step of the business development. When the indigenous
peoples go to CDOs (Chief District Officers) to register a business, company or
organization or to apply for a citizenship card and passport, they are often asked for bribe
money. They are delayed or even denied their basic civil rights to citizenship and to form
social organizations if they do not offer a bribe or if they do not please the government
officials in some other ways. Over 93% (70 out of 75 districts) of the CDOs were
Brahmin or Chetries in 2011.
The findings further reveal that the indigenous peoples who live in closer
proximity to Brahmins appear to be at a higher risk of poverty than those who do not. In
other words, indigenous peoples living in the villages which were invaded by the
Brahmins (or Khas peoples) and in which Brahmins still live today seem worse off than
185
those living in villages without Brahmin residents. This finding was counter-intuitive
because conventional wisdom suggests that the presence of a few wealthy or educated
persons in a community have spillover effects on those who are on the periphery. What
appears to happen in Nepal, in contrast, is that a few clever or wealthy Brahmins exploit
the masses of indigenous peoples, who are less educated or less wealthy. Consequently,
the closer these groups live in geographic proximity, the easier it is for exploitation to
occur. The findings provide further support to the discrimination hypothesis. They
suggest that the culture of discrimination and other social ills, including the caste-system,
arrived in Nepal with the arrival of Brahmin/Chetries, and this discrimination appears to
be the reason why indigenous peoples in Nepal are poor.
The findings also suggest that the culture of discrimination in Nepal is
institutionalized and systematic, and the root of this system may be traced back to the
first national law or constitution of Nepal, Muluki Ain 1854. The study revealed that the
indigenous groups that are at the highest risk of poverty were the groups which were
designated as “excluded or lower” by Muluki Ain 1854. The findings suggest that Muluki
Ain 1854 effectively provided license to Brahmins and Chetries to commit all forms of
injustices and atrocities against the indigenous peoples and the nation without legal
consequences. Although counterfactuals are difficult to prove, the evidence suggests that
Khas migration and subsequent establishment of a Nepali state and its institutions,
particularly Muluki Ain 1854, appear to be the only plausible explanation to why
indigenous peoples and their territories in Nepal are poor.
Although there have been several amendments to the Muluki Ain 1854 since its
inception and new constitutions have been written, this document seems to have a
186
significant and lasting impact on the impoverishment of the indigenous peoples and the
county as a whole. The Muluki Ain 1854 and its subsequent amendments appear to have
been designed by the Brahmins and Chetries to keep the indigenous peoples and the
Dalits of Nepal uneducated and in poverty– most likely so that they can be easily
exploited. To some extent, the historical processes by which the indigenous peoples of
Nepal became poor appear to parallel the colonization of their counterparts in the
Americas, Africa and Australia. These findings provide some empirical support to the
theory of institutional design (North, 1990).
Given these historical and institutional discriminatory structures against the
indigenous peoples, economic policies and programs that are focused on changing the
productivity characteristics of the poor are less likely to bring about needed change in the
socioeconomic status of the indigenous peoples. Under current institutional conditions,
even if the indigenous peoples have the same level of human capital endowment or other
productivity characteristics, they are unlikely to be as economically and politically
welloff as the non-indigenous people (Brahmin). Restructuring of the institutions,
particularly the constitution of Nepal, seems imperative if a substantive gain in the well-
being of the indigenous peoples and the country as whole is to be achieved.
How Nepal as a state will respond to the poverty of the indigenous peoples will
determine the future of the country as a whole. Certainly Nepal cannot afford to continue
with its racist institutions and policies against the native people who constitute the
majority of the population. Likewise the government as an organization can no longer
continue to be used as the exclusive club of the Brahmin/Chetries. For Nepal to survive as
a nation-sate, it needs to redesign its institutions, particularly the constitution.
187
As Nepal is currently undergoing the process of restructuring the country, how it
redesigns its institutions remains to be seen. For the moment, it seems imperative that this
process be informed by empirical evidences of what works best for all people of Nepal--
not just for the Brahmin/Chetries. At the minimum, the government needs to be inclusive
and should respect and tolerate different ethnic groups, cultures, languages and faiths. It
should promote uniform development of all ethnic groups. One approach to achieve this
goal might be to redesign the constitution in ways that allow indigenous peoples to
selfgovern their own territories (Cornel, 2002; 2005), protect their property rights
(Sened,1997), and promote an indigenous education system and health system. Such an
institution will reduce external constraints on the indigenous peoples and provide
incentives to develop their own communities. When each ethnic and caste group develops
its own villages in its own way, the country as a whole will be developed and the wealth
of the nation will grow. Evidence suggests that people, including indigenous peoples, are
capable of governing themselves, managing their common pool resources, and
determining their own future without the dictate of the state (Ostrom, 1990). The new
institutions should be unbiased and should provide incentive structures such that each
person living in that society can advance his/her well-being to the fullest potential
according to his/her own culture. Only such institutions can guarantee to all people the
capacity to live a good life.
