Open Collections

Digital Himalaya Journals

Inter-Regional Variations in the Inequality and Poverty in Bhutan Mehta, Sanjeev between 2007-06 and 2007-08

Item Metadata

Download

Media
dhimjournal-1.0365199.pdf
Metadata
JSON: dhimjournal-1.0365199.json
JSON-LD: dhimjournal-1.0365199-ld.json
RDF/XML (Pretty): dhimjournal-1.0365199-rdf.xml
RDF/JSON: dhimjournal-1.0365199-rdf.json
Turtle: dhimjournal-1.0365199-turtle.txt
N-Triples: dhimjournal-1.0365199-rdf-ntriples.txt
Original Record: dhimjournal-1.0365199-source.json
Full Text
dhimjournal-1.0365199-fulltext.txt
Citation
dhimjournal-1.0365199.ris

Full Text

 Inter-Regional Variations in the Inequality and Poverty in
Bhutan
Sanjeev Mehta
Abstract
The findings of this sample study suggest existence of high income
disparities between the urban and rural areas and across
dzongkhags. Urban areas contribute about 69% of the total
income, and Average Monthly Per Capita Income of urban areas is
almost four and a half times higher than that of rural areas. The
Gini coefficient value is higher in the urban areas (0.58) as
compared to the rural areas (0.36) reflecting higher income
inequality in the urban areas. Largely, income disparities can be
explained in terms of the pattern of productive assets ownership.
81 % of the productive assets are found to be concentrated in the
urban areas. The skewed pattern of the disbursement of bank
loans indicates that income inequality is also policy induced.
The finding suggests that the head count ratio is 66.23 if
measured in terms of upper poverty line and 50.66% on the basis
of the lower poverty line. Poverty is more a rural phenomenon as
about 86% of the poor live in the rural areas. FGT index of
normalized poverty gap is 22.61%) and poverty gap in the rural
areas is more than three times the poverty gap in urban areas.
Dzongkhag-wise, the highest incidence of rural poverty is found in
Pemagatshel and Samdrup Jongkhar and lowest incidence
poverty is recorded in Chukha. The poverty decomposition study
conveys that farmers, private sector employees and illiterates are
among the most vulnerable groups to the incidence of poverty.
One important policy implication that emerges from this analysis is
that poverty alleviation measures should be concentrated in those
areas where the ratio of ANPG/AHCR is higher. An appropriate
data base on poverty would make the poverty alleviation
measures targeted and consequently more effective in terms of
reducing the magnitude of        absolute poverty.
Senior lecturer, Sherubte College, Kanglung
38
 Journal of Bhutan Studies
Background
At the core of modern development economics is the issue of
wide spread poverty and growing inequality. Simon Kuznets
(1955) in his 'inverted U curve' hypothesis suggested that in
the early stage of economic growth income distribution tends
to worsen and in later stages it tends to improve. As modern
economic growth is spreading across the globe, the problem is
to not only to increase the size of the cake but also to ensure
that it is equitably distributed. In the initial phase, economic
growth tends to accentuate distributional disparity.
Economic growth is essential for improvement in the living
standards of the population and to reduce absolute
deprivation (poverty). It is through the process of trickle down
that growth benefits percolate to the lowest strata of the
society. The increased disparities in the distribution of
income both across the population groups and between
different regions, which are widely experienced in the
developing countries reflect the failure of the trickle down
process. Income inequality is an outcome of skewed
distribution of factors of production both in terms of quantity
and quality, strategy of economic growth, inappropriate social
and political institutions, lack of or inadequate capabilities
and functioning of the population, etc. A.K. Sen (1984)
maintained that absolute deprivation in terms of personal
capability relates to relative deprivation in terms of
commodities, income and resources.
In the 1990s, Bhutan witnessed acceleration in the growth
rate of GDP. Bhutan's economy has grown significantly since
1990. It registered an average annual real growth rate of GDP
of 6.07% in the last decade (NAS 1980-2000). During the
given period Bhutan's population also expanded at a rate of
3.4 % annuaUy. The very high growth rate of the population
caused GDP per capita to grow moderately close to 2.6%
annuaUy. Bhutan's per capita GNP is about US $640 (Source:
39
 Variations in the Inequality and Poverty in Bhutan
State of world's children 2003). If the inverted U curve
hypothesis is to be believed it would mean that this growth is
accompanied by growing inequality. According to UNICEF
(2003) in Bhutan, the poorest 40% of the population receives
as tittle as 13% share in household income, whereas the top
20% population receives as much as 49% of the household
income. The UNICEF database highlights the high income
disparities. But this database cannot be disaggregated further
to review interregional differences.
Graph No. 1: Growth rate in 1990s
Source: NAS 1980-2000 (CSO, Planning Commission)
Various development literatures indicate that relative poverty
degenerates into absolute poverty. If a country has large
income disparity, there is a greater possibility of higher
incidence of absolute deprivation. Regional disparity in terms
of economic growth tends to accentuate income disparity and
the laggards have greater incidence and extent of absolute
poverty.
There are two important sources of information on the extent
of absolute poverty in Bhutan, one is Household Income and
Expenditure Survey (HIES) 2000 and another is UNDP
estimates. According to RGOB (2001) average monthly per
capita income of Bhutan is just Nu. 1200 (which means less
than $1 a day, which is considered below the global poverty
tine) and approximately 27% of the population lives below the
40
 Journal of Bhutan Studies
poverty tine. A poverty analysis report on Bhutan published
by UNDP in 2004 suggests that about 31.7% of the
population lives below the poverty tinei. As far as absolute
poverty is concerned, regional disparities are very high. Head
count ratio shows high variation across the regions- 48% in
the Eastern Bhutan, 29.5% in the Central Bhutan and 18.7%
in the Western Bhutan. Even urban-rural differences in the
absolute poverty are very high as 38.3% of the rural
population lives below poverty line as compared to just 4.2%
of the urban population.
The poverty analysis undertaken by UNDP has highlighted
the fact that the problem of poverty and inequality in Bhutan
is not only existent but is also significant. StiU, the UNDP
report cannot be disaggregated to the regional level, hence it
cannot be used to draw inferences about the prevalence of the
twin problems of poverty and inequality at the micro level.
The central objective of this study is to identify the extent of
inter-regional variation in the magnitude of absolute and
relative poverty and to find out the possible explanatory
variables affecting the disparity.
Methodology
This sample study is primarily based on primary data
coUection from 6 dzongkhags: Chukha, Haa, Bumthang,
Lhuntshe, Pemagatshel and Samdrup Jongkhar. These 6
dzongkhags are so selected as to provide representation to
Western, Central and Eastern Bhutan. From selected
dzongkhags samples were coUected from rural and urban
areas. The rural urban samples were planned to be collected
in a ratio of 3:1, as about 21% of Bhutan's population lives in
urban areas. But finaUy the proportion of rural samples
declined to 65% due to low response rate and greater
rejection of the questionnaires due to incomplete or
inconsistent information. A stratified convenient sampling
process was used in this study as an appropriate sampling
i   Finding  of this   study were  reported  in  Kuensel  (the   national
newspaper of Bhutan), dated 25 October, 2004.
41
 Variations in the Inequality and Poverty in Bhutan
frame was not available. Stratification was done to
incorporate appropriate size of rural and urban samples as
weU as to provide appropriate representation to different
categories of occupation. The names of the dzongkhags and
the rural and urban areas covered in this study are given in
the table no.l.
Table No. 1: Regions covered in the study
Dzongkhag
Urban
Rural
Bumthang
-
Chumey, UraTrabi, UraTroepa,
UraTarsang and Ura Chari.
Chukha
Phuntsholing
Phuntsholing goenpa
Haa
Haa, Kastho
Yangthang, Hatey, Paytasima,
Tokey, Ingo, Chimpa, Kibri,
Takchu Goenpa, Bagana, Bjang
Goenpa, Kana and Jyemkhana.
Lhuntshe
Lhuntshe
Gangzur, Khoma, Phasidung,
Budur and Chokhor
Pemagatshel
Pemagatshel
Bartsheri, Moshizor, Dungjung,
Shumar, Gopini, Bangdala, Kheri
Goenpa, Lower and upper
Gypsem.
Samdrup-
Jongkhar
Samdrup Jongkhar
Devathang, Lamsarang, Wooling
and Sekpasang
The data were coUected through personal interview and
questionnaire. Sample units for this study are households.
For the study size distribution of personal income, individuals
in the working age group from each household were
identified. Working age group is defined as the age group
above 18 years subjected to the condition that either being
employed for any period in last 365 days on 31 December
2004 or sought employment during the same period. For the
poverty analysis we have used the concept of household
income and for the personal income distribution we have
used the concept of personal income.
In many occupations (such as agriculture and business)
income is contributed by combined labour of the household
members, in this case this income is treated as the income
42
 Journal of Bhutan Studies
occurring to the head of the household.
For the poverty analysis we have used two criteria: 1) average
monthly per capita income of Nu.750. This is based on the
upper poverty tine criterion of Nu.748.1 per capita per month
as used by UNDP for its poverty analysis report. But in this
study we will define it as the lower poverty line (LPL). 2)
Average monthly per capita income (AMPCI) of Nu. 1200,
which was calculated by HIES (2000) as AMPCI of Bhutan,
which is even lower than $1 per person per day criteria of
defining international poverty line. This criterion would be
used to define the upper poverty line (UPL). This is to
measure the income shortfall from the average level of per
capita income per month. This is an arbitrary criterion based
on our assumption that a person should at least acquire a
decent minimum standard of living comparable to the average
living standard in order to avoid any form of deprivation and
discrimination. This assumption is basicaUy drawn from our
individual assessment of Sen's (1993) writing on well-being,
especiaUy from one statement: 'The functioning of well-being
vary from such elementary ones ...to the complex one such as
being happy, achieving self-respect, taking part in the life of
the community, appearing in public without shame."
Occupationally, individuals are divided in 9 categories. The
list of the occupation and their respective codes is given in
the table no. 2. In case if a person is engaged in more than
one occupation, the occupation of that individual is further
divided in two categories: primary and secondary. Primary
occupation is defined as the occupation which earns greater
income to a person and the occupation from which a person
derives lower % of its total income is termed as secondary
occupation.
Table No. 2: Occupation categories
Occupation code
Occupation
0
Unemployed
1
Farming
2
Artisanship
4
Business (Trade and manufacturing)
43
 Variations in the Inequality and Poverty in Bhutan
5
Government employees
6
Semi govt. Employees
7
Private sector employee
8
Self employment in the informal activities
9
Hired employees in informal sector (daily
wage earners)
10
Religious occupation (monks)
We interviewed more than 500 individuals across the 6
dzongkhags but after applying GIGO method the effective
sample size became 456.
The primary data collected are complemented with secondary
data for further analysis. The type of the secondary data used
and its sources are identified at the appropriate places in the
report.
AU the statistical tests done in this report are either done
manually or carried out using Excel worksheet of Microsoft
office XP.
Concepts Used
In any study related to income distribution it is necessary to
select an income concept, which is theoreticaUy acceptable
and practically applicable. In this paper the concept of earned
income is used. The earned income is pre tax income that
excludes transfer payments. The concept of earned income is
based on SNA guidelines that include both the actual and
imputed income from aU the sources, earned in cash or in
kind. The reference period for the income estimates is the
calendar year ending in December 2004. Average Monthly Per
Capita Income (AMPCI) is calculated from the monthly
household income by dividing it from number of the members
in the household. Total personal income is defined as the sum
of factor income earned from varied sources by an individual.
Assets are defined as productive real assets which include:
land, other fixed capital and financial stocks. Value of the
land is estimated at a blanket rate of Nu. 50,000 per acre for
44
 Journal of Bhutan Studies
wet land and Nu. 10,000 per acre for dry land. This is done to
avoid regional variations in the real estate prices and to make
the data comparable across the regions. Other fixed capital
and financial stocks are valued at their current market price.
For this information we have solely depended on the
information rendered by the respondents.
Findings of the study
Findings of the study are divided into three parts. Part-1 is
discussion about sample characteristics. Part-2 deals with the
disparity in the size distribution of income. In part-3 the
magnitude and extent of absolute poverty is discussed.
Part 1: Sample Characteristics
Regional distribution of the total 456 samples is given in the
table no. 3. 298 samples (65.35% of the total) are from rural
areas and 158 samples (34.65% of the total) are from urban
areas. Dzongkhag-wise sample distribution is not based on
the weight of their respective population share because
dzongkhag-wise population figures are unavailable.
Table No. 3: Region-wise distribution of samples
Dzongkhag
Rural
Urban
Total
Bumthang
70
0
70 (15.35%)
Chukha
4
93
97 (21.27%)
Haa
65
24
89 (19.51)
Lhuntshe
99
1
100 (21.92%)
Pemagatshel
32
6
38 (8.33%)
Samdrup Jongkhar
28
34
62 (13.60%)
Total
298 (65.35%)
158 (34.65%)
456
311 samples (68.2%) are male and the remaining are female
samples. In the rural areas 65.1% samples are male and
74.05% of the urban samples are male. Larger LFPR among
the rural females as compared to their urban counterpart is a
common characteristic in the predominantly agrarian
societies. Occupational profile of the samples is given in the
table no 4.
45
 Variations in the Inequality and Poverty in Bhutan
Table No.4: Occupational distribution
Occupation
Code
Bumthang
Chukha
Haa
Lhuntshe
Pemagatshel
S/Jongkhar
Total
R
U
R
U
R
U
R
U
R
U
R
U
R
U
GT
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
1
1
45
0
1
1
42
1
79
0
24
0
18
0
209
2
211
2
9
0
0
0
0
0
0
0
1
0
0
0
10
0
10
4
11
0
1
28
6
20
5
1
3
3
8
14
33
66
99
5
3
0
0
17
8
0
6
0
2
1
2
9
22
27
49
6
1
0
0
9
0
1
0
0
0
2
0
2
1
14
15
7
1
0
2
32
2
0
0
0
1
0
0
5
6
37
43
8
0
0
0
6
3
2
9
0
0
0
0
3
12
11
23
9
0
0
0
0
0
0
0
0
0
0
0
1
0
1
1
10
0
0
0
0
4
0
0
0
0
0
0
0
4
0
4
Total
70
0
4
93
65
24
99
1
32
6
28
34
298
158
456
R= Rural
U=Urban
46
 Journal of Bhutan Studies
Farming (code 1) is the largest source of occupation as
46.27% samples are farmers. Farming is virtuaUy the
predominant source of occupation in rural areas as it is the
main source of livelihood for 70.13% of the rural samples.
Business including trade is the second most important form
of occupation as it provides occupation to 21.71% samples. In
the urban areas business is the main source of occupation as
it involved 41.77% of the urban samples and in rural areas it
provides employment to 22.14% ofthe rural samples. There is
a greater variation in the occupation profile between the
dzongkhags. In the rural areas of Lhuntshe and Pemagatshel
dzongkhags the percentage of farmers is 79.79 and 75
respectively. In the rural areas of Haa, Bumthang and
Samdrup Jongkhar the share of farm based activities is about
64%. This indicates that rural areas in the eastern
dzongkhags provide fewer opportunities for occupational
diversification. Government sector provides employment to
about 10.7% ofthe total samples.
Greater rural urban difference in the scope of government
employment is reflected in the higher percentage of urban
samples ie. 17.08% are engaged in government sector jobs as
compared to only 7.38% of the samples in the rural areas.
The private sector plays a very marginal role in creating jobs
in rural areas as it employs only 2% of rural samples; in
urban centres the role of private sector in creating jobs is
more significant as it provides employment to about 23% of
the urban samples. The inter-regional disparity in the growth
of private sector is seen from the different scope of private
sector in creating jobs across the dzongkhags. In Chukha,
the private sector employs about 35% of the samples,
whereas its proportion ranges between 0 to 2.5% in other
dzongkhags except Samdrup Jongkhar where it provides
employment to 8% samples. Low job creating capacity of the
private sector reflects highly inadequate development of the
private sector everywhere except in the case of Chukha
Dzongkhag. Other occupations are of lesser significance
across the regions.
47
 Variations in the Inequality and Poverty in Bhutan
About 27.19% of the total samples also take up secondary
occupation to supplement their income from primary activity.
The percentage of the individuals undertaking secondary
occupation is higher in rural areas (33.56%) as compared to
that in urban areas (13.92%). The inter-dzongkhag variation
is still greater. The most common form of secondary
occupation in rural areas is in the informal sector as daily
wage labour followed by handicraft. In the urban areas
agriculture is the most common secondary occupation. This
is because many urbanites have agricultural property in the
rural areas.
Average household size for the samples is given in table no. 5.
Average household size for the entire sample is 5.7 and the
average size of sampled households in urban and rural areas
is 5.13 and 6.01 respectively.
Table No.5: Average household
size
Dzongkhags
Rural
Urban
Total
Bumthang
6.22
-
6.22
Chukha
4.75
4.72
4.72
Haa
6.58
7.5
6.84
Lhuntshe
5.41
-
5.4
Pemagatshel
5.53
4.83
5.42
S/Jongkhar
7
4.65
5.71
Total
6.01
5.13
5.7
Table no. 6 further highlights sharp rural-urban differences
in the literacy rates. In the rural areas the literacy rate is
30.87%, that is about a third of the urban literacy rate.
Bumthang, Haa, Lhuntshe and Pemagatshel are below
average performers. Lhuntshe fares the poorest in the literacy
front with a literacy rate of just 17.2%. But this finding
cannot be used for generalization as the literacy level is
calculated only for the persons who are in the working age
group.
48
 Journal of Bhutan Studies
Table No. 6: Literacy rate
(In%)
Dzongkhag
Rural
Urban
Total
Bumthang
44
-
44
Chukha
50
89.25
87.63
Haa
33.85
66.67
42.7
Lhuntshe
18
-
17.2
Pemagatshel
31.2
0
42.1
Samdrup Jongkhar
46.43
88.2
69.35
Total
30.87
86.08
50
Part 2: Disparity in the Distribution of Income
Regional Disparity of Income
The most commonly used measure of relative poverty or
inequality is the personal or size distribution of income. It
deals with persons or households and the total income they
receive. This measure of inequality is most conveniently
reflected through the Gini coefficient. Functional distribution
of income is another method of measuring income inequality.
In this work we have analysed disparity in the size
distribution of personal income and its regional variation
through the Gini coefficient.
At the beginning it would be coherent to look at share of each
sample dzongkhag and its rural urban components in the
Total Personal Income.
Share of different regions in the total personal income is given
in the table no. 7. Regional distribution of income reflects a
high degree of disparity between different dzongkhags and
between urban and the rural areas. Chukha accounts for
almost half of the total personal income where as its share in
total samples is just 21.27% The share of Haa's Total
Personal Income is 17.5% whereas its share in total samples
is 19.51%. The remaining four dzongkhags collectively
contribute about 32% of the total personal income where as
they coUectively account for about 60% of the total samples.
49
 Variations in the Inequality and Poverty in Bhutan
Another angle of looking at the regional disparity in the
income shares is urban rural differences. Urban centres
contributed to 69.38% of the total personal income and they
account for almost 35% in the total sample size. On the other
hand, the rural areas which command 65% share in total
samples, contribute as tittle as about 31% of the Total
Personal Income. The urban centres are relatively more
affluent than their rural counterparts.
Table No. 7: Share of different regions in total Personal Income
(Figures in ngultrum thousands)
Dzongkhag
Rural
Urban
Total
Bumthang
3770
-
3770 (7.77%)
Chukha
483
23981
24464(50.45%)
Haa
3683
4810
8493 (17.51)
Lhuntshe
4621
274
4895 (10.09%)
Pemagatshel
1140.5
1015
2155.5 (4.45)
Samdrup Jongkhar
1152
3561
4713 (9.72%)
Total
14849.5 (30.62%)
33641 (69.38%)
48490.5
Graph No. 2: Dzongkhag-wise distribution of total personal
income
Dzongkhag-v\
Per
rise Distribute
sonal Income
8%
>n of Total
10%
/10/_
□ Bumtang
■ Chukkha
■ Haa
□ Lhuntshe
■ Pema Gatshel
□ Samdrup
Jongkhar
1°0/v^''~^^s^^^~.
D%
18%"—~-_U-^^"^    I
—-—       5
50
 Journal of Bhutan Studies
The evidence of the existence of high disparity in economic
growth and consequent disparity in incomes across the
regions can be viewed from the regional variations in the
Average Monthly Per Capita Income (AMPCI) as shown in
table no. 8. These differences are sharp across the
dzongkhags and are sharper within the dzongkhags between
the urban and rural areas. The income disparity across the
dzongkhags is an outcome of differential economic growth
rate and disparate economic opportunities offered by different
locations.
Total combined AMPCI for aU the samples is Nu. 1809.56,
which is almost 50% greater than UNDP estimates at about
AMPCI (Nu. 1200) of Bhutan. AMPCI of Chukha is Nu.4353,
which is more than double the combined AMPCI. Dzongkhags
tike Bumthang, Lhuntshe and Pemagatshel are not only
below average but their AMPCI is about half of the total
combined AMPCI. The performance of Samdrup Jongkhar
and Haa are a little below the average. Chukha's AMPCI is
almost 5 times greater than that of Bumthang and Lhuntshe.
Both the urban and the rural samples from Pemagatshel and
Samdrup Jongkhar have the least AMPCI amongst aU urban
and rural centres from aU the dzongkhags.
The urban rural difference in AMPCI is also very large. The
AMPCI of the urban samples is almost four and half times
greater than the AMPCI of rural samples. F-test was
conducted to verify whether difference in the Monthly Per
Capita Income (MPCI) between rural urban areas is
significant. F-test value is 3.708E-221; differences in the
variability of the urban and rural MPCI is not at aU
significant.
Table No. 8: AMPCI across dzongkhags (In Nu.)
Rural
Urban
Combined
Bumthang
833.83
833.83
Chukha
1677.08
4353.02
4247.23
Haa
847.65
3270.03
1498.38
51
 Variations in the Inequality and Poverty in Bhutan
Lhuntshe
863.87
863.87
Pemagatshel
573.82
2732.91
971.55
Samdrup Jongkhar
544.61
2228.21
1467.88
Total
805.40
3701.36
1809.56
The region-wise physical asset ownership pattern is highly
skewed in the favour of urban areas, which account for
almost 81% of the total physical assets (see table no. 9). This
is probably the main reason for income disparity between
urban and the rural areas. One interesting finding is that the
correlation between assets value and income earned is
dramatically different between urban and rural areas. In
urban areas the r value is +0.9591 and in the rural areas the
r value is +0.5405. This is because a greater part of rural
income is contributed by human labour. Given the low
literacy rates in rural areas and lower share of rural areas in
the physical assets the productivity of labour in rural areas
would definitely be lower. There is also an evidence of
diminishing returns to scale in the use of physical assets in
the urban areas. Almost 81% of the total assets are owned by
urban samples, but the share of urban centres in total
income is 69.38% (see table no.7). On the other hand, 19% of
the total physical assets are owned by the rural samples, but
the share of rural centres in the total income is 30.62%.
Dzongkhag-wise disparity in the physical ownership assets is
equally sharp. Chukha accounts for 57.51% ofthe productive
assets and Pemagatshel accounts for only 1.77% of the total
physical assets. Chukha and Haa together own about 78% of
the total physical assets and the collective share of the
remaining four dzongkhags is just 22% while they together
constitute about 60% share in the total sample size. The asset
ownership disparity also explains the income wise disparity
among the dzongkhags.
52
 Journal of Bhutan Studies
Table No. 9: Physical asset ownership region-wise
(Figures in Nu. 000)
Dzongkhag
Rural
Urban
Total
Bumthang
14704
14704
(3.90%)
Chukha
2055
214757
216812
(57.51%)
Haa
21337
57150
78487
(20.82%)
Lhuntshe
25111
500
25611
(6.79%)
Pemagatshel
4612
2075
6687
(1.77%)
Samdrup Jongkhar
3885
30811
34696
(9.20%)
Total
71704
(19.02%)
305293
(80.98%)
376997
(100%)
Graph No.3: Dzongkhag-wise distribution of productive assets
Dzongkhag-wise Distribution of
Productive Assets
Q0/„                  40/„
~A/——-21%
□ Bumtang
■ Haa
□ Lhuntshe
■ Pema Gatshel
□ Chukkha
□ Samdrup Jongkhar
>/%
Size distribution of personal income
In this study we have used the Gini coefficient to measure the
magnitude of disparity in the size distribution of personal
income. The value of the Gini coefficient is calculated for all
the samples taken together, for each dzongkhag and for their
rural and urban constituents. This analysis will give us a
deeper understanding of the magnitude of personal income
distribution disparity as weU as its rural/urban differences.
53
 Variations in the Inequality and Poverty in Bhutan
Overall situation
As already pointed out earlier, total personal income is
heavtiy biased in the favour of urban areas, it would be
correct to infer that size distribution of personal income
would be highly skewed. It is not wrong to believe so because
generally inequality tends be greater in the urban centres,
given the operation of the 'inverted U curve' hypothesis. In
dualistic economies urban centres experience faster economic
growth; consequently, not only urban/rural divide grows but
also inequality within the urban centres widens.
Table No. 10:  Overall size distribution of personal income
Sample Quintile
Absolute Income
(in Nu.,000)
% Share
Ql
(0-20%)
1569.9
3.24
Q2
(20-40%)
2826.5
5.83
Q3
(40-60%)
4307
8.88
Q4
(60-80%)
6816.2
14.06
Q5
(80-100%)
32970.9
67.99
Total
48490.5
100
The share of sample quintiles in the total personal income is
as reflected in table no. 10. The poorest 20% ofthe samples
receive just 3.24% share in total personal income and the
share of the richest 20% samples receive as much as 68% of
the total personal income. The ratio of the income share of
the richest 20% to the poorest 20% is 20.98. Income disparity
is wider in the urban areas and narrower in the rural areas.
In the rural areas the ratio of the share in total income of the
richest 20% to the share of poorest 20% is 7.09 as compared
to 32.01 in the urban areas.
54
 Journal of Bhutan Studies
Table No. 11: Gini
coefficient
Dzongkhag
Rural
Urban
Total
Bumthang
0.3558
0.3558
Chukha
0.6245
0.6245
Haa
0.3653
0.4319
0.4795
Lhuntshe
0.3964
Pemagatshel
0.1965
0.4853
0.3827
Samdrup Jongkhar
0.2184
0.445
0.433
Total
0.3612
0.5801
0.551
Gini coefficient values are given in the table no. 11. Gini
coefficient measures income inequality in a range of 0-1. If
the Gini coefficient value is 0, it means complete equality,
where aU the persons receive simtiar income. On the contrary
value 1 denotes complete inequality, where only one person
receives aU the income. As the income inequality widens, the
value of the Gini coefficient rises.
The overall value ofthe Gini coefficient is 0.551. Its value for
the urban and rural areas is 0.5801 and 0.3612 respectively.
From this we can infer that income is heavily concentrated in
the hand of a few persons; consequently, inequality is greater
in the urban areas as compared to that in the rural areas.
The highest value of the Gini coefficient (0.6245) is recorded
in urban areas of Chukha Dzongkhag, which implies that size
distribution of income is widest there. As we have already
noted that AMPCI is highest in Chukha Dzongkhag, the
highest degree of inequality there is consistent with the
inverted U curve hypothesis. Urban centres from Haa
Dzongkhag exhibit the most equitable income distribution
from amongst all the urban centres. Urban areas in Haa have
the lowest calculated value of Gini coefficient -(0.4319).
Pemagatshel and Samdrup Jongkhar have recorded the
lowest Gini coefficient value amongst the rural areas at
0.1965 and 0.2184 respectively. It is not sheer coincidence
that the rural areas of these dzongkhags have also recorded
55
 Variations in the Inequality and Poverty in Bhutan
the lowest AMPCI. We can deduce that in the rural areas
there is more equitable distribution of poverty.
The relationship between income measured as AMPCI and
inequality measured through the Gini coefficient is shown in
the table 12 and also in chart no. 4. This chart is drawn for
the combined samples. In chart no. 4 the scattered diagram
with a best fit shows that value of the Gini coefficient
increases as AMPCI increases. The best fit deflects
downwards later, implying that after a threshold level of
AMPCI or PCI is reached the value of the Gini coefficient
declines i.e. - inequality reduces. As Bhutan is in the initial
phase of economic growth, it is natural that inequality in the
personal distribution of income would grow and only in later
phases the growth would be combined with the narrowing of
disparities in the personal income distribution.
Table No. 12
• Relation between level of
income
and inequality
Rural
Urban
Combined
AMPCI
(in Nu.)
Gini
AMPCI
(in Nu.)
Gini
AMPCI
(in Nu.)
Gini
Bumthang
833.83
0.3558
833.83
0.3558
Chukha
1677.08
4353.02
0.6245
4247.23
0.6245
Haa
847.65
0.3653
3270.03
0.4319
1498.38
0.4795
Lhuntshe
863.87
0.3964
863.87
0.3964
Pemagatshel
573.82
0.1965
2732.91
0.4853
971.55
0.3827
Samdrup
Jongkhar
544.61
0.2184
2228.21
0.445
1467.88
0.433
Total
805.4
0.3612
3701.36
0.5801
1809.56
0.551
Graph No. 4: Trends in income and inequality
56
 Journal of Bhutan Studies
Trends in Income and Inequality
2000 4000
AMPCI (in Nu.)
6000
The correlation between combined AMPCI and the Gini
coefficient for different dzongkhags is 0.9649, which is very
high. This implies that rise in AMPCI would be combined
with greater inequality in the distribution of personal income.
Personal income distribution can be explained in terms of size
distribution of productive asset. Chart no. 5 shows that as
the value of assets owned increases the income also
increases. The scattered diagram reflects that the majority of
the points are very near to the best fit; there seems to be high
degree of association between the two variables. The table no.
13 reflects the correlation coefficient values between the value
of the productive assets and the total income earned.
Table No. 13: Correlation between the value of productive
assets and income earned
Rural
Urban
Combined
Correlation coefficient (r)
0.5405
0.9591
0.9554
The correlation coefficient between the value of productive
assets and total income earned is 0.9554 for aU the samples
taken together and the value vary between the urban and the
rural centres. Though there is positive correlation between
the two in the urban and the rural areas, the coefficient value
is much higher in the urban centres. Though higher
correlation does not indicate causality, it is a definite pointer
towards the fact that there is a greater association between
57
 Variations in the Inequality and Poverty in Bhutan
the two.
Graph No. 5: Relation between income and assets
Relation between income and assets
6000
Income
-Poly. (Income)
20000 40000 60000
Value of assets (In Nu.000)
80000
Table no. 14 conveys that there is a high degree of
concentration of productive assets in a few hands. The first
quintile owns as less as 0.005% ofthe total productive assets
and the 5th quinttie owns an overwhelming 86.78% share. In
the urban areas, the concentration of physical assets is much
sharper as the first 40% of the samples do not own any assets
and the top 20% samples own as much as 93.84% of the
assets. In the rural areas asset ownership is more equitable
as the 1st quintile owns 2.2% of the total assets and the 5th
quintile owns 49.83% ofthe assets.
Table No. 14: Size distribution c
f productive assets
Sample quintile
Rural % share
Urban % Share
Combined % Share
Ql
(0-20%)
2.19
0
0.005
Q2
(20-40%)
9.46
0
1.739
Q3
(40-60%)
15.39
1.02
4.221
Q4
(60-80%)
23.14
5.14
7.246
Q5
(80-100%)
49.83
93.84
86.786
Total
100
100
100
Dzongkhag-wise size distribution of productive assets reflects
58
 Journal of Bhutan Studies
high variability between and within the dzongkhags. Highest
disparity is witnessed in the urban areas of Chukha
Dzongkhag where the share of the top 20% of the samples is
96.13%. It means that virtually aU the productive assets are
concentrated in a few hands. The remaining 80% of the
samples own less than 4% of the productive assets. In the
urban areas, the most equitable distribution of the productive
assets is in the urban centres of Pemagatshel Dzongkhag
where the share of top 20% of the samples is 44.58%. In all
urban centres the share of the bottom 40% of the samples in
the total productive assets is very low across the dzongkhags
ranging from 0% in Chukha, Pemagatshel and Samdrup
Jongkhar to 2.45% in Haa.
In the rural areas across the dzongkhags, size distribution of
assets is less skewed than that in the urban areas, but inter-
dzongkhag variation is stiU large. In the rural areas of Haa,
the top 20% of the samples own as much as 50.41% of the
total productive assets and the share of the bottom 40% is
just 10.32% In the rural areas of Bumthang and Lhuntshe
dzongkhags the share of the top 20% samples in the total
productive assets is about 46% and the share of the bottom
40% samples is 13.84% and 12.96% respectively in these two
dzongkhags.
The rural areas of Pemagatshel have the most equitable
distribution of productive assets, foUowed by Samdrup
Jongkhar. In Pemagatshel and Samdrup Jongkhar the share
of the top 20% of the samples is 36.64% and 37.32%
respectively. The share of the bottom 40% of the samples in
the total productive assets is 22.87% in Pemagatshel and
20.21% in Samdrup Jongkhar.
The high concentration of productive assets in few hands in
Chukha perhaps explains why the Gini coefficient value is
high there. This explanation is also probably true in the case
of rural areas where productive asset distribution pattern is
closely related to inequality in the personal income
distribution.    The    higher    the    inequality    in    the    asset
59
 Variations in the Inequality and Poverty in Bhutan
distribution pattern the higher the value of the Gini
coefficient and vice versa.
But there are certain interesting trends in the Gini coefficient
values in the urban areas which cannot be explained in terms
of asset distribution pattern. The value of the Gini coefficient
in the urban areas of Haa is 0.4319 and in Pemagatshel it is
0.4853 that means inequality in the size distribution of
personal income is higher in Pemagatshel. But the productive
assets are more equitably distributed in Pemagatshel as the
top 20% of urban samples own only 44.58% of the assets
than they are in Haa, where the top 20% urban samples own
88.26% of the assets. Why does the more unequal
distribution of productive assets result in more equitable
distribution of income? This question is left to be answered by
future researchers.
Another important determinant of the size distribution of
personal income and regional income disparity is level of
education that affects the quality of the labour force and
makes it more productive. As far as urban rural differences in
the level of income are concerned, educational attainment is
considered to be a significant factor. We wiU consider whether
this theoretical postulate is relevant. The average literacy rate
in the urban samples is 76.18% and for the rural samples it
is 30.87% But despite higher literacy rates the disparity in
the size distribution of income is higher in the urban centres.
On the other hand, urban areas have both higher literacy
rates and a higher level of AMPCI. This implies that higher
educational attainment enables an individual to be more
productive and earn higher income. The value of the
correlation coefficient between education level and personal
income is low but positive i.e.: 0.1577. Interestingly, the value
of the correlation coefficient is lower in the urban areas
(+0.1212) than in the rural areas (+0.2651). The value of
productive physical assets owned has more significant
association with income earning capability than the level of
education in the urban areas. The higher value of the
correlation coefficient between assets value and the size of
60
 Journal of Bhutan Studies
personal income (+0.9561) implies that education attainment
plays a less significant role. The implications are similar for
the rural areas, where the correlation coefficient between
value of physical productive assets and the size of personal
income is (+0.5405), higher than between education level and
size of personal income (+0.2651). Education can play a more
important role in eliminating absolute poverty but not the
same role in removing income inequality.
FinaUy it can be stated that the promotion of education along
with more equal redistribution of productive physical assets
can narrow the inequalities of personal income distribution
and regional income distribution.
It would also be pertinent to explore whether regional
disparity is policy induced. Inappropriate government
policies, inequitable allocation of the public expenditure and
other financial resources, and inadequate development of
infrastructure are some important factors that create policy
bias against certain regions.
In this study we have used bank loans as a proxy variable for
government policy. We coUected secondary data pertaining to
loan advanced by Bank of Bhutan during the financial year
2004. Of the total loans advanced by the Bank of Bhutan in
these six dzongkhags, Chukha received a predominant share
of 94.48%. On the other hand Lhuntshe and Pemagatshel
received even less than 0.24% and less than 0.41%
respectively. This indicates that the rate of private investment
must be significantly lower in the relatively backward
dzongkhags, which results in regional disparity in the level of
per capita income.
Table No. 15: Loan advanced by BOB in year 2004
Dzongkhag
Amount (In Nu. million)
%
Bumthang
31.39
2.30
Chukha
1291.33
94.48
Haa
21.25
1.55
61
 Variations in the Inequality and Poverty in Bhutan
Lhuntshe
3.30
0.24
Pemagatshel
5.65
0.41
Samdrup Jongkhar
13.85
1.01
Total
1366.77
100
Source: Bank of Bhutan
Part 3:     Magnitude and the Extent of Absolute Poverty
Poverty Criterion
Absolute poverty is defined as the inability of a person to
command necessary resources to meet basic minimum needs.
This would require setting up a minimum income criterion
that enables a person to satisfy the basic minimum needs.
As we have mentioned earlier in the methodology section, our
definition of absolute poverty is based on poverty estimates of
Bhutan by UNDP and the HIES (2000) estimates of AMPCI.
In this section we wiU explore the micro level magnitude of
poverty and its regional variation. Magnitude of poverty is
calculated through two indices:
1) Head Count Ratio (HCR): measures the fraction of total
population which faUs below the poverty line. We have
calculated two HCR based on our definition of lower poverty
tine (LPL) and the upper poverty tine (UPL), which were
already defined earlier.
2) Normalized Poverty Gap (NPG): based on Foster Greer
Thorbecke (1984) (FGT) index. Poverty gap is a better
measure of the magnitude of absolute poverty than HCR.
HCR only measures the fraction of total population that fall
below the poverty line irrespective of the shortfall of the
income from the poverty tine, and all are given equal weight.
Suppose the poverty tine is Nu.1200, there are some persons
who earn Nu. 1100 and there might be others who earn only
Nu. 400, but these differences are not taken into account in
HCR. Poverty gap measures the amount of income transfer
62
 Journal of Bhutan Studies
necessary to bring the poor people above the poverty tine i.e.:
enable them to acquire the income that defines the poverty
tine.
A normalized poverty gap provides information regarding how
far the households are from the poverty tine. This measure
captures the mean aggregate income shortfall relative to the
poverty tine across the whole population. It measures the
depth and severity of the poverty. The measures of depth and
severity of poverty are complementary to the incidence of
poverty. This concept is also particularly important for the
evaluation of the programmes and policies aimed at
aUeviating poverty.
In this study we wiU also explore the urban/rural and inter-
dzongkhag differences in the magnitude of absolute poverty.
Poverty Analysis Based on Upper Poverty Line
As mentioned earlier in this study our measure of upper
poverty line (UPL) is AMPCI of Nu.1200. This criterion is used
to estimate the extent and the magnitude of the shortfall of
individual's monthly per capita income from the AMPCI. This
criterion is also close to the criterion of international poverty
tine i.e.: $1 per person per day which comes out to be less
than Nu. 1500 per person per month. The findings about head
count ratio are given in the table no. 16.
Table No. 16: Head Count Ratio based on UPL
Dzongkhag
No. of Poor
HCR (In %)
Bumthang
55
78.57
Chukha
34
35.05
Haa
62
69.66
Lhutnshe
79
79
Pemagatshel
32
84.21
Samdrup Jongkhar
40
64.52
Total
302
66.23
63
 Variations in the Inequality and Poverty in Bhutan
302 samples out of a total of 456 samples have monthly per
capita income less than Nu.1200 that means the overall head
count ratio based on the UPL criterion is 66.23%
Pemagatshel Dzongkhag has the highest poverty ratio as its
HCR base is 84.21%. Headcount ration in Bumthang and
Lhuntshe dzongkhags is 79% and 78.57% respectively. 4 out
of 6 dzongkhags have higher HCR than average HCR. In Haa
Dzongkhag HCR based on UPL is 64.52% Chukha
Dzongkhag experienced the lowest poverty rate at 35.05%
Graph No. 6: Overall Head Count Ratio
90
80
70 -I
?        60 \
£        50
K        40
2        30
20 -j
10 -
o
HCR based on UPL criterion
84.21	
78757^
-rs-
09.00
64.52
' / *s/.*
o-
Dzongkhags
vT
0>v
Graph No. 7: Rural-urban dispersal of the poor
Rural-Urban Dispersal ofthe Poor
19%
■ Rural
■ Urban
81%
Graph No. 8: Regional dispersion of the poor
Regional Dispersion of Poor
(Based on UPL)
■ Bum thang
13%
18%
S 1 1 %
■ Chukha
n Haa
1 1 % ^---~'^====:=__
n Lhutnshe
26%  """■                       '
21%
□ Pern agatshel
□ S am drup
Jonakhar
 Journal of Bhutan Studies
Estimates of rural poverty based on UPL
We have already analyzed that the AMPCI in the rural areas is
lower than that in the urban areas. The AMPCI in rural areas
is Nu.805.4 and in urban areas it is Nu.3701.36. From the
magnitude of the difference between the urban and rural
AMPCI it can be inferred that poverty must be much more
concentrated in the rural areas. Analysis of the results in the
table no. 17 authenticates the inference.
Table No. 17: HCR in the rural areas (based on UPL)
Dzongkhag
No. of Poors
HCR (in %)
Bumthang
55
78.57
Chukha
3
75
Haa
51
78.46
Lhuntshe
79
79.8
Pemagatshel
31
96.87
Samdrup Jongkhar
26
92.85
Total
245
82.21
In this study we found that the total number of rural poor is
245 that mean about 81.13% of total poor persons live in the
rural areas. Obviously poverty is a rural phenomenon.
HCR for the rural samples is 82.21% which implies that
82.21% of the rural samples live below poverty line based on
the upper poverty line criterion. The geographical distribution
of rural poverty is also skewed.
65
 Variations in the Inequality and Poverty in Bhutan
Rural areas in Pemagatshel and Samdrup Jongkhar
dzongkhags have very high concentrations of poverty. In the
rural areas of Pemagatshel 96.87% of the samples live below
a monthly income of Nu.1200 and 92.85% of the rural
samples in Samdrup Jongkhar Dzongkhag are poor if defined
on the basis of the upper poverty line. This is not surprising
given that AMPCI in the rural areas of these dzongkhags is
below Nu. 575. In the remaining dzongkhags, rural poverty
ratios are below average and between 78-80% of the samples
are poor.
Estimates of urban poverty based on UPL
Since AMPCI in the urban areas is almost four and half times
higher than that in the rural areas, the urban poverty ratio
must be lower than the rural poverty ratio. The magnitude of
urban poverty is analyzed on the basis of information given in
the table no. 18.
Table No. 18: HCR in urban areas (based on UPL)
Dzongkhag
No. of Poors
HCR (in%)
Bumthang
-
-
Chukha
31
33.33
Haa
11
45.83
Lhuntshe
-
-
Pemagatshel
1
16.67
Samdrup Jongkhar
14
41.18
Total
57
36.07
The total number of urban poor is 57, which mean the urban
poor accounts for only 18.87% ofthe total number of poor. In
the urban areas poverty is less concentrated than in the rural
areas. The HCR in urban areas is only 36.07% compared to
82.21% in the rural areas. In Pemagatshel Dzongkhag the
urban poverty ratio is just 16.67% Due to the small sample
size there might be greater sample error and no inference
66
 Journal of Bhutan Studies
should be drawn from this. From the remaining samples, the
urban areas of Chukha Dzongkhag have the lowest poverty
rate, where a third of the samples faU below the poverty line.
In the urban areas of Haa Dzongkhag the poverty ratio is the
highest as 45.83% of the samples are poor based on the
upper poverty line criterion. In Samdrup Jongkhar
Dzongkhag the urban poverty ratio is 41.18%
Based on the upper poverty line criterion it can be concluded
that the average poverty rate is very high as almost 2/3 ofthe
samples live below the poverty line. The rural/urban
differences in the head count ratios are very sharp and
poverty is more concentrated in the rural areas as 81.13% of
the total poor live in the rural areas. Urban poverty is also
significant as more than a third of the urban samples live
below the poverty tine.
Graph No. 9: HCR in the rural and urban areas
HCR in the rural and urban areas
120 n
100
80
60 4
o     40
20
04
■
m i m m i n iii ii
^ ^ j? .*?
3> sr&
«*v
'N ^    .^
0<>
c&~   xsr Dzongkhag
□ Rural
■ Urban
Normalized Poverty Gap based on UPL
On average, the shortfall of the poor's income (which is the
67
 Variations in the Inequality and Poverty in Bhutan
measure of normalized poverty gap) from the upper poverty
tine is 36.42%. The poverty gap in rural areas is more than
two and a half times larger than the poverty gap in the urban
areas. The variance in the poverty gap between the rural and
urban area is not significant as suggested by the F-test value
of 0.398. Standard deviation for the combined value of the
poverty gap for different dzongkhags is 11.05% If we take
±3SD from the average poverty gap, income shortfaU range for
the 89% of the poor people is from 3.27% to 69.57%. In the
rural areas where almost 80% of the poor reside, the average
poverty gap is 47% with a standard deviation of 4.58%. If we
take ±3SD from the average poverty gap in the rural areas,
the poverty gap range for the 89% of the rural poor would be
33.25% to 60.74%.
Table No. 19: Dzongkhag-wise normalized Poverty Gap based
on UPL (in %)
Dzongkhag
Rural
Urban
Combined
Bumthang
43.46
43.46
Chukha
49.82
15.14
16.20
Haa
45.15
18.36
38.14
Lhuntshe
46.70
46.70
Pemagatshel
50.65
5.42
43.52
Samdrup Jongkhar
56.17
21.27
37.03
Total
47.00
16.49
36.42
The largest poverty gap (56.17%) in the rural areas exists in
Samdrup Jongkhar Dzongkhag and the lowest poverty gap in
the rural areas exists in Bumthang Dzongkhag. Also, there is
much less variation in the poverty gap in the rural areas of
different dzongkhags as compared to that in the urban areas.
Graph No. 10: Poverty gap
68
 Journal of Bhutan Studies
50
40
o. 30
ro
O
£ 20
v
o   10
Q.
0
Poverty Gap: Area of Residence-wise
AL	
36.42
16.49
Rural Urban combined
Area of Residence
DSeriesI
Graph No. 11: Normalized poverty gap
Normalised Poverty Gap
60 -p
50 —
40 --T
30 -
20 -
10 -
0 V
■ Rural
■ Urban
-I    □ Combined
jr
^    J>-   jJ'
«T  J^   ^
cf
Dzongkhags
Poverty Analysis Based on Lower Poverty Line
Our lower poverty tine estimates are based on the HIES 2000
criterion of an upper poverty tine fixed at Nu.748.10 per
capita per month. We have rounded off to Nu.750 per capita
per month. Application of this criterion would give us more
realistic estimates of the absolute poverty as compared to the
upper poverty line based estimates. 231 samples were
identified as poor because their monthly per capita income
feU below Nu.750. As reflected in the Table no.  20, overaU
69
 Variations in the Inequality and Poverty in Bhutan
head count ratio is 50.66% i.e. almost half of the total
samples live below the poverty line. The highest incidence of
poverty occurs in Lhuntshe Dzongkhag where the head count
ratio is 66% In Bumthang and Pemagatshel dzongkhags the
head count ratio is more than 60%. The lowest incidence of
poverty is found in Chukha Dzongkhag.
Our study shows that the incidence of poverty measured as
head count ratio is higher than the HIES 2000 estimates of
31.75%.
Table No. 20: Head Count Ratio based on LPL
Dzongkhag
No. of Poor
HCR (In %)
Bumthang
43
61.42
Chukha
19
19.59
Haa
48
53.93
Lhuntshe
66
66
Pemagatshel
23
60.52
Samdrup Jongkhar
32
51.61
Total
231
50.66
Graph No. 12: HCR based on LPL criterion
HCR based on LPL criterion
<f cf
cf
Dzongkhag
Dzongkhag-wise dispersal of poverty shows that the largest
70
 Journal of Bhutan Studies
number of poor persons live in Lhuntshe, which constitutes
about 28% of the total number of poor. 8% of the total poor
live in Chukha Dzongkhag. Shifting of the poverty tine
criterion downwards (from UPL to LPL) has changed the
pattern of regional dispersion of poverty, though very
margrnaUy. With this shift, the share of Lhuntshe Dzongkhag
in total poor population increased from 26% to 28% and the
share of Pemagatshel declined by 1 percentage point. The
share of Samdrup Jongkhar and Bumthang in the total
number of poor increased by 1 percentage point, while the
share of Chukha Dzongkhag declined by 2 percentage points.
The share of Haa in the total number of poor remained
constant.
Graph No. 13: Regional dispersion of the poor based on LPL
Regional Dispersion of Poor
(Based On LPL)
14%
19%
10%
28%
21%
■ Burrtang
■ Chukkha
□ Haa
■ Lhutnshe
■ Ftemagatshel
■ Samdrupjongkhar
Estimates of rural poverty based on LPL
Shifting the poverty tine from UPL to LPL criterion has
resulted in an increase in the ratio of rural poor to total poor.
Basing the poverty estimates on the upper poverty line we
found that 81.13% of poor reside in the rural areas. On the
basis of LPL estimate of poverty we found that percentage
share of rural poor increased to 86% and correspondingly the
share of urban poor declined. This is due to a higher poverty
71
 Variations in the Inequality and Poverty in Bhutan
gap in the rural areas as compared to that in the urban
areas. In the poverty gap estimates based on UPL, normalized
poverty gap in the rural and urban areas is 47% and 16.5%
respectively. This further strengthens our earlier conclusion
that poverty is more a rural phenomenon. Since almost 70%
of the rural samples are farmers it can be inferred that
poverty is more intensely spread among the farmers. As a
majority of the farmers in the sample study practice
traditional subsistence farming and in certain areas, like in
the rural areas of Haa, farming is a seasonal occupation, it is
quite probable that poverty is more concentrated in these
groups.
Graph No. 14: Rural-urban dispersion of the poor based on LPL
Rural-Urban Dispersion of Poor
14%  (Based on LPL)
■ Rural
■ Urban
86%
The head count ratio for the rural areas on the basis of LPL is
66.77%, 199 of the 298 rural samples have monthly per
capita income less than Nu.750. Table No. 21 provides
information about the head count ratios in rural areas of
different dzongkhags. The highest head count ratio (82.14%)
is found in the rural areas of Samdrup Jongkhar Dzongkhang
followed by the rural areas of Pemagatshel Dzongkhag where
the head count ratio is 71.87% Chukha Dzongkhag has the
lowest head count ratio amongst the rural areas.
Table No. 21: HCR in the rural areas (based on LPL)
72
 Journal of Bhutan Studies
Dzongkhag
No. of Poor
HCR (in %)
Bumthang
43
61.42
Chukha
2
50
Haa
42
64.61
Lhuntshe
66
66
Pemagatshel
23
71.87
Samdrup Jongkhar
23
82.14
Total
199
66.77
Estimates of urban poverty based on lower poverty line criterion
The head count ratio in the urban areas is much lower than
the head count ratio in the rural areas. As already noted
almost 2/3 of the rural samples are poor, where as only
about 1/5 of the urban population lives below the poverty
tine. Table no. 22 provides dzongkhag-wise head count ratios
in the urban areas. More than 50% of the urban poor are
concentrated in the Chukha Dzongkhag. The very small
sample size for the urban areas of Pemagatshel Dzongkhag is
the main reason for the unusual result of the absence of
poverty there. Despite the huge difference between rural and
urban HCR, the variance in the two HCR is not significant as
the F-test value of 0.7846 suggests.
Table No. 22: HCR in the urban areas (based on LPL)
Dzongkhag
No. of Poor
HCR (in %)
Bumthang
-
-
Chukha
17
18.28
Haa
6
25
Lhuntshe
-
-
Pemagatshel
0
0
Samdrup Jongkhar
9
26.47
Total
32
20.25
Graph No. 15: HCR in the rural and urban areas based on LPL
73
 Variations in the Inequality and Poverty in Bhutan
HCR in the Rural and Urban Areas (Based On LPL)
100 i
OS?
f   #     ^   jr    jf    *
v^
■ Rural
~    /   / ^rban
x      ^ Dzongkhags
The highest incidence of urban poverty is witnessed in
Samdrup Jongkhar Dzongkhag where the headcount ratio is
26.47%. The lowest incidence of urban poverty is in Chukha
Dzongkhag where the head count ratio is 18.28%. The size of
urban samples from Pemagatshel Dzongkhag is too small to
be of any analytical importance.
Normalized Poverty Gap Based on LPL
The average shortfall of the poor's income (normalized poverty
gap) from the lower poverty tine (defined as per capita
monthly income of Nu.750) is 22.23%. The poverty gap for all
the rural areas is 29.89% which is 3.35 times larger than the
poverty gap of 8.9% in the urban areas. The variance in the
poverty gap between the rural and urban area is not
significant as suggested by the F-test value 0.687.
Highest poverty gap (30.6%) is found in Lhuntshe Dzongkhag
and the lowest poverty gap is found in the Chukha
Dzongkhag where the poverty gap is 9.44% In the rural
areas, largest poverty gap (36.02%) is in Samdrup Jongkhar
Dzongkhag. In the rural areas of Chukha and Lhuntshe
poverty gap is over 30%. In the remaining dzongkhags the
range of poverty gap in the rural areas is 27-30%. The urban
areas have a lower poverty gap. Poverty gaps in the urban
areas range from a low of 8.77% in Chukha Dzongkhag to a
high   of   12.04%   in   Samdrup   Jongkhar   Dzongkhag.   The
74
 Journal of Bhutan Studies
poverty gap for the rural and urban areas of the different
dzongkhags are given in the table no.23.
Table No. 23: Dzongkhag-wise Normalized Poverty Gap based
on LPL criterion (in %)
Dzongkhag
Rural
Urban
Combined
Bumthang
28.46
-
28.46
Chukha
34.72
8.77
9.44
Haa
27.50
7.47
22.43
Lhuntshe
30.60
-
30.60
Pemagatshel
29.36
-
25.20
Samdrup Jongkhar
36.02
12.04
22.88
Total
29.89
8.90
22.61
Impact ofthe shift in the poverty line
The shift from the upper poverty line to the lower poverty tine
has resulted in a faU in the head count ratio and the level of
poverty gap. Table no. 24 reflects the impact of the lowering of
the poverty tine on the head count ratio and on the
normalized poverty gap. The shift in the poverty tine brought
about a greater fall in HCR (- 44.11%) in Chukha Dzongkhag
Graph No. 16: Normalized Poverty Gap (LPL)
Normalised Poverty Gap (LPL)
c
Q.
I
40
35
30
25
20 -
15 -
10 -
5 -
0
IT
■ Rural
■ Urban
□ Combined
/>
&
<?
J
cf
Dzongkhgs
and the least fall in head count ratio (-16.46%) in Lhuntshe.
75
 Variations in the Inequality and Poverty in Bhutan
Pemagatshel Dzongkhag experienced the largest decline in the
poverty gap equivalent of 42.10% and the least was
experienced in Lhuntshe Dzongkhag (-34.48%). For aU the
samples taken together, the decline in head count ratio and
normalized poverty gap was 23.51% and 37.92% respectively.
On average, for a percentage decline in HCR, the poverty gap
feU by 1.61%. The higher the ratio the greater the intensity of
poverty aUeviation measures needed to reduce the poverty
gap-
Table No. 24: Effect ofthe shift in the poverty line from UPL to
LPL.
Dzongkhags
% change
in HCR
% change
inlMPG
A1NPG/AHCR
Bumthang
-21.83
-34.51
1.58
Chukha
-44.11
-41.73
0.95
Haa
-22.58
-41.19
1.82
Lhuntshe
-16.46
-34.48
2.10
Pemagatshel
-28.13
-42.10
1.50
Samdrup Jongkhar
-20.01
-38.21
1.91
Total
-23.51
-37.92
1.61
One important policy implication that emerges from this
analysis is that poverty alleviation measures should be
concentrated in those areas where the ratio of ANPG/AHCR is
higher. Such a data base would make the poverty alleviation
measures targeted and consequently more effective in terms
of reducing the magnitude of absolute poverty.
Poverty decomposition
Factors determining the income generating capacity of an
individual will also determine absolute deprivation. For
understanding the dynamics of absolute poverty these
variables should be identified. Identification of these factors is
termed as decomposition analysis. Recently Deaton and Dreze
(2002)     and     Shorroks     and     Wan     (2004)     have     used
76
 Journal of Bhutan Studies
decomposition analysis. Decomposition can be done in
different ways such as by population sub-groups or by factor
components. It can be used to analyze the impact of gender,
occupation, area of residence etc. on the head count ratio. In
this study we have used simple decomposition analysis and
not the regression-based decomposition analysis.
As far as the area of residence-based decomposition of
absolute poverty is concerned, we have already established
that the incidence of poverty is much higher in rural areas.
The urban areas have greater income disparity but lower
incidence of absolute poverty.
Occupation based decomposition of poverty
In this part, occupation-wise decomposition of poverty
analysis is done to see which category of the occupation has a
larger incidence of poverty. This study will help us to identify
the groups which are more vulnerable to the problem of
absolute poverty. Results of these findings are shown in the
table no. 25.
Table No. 25 Occupation based decomposition of HCR (in %)
Occupation
LPL estimates
UPL estimates
Farmers
78.5
98.3
Businessmen
25.3
37.4
Govt. /Semi Govt,
employees
8.6
31
Pvt. Employees
(formal sector)
41.9
65.1
Informal activities
39.1
56.5
The largest incidence of poverty is found amongst the farmers
as 78.5% ofthe farmers are poor, based on lower poverty tine
estimates. Hired employees in the private sector experience
second largest incidence of poverty. The government and
semi-government sector employees are the least poor group.
Shifting from the lower to the upper poverty tine does not
alter the results. This analysis highlights the fact that poverty
is more concentrated in the rural areas and farmers are the
77
 Variations in the Inequality and Poverty in Bhutan
most vulnerable section. Private sector employees are the next
most vulnerable group.
Literacy status based decomposition of poverty
Literacy status of a person has an important bearing on his
or her functioning. Table no. 26 provides information about
the head count ratio for literate and iUiterate samples. The
head count ratio for the Utiterate group is almost two and a
half times greater than that for the literate group if we base
our estimates on the lower poverty tine. On the basis of upper
poverty line also, the iUiterates have a much higher incidence
of poverty. The illiterates are more vulnerable to absolute
poverty.
Table No. 26: Literacy based decomposition of HCR (in %)
Literacy status
LPL estimates
UPL estimates
Illiterate
72.93
88.21
Literate
28.19
44.05
Gender based decomposition of poverty
Evidence in this study suggests that poverty incidence is
partly affected by the sex of the individual (see table no. 27).
Females have a higher incidence of poverty than male
members. The gender-based differences in poverty incidence
can be explained in terms of lack of equal opportunities and
capabilities. The shift from lower poverty line to upper
poverty line does not alter the picture. A mark of caution is
that these differences do not necessarily originate from
gender biases and there could be other factors explaining the
differences. This is left to the future researchers to identify
the explanatory variables.
Table No. 27: Gender based decomposition of HCR (in %)
Gender
LPL estimates
UPL estimates
Male
46.95
63.67
Female
58.62
71.72
Area of residence based decomposition of poverty
78
 Journal of Bhutan Studies
Urban rural differences in the incidence of poverty have
already been mentioned in the part 3 of the section 2 of this
paper. In rural areas the head count ratio is 82.2% and
66.77% respectively on the basis of UPL and LPL criterion.
The urban areas have much a lower poverty incidence of 36%
and 20.2% respectively on the basis of UPL and LPL criterion.
Conclusions
Findings of this sample  study can also be used to  make
analysis about population characteristics.
Important conclusions of the study are:
1) AMPCI is calculated to be Nu. 1809.5, which is almost 50%
higher than the UNDP estimated AMPCI of Nu.1200. On
average, Bhutanese live on more than a $1 a day which is
known as international poverty tine.
2) There is high degree of inequality in Bhutan both in terms
of regional and personal distribution of income. Rural/urban
disparities are very high. AMPCI in urban areas is almost four
and half times more than the AMPCI in the rural areas.
Urban areas contribute almost 69% of the total income and
the rural areas having a share of 65% in total samples
contribute just 31% of the total income.
3) Disparities are wider across the dzongkhags. Both the rural
and urban AMPCI are highest in Chukha Dzongkhag. In the
urban areas AMPCI ranges from Nu.4353 in Chukha
Dzongkhag to Nu.2228 in Samdrup Jongkhar Dzongkhag. In
the rural areas AMPCI ranges from Nu.1677 in Chukha
Dzongkhag to Nu. 544 in Samdrup Jongkhar Dzongkhag.
4) The value of the productive assets is found to have very
high association with the level of AMPCI. Level of education
tends to have less significant association with the AMPCI.
Bank loan disbursements across the dzongkhags is heavily
biased in the favour of Chukha Dzongkhag, which is probably
one important reason of asset holding and consequent
personal income disparities across the dzongkhags.  Income
79
 Variations in the Inequality and Poverty in Bhutan
disparities thus can be explained in terms of disparities in the
productive assets ownership pattern. Urban samples own
almost 81% of the value total physical assets. Dzongkhag-
wise disparities in the asset holdings are very high. Percentile
share of different dzongkhags in the asset holding ranges
from 57.6% in Chukha to 1.7% in Pemagatshel.
5) Size distribution of income within the dzongkhags and
within the urban rural areas, measured in terms of the Gini
coefficient, shows that income inequalities are higher in
urban areas and lower in rural areas. The Gini coefficient
value tends to rise across the dzongkhags as their AMPCI
rises. There is a strong evidence of initial trade-off between
income and equity.
6) Our estimates of absolute poverty on the basis of upper
poverty line (Nu.1200 per capita per month) convey that
66.2% of the samples live below the poverty line. On the basis
of lower poverty tine (Nu.750 per capita per month) 50.67% of
the samples are found to be poor. The rural areas have a
higher incidence of poverty; a large number of poor live in the
rural areas and also a majority of the poor are farmers. It can
be stated that there is a high concentration of poverty in the
agricultural sector which is characterized by traditional
subsistence, and in certain regions it is a seasonal practice
and consequently less productive. Hired employees of the
private sector are the second most vulnerable section. Inter-
dzongkhag variation in the head count ratio is very high as it
varies from 19.6% in Chukha to 66% in Lhuntshe Dzongkhag
(on the basis of the lower poverty line) and from 35% in
Chukha to 79% in Lhuntshe Dzongkhag (based on upper
poverty tine).
7) Not only is the head count ratio high but the poverty gap is
also high. The average shortfall of the income of the poor is
37.03% if calculated on the basis of the upper poverty line
and it is 22.61% if calculated from the lower poverty tine. The
poverty gap calculated from either criterion suggests that
there   are   high   urban/rural   differences   in   the   depth   of
80
 Journal of Bhutan Studies
poverty. In the rural areas, the poverty gap is almost three
times larger than the poverty gap in urban areas. The poverty
gap is the largest in Samdrup Jongkhar Dzongkhag both in
the case of urban and rural areas. Chukha Dzongkhag has
the least poverty gap.
8) Decomposition analysis suggests that incidence of absolute
poverty is likely to be much greater amongst farmers, hired
employees of the private sector and illiterates. These are the
most vulnerable sections of the population and significant
factors to the incidence of absolute poverty. Male citizens are
less prone to poverty as compared to female citizens. The
rural people are much more vulnerable to poverty than their
urban counterparts.
9) Targeted measures of poverty aUeviation wiU have a strong
influence on the reduction in the poverty gap. In the areas
where incidence of poverty is high, poverty aUeviation
measures would have better effect on reducing the poverty
gap. Since poverty is primarily a rural phenomenon, a rural-
centric development strategy would have a more positive
effect on the magnitude of the poverty.
References
CSO, RGOB (2000). National Account Statistics 1980-2000,
Thimphu.
   (2001).   "Household  Income  and  Expenditure   Survey
2000 (Pilot): Report on Income and Expenditure, Poverty
Measurement, and Socioeconomic Profile of the Households,"
Thimphu.
Deaton, A. and Dreze, J. (2002). "Poverty and Inequality in
India: A Re Examination," Economic and Political Weekly,
September 7, 2002.
Foster, J., Greer, J. & Thorbecke, E. (1984). "A Class of
decomposable Poverty Measures," Econometrica, 52(3), 761-
766.
81
 Variations in the Inequality and Poverty in Bhutan
Kuensel,   "National   Poverty   Line   established   at   Nu.740   a
month," October 25, 2004.
Kuznets,    Simon   (1955).    "Economic   Growth   and   Income
Inequality," American Economic Review, Vol 49, March 1955.
Sen,   A.K.   (1984).   "Poor,   Relatively   Speaking,"   Resources,
Values and Development, Oxford University Press.
  (1993). "Capability and Weil-Being," in Nussbaum, M.
C,    Sen,    A.    (Eds),    The   Quality   Of  Life,    Oxford   India
Paperbacks.
Shorrocks, A.F. and Wan, G. (2004). "Spatial Decomposition
of Inequality," Journal of Economic geography.
UNICEF (2003). State of Word's Children.
82
 83

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/cdm.dhimjournal.1-0365199/manifest

Comment

Related Items