ROYAL GOVERNMENT OF BHUTAN BHUTAN POVERTY ANALYSIS REPORT 2017 © 2017 National Statistics Bureau of Bhutan All rights reserved. Published in 2017. Printed in Bhutan. ISBN 978-99936-28-48-4 Layout and design: Loday Natshog Communications, Thimphu Photos contributed by Tshering Penjor and BS Thapa National Statistics Bureau Royal Government of Bhutan PO Box No 338 Thimphu, Bhutan Tel: +975 2 333296, +975 2 335848 Fax: +975 2 323069 www.nsb.gov.bt CONTENTS Acknowledgements iv Foreword v Executive Summary vii Chapter 1. Demographic Characteristics 1 1.1. Background 1 1.2. Objectives 1 1.3.  Data Source 2 Chapter 2. Updating the Poverty Line 5 2.1.  Updated Food Poverty Line 5 2.2.  Updated Non-food Allowance and Total Poverty Line 6 2.3.  Spatial Price Index 6 Chapter 3. Patterns in Consumption Poverty 9 3.1.  Poverty Rate 9 3.2.  Depth and Severity of Poverty 13 3.3.  Poverty trend 14 3.4.  Poverty by Household Characteristics 15 Chapter 4. Basic Needs 21 4.1. Education 21 4.2. Health 22 4.3.  Household Amenities, Assets, and Access to Services 24 4.4.  Perception and Priorities 25 Chapter 5. Inequality 29 5.1.  Consumption Quintiles 29 5.2.  Gini Index 30 Chapter 6. Conclusion 33 Annex I: Additional Statistical Tables 35 Annex II: Technical Notes 45 iii ACKNOWLEDGEMENTS The National Statistics Bureau (NSB), Investigator), and from the World Bank, Royal Government of Bhutan, and the Dr. Hiroki Uematsu (Senior Economist), World Bank jointly prepared this Poverty formed the core team. Dr. Hiroki conducted Analysis Report. From the NSB, Mr. the validation exercises and provided tech- Cheku Dorji (Chief Statistical Officer), Mr. nical support throughout the preparation Phub Sangay (Chief Statistical Officer), of the Report. The Director of the NSB, Mr. Dorji Lethro (Sr. Statistical Officer), Mr. Chhime Tshering, provided advice and Mr. Cheda Jamtsho (Research Officer), guidance that helped improve the Report. and Mr. Rinchen Tshering (Sr. Statistical iv FOREWORD The National Statistics Bureau (NSB) is It is our earnest hope that the Report pleased to present Poverty Analysis Report will serve to be a useful input for the 12th (PAR) 2017. It is based on data collected Five Year Plan, and at the same time, con- during the latest Bhutan Living Standards tribute meaningfully to the formulation of Survey (BLSS), which was carried out policies and programmes aimed at impro- between March and April in 2017, by the ving the living standards of the poor and NSB with support from the World Bank. the underprivileged. Furthermore, we are A significant amount of poverty related hopeful that the Report will be a useful statistics for Bhutan have been available reference for development partners, civil since the 2004 PAR, which was based on society organizations, academia, and rese- BLSS 2003. In 2007, the NSB updated archers in Bhutan and beyond. the baseline for poverty to establish a more Finally, the NSB would like to thank accurate and reliable baseline. The 2007 the World Bank for its financial and techni- and 2012 PARs contain poverty estimates cal support in bringing out PAR 2017. The at the Dzongkhag level, while the current contributions, commitment, and support Report contains poverty estimates at the that we received from Dr. Hiroki Uematsu Dzongkhag level as well as for the four exist- of the World Bank are highly commenda- ing Thromdes. ble and appreciated. The main objective of this Report is to update the poverty estimates, paying close attention to the methods used to ensure consistency and comparability of poverty estimates overtime. We are pleased to inform that the poverty rate has decrea- sed from 12% in 2012 to 8.2% in 2017. Chhime Tshering It is encouraging to note that Bhutan Director was able to reduce poverty significantly in National Statistics Bureau the last five years. This could be attribu- ted to the effectiveness of the Royal Kidu Programme targeting the needy and the impoverished population, works by various NGOs, and the successful execution of the 11th Five Year Plan programmes. v EXECUTIVE SUMMARY The Poverty Analysis Report (PAR) 2017 Household Characteristics is being prepared with the objective to In both urban and rural areas, a poor provide a focused picture of poverty at the household has a much larger family size National, Dzongkhags, and Thromde levels, than a non-poor household. On average, based on the Bhutan Living Standards female-headed households are observed to Survey (BLSS) 2017. be less poor than male-headed households. Persons living in households where the Poverty Rate head is currently working have higher living The PAR 2017 makes use of a poverty standards than those living in a households line, estimated for 2017 at Nu2,195.95 whose head is either unemployed or not per person per month. The poverty line, in the labour force. Among the employed, representing the level of consumption poverty levels are higher in households needed to secure the necessities of life, is whose head works in agriculture. obtained by adding estimated food and Around 68% of the household heads non-food requirements of Nu1,473.45 and in Bhutan are aged between 25 and 54 years, Nu722.50, respectively. Using this poverty while less than 3% are below 25 years, and line, an estimated 8.2% of the population about 11% are 65 years and above. The is found to be poor. Thus, poverty has poverty rate is about 2% for those house- declined by about a third from the estimate hold members whose household head is of 12% in 2012. below 25 years of age as compared to Poverty in rural areas (11.9%) is sig- 10% for those household members whose nificantly higher than urban areas (0.8%). household head is 65 years and older. This Further, only 1.5% of the population is indicates that the elderly requires social subsistence poor, i.e., persons belonging to protection, given their limited capacities to households with per capita consumption engage productively in economic activity. below food requirements of Nu1,473.45. Subsistence poverty is lower than the esti- Basic Needs mate rate in 2012 of 2.8%. Poverty rates, A marked disparity in aggregate simple according to PAR 2017 are observed to literacy can be observed between the poor be high in Dagana, Zhemgang, Monggar, and non-poor in 2017 with the poor having Trongsa, and Pema Gatshel, compared to a literacy rate of 57% compared to that other Dzongkhags, while Haa, Thimphu and of the non-poor (66.8%). Disparities are Paro have the least poverty. further observed within urban and rural vii Bhutan Poverty Analysis Report 2017 areas: the literacy rate of the poor in urban compared to only 29% among poor house- areas is 16 percentage points lower than holds. However, ordinary phone ownership the rate for the urban non-poor, while in among poor households (80.5%) is signifi- the rural areas the rate for the poor is just cantly higher than non-poor households two percentage points lower than the rural (54.3%). Nationally, only 39% among the non-poor. poor households have television, compared Just about 50% of the non-poor to 76% of the non-poor households. adult population (15+) have not attended Most of the poor suggest that road school/institute, compared to about 67% infrastructure, water supply, and medical of the adult poor population. facilities should be the priorities of the Around 12% of the surveyed popu- Government. However, in rural areas, poor lation reported that they had suffered from households specify public transport, while sickness or injury in the four weeks prior to in urban areas, employment creation was the Survey, with no significant difference specified as priority concern. between the poor and non-poor. However, within this population, only a little over half Inequality (61.0%) of the poor visited a medical facility, On average, a person in the top 20% of the compared to about 70% of the non-poor. national population consumes 6.7 times The majority (99.5%) of the popula- more than a person in the bottom 20% of tion have access to improved water source the population. However, a person in the with hardly any disparity between the poor top 10% consumes 1.6 times more than and the non-poor households. At least 92% a person in the bottom 40% of the pop- of households have access to improved ulation. The Gini index, which measures sanitation; between poor and non-poor inequality, has remained almost the same households, both in urban and rural areas, at the national level (0.36 in 2012 and 0.38 the disparity is around 8%. in 2017). Among the non-poor households, 67% have at least one smart phone, viii Chapter 1.  Demographic Characteristics 1.1.  Background • New estimates of per capita house- The purpose of this report is to provide hold consumption are prepared that updated poverty estimates for Bhutan are as comparable as possible with using newly available data from the Bhutan the consumption estimates prepared Living Standards Survey (BLSS) of 2017. in 2012; Poverty estimates were produced in 2003, • The per capita consumption of each 2007, and 2012 using corresponding BLSS household in the sample is compared data. The updated poverty estimates in this to the updated poverty lines to iden- report can be used to monitor Bhutan’s tify the poor and to calculate the success in reducing poverty during the past relevant poverty indicators. five years since the last poverty estimates Chapter one briefly describes BLSS in 2012. It is also useful for broadening 2017, which is the primary data source and deepening our understanding of the used in preparing the updated 2017 pov- changing dimensions of Bhutan’s poverty erty estimates. Chapter two summarizes and for designing appropriate interven- the work carried out to update the 2017 tions for poverty reduction and monitoring poverty lines for inflation. Chapter three efforts. presents patterns in consumption pov- erty. Chapter four presents an analysis of 1.2.  Objectives socio-economic indicators that provide an The key objective of this report is to update independent source of information on pov- poverty estimates that are as comparable erty reduction during the period 2012-2017. as possible with the estimates prepared for Chapter five provides measures of income 2012. This involves the following steps: inequality (for example, estimates of Gini • The 2017 poverty lines are updated coefficient). Chapter six provides the for inflation in food and non-food report’s conclusions and recommendations. prices during the 2012-2017 period; 1 Bhutan Poverty Analysis Report 2017 1.3.  Data Source collecting consumption expenditure data, it The data used for this report is from BLSS also collected data on demographic charac- 2017, which is the latest, and fourth in a teristics of household members, household series of national household surveys that assets, credit and income, remittances, have been conducted by the NSB. Like in housing, access to public facilities and previous surveys, BLSS 2017 followed the services, education, employment, health World Bank’s Living Standard Measure- of household members, and prices paid for ment Study (LSMS) methodology. The commodities. Also, it included questions sample size has been increased to nearly on happiness and self-rated poverty. 12,000, compared to about 10,000 in 2012 The sample households for BLSS and 2007 and about three times the size of 2017 were selected on the basis of two the survey in 2003. BLSS 2017 surveyed mutually exclusive sampling frames for 11,660 households across the country from rural and urban areas. The total sample size a planned sample size of 11,812. It pro- was set to about 11,812 (more than BLSS vides greater levels of detailed information 2012) and sample sizes of urban and rural needed to prepare the updated poverty areas were allocated across all Dzongkhags estimates. The questionnaires that were and strata in proportion to the number of administered in BLSS 2012 and BLSS households. The primary sampling units 2017 were similar. (PSUs) were Enumeration Areas (EAs) for Using the BLSS 2017 data, an urban (towns) and Chiwogs for rural areas aggregate of household consumption was while the secondary sampling units (SSUs) generated and subsequently analysed. This were the households within the selected aggregate excludes household expendi- EAs/ Chiwogs. tures on durables, irregular expenses, and A set of household weights is needed health expenses (on consultations and when interpreting statistics from the BLSS hospitalization) from the total household 2017 household data. These weights are consumption expenditures (found in the needed to correct for the varying area and BLSS 2017 report), but includes expenses household in the survey design. They are on medicines. Details on the computation made up of three components: (a) a cor- of this consumption aggregate are pro- rection for the differing sampling rates of vided in Technical Note 1 of Annex-II. PSUs used in the strata at the area stage BLSS 2017 gathered data on house- of sampling; (b) a correction for varying hold consumption expenditure, and as such, numbers of households selected in each provides a means of assessing the level of PSU; and, (c) a correction for non-response. poverty and well-being in Bhutan. Besides The survey population coverage included 2 Demographic Characteristics all households in the country except (a) diplomatic and expatriates households; (b) institutional households, i.e., residents of hotels, boarding and lodging houses, monasteries, nunneries, school hostels, orphanages, rescue homes, those under trials in jails, and in-house patients of hospitals; and, (c) barracks of military and para-military forces, including the police. 3 Chapter 2.  Updating the Poverty Line Bhutan’s poverty lines, defined in 2007, The food poverty line is based on the consist of a single national food poverty estimated cost of a single national reference line and non-food allowance, and refer to food bundle providing an average subsistence the monthly per capita levels of food and diet of 2,124 Kcal per day (i.e., averaged non-food consumption. Both the food over persons of all ages and both sexes).1 poverty line and the non-food allowance The reference food bundle was designed to measured in current prices must, therefore, reflect the actual food consumption patterns be updated for inflation, i.e., they need to of Bhutanese in 2007 who consumed a diet be converted into 2017 prices. This Chap- yielding approximately 2,124 Kcal per ter discusses the procedures used to update day. The food basket used in this report the 2012 poverty lines. is representative of the diet of a reference population, namely population in the sec- 2.1.  Updated Food Poverty Line ond, third or fourth decile based on nominal The poverty line, which is the minimum per capita consumption. The selection of acceptable standard of per capita consump- households in the second to the fourth deciles tion needed to assure a minimum standard of the per capita expenditure distribution of living, is obtained using the Cost of ensures that neither expensive nor cheap Basic Needs (CBN) approach, a com- food items are heavily represented in the monly used methodology for constructing basket. After all, prices paid even of the same the poverty lines in many countries. This items could differ across the population. approach estimates the food component of Although food consumption patterns differ the poverty line as the cost of a food bundle across the country, a single food basket was that provides a predetermined minimum used to ensure a consistent comparison of required level of food energy. The total welfare levels of people living in different poverty line is obtained by adding to the food component the cost of the non-food 1  There are 53 food items in the food bundle allowance. 5 Bhutan Poverty Analysis Report 2017 areas of Bhutan. The 2012 poverty line is In order to update the non-food updated for inflation to the year 2017. The allowance for inflation in different regions, methodology used to update for inflation it is necessary to develop regional non- involves (1) updating the food poverty line food price indices similar to the food price using the ratio of the food information of index. Estimates of inflation in non-food 2017 to the food inflation of 2012, (2) using prices developed in this report are based on the food price data collected in BLSS 2017 non-food price data collected for 2012 and to estimate spatial (regional) differences in 2017 inflations. food prices in the survey year. The Consumer Nationwide, the non-food allowance Price Index (CPI) is believed to be a reli- was estimated at Nu722.50 per person per able source of information about inflation month. Adding this non-food allowance to because of its rigorous collection. the food poverty line yields the total pov- Households (and their members) erty line, estimated to be Nu2,195.95 per consuming (in real terms) less than person per month, at 2017 prices. the food poverty line, of Nu1,473.45 Households (and their members) per person per month, are considered subsistence poor. consuming (in real terms) less than the total poverty line, of Nu2,195.95 2.2.  Updated Non-food per person per month are considered Allowance and Total Poverty Line poor. The 2007 baseline non-food allowance was estimated as the per capita monthly non- Table 2.1 shows the comparison of food consumption of households in the poverty lines (food poverty line, non-food reference population whose food spending allowance and poverty line) for 2007, 2012 was near the food poverty line. This is a and 2017. As mentioned, the 2017 food conservative non-food allowance because and nonfood poverty lines are derived from it represents non-food consumption that is the 2012 values by adjusting for inflation at the expense of food consumption, which that occurred between 2017 and 2012. could otherwise be used to achieve the ref- erence food bundle of 2,124 Kcal per day 2.3.  Spatial Price Index per person.2 Prices differ across the country and, there- fore, per capita consumption expenditures (in nominal terms) across regions are 2  Although persons with total per capita consumption below the food poverty line would have to sacrifice not directly comparable. An important some food consumption to purchase non-food items, staple food like rice is found to be a lot they would presumably substitute cheaper foods for more expensive foods within the reference food bundle more expensive in Gasa than in Wangdue 6 Updating the Poverty Line Table 2.1  Poverty Lines of 2007, 2012 and 2017 true cost-of-living index. One possible spa- Poverty lines 2007 2012 2017 tial price index is the Paasche index, which Food poverty line 688.96 1,154.74 1,473.45 calculates the cost of buying a region’s Non-food allowance 407.98 550.10 722.50 Total poverty line 1,096.94 1,704.84 2,195.95 basket of goods using base reference prices. A Paasche index was computed with food Table 2.2  Regional Price Deflator (Median of items using the BLSS 2017 median price Household-level Paasche Indices), by Dzongkhag and Area data. Details on these computations are Dzongkhag Urban Rural provided in Technical Note 1 (d). Bumthang 1.25 1.17 Consequently, the average monthly Chukha 0.94 0.97 household consumption in 2017 for Bhutan Dagana 0.93 0.92 Gasa 1.09 1.24 was estimated at Nu28,550 in real terms as Haa 1.12 1.09 a result of adjustments in the differences Lhuentse 1.03 0.96 in cost of living (with exclusion of some Mongar 1.08 1.05 Paro 1.09 1.05 non-food expenditures on durable items Pema Gatshel 1.04 1.15 and other irregular expenses). Average Punakha 1.20 1.13 monthly per capita consumption in real Samdrup Jongkhar 0.93 0.72 Samtse 0.87 0.86 terms was estimated at Nu6,758 per person Sarpang 0.92 0.89 per month. In 2007 and 2012, average per Thimphu 1.10 1.11 capita consumption in real terms was esti- Trashigang 0.97 0.95 Trashi Yangtse 1.08 1.00 mated at Nu2,745 and Nu5,493 per person Trongsa 1.12 1.11 per month, respectively (Figure 2.1). Tsirang 0.86 0.95 Wangdue Phodrang 1.08 1.02 Figure 2.1  Household and Per Capita Zhemgang 1.04 0.95 Consumption Expenditure (in real terms) in 2007, Bhutan 1.06 0.99 2012 and 2017 30,000 28,550 Phodrang, so that a household in Gasa 25,000 consumes less with the same nominal 20,000 20,913 Ngultrum consumption expenditure on rice than a 15,000 11,777 household in Wangdue Phodrang. To make 10,000 6,758 per capita consumption between regions 5,000 5,493 2,745 comparable, values must be deflated using - a cost of living index. However, no such index is available. The usual approach to Year controlling for spatial price differences is to use a price index that approximates the Househol d Per capita 7 Chapter 3.  Patterns in Consumption Poverty Households with per capita real consump- • Poverty Squared Gap (or Severity of tions below the poverty line are said to be Poverty) – a measure of the inequal- poor and those with per capita real con- ity among the poor. sumption below the food poverty line are The above poverty measures are considered subsistence poor. Subsistence presented in this report for the country as poverty may be viewed as extreme poverty, a whole, and for certain groups of the pop- i.e., those whose consumption expenditure ulation, such as for households in urban is insufficient even to meet basic food needs and rural areas, and in Dzongkhags, and even if they devote their entire consump- by the sex of the household head, among tion expenditure to food alone. others. For more information on indices of Consumption poverty in this report poverty, see Technical Note 4. is measured at the household level since data from BLSS 2017 does not allow 3.1.  Poverty Rate intra-household analysis. Consequently, if The food poverty line and total poverty a household is considered poor, then all its line are used to compute subsistence and members are considered poor. Similarly, if poverty incidence, respectively. Figure 3.1 a household is non-poor, none of its mem- illustrates subsistence and poverty rates for bers is poor. population across urban and rural areas. Three aspects of consumption pov- These rates are poverty head counts, erty are of particular interest: i.e., the percentage of poor persons. For • Poverty Incidence – the proportion the year 2017, the total poverty rate3 for of persons (or households) identified Bhutan is estimated at 8.2%. This means as poor; • Poverty Gap (or Depth of Poverty) – 3  However, if we do not update the poverty line in the extent to which those identified as 2017, i.e., we keep the poverty line in 2017 same as in poor fall below the poverty line; 2012 (Nu1,704.84 per person per month), the poverty rate would be 3.2% and the subsistence poverty rate would be 0.37% 9 Bhutan Poverty Analysis Report 2017 Figure 3.1  Population Poverty and Subsistence precision of these estimates. The best Poverty in Bhutan estimate of poverty rate in Bhutan in 14.00 11.94 2017 is 8.2%. However, this estimate has 12.00 a margin of error of nearly 0.5%.i.e., if we 10.00 8.21 conduct a similar survey a 100 times, 95% Percenta ge 8.00 6.00 of the time, the true poverty rate will fall 4.00 2.31 between 7.3% and 9.1%. Similarly, urban 1.54 2.00 0.78 poverty, estimated at 0.8% (but could range 0.01 0.00 between 0.4% to 1.2%), is much lower Urban Rural Bhutan than rural poverty of 11.9% (could range Povert y Rate Subsistenc e Povery Rate between 10.6% to 13.2%). About 97% of poor persons throughout the country that around one out of 12 persons belongs reside in rural areas. Among the extremely to households whose per capita real con- poor, practically everyone resides in rural sumption is below the total poverty line of areas. Consequently, efforts toward poverty Nu2,195.95 per person per month. Poverty reduction ought to continue with a strong in Bhutan is still a rural phenomenon with focus on rural development. The poverty about 11.9% of the rural population being estimates of 2017 are comparable with poor against only 0.8% in the urban areas. previous estimates of 12.0% poor and It is observed that subsistence inci- 2.8% subsistence poor in 2012. From Table dence, i.e., extreme poverty, is relatively 3.1, it can be derived that, of the estimated small in the country; only around 1.5% of surveyed population of 692,895 persons the population in Bhutan belong to house- in the country, 56,855 are estimated to be holds that spend less per person per month poor and 10,687 are subsistence poor. than the food poverty line of Nu1,473.45. Table 3.2 presents poverty incidence While nearly 0.8% of extremely poor per- and subsistence incidence as a percent sons in rural areas is small, it is significantly of households. About 6% of households large in relation to that of the urban areas are poor, and 1% are subsistence poor (0.01%). households. Out of the estimated 164,011 The poverty and subsistence poverty households, 9,424 are poor, and 1,677 are statistics are shown in Table 3.1 together extremely poor. with their standard errors. Because the pov- A comparison of the poverty statistics erty incidence figures are estimates from a in Table 3.1 and Table 3.2 indicates that sample survey, it is important to consider poverty measures based on population are their standard error when evaluating the larger than those based on the number of 10 Patterns in Consumption Poverty Table 3.1  Population Poverty and Subsistence Poverty by Area Poverty Subsistence Poverty Standard Contribution to Standard Contribution to Area Rate error National Rate error National Population share Urban 0.78 0.20 3.16 0.01 0.01 0.23 33.45 Rural 11.94 0.70 96.84 2.31 0.29 99.77 66.55 Bhutan 8.21 0.48 100.00 1.54 0.19 100.00 100.00 Table 3.2  Household Poverty and Subsistence Poverty by Area Poverty Subsistence Poverty Standard Contribution to Standard Contribution to Area Rate error National Rate error National Household share Urban 0.48 0.11 2.97 0.02 0.02 0.74 35.57 Rural 8.65 0.53 97.03 1.58 0.19 99.26 64.43 Bhutan 5.75 0.34 100.00 1.02 0.12 100.00 100.00 Table 3.3  Population Poverty by Dzongkhag Distribution of Distribution of Dzongkhag Poverty rate Standard error the Poor Population Bumthang 2.1 0.8 0.6 15,959 Chhukha 3.5 0.8 3.9 63,355 Phuentsholing Thromde 0.9 0.4 0.3 20,560 Other than Phuentsholing Thromde 4.8 1.2 3.6 42,795 Dagana 33.3 5.6 13.7 23,453 Gasa 12.6 5.5 0.8 3,575 Haa 0.9 0.7 0.2 10,995 Lhuentse 6.7 2.5 1.8 15,552 Monggar 17.1 3.0 12.6 41,956 Paro 0.3 0.3 0.2 36,329 Pema Gatshel 13.7 3.3 6.7 27,636 Punakha 2.6 1.3 1.2 26,724 Samdrup Jongkhar 6.2 1.4 4.0 36,154 Samdrup Jongkhar Thromde 0.3 0.2 0.0 9,376 Other than Samdrup Jongkhar Thromde 8.3 1.9 3.9 26,778 Samtse 12.3 2.1 13.6 63,132 Sarpang 12.1 1.8 8.8 41,254 Gelephu Thromde 1.1 0.6 0.2 8,015 Other than Gelephu Thromde 14.7 2.2 8.6 33,238 Thimphu 0.6 0.2 1.3 125,551 Thimphu Thromde 0.4 0.2 0.7 98,148 Other than Thimphu Thromde 1.1 0.6 0.5 27,403 Trashigang 10.7 2.6 8.9 47,102 Trashi Yangtse 11.9 2.0 3.2 15,363 Trongsa 14.0 2.6 4.4 17,768 Tsirang 4.8 2.0 1.7 20,409 Wangdue Phodrang 5.4 1.7 3.9 41,405 Zhemgang 25.1 4.2 8.5 19,224 Bhutan 8.2 0.5 100.0 692,895 11 Bhutan Poverty Analysis Report 2017 households because poor households, on Figure 3.2  Distribution of Population Poverty and Subsistence Poverty by Dzongkhag average, have more household members. 25.00 Dzongkhag level estimates of poverty incidence and subsistence poverty for the 20.00 population and for households are shown Percenta ge 15.00 in Table 3.3 (together with their standard errors). Ranks for Dzongkhags are difficult to 10.00 determine due to overlapping confidence 5.00 intervals, but it is observed that poverty rates 0.00 are highest in Dagana, Zhemgang, Mong- Trongsa Mong gar Sarpan g Chhukh a Tsirang Thimphu Haa Trashi Yangtse Wangdue Phodrang Gasa Paro Pema Gatshel Lhuentse Dagana Samtse Punakha Trashigang Zhemgang Bumthang Samdru p Jongkhar gar, Trongsa, and Pema Gatshel. However, the Survey shows that Haa, Thimphu and Paro have the least poverty rates. The four Thromdes (Phuentsholing Thromde, Samdrup Poor Subsistenc e Poor Jongkhar Thromde, Gelephu Thromde and Thimphu Thromde) have poverty rates of at most 1% of their respective populations. households, with Dagana and Zhemgang In terms of subsistence poverty, the also contributing a big share to total house- highest rate is observed in Dagana with hold poverty in the country. 11% of the population being extremely It is also important to observe the dis- poor. Further, about a quarter (23.2%) tribution of the poor population (Fig. 3.2). of all the extremely poor in Bhutan Among the Dzongkhags, Dagana (13.7%), reside in Dagana. Some Dzongkhags such Samtse (13.6%) and Monggar (12.6%) have as Bumthang, Paro and Thimphu have the highest shares of the entire poor pop- virtually no subsistence poverty. Among ulation in country; with 40% of the poor Thromdes, Phuentsholing Thromde, Gelephu residing in these three Dzongkhags alone. Thromde and Thimphu Thromde also have In terms of the distribution of subsistence no subsistence poverty (Table 3.4). poor, again the Dzongkhags of Dagana The estimated number of poor (23.2%), Monggar (15.3%) and Samtse households across Dzongkhags is provided in (12.4%) have the highest proportion of the Table A.1 (Annex I). These tables include subsistence poor population. In fact, half the contribution of each Dzongkhag to total of the subsistence poor live in these three household poverty in the country. Dagana Dzongkhags. (23.7%), Zhemgang (16.3%), Monggar (14.0%) have a higher proportion of poor 12 Patterns in Consumption Poverty Table 3.4  Population Subsistence Poverty by Dzongkhag Distribution of Distribution of Dzongkhag Poverty rate Standard error the Poor Population Bumthang 0.0 0.0 0.0 15,959 Chhukha 0.2 0.2 1.3 63,355 Phuentsholing Thromde 0.0 0.0 0.0 20,560 Other than Phuentsholing Thromde 0.3 0.3 1.3 42,795 Dagana 10.6 3.2 23.2 23,453 Gasa 1.0 1.0 0.3 3,575 Haa 0.4 0.4 0.4 10,995 Lhuentse 1.5 1.2 2.2 15,552 Monggar 0.0 0.0 15.3 41,956 Paro 0.0 0.0 0.0 36,329 Pema Gatshel 1.8 1.2 4.6 27,636 Punakha 0.1 0.1 0.2 26,724 Samdrup Jongkhar 2.0 0.8 6.6 36,154 Samdrup Jongkhar Thromde 0.3 0.2 0.2 9,376 Other than Samdrup Jongkhar Thromde 2.5 1.0 6.4 26,778 Samtse 2.1 0.8 12.4 63,132 Sarpang 0.0 0.0 7.7 41,254 Gelephu Thromde 0.0 0.0 0.0 8,015 Other than Gelephu Thromde 2.5 0.7 7.7 33,238 Thimphu 0.0 0.0 0.0 125,551 Thimphu Thromde 0.0 0.0 0.0 98,148 Other than Thimphu Thromde 0.0 0.0 0.0 27,403 Trashigang 1.6 0.8 7.1 47,102 Trashi Yangtse 1.2 0.6 1.7 15,363 Trongsa 3.9 1.5 6.6 17,768 Tsirang 0.4 0.4 0.7 20,409 Wangdue Phodrang 0.4 0.4 1.7 41,405 Zhemgang 4.4 2.1 8.0 19,224 Bhutan 1.5 0.2 100.0 692,895 3.2.  Depth and Severity of difference between the poverty line and the Poverty actual per capita expenditure (the gap is Poverty analysis is not limited to examining zero for all non-poor individuals). The pov- poverty rates and comparing the statistics erty gap index measures the average extent across sub-groups of the population. It is to which individuals in a population fall important to also look into the depth and below the poverty line and expresses it as severity of poverty. The poverty gap and a percentage of the poverty line. The pov- poverty squared gap indices measure the erty squared gap index gives more weight depth and severity of poverty, respectively. to the very poor than those who are less For an individual, the poverty gap is the poor. It is the average value of the square 13 Bhutan Poverty Analysis Report 2017 of the depth of poverty for each individual The table also includes the contribution of measured relative to the poverty line. More the Dzongkhags to the national poverty mea- explanation on these indices is available in sures. Some Dzongkhags such as Monggar Technical Note 4. and Samtse have very high poverty mea- For both the poverty gap and poverty sures (whether in terms of gap or severity) squared gap, as well as for poverty rate, but Dagana has the highest contribution to the larger the value of the index, and the the national poverty measures. greater the degree of poverty. These pov- erty measures are important for planning 3.3.  Poverty trend poverty reduction programmes. All things Figure 3.4 shows that the overall poverty being equal, sub-groups of the population rate in the country reduced from 23.2% with higher measures should receive prior- in 2007 to 12% in 2012 and further to ity for poverty reduction programmes. 8.2% in 2017. Rural poverty reduced Figure 3.3 shows that poverty is from 30.9% in 2007 to 16.7% in 2012 and deeper and more severe in rural areas 11.9% in 2017. However, the proportion of than in urban areas. The poverty gap in poor in urban areas remained practically rural areas is almost 2.4% as compared to unchanged at about 2% between 2007 and just below 0.2% in urban areas. Poverty 2012, but significantly reduced to 0.7% in squared gap in rural areas is a little over 2017. 0.7% while it is just 0.04% in urban areas. Figure 3.4  Population Povery Rates for 2007, 2012 and 2017 Figure 3.3  Depth and Severity of Poverty by Area 35 3.00 2.38 30.9 30 2.50 2.00 1.64 25 Percenta ge 23.2 1.50 Percenta ge 20 1.00 0.73 0.50 16.7 15 0.50 0.15 0.04 12.0 11.9 0.00 10 Urban Rural Bhutan 8.2 Povert y Gap Povert y Square d Gap 5 1.7 1.8 0 0.7 The poverty gap and poverty squared gap (with their standard errors) across Urban Rural Bhutan Dzongkhags are listed in Annex I (Table A-3). 14 Patterns in Consumption Poverty As shown in Figure 3.5, subsistence Table 3.5  Average Household Size by Area, Poverty Status, and Sex of Head poverty decreased from 5.9% in 2007 to Household Head about 2.8 % in 2012. In 2017, the subsis- Area/Poverty Status Male Female Total tence poverty rate further reduced to 1.6%. Urban 4.1 3.7 4.0 In the rural areas, the rate was reduced Poor 7.1 3.4 6.4 Non-poor 4.1 3.7 4.0 from 8% in 2007 to 3.9% in 2012, and it Rural 4.4 4.4 4.4 is 2.5% in 2017. In the urban areas, the Poor 5.9 6.3 6.0 subsistence poverty rate is significantly low Non-poor 4.2 4.2 4.2 Bhutan 4.3 4.2 4.2 (around three in 10,000 persons). Poor 5.9 6.2 6.0 Non-poor 4.1 4.1 4.1 Figure 3.5  Population Subsistence Povery Rates for 2007, 2012 and 2017 9.0 poor and non-poor households is slightly 8.0 8.0 larger in urban areas as compared to rural 7.0 areas. The average household size is almost the same irrespective of the sex of the 6.0 5.9 household head. However, the difference Percenta ge 5.0 in household size between the poor and 4.0 3.9 non-poor households is slightly larger for 3.0 2.8 female-headed households. 2.5 2.0 Figure 3.6  Household Poverty and Subsistence 1.6 Poverty Rates by Household Size 1.0 45.0 40.7 0.2 0.3 0.0 0.0 40.0 35.0 31.4 Urban Rural Bhutan 28.2 30.0 Percenta ge 3.4.  Poverty by Household 25.0 19.4 Characteristics 20.0 Households differ in their demographic 15.0 11.8 composition and characteristics. On 10.0 5.8 5.3 6.7 average, household sizes in Bhutan are 5.0 0.8 1.6 0.7 2.3 2.7 slightly larger in rural (4.4) than in urban 0.0 0.5 0.2 1 HH 2-3 HH 4-5 HH 6-8 HH 9 and more (4.0) areas. As shown in Table 3.5, a poor member member s member s member s HH members household typically has a much larger Povert y Rate family size (6.0) than a non-poor household Share of total household s (4.1). The difference in family sizes among Subsistenc e Povert y Rate 15 Bhutan Poverty Analysis Report 2017 As shown in Figure 3.6, both pov- Figure 3.7  Population Poverty Rate by Economic Activity of the Household Head erty rates and subsistence poverty rates increase with the size of the households. Bhutan 100.0 2.0 The increase in poverty rate is faster than 20.5 subsistence poverty rate as the household Economi cally inactive 7.8 size increases. The share of households 0.1 increases rapidly reaching a maximum of Unemployed 2.3 41% for households with four or five mem- 6.7 Service bers. However, the share then decreases 1.2 and reaches a minimum of 3% for house- Indust ry 3.9 2.2 holds with nine or more members. This 68.8 indicates that, although the poverty rates Agriculture 9.6 are higher among households with a larger household size, the corresponding share of these households to total households is Contribution to national pover ty rate much less. Povert y incidenc e Typically, welfare and household demographic composition are observed urban and rural areas. However, the sex of to have a relationship with the character- the household head does not have much istics of the household head. On average, influence on subsistence poverty (Table 3.6). female-headed households are observed to Figure 3.7 combines information on be less poor than male-headed households. poverty, participation in the labour force, The trend is observed to be similar in both and main sectors of employment of the Table 3.6  Household Poverty and Subsistence Poverty Rates by Area and Sex of Household Head Area/ Poverty Rate Subsistence Poverty Rate Household Contribution to Contribution to Share of Total Household Head Index National Index National Heads Urban 0.5 3.0 0.0 0.7 35.6 Male 0.6 2.4 0.0 0.0 25.2 Female 0.3 0.5 0.1 0.7 10.3 Rural 8.7 97.0 1.6 99.3 64.4 Male 9.4 64.4 1.7 63.8 39.4 Female 7.5 32.6 1.4 35.4 25.0 Bhutan 5.7 100.0 1.0 100.0 100.0 Male 5.9 66.8 1.0 63.8 64.7 Female 5.4 33.2 1.0 36.2 35.3 16 Patterns in Consumption Poverty households. Living standard of a person Figure 3.8  Household Poverty Rate by Educational Attainment of Household Head by Area is higher among those households whose 12.0 heads are currently working as compared 10.5 to those whose heads are either unem- 10.0 8.8 ployed or not in the labour force. Among 8.0 the employed, poverty rates are higher in Povert y Rate households whose heads are working in 6.0 4.9 agriculture (9.6%), though this is a decrease by almost half from the 2012 (18.5%) fig- 4.0 2.9 ure for the same. Most of the poor live in 2.0 1.2 1.1 households whose head is either engaged 0.3 0.0 in agriculture (68.8%) or whose head is not 0.0 0.1 0.4 actively participating in the labour force (20.5%). Urban Rural Bhutan Figure 3.8 shows household poverty rates by educational attainment levels of the household head. As expected, the to education increase considerably if the higher the level of education completed head had attended middle secondary level by the household head, the lower the of education irrespective of whether the poverty rate for the household. In other household is in an urban or rural area. words, the level of poverty decreases as Table 3.7 shows that the poverty rates the educational level of the household increase with the age of the household head increases. About 9% of the house- head. The poverty rate is about 2% for holds with household heads who had not those below 25 years as compared to 10% attended a school are poor. The returns for those aged 65 years and older. This Table 3.7  Household Poverty and Subsistence Poverty Rates by Age of Household Head Age of Poverty Rate Subsistence Poverty Rate Household Contribution to Contribution to Share of Total Household Head Index National Poverty Index National Poverty Heads <25 1.9 0.9 0.0 0.0 2.8 25-34 3.0 12.0 0.6 14.3 22.8 35-44 5.3 22.2 0.8 18.4 24.3 45-54 5.7 21.2 1.0 20.0 21.4 55-64 7.5 21.2 1.4 21.7 16.3 65+ 10.4 22.5 2.1 25.7 12.5 All ages 5.7 100.0 1.0 100.0 100.0 17 Bhutan Poverty Analysis Report 2017 suggests in the importance of providing households (11.8%). Overall, there is an social protection for elderly person. It is increase in the use of wood (42.4%) and noticed that most household heads (68.4%) cement/tiles (30.6%) types of floor mate- in Bhutan are aged between 25 to 54 years, rials and a decrease in the use of plank/ while less than 3% are below the age of shingles (12.1%) as compared to BLSS 25 years, and about 11% are 65 years and 2012. above. Figure 3.10 shows the distribution Figure 3.9 shows the distribution of of main materials used for walls by pov- floor materials by household poverty status. erty status. More than half (54.9%) of There is no significant difference in the the poor households have residences with types of floor materials used by the poor mud-bounded walls while a slightly more and the non-poor households, except in than one-third (34.7%) of the non-poor the use of cement/tiles, clay/earthen and households have dwellings that have plank/shingles. About 12% of the poor mud-bounded walls. Only 12% of the households have cement/tile floors, com- poor households have cement-bounded or pared to 32% of the non-poor households. concrete walls as compared to 40% of the Figure 3.9  Distribution of Type of Floor by Figure 3.10  Distribution of Type of External Household Poverty Status Walls by Household Poverty Status 60 60 54.9 50 50 42.6 42.4 42.4 Percenta ge Percenta ge 40 40 35.9 34.7 31.8 30.6 30 30 25.5 24.6 18.2 19.6 17.6 20 20 14.6 13.9 12.1 12.3 11.8 12.9 12.5 10.0 9.0 10 8.8 8.8 7.7 7.1 7.2 10 7.0 6.3 5.2 4.0 3.7 1.6 0.7 0.7 1.5 0 0 Poor Non- Poor Total Poor Non- Poor Total 18% of the poor households have clay/ non-poor households. The proportion of earthen floors while only 6% of the non- household with wood/branches is higher in poor households have clay/earthen floors. the poor households (19.6%) as compared A higher proportion of the poor house- to non-poor households (12.5%). holds (17.6%) uses plank/singles for floors Figure 3.11 shows the distribution of in their residences compared to non-poor households and poverty rates by the size of 18 Patterns in Consumption Poverty Figure 3.11  Household Distribution and Poverty Table 3.8  Household Land ownership by Area in Rural Areas and Poverty Status 25 Area Poor Non-Poor Total 22.7 Urban 42.2 41.0 41.0 Percenta ge of households 20 19.7 Rural 95.7 86.7 87.5 Bhutan 94.1 69.6 71.0 15 15.0 12.6 13.8 13.2 and the highest for those who own four to 10 10.3 five acres (13.2%) of land. The incidence 9.2 9.3 9.1 7.9 7.9 8.3 of poverty is almost similar for those who 5 own up-to one acre (9.2%), two to three 3.0 acres (9.3%) and three to four acres (9.1%) 0 of land. Table 3.8 illustrates land ownership in urban and rural areas by poverty status. Povert y inciden ce Distri bution of household Across the country, 71% of households land holdings in rural areas. More than a own land with a higher proportion owned fifth (22.7%) of rural households own up by poor households (94.1%). The propor- to one acre of land with the proportion of tion of households owning land in rural rural households owning land decreasing areas (87.5%) is more than two times with the size of land holding. Compared that of urban areas (41.0%). Compared to PAR 2012, the proportion of landless to 2012, the proportion of households in households (12.6%) and those households 2017 who own land has increased in both who own up-to one acre of land have urban (32.3% in 2012) and rural (83.6% in decreased while households in other cate- 2012) areas, resulting in an overall increase gories have increased. The poverty rate is in land ownership in the country in 2017. the lowest (3.0%) for landless households 19 Chapter 4. Basic Needs Other non-monetary dimensions of wel- Figure 4.1  Literacy Rate by Area and Poverty Status fare, such as health and education status, 81.7 81.8 90 that pertain to basic needs, are comple- 80 66.8 66.0 66.2 mentary to consumption poverty. The 70 58.5 58.3 56.8 56.5 health status of an individual undoubtedly 60 Literacy Rate 50 determines her/his quality of life. Literacy 40 and education attainment are widely rec- 30 20 ognized to be important for improving the 10 living standards of the population. People 0 Urban Rural Bhutan with little or no education are likely to be Poor Non-p oor Total unemployed, or if they do get employed, they often have low-paying, labour-inten- between the poor and the non-poor exists sive occupation, especially in the informal in both urban and rural areas, the disparity economy. Such vulnerable employment is much higher in urban areas. The liter- often put them at risk of staying poor. More acy rate of the poor in urban areas is 16 education provides individuals with the percentage points lower than the urban basic knowledge, skills, and competence non-poor, while in the rural areas, the lit- required for economic productivity, which eracy rate of the poor is just 2 percentage in turn, will provide them with assets and points lower than the non-poor. other capabilities for further improving Figure 4.2 shows the distribution of edu- their living standards, and consequently, cational attainment of adults aged 15 some degree of social mobility. years and older by poverty status. About 67% of the poor population aged 15 years 4.1.  Education and older had never attended a school/ As shown in Figure 4.1, poor persons in institute as compared to just over half of Bhutan have a lower literacy rate than the non-poor in the same age group have non-poor persons; 57% of the poor are not attended any schooling. There are literate as compared to 69% of the non- almost equal proportions of the poor and poor. Though a disparity in literacy rates non-poor adult population that had studied 21 Bhutan Poverty Analysis Report 2017 up to class eight. However, the proportion last four weeks prior to the Survey. Over a of poor persons is much smaller at higher tenth (12%) of the population reported that levels of educational attainment. Just about they had suffered from sickness or injury in 1% of the poor population had studied the last four weeks, with no significant dif- beyond the secondary school level while the ference between the poor and the non-poor adult population among the non-poor had (Figure 4.3). However, among those who 9% within the same educational attain- reported some illness or injury, 61% of the ment bracket. poor visited a medical facility, compared to Figure 4.2  Distribution of Adult (15+) Education about 70% of the non-poor. Attainment by Poverty Status Table 4.1 shows the distribution of 100 1.3 persons who suffered from sickness/injury 8.5 7.9 90 15.6 during the four weeks prior to the Survey with health seeking behavior by area and 80 23.7 23.0 poverty status. There is wide disparity 16.7 70 between the percentage of the poor (7.0%) and the non-poor (24.3%) among those 60 18.1 18.0 who first visited the Jigme Dorji Wangchuck Percenta ge 50 National Referral Hospital (JDWNRH), when they suffered from sickness or injury 40 in the four weeks before the Survey. Almost 66.5 30 equal proportions of the poor and the 49.7 51.0 20 non-poor population had visited regional referral hospitals. Just about 16% of the 10 poor visited a district hospital, compared to 0 25% of the non-poor. The majority of the poor (58.9%) visited a Basic Health Unit None All most grade VIII (BHU)/Satellite clinic/Sub-post, com- Grade IX-XII Beyond grade XII pared to only 27% of the non-poor. The disparity is wider in urban areas. 4.2.  Health Among women who gave birth during BLSS 2017 collected information on the the 12 months prior to the BLSS 2017, a health conditions and access to health slight difference between the percentage of services from all household members. the poor and non-poor women who received Household members were asked whether antenatal care is observed (Figure 4.4). The they suffered from sickness or injury in the difference is significant in urban areas, 22 Basic Needs Figure 4.3  Health Seeking Behaviour by Area Figure 4.4  Proportion of Women Who Received and Poverty Status Antenatal Care by Area and Poverty Status 80 71.5 68.3 69.6 94.3 62.2 100 94.1 89.7 91.9 70 60.9 61.0 88.3 90.9 90 60 75.6 80 74.7 50 40 70 30 60 14.0 13.2 11.5 10.5 12.1 46.6 20 10.4 50 10 40 0 30 20 Urban Rural Bhutan 10 0 Urban Rural Bhutan Percenta ge of persons Who Reported Sick During the Four Weeks Prior to the Enumer ation Data Poor Non-poor Total Proportion of persons Who Reported Sick During the Four Weeks and Consulted Health Provide r Table 4.1  Distribution of Persons who Suffered from Sickness/Injury Four Weeks prior to the Survey with Health Seeking Behaviour by Area and Poverty Status Health Service Provider Urban Rural Bhutan Consulted Poor Non-Poor Total Poor Non-Poor Total Poor Non-Poor Total JDWNRH 23.3 43.5 43.3 6.3 11.3 10.8 7.0 24.3 23.2 Regional Referral Hospital 21.3 22.6 22.6 16.6 15.5 15.6 16.8 18.4 18.3 District Hospital 5.9 15.7 15.6 16.3 31.9 30.3 15.8 25.3 24.7 BHU/Satellite Clinic/ Sub post 43.1 13.0 13.3 59.6 37.0 39.2 58.9 27.3 29.3 ORC 0.0 0.0 0.0 1.3 1.2 1.2 1.3 0.7 0.8 Private Diagnostics Centers 0.0 0.4 0.4 0.0 0.5 0.4 0.0 0.5 0.4 Indigenous centres (Sowa Rigpa) 0.0 0.5 0.5 0.0 0.6 0.6 0.0 0.6 0.6 Chemist/Pharmacy/Retail 0.0 0.1 0.1 0.0 0.1 0.1 0.0 0.1 0.1 pharmacy shop Other private hospital/clinic 0.0 0.3 0.3 0.0 0.0 0.0 0.0 0.1 0.1 Lama/pandit/Priest (Rimdo/Puja) 0.0 0.0 0.0 0.0 0.1 0.1 0.0 0.1 0.1 Traditional Practitioner (Pow/ 0.0 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0 Pam,Shaman, Terda etc) Indian hospital paid by Govt. 0.0 0.8 0.8 0.0 0.5 0.5 0.0 0.7 0.6 Indian hospital paid by self 0.0 0.4 0.4 0.0 0.4 0.4 0.0 0.4 0.4 Outside Bhutan hospital paid by 0.0 0.1 0.1 0.0 0.1 0.1 0.0 0.1 0.1 Govt. (Other than India) Outside Bhutan hospital paid by 0.0 0.2 0.2 0.0 0.1 0.1 0.0 0.1 0.1 self/private (Other than India) Others 6.3 2.4 2.4 0.0 0.7 0.7 0.3 1.4 1.3 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 23 Bhutan Poverty Analysis Report 2017 where a smaller proportion of poor women Figure 4.5  Proportion of Households with Access to Improved Water Source by Poverty Status and Area (46.6%) received antenatal care as compared 100.0 99.5 99.5 to non-poor women (94.4%). In rural areas, 99.6 99.6 99.4 9 9.4 99.5 99.5 100 the disparity is less wide. 90 80 4.3.  Household Amenities, 70 60 Percenta ge Assets, and Access to Services 50 The living conditions of a household are 40 often highly correlated with its amenities, 30 assets, and access to services. Household 20 amenities, including suitable sanitation 10 facilities, and access to safe water sources, 0 are not only wealth indicators, but also Urban Rural Bhutan improve welfare conditions of the house- Poor Non- Poor Total hold. Lack of safe water or basic sanitation affects an individual’s health by increasing tap, protected well, protected spring, and her/his chances of contracting diseases rainwater collection). There is hardly that are transmitted in unsanitary environ- any disparity in access to improved water ments. Some assets may allow households source between the poor and the non-poor. to cope with the risks brought about by Figure 4.6 shows that 91.5% of seasonal variations in incomes from farm- households have access to improved ing, or other sources of vulnerability. If the Figure 4.6  Proportion of Households with Access head of the household suddenly becomes to Improved Sanitation by Poverty Status and Area 97.6 97.6 unemployed, or dies, or if a natural disaster 100 94.5 91.7 87.5 88.1 88.1 87.7 91.5 occurs, the household could use its assets 90 to smoothen consumption. Consequently, 80 70 it is important to look at the amenities and 60 Percenta ge assets of a household, as well as its access to 50 basic social services to get a comprehensive 40 assessment of its welfare conditions. 30 Figure 4.5 illustrates that, across the 20 country, nearly all (99.5%) households 10 in 2017 have access to improved water 0 source (i.e., piped in dwelling, pipe in com- Urban Rural Bhutan pound, neighbors’ pipe, public outdoor Poor Non- Poor Total 24 Basic Needs sanitation (sewers or septic tanks, flush-la- Figure 4.8  Proportion of the Population with Ownership of Mobile phone, TV and Internet trines, pit with slab, or ventilated improved Connection by Poverty Status pit latrines). There is not much disparity in 90 80.5 access to improved sanitation between the 80 76.4 74.3 poor (87.7%) and the non-poor (91.7%). 70 66.8 64.6 60.2 58.1 54.3 55.8 BLSS 2017 found that the main 60 Percenta ge source of energy for lighting throughout 50 38.8 40 the country is electricity (98.6%), which 29.3 30 23.4 is proportionally higher in urban areas 20 (99.2%) than in rural areas (98.3%). Figure 10 4.7 illustrates that all poor (100.0%) and 0 non-poor (99.2%) households in urban areas depend on electricity for lighting. Poor Non- Poor Total However, in rural areas, only a little over 97% of the poor households have electric- The disparity between the poor and the ity as their main source of lighting. non-poor is also evident for ownership of television and for access to Internet connec- Figure 4.7  Distribution of Households’ Fuel Use for Lighting by Poverty Status and Area tion in their homes. Only 39% among the Non-Poor 99.2 0.5 0.3 poor households have television, compared to 76% for non-poor households. There is Poor 100.0 a marked increase in Internet connection Non-Poor 98.4 1.1 0.2 0.3 at home compared to BLSS 2012 (11.6%). Poor 97.3 1.1 1.6 According to BLSS 2017, 58% of house- Non-Poor 98.7 0.9 0.2 0.2 holds have Internet connection at home, Poor 97.4 1.1 1.6 with a higher proportion among non-poor households (60.2%), compared to poor Electricit y Kerosene Candle s Others households (23.4%). Figure 4.8 illustrates that 67% of 4.4.  Perception and Priorities non-poor households owns a smart phones In BLSS 2017, questions about the per- while only 29% of poor households own ception of poverty and happiness were the same. However, the proportion of asked among households interviewed. other phone (ordinary phone) ownership The household head was asked if he/she is higher for poor households (80.5%) as considered the household to be poor. This compared to non-poor households (54.3%). is considered to be a measure of perceived 25 Bhutan Poverty Analysis Report 2017 poverty. Across the country, at least one poor is Nu4,606, compared to Nu1,770 for in six (15.3%) of the household heads the poor. considers their households to be either Regarding perception on happiness poor or very poor. In the urban areas, the (Table 4.3), a majority of the household perceived poverty rate (poor or very poor) heads reported that they were moderately is 8.4%, which is mostly driven by the poor happy (39.2%) or very happy (36.3%). households (54.6%). There are at least 11% Although there is hardly any difference of household heads who do not consider between the poor and the non-poor their households to be poor, yet the analysis who reported to be moderately happy, of survey data shows that they are actually the proportion of household heads who poor. About one in seven (14%) household reported being very happy is much higher heads belonging to non-poor households among the non-poor households; around consider their households to be poor or 24% and 13% higher in urban and rural very poor, and the proportion is more than areas, respectively. More poor households double in rural areas (17.6%) as compared reported being neither happy nor unhappy to urban areas (8.2%). as compared to non-poor households in It could be useful to develop a dif- both urban and rural areas. ferent poverty profile based on perceived BLSS 2017 respondents were asked (subjective) poverty, which is far easier and to identify an action agenda for the simpler to ask. It is generally true that the Government that would improve their subjective poverty line is much higher than welfare. Most of the poor, especially in the poverty line. The mean per-capita the rural areas, suggest that building of expenditure (in real terms) of the perceived roads, water supply, public transport, and Table 4.2  Household Distribution of Subjective Table 4.3  Household Distribution of Subjective Poverty by Area and Poverty Status Happiness by Area and Poverty Status Area/ Neither Area/ Neither Poverty Not poor nor Very Don't Poverty Very Moderately happy/ Moderately Very Status poor non-poor Poor poor know Status Happy happy unhappy unhappy unhappy Urban 29.0 60.6 7.5 0.9 2.0 Urban 38.2 40.1 16.3 2.3 3.0 poor 18.4 26.9 25.5 29.2 0.0 poor 14.2 38.4 22.5 19.4 5.5 Non-poor 29.0 60.8 7.4 0.8 2.0 Non-poor 38.4 40.2 16.3 2.3 3.0 Rural 15.2 63.7 16.5 2.6 2.0 Rural 35.3 38.7 19.2 4.4 2.4 poor 11.0 53.0 28.2 7.1 0.8 poor 23.4 36.5 28.2 7.3 4.5 Non-poor 15.6 64.7 15.4 2.2 2.1 Non-poor 36.4 38.9 18.4 4.1 2.2 Bhutan 20.1 62.6 13.3 2.0 2.0 Bhutan 36.3 39.2 18.2 3.7 2.6 poor 11.2 52.2 28.1 7.7 0.8 poor 23.1 36.6 28.0 7.7 4.6 Non-poor 20.7 63.2 12.4 1.7 2.1 Non-poor 37.2 39.4 17.6 3.4 2.5 26 Basic Needs medical facilities should be the priorities Figure 4.9  Rate of Characteristics for Perceived Poor and Poor Households of the Government. In urban areas, poor 94.1 87.7 100 households specified job creation, water 90 85.7 80 supply, roads, and medical facilities as Percenta ge 70 78.9 80.8 60 50.5 74.7 priority concerns. 50 40 Figure 4.9 shows the difference for 30 21.2 38.8 20 some indicators among the poor and those 10 10.6 0 who perceived themselves as poor. The lit- Literac y rate Land TV Antenata l Safe of household ownership ownership care sanitation eracy rate of the household heads (10.6%), head TV ownership (38.8%), and access to Perceived Poor Poor antenatal care (74.7%) are lower for the poor than for the perceived poor while land ownership (94.1%) and safe sanitation (87.7%) are higher for the poor as com- pared to the perceived poor. 27 Chapter 5. Inequality Poverty indicators focus on the population quintile (6.7%) 4 is only one sixth that or households at the bottom of the per of the share of the richest quintile of capita consumption distribution, but it the population (Figure 5.1). Although the is also important to look at the spread of shares of the poorest quintile in urban and consumption over the entire population rural areas are almost similar, the share using inequality indicators. There is much of the richest quintile in the rural areas interest in measuring inequality since high is higher than that of the urban areas. levels of inequality may contribute to, if Figure 5.1  Per Capita Consumption Quintiles not exacerbate, poverty. Growth is known to be important for poverty reduction. High Urban 7.9 12.4 16.8 22.9 40.0 inequality may result to lower subsequent Rural 7.6 11.5 15.5 21.9 43.5 and sustained economic growth and, con- sequently, in less poverty reduction. A high Bhutan 6.7 10.7 15.3 22.6 44.7 level of inequality may make it difficult for the poor to have a substantial share of the Percenta ge benefits of subsequent economic growth. First Second Third Four th Fifth Inequality indicators attempt to measure the deviation of a given consumption dis- Table 5.1 shows that a person belong- tribution from the ideal distribution, called ing to the richest 20% of the national perfect equality. population consumes on average 6.8 times more than a person belonging to the bot- 5.1.  Consumption Quintiles tom 20% of the population. This difference Typically, the population is ranked by is similar to the estimates in PAR 2012, sug- ascending order of per capita consump- gesting that there are no improvements in tion and the distribution is divided into fifths, i.e., 20% of the population wor 4  The consumption aggregates for the poverty equivalent quintiles. In Bhutan, the share analysis is different than the one reported in the BLSS of national consumption of the poorest Report 2017 and therefore the share may not exactly correspond. Refer Technical Note 1 of this Report for further information. 29 Bhutan Poverty Analysis Report 2017 Table 5.1  Average Monthly Real Per Capita Consumption (Nu), Share in National Consumption, Average Share of Food to Total Consumption, Average Household Size by Consumption Quintile Lower Upper Indicator Lowest Middle Middle Middle Upper Overall Average Per Capita Consumption 2,443.8 4,075.6 5,928.2 8,769.2 16,733.1 6,758.0 Share of National Consumption 6.7 10.7 15.3 22.6 44.7 100 Average Share of Food Consumption to Total Consumption 61.5 59.4 55.4 51 45.4 54.5 Average Household Size 5.4 4.6 4.1 3.8 3.2 4.2 consumption inequality. A person in the top remittances, or households residing in 10% consumes 1.6 times more than a per- ‘rural’ areas that have some members son in the bottom 40% of the population who are earning in ‘urban’ areas, thereby, (which is also referred to as Palma ratio). contributing to the inequality observed. In As is to be expected from Engel’s Law, the addition, it may suggest the need to exam- proportion of total consumption allocated ine the current definition of urban and to food tends to decrease as the level of per rural areas. capita real consumption increases. The Gini coefficient, measured by the ratio of the area between the line of 5.2.  Gini Index perfect equality to the Lorenz curve, to Consumption inequality can also be the area (of the triangle) under the line examined using graphical tools such as the of perfect equality, is a commonly used Lorenz curve, which maps the cumulative indicator of inequality. The Gini index consumption share on the vertical axis ranges between 0 and 1 (with zero meaning against the distribution of the population perfect equality and one meaning perfect on the horizontal axis. If each household inequality). The typical values of the Gini had the same consumption, the resulting Figure 5.2  Lorenz Curve of Per Capita curve would be a 45-degree line known Consumption by Area as the line of perfect equality. Figure Lorenz Curves 5.2 illustrates the Lorenz curve of total 81 household consumption in Bhutan. The further the Lorenz curve is from the line 6. of perfect equality, the higher the level of 4. inequality. The Lorenz curve here indicates 2. that inequality in urban and rural areas is 0. very pronounced. The degree of inequality 0. 2. 4. 6. 81 is similar in urban and rural areas. This Lorenz curve for Bhutan Line of Perfect Equality Urban Area s Rural Area s similarity may be the result of in-country 30 Inequality Figure 5.3  Gini Coefficient by Area 0.40 0.38 0.35 0.35 0.32 0.30 0.25 0.20 0.15 0.10 0.05 0.00 Urban Rural Bhutan coefficient range between 0.2 and 0.5. While comparisons with previous estimates and international figures may be carried out, but such comparisons should be done with much caution. Comparisons are more meaningful across groups within the coun- try. Figure 5.3 provides the Gini index at the national level and within urban and rural areas. The Gini at the national level (0.38) is observed to be higher than that of urban (0.32) or rural (0.35) areas. 31 Chapter 6. Conclusion The Royal Government of Bhutan has in this Report conveys information neces- been, over the past years, implementing sary to guide the implementation of plans sustainable development activities with and programmes aimed at eradicating pov- the focus of increasing the living standards erty and improving the living standards of of its citizens. The 10th FYP in particular the poor in Bhutan. This Report shows that aimed to alleviate poverty under the theme, poverty in Bhutan is still very much a rural ‘Poverty Reduction.’ The 11th FYP also phenomenon, and that living standards had plans and programmes geared towards vary considerably across the Dzongkhags. the reduction of poverty. The Millennium While understanding drivers of Declaration, signed by the global commu- poverty reduction requires extensive data nity in 2000 at the United Nations, was a analyses, our preliminary analyses show commitment to ensuring that poverty is that most of the poverty reduction between reduced to half its 1990 status by 2015. The 2012 and 2017 was due to increasing non- Sustainable Development Goals (SDGs) food consumption with no major change in further reaffirmed the global commitment food consumption patterns. For example, to poverty reduction. surveyed households on average spend This is the fourth Poverty Analysis more on transportation, clothing, and rec- Report produced by the NSB. From 2007 reation in 2017, compared to 2012, after onwards, poverty indicators were produced adjusting for inflation. The NSB plans to at the Dzongkhag level. The poverty rates conduct a thorough assessment of poverty have been decreasing consistently from reduction in the near future. 31.7% in 2003 to 23.2% in 2007 and The pace of poverty reduction 12.0% in 2012. appears to have slowed down between 2012 Besides providing comparable and and 2017, relative to the period between updated poverty profiles, PAR 2017 also 2007 and 2012. However, an analysis using presents a spatial distribution of poverty in the World Development Indicators by the Bhutan at the Dzongkhag level, and includes World Bank shows that Bhutan’s poverty the four Thromdes. Updated information reduction over the last 10 years is still about the conditions of the poor presented remarkable from a global perspective. Of 33 Bhutan Poverty Analysis Report 2017 the 38 countries for which there are more represents a social problem that requires than three national poverty estimates since joint efforts by the Government, the 2005 5, Bhutan ranks 7th in terms of the private sector, and the development part- rate of poverty reduction (23.2% to 8.2% ners in addressing it. Development plans in 10 years or 9.9% reduction in poverty should promote sustained, broad-based headcount rate every year). inclusive growth, speeding up growth in Using the same dataset, but look- lagging regions, and reducing poverty in ing at the two episodes (2007-2012 and more deprived population groups. There 2012-2017) separately, Bhutan’s poverty is a need to learn from the successes and reduction ranked in the 85th percentile failures in poverty reduction of other coun- between 2007 and 2012, and in the 67th tries, and customize plans for Bhutan. It is percentile between 2012 and 2017. Even a hoped that this report will help all develop- seemingly slowed rate of poverty reduction ment stakeholders to understand the living between 2012 and 2017 outperformed conditions of the poor in the country, and approximately two thirds of all available to listen to their often unheard voices in episodes since 2005. order to generate informed discussions and Poverty is an important concern policy actions. not only for those who are poor but also 5  This also excludes countries in Europe and Central Asia region where many countries use relative poverty lines to track national poverty 34 Annex I: Additional Statistical Tables Annex I: Additional Statistical Tables Table A.1  Household Poverty by Dzongkhag Distribution of Distribution of Dzongkhag Poverty rate Standard error the Poor Households Bumthang 1.7 0.6 0.7 3,836 Chhukha 2.2 0.5 3.5 14,865 Phuentsholing Thromde 0.7 0.4 0.4 5,125 Other than Phuentsholing Thromde 3.0 0.7 3.1 9,740 Dagana 23.7 4.5 15.0 5,974 Gasa 7.4 3.0 0.7 873 Haa 1.1 0.7 0.3 2,752 Lhuentse 5.2 1.9 2.1 3,754 Monggar 14.0 2.5 13.4 9,049 Paro 0.2 0.2 0.2 8,969 Pema Gatshel 10.1 2.4 7.0 6,536 Punakha 1.8 0.6 1.2 6,450 Samdrup Jongkhar 4.5 1.1 4.1 8,502 Samdrup Jongkhar Thromde 0.5 0.4 0.1 2,379 Other than Samdrup Jongkhar Thromde 6.1 1.5 3.9 6,123 Samtse 8.5 1.5 13.1 14,503 Sarpang 8.4 1.2 9.4 10,537 Gelephu Thromde 1.1 0.7 9.1 2,506 Other than Gelephu Thromde 10.7 1.6 0.4 8,031 Thimphu 0.3 0.1 1.1 30,147 Thimphu Thromde 0.2 0.1 0.6 24,266 Other than Thimphu Thromde 0.8 0.5 0.5 5,882 Trashigang 7.8 1.9 9.3 11,228 Trashi Yangtse 8.7 1.5 3.9 4,228 Trongsa 9.6 2.1 4.0 3,899 Tsirang 2.6 1.2 1.4 5,074 Wangdue Phodrang 3.0 1.0 2.8 8,847 Zhemgang 16.3 3.0 6.9 3,988 Bhutan 5.7 0.3 100.0 164,011 35 Bhutan Poverty Analysis Report 2017 Table A.2  Household Subsistence Poverty by Dzongkhag Distribution of Distribution of Dzongkhag Poverty rate Standard error the Poor Households Bumthang 0.0 0.0 0.0 3,836 Chhukha 0.1 0.1 1.7 14,865 Phuentsholing Thromde 0.0 0.0 0.0 5,125 Other than Phuentsholing Thromde 0.3 0.3 1.7 9,740 Dagana 7.0 2.1 25.1 5,974 Gasa 0.6 0.6 0.3 873 Haa 0.5 0.5 0.9 2,752 Lhuentse 0.9 0.7 2.0 3,754 Monggar 3.2 0.9 17.5 9,049 Paro 0.0 0.0 0.0 8,969 Pema Gatshel 1.1 0.6 4.2 6,536 Punakha 0.4 0.4 1.5 6,450 Samdrup Jongkhar 1.4 0.5 6.9 8,502 Samdrup Jongkhar Thromde 0.3 0.2 0.2 2,379 Other than Samdrup Jongkhar Thromde 2.5 1.0 6.4 6,123 Samtse 1.0 0.4 9.0 14,503 Sarpang 1.3 0.4 7.9 10,537 Gelephu Thromde 0.0 0.0 0.0 2,506 Other than Gelephu Thromde 2.5 0.7 7.9 8,031 Thimphu 0.0 0.0 0.0 30,147 Thimphu Thromde 0.0 0.0 0.0 24,266 Other than Thimphu Thromde 0.0 0.0 0.0 5,882 Trashigang 1.1 0.6 7.6 11,228 Trashi Yangtse 1.0 0.5 2.5 4,228 Trongsa 2.5 1.1 5.7 3,899 Tsirang 0.2 0.2 0.6 5,074 Wangdue Phodrang 0.2 0.2 1.0 8,847 Zhemgang 2.7 1.1 6.4 3,988 Bhutan 1.5 0.2 100.0 164,011 36 Annex I: Additional Statistical Tables Table A.3  Population Poverty Gap and Poverty Squared Gap by Dzongkhag Standard Contribution Standard Contribution Distribution of Dzongkhag Index error to total Index error to total Population Bumthang 0.2 0.1 0.3 0.0 0.0 0.2 15,959 Chhukha 0.7 0.2 3.7 0.2 0.1 3.0 63,355 Phuentsholing Thromde 0.1 0.1 0.2 0.0 0.0 0.1 20,560 Other than Phuentsholing 0.9 0.2 3.5 0.2 0.1 2.8 42,795 Thromde Dagana 9.1 2.1 18.7 3.2 0.9 21.5 23,453 Gasa 1.7 0.7 0.5 0.3 0.2 0.4 3,575 Haa 0.2 0.2 0.2 0.1 0.1 0.2 10,995 Lhuentse 1.4 0.7 1.9 0.5 0.4 2.4 15,552 Monggar 3.6 0.7 13.2 1.0 0.2 12.6 41,956 Paro 0.0 0.0 0.1 0.0 0.0 0.1 36,329 Pema Gatshel 2.3 0.8 5.7 0.6 0.3 5.1 27,636 Punakha 0.4 0.2 0.9 0.1 0.0 0.7 26,724 Samdrup Jongkhar 1.4 0.4 4.4 0.5 0.1 4.8 36,154 Samdrup Jongkhar 0.2 0.1 0.1 0.1 0.1 0.3 9,376 Thromde Other than Samdrup 1.8 0.5 4.3 0.6 0.2 4.5 26,778 Jongkhar Thromde Samtse 2.2 0.4 12.1 0.7 0.2 12.4 63,132 Sarpang 2.3 0.4 8.5 0.6 0.1 7.6 41,254 Gelephu Thromde 0.2 0.1 0.1 0.0 0.0 0.1 8,015 Other than Gelephu 2.9 0.5 8.4 0.8 0.2 7.5 33,238 Thromde Thimphu 0.1 0.0 0.9 0.0 0.0 0.5 125,551 Thimphu Thromde 0.1 0.0 0.6 0.0 0.0 0.5 98,148 Other than Thimphu 0.1 0.1 0.2 0.0 0.0 0.1 27,403 Thromde Trashigang 2.2 0.7 9.0 0.7 0.4 10.2 47,102 Trashi Yangtse 1.9 0.4 2.6 0.5 0.2 2.4 15,363 Trongsa 3.4 0.8 5.3 1.1 0.3 5.9 17,768 Tsirang 0.6 0.3 1.1 0.2 0.1 1.0 20,409 Wangdue Phodrang 0.8 0.3 2.9 0.2 0.1 2.1 41,405 Zhemgang 4.7 1.0 8.0 1.3 0.3 7.2 19,224 Bhutan 1.6 0.1 100 0.5 0.1 100 692,895 37 Bhutan Poverty Analysis Report 2017 Table A.4  Population Subsistence Poverty Gap and Subsistence Poverty Squared Gap by Dzongkhag Standard Contribution Standard Contribution Distribution of Dzongkhag Index error to total Index error to total Population Bumthang 0.0 0.0 0.0 0.0 0.0 0.0 15,959 Chhukha 0.0 0.0 0.7 0.0 0.0 0.2 63,355 Phuentsholing Thromde 0.0 0.0 0.0 0.0 0.0 0.0 20,560 Other than Phuentsholing 0.0 0.0 0.7 0.0 0.0 0.2 42,795 Thromde Dagana 1.7 0.7 24.1 0.4 0.2 21.8 23,453 Gasa 0.0 0.0 0.0 0.0 0.0 0.0 3,575 Haa 0.1 0.1 0.4 0.0 0.0 0.2 10,995 Lhuentse 0.4 0.4 4.2 0.1 0.1 4.8 15,552 Monggar 0.4 0.2 10.4 0.1 0.1 8.9 41,956 Paro 0.0 0.0 0.0 0.0 0.0 0.0 36,329 Pema Gatshel 0.3 0.2 4.6 0.0 0.0 3.0 27,636 Punakha 0.0 0.0 0.6 0.0 0.0 1.0 26,724 Samdrup Jongkhar 0.2 0.1 5.5 0.1 0.0 4.5 36,154 Samdrup Jongkhar Thromde 0.1 0.1 0.6 0.0 0.0 1.0 9,376 Other than Samdrup 0.3 0.1 4.9 0.1 0.0 3.5 26,778 Jongkhar Thromde Samtse 0.4 0.2 15.4 0.1 0.1 16.7 63,132 Sarpang 0.2 0.1 5.5 0.0 0.0 3.0 41,254 Gelephu Thromde 0.0 0.0 0.0 0.0 0.0 0.0 8,015 Other than Gelephu 0.3 0.1 5.5 0.0 0.0 3.0 33,238 Thromde Thimphu 0.0 0.0 0.0 0.0 0.0 0.0 125,551 Thimphu Thromde 0.0 0.0 0.0 0.0 0.0 0.0 98,148 Other than Thimphu 0.0 0.0 0.0 0.0 0.0 0.0 27,403 Thromde Trashigang 0.4 0.3 11.8 0.2 0.2 23.1 47,102 Trashi Yangtse 0.3 0.1 2.5 0.1 0.1 3.4 15,363 Trongsa 0.7 0.3 7.6 0.1 0.1 5.7 17,768 Tsirang 0.1 0.1 1.2 0.0 0.0 1.2 20,409 Wangdue Phodrang 0.0 0.0 0.2 0.0 0.0 0.0 41,405 Zhemgang 0.4 0.2 5.2 0.1 0.0 2.4 19,224 Bhutan 0.2 0.0 100.0 0.1 0.0 100.0 692,895 Table A.5  Population Poverty Gap and Poverty Squared Gap by Area Poverty Gap Poverty Squared Gap Standard Contribution Standard Contribution Distribution of Area Index error to Total Index error to Total Population Urban 0.2 0.1 3.1 0.0 0.0 2.6 231,805 Rural 2.4 0.2 96.9 0.7 0.1 97.4 461,090 Bhutan 1.6 0.1 100.0 0.5 0.1 100.0 692,895 38 Annex I: Additional Statistical Tables Table A.6  Population Subsistence Poverty Gap and Subsistence Poverty Squared Gap by Area Poverty Gap Poverty Squared Gap Standard Contribution Standard Contribution Distribution of Area Index error to Total Index error to Total Population Urban 0.0 0.0 0.6 0.0 0.0 1.0 231,805 Rural 0.4 0.1 99.4 0.1 0.0 99.0 461,090 Bhutan 0.2 0.0 100.0 0.1 0.0 100.0 692,895 Table A.7  Household Poverty Rate, Poverty Gap and Poverty Squared Gap by Area and Sex of Household Head Area/Sex of Poverty Rate Poverty Gap Poverty Squared Gap Household Contribution Contribution Contribution Distribution of Head Index to Total Index to Total Index to Total Househods Urban 0.5 100.0 0.1 100.0 0.0 100.0 58,333 Male 0.6 82.0 0.1 76.2 0.0 63.7 41,373 Female 0.3 18.0 0.1 23.8 0.0 36.3 16,960 Rural 8.7 100.0 1.7 100.0 0.5 100.0 105,678 Male 9.4 66.4 1.8 65.6 0.6 66.4 64,691 Female 7.5 33.6 1.5 34.4 0.4 33.6 40,987 Bhutan 5.7 100.0 1.1 100.0 0.3 100.0 164,011 Male 5.9 66.8 1.1 66.0 0.4 66.4 106,064 Female 5.4 33.2 1.1 34.0 0.3 33.6 57,947 Table A.8  Household Poverty Rate Poverty Gap and Poverty Squared Gap by Area and Age of Household Head Area/Age of Poverty Rate Poverty Gap Poverty Squared Gap Household Contribution Contribution Contribution Distribution of Head Index to Total Index to Total Index to Total Househods Urban 0.5 100.0 0.1 100.0 0.0 100.0 58,333 < 25 0.0 0.0 0.0 0.0 0.0 0.0 2,802 25-34 0.3 18.5 0.0 14.5 0.0 9.8 20,805 35-44 0.4 22.9 0.1 30.0 0.0 28.2 16,071 45-54 0.6 26.3 0.1 17.2 0.0 11.8 11,502 55-64 1.6 25.3 0.3 24.1 0.1 21.6 4,535 65 + 0.7 6.9 0.3 14.3 0.2 28.7 2,618 Rural 8.7 100.0 1.7 100.0 0.5 100.0 105,678 < 25 4.9 1.0 0.6 0.6 0.1 0.4 1,821 25-34 6.5 11.8 1.3 11.7 0.4 11.6 16,503 35-44 8.5 22.2 1.6 21.7 0.5 20.2 23,723 45-54 8.2 21.0 1.6 21.0 0.5 20.0 23,567 55-64 8.7 21.0 1.7 20.8 0.5 22.2 22,214 65 + 11.8 23.0 2.4 24.2 0.8 25.6 17,851 Bhutan 5.7 100.0 1.1 100.0 0.3 100.0 164,011 < 25 1.9 0.9 0.2 0.6 0.0 0.4 4,623 25-34 3.0 12.0 0.6 11.8 0.2 11.5 37,308 35-44 5.3 22.2 1.0 21.9 0.3 20.4 39,793 45-54 5.7 21.2 1.1 20.9 0.3 19.8 35,069 55-64 7.5 21.2 1.4 20.9 0.5 22.2 26,749 65 + 10.4 22.5 2.2 23.9 0.7 25.7 20,469 39 Bhutan Poverty Analysis Report 2017 Table A.9  Household Poverty and Subsistence Poverty Rate by area and Household Size Poverty Rate Subsistence Rate Area/ Contribution to Contribution to Distribution of Household Size Index National Index National Househods Urban 0.5 3.0 0.0 1.3 58,333 1 0.0 0.0 0.0 0.0 3,814 2-3 0.2 0.4 0.1 0.7 19,417 3-4 0.4 1.0 0.0 0.6 26,235 5-8 1.1 1.0 0.0 0.0 8,034 9+ 7.3 0.6 0.0 0.0 833 Rural 8.7 97.0 1.6 98.7 105,678 1 1.4 0.8 0.8 2.7 5,729 2-3 2.4 8.1 0.3 6.3 32,043 3-4 8.5 36.4 1.2 29.6 40,524 5-8 15.5 39.0 3.1 43.9 23,752 9+ 33.0 12.7 8.2 16.3 3,630 Bhutan 5.7 100.0 1.0 100.0 164,011 1 0.8 0.8 0.5 2.7 9,543 2-3 1.6 8.5 0.2 7.0 51,460 3-4 5.3 37.4 0.7 30.1 66,759 5-8 11.8 39.9 2.3 43.9 31,786 9+ 28.2 13.4 6.7 16.3 4,463 40 Annex I: Additional Statistical Tables Table A.10  Population Literacy Rate for Aged Six Table A.11  Proportion of Women (15-49 years) Years and Above by Dzongkhag and Poverty Status Who Received Antenatal Care by Dzongkhag and Poverty Status Dzongkhag Poor Non-poor Total Bumthang 89.4 85.2 85.3 Dzongkhag Poor Non-poor Total Chhukha 50.1 66.9 66.3 Bumthang 0.0 83.5 83.5 Phuentsholing Chhukha 100.0 88.7 89.1 72.5 83.3 83.2 Thromde Phuentsholing 0.0 93.1 93.1 Other than Thromde Phuentsholing 47.9 58.8 58.3 Other than Thromde Phuentsholing 100.0 84.0 85.2 Dagana 65.1 66.1 65.8 Thromde Gasa 65.9 63.9 64.1 Dagana 100.0 70.8 74.5 Haa 42.9 59.8 59.6 Gasa 100.0 100.0 100.0 Lhuentse 0.0 62.9 62.6 Haa 0.0 92.0 92.0 Monggar 52.4 63.2 61.4 Lhuentse 0.0 84.9 84.9 Paro 60.0 63.0 63.0 Monggar 80.0 98.1 95.3 Pema Gatshel 60.7 65.4 64.8 Paro 0.0 95.0 92.6 Punakha 41.2 61.4 60.8 Pema Gatshel 50.0 72.2 67.6 Samdrup Punakha 0.0 92.5 92.5 58.4 68.8 68.1 Jongkhar Samdrup 100.0 97.8 97.8 Samdrup Jongkhar Jongkhar 0.0 78.7 78.5 Thromde Samdrup Jongkhar 0.0 91.7 91.7 Other than Thromde Samdrup Jongkhar 59.2 65.1 64.6 Other than Thromde Samdrup Jongkhar 100.0 100.0 100.0 Samtse 54.0 62.1 61.1 Thromde Sarpang 58.2 69.5 68.1 Samtse 100.0 100.0 100.0 Gelephu Thromde 60.0 85.7 85.4 Sarpang 76.9 96.9 95.1 Other than Gelephu Thromde 0.0 94.8 86.5 58.1 64.9 63.9 Gelephu Thromde Other than 100.0 97.5 97.7 Thimphu 74.7 80.2 80.2 Gelephu Thromde Thimphu Thromde 75.9 84.2 84.2 Thimphu 0.0 95.3 95.3 Other than Thimphu Thromde 0.0 96.6 96.6 73.1 66.3 66.4 Thimphu Thromde Other than 0.0 84.0 84.0 Trashigang 49.1 58.4 57.5 Thimphu Thromde Trashi Yangtse 50.3 60.3 59.1 Trashigang 75.0 95.6 93.9 Trongsa 60.4 66.3 65.5 Trashi Yangtse 100.0 100.0 100.0 Tsirang 55.4 63.2 62.8 Trongsa 100.0 100.0 100.0 Wangdue Tsirang 0.0 79.0 79.0 47.1 45.7 45.7 Phodrang Wangdue 100.0 60.6 66.2 Zhemgang 61.1 73.7 70.5 Phodrang Bhutan 56.8 66.8 66.0 Zhemgang 0.0 83.8 59.7 Bhutan 74.7 91.9 90.9 41 Bhutan Poverty Analysis Report 2017 Table A.12  Proportion of Population Who Table A.13  Proportion of Population with Access Reported Sick/Injured Four Weeks Prior to the to Improved Water Source by Dzongkhag and Survey by Dzongkhag and Poverty Status Poverty Status Dzonkhag Poor Non-poor Total Dzongkhag Poor Non-Poor Total Bumthang 2.3 15.4 15.2 Bumthang 100.0 99.8 99.8 Chhukha 3.1 7.8 7.6 Chhukha 100.0 99.5 99.5 Phuentsholing Phuentsholing 7.0 14.4 14.4 100.0 100.0 100.0 Thromde Thromde Other than Other than Phuentsholing 2.7 4.5 4.4 Phuentsholing 100.0 99.3 99.3 Thromde Thromde Dagana 3.2 3.1 3.1 Dagana 100.0 100.0 100.0 Gasa 38.5 18.3 20.9 Gasa 100.0 100.0 100.0 Haa 0.0 7.5 7.5 Haa 50.0 100.0 99.5 Lhuntse 16.3 10.8 11.2 Lhuentse 100.0 100.0 100.0 Monggar 27.3 20.8 21.9 Monggar 98.2 99.7 99.5 Paro 16.7 13.1 13.1 Paro 100.0 99.6 99.6 Pema Gatshel 5.2 11.2 10.4 Pema Gatshel 100.0 100.0 100.0 Punakha 16.4 12.2 12.3 Punakha 100.0 99.3 99.3 Samdrup Jongkhar 9.2 8.1 8.2 Samdrup 100.0 99.3 99.3 Samdrup Jongkhar Jongkhar 0.0 4.6 4.5 Thromde Samdrup Jongkhar 100.0 99.0 99.0 Other than Samdrup Thromde 9.3 9.4 9.4 Jongkhar Thromde Other than Samtse 9.7 15.2 14.6 Samdrup Jongkhar 100.0 99.4 99.4 Thromde Sarpang 8.3 12.4 11.9 Samtse 100.0 98.7 98.8 Gelephu Thromde 0.0 6.9 6.8 Sarpang 98.0 99.7 99.6 Other than Gelephu 8.5 14.0 13.1 Gelephu Thromde 100.0 100.0 100.0 Thromde Thimphu 13.9 14.0 14.0 Other than 97.9 99.6 99.5 Gelephu Thromde Thimphu Thromde 18.7 15.4 15.4 Thimphu 100.0 99.5 99.5 Other than Thimphu 7.4 9.0 9.0 Thimphu Thromde 100.0 99.4 99.4 Thromde Trashigang 13.8 10.9 11.2 Other than 100.0 99.8 99.8 Thimphu Thromde Trashi Yangtse 10.8 11.5 11.4 Trashigang 100.0 99.6 99.6 Trongsa 16.5 12.1 12.7 Trashi Yangtse 100.0 100.0 100.0 Tsirang 6.3 12.1 11.8 Trongsa 100.0 98.8 98.9 Wangdue Phodrang 5.9 12.0 11.7 Tsirang 100.0 99.2 99.2 Zhemgang 0.3 4.0 3.1 Wangdue 100.0 100.0 100.0 Total 10.5 12.1 12.0 Phodrang Zhemgang 100.0 99.2 99.3 Bhutan 99.4 99.5 99.5 42 Annex I: Additional Statistical Tables Table A.14  Proportion of Population with Access Table A.15  Proportion of Population using Solid to Improved Sanitation by Dzongkhag and Fuels by Dzongkhag and Poverty Status Poverty Status Dzongkhag Poor Non-Poor Total Dzongkhag Poor Non-Poor Total Bumthang 11.6 7.8 7.9 Bumthang 57.9 92.8 92.3 Chhukha 59.4 24.4 25.2 Chhukha 74.5 91.7 91.3 Phuentsholing 0.0 0.6 0.6 Phuentsholing Thromde 100.0 100.0 100.0 Thromde Other than Other than Phuentsholing 66.7 37.2 38.1 Phuentsholing 71.4 87.3 86.8 Thromde Thromde Dagana 47.3 11.5 19.9 Dagana 97.8 89.1 91.1 Gasa 53.8 35.5 36.9 Gasa 100.0 76.2 77.9 Haa 100.0 17.0 17.8 Haa 100.0 78.7 78.9 Lhuentse 58.8 25.6 27.4 Lhuentse 88.2 92.2 92.0 Monggar 51.0 22.3 26.3 Monggar 96.4 99.5 99.0 Paro 0.0 1.9 1.9 Paro 100.0 88.5 88.5 Pema Gatshel 32.0 27.2 27.7 Pema Gatshel 95.7 96.4 96.3 Punakha 44.4 7.8 8.5 Punakha 0.0 73.9 72.6 Samdrup 79.6 36.7 38.6 Samdrup Jongkhar 91.0 95.6 95.4 Jongkhar Samdrup Jongkhar 100.0 0.0 0.5 Samdrup Jongkhar Thromde 100.0 91.0 91.0 Thromde Other than Other than Samdrup Jongkhar 78.9 51.8 53.4 Samdrup Jongkhar 90.7 97.4 97.0 Thromde Thromde Samtse 94.6 49.4 53.2 Samtse 91.9 98.1 97.6 Sarpang 57.0 20.0 23.1 Sarpang 72.0 92.2 90.5 Gelephu Thromde 0.0 0.0 0.0 Gelephu Thromde 100.0 98.6 98.6 Other than 58.8 26.9 30.3 Other than Gelephu Thromde 71.1 90.0 88.0 Gelephu Thromde Thimphu 20.6 3.3 3.3 Thimphu 89.7 97.0 97.0 Thimphu Thromde 17.0 0.1 0.2 Thimphu Thromde 81.4 98.0 98.0 Other than 25.0 16.4 16.4 Other than Thimphu Thromde 100.0 93.0 93.0 Thimphu Thromde Trashigang 72.3 37.8 40.5 Trashigang 85.3 89.0 88.7 Trashi Yangtse 24.8 19.1 19.6 Trashi Yangtse 76.7 85.1 84.4 Trongsa 25.7 9.7 11.2 Trongsa 64.9 78.4 77.1 Tsirang 84.6 28.9 30.4 Tsirang 61.5 89.1 88.4 Wangdue 62.5 14.6 16.0 Wangdue Phodrang 93.8 82.8 83.1 Phodrang Zhemgang 8.2 14.0 13.1 Zhemgang 100.0 95.9 96.6 Bhutan 54.2 19.8 21.8 Bhutan 87.7 91.7 91.5 43 Bhutan Poverty Analysis Report 2017 Table A.16  Proportion of Households Who Have TV by Dzongkhag and Poverty Status Dzongkhag Poor Non-Poor Total Bumthang 53.7 86.1 85.6 Chhukha 49.1 77.3 76.7 Phuentsholing 100.0 92.8 92.8 Thromde Other than Phuentsholing 42.9 69.0 68.2 Thromde Dagana 45.1 61.4 57.5 Gasa 76.9 61.3 62.4 Haa 50.0 81.2 80.9 Lhuentse 58.8 69.1 68.5 Monggar 27.7 61.0 56.3 Paro 0.0 91.0 90.8 Pema Gatshel 47.1 63.9 62.2 Punakha 0.0 82.7 81.3 Samdrup 22.6 66.4 64.4 Jongkhar Samdrup Jongkhar 0.0 93.9 93.4 Thromde Other than Samdrup Jongkhar 23.4 55.0 53.1 Thromde Samtse 25.7 66.4 62.9 Sarpang 41.6 74.7 71.9 Gelephu Thromde 20.0 91.9 91.2 Other than 42.3 68.7 65.9 Gelephu Thromde Thimphu 60.6 91.9 91.8 Thimphu Thromde 69.1 95.4 95.4 Other than 50.0 77.3 77.1 Thimphu Thromde Trashigang 29.9 69.4 66.4 Trashi Yangtse 51.9 67.3 65.9 Trongsa 51.4 71.0 69.1 Tsirang 53.8 72.1 71.6 Wangdue 43.8 78.9 77.9 Phodrang Zhemgang 48.7 64.1 61.6 Bhutan 38.8 76.4 74.3 44 Annex II: Technical Notes Annex II: Technical Notes Technical Note 1 (Measuring or indeed to separate business transactions Aggregate Consumption) from consumption transactions. Aggregations of consumption and expen- b) Food consumption diture data were made following the Households consume food obtained from recommendations by A. Deaton and S. a variety of different sources, and so in Zaidi (2002). Most of the information computing a measure of total food con- below is quoted from their paper. sumption to include as part of an aggregate welfare measure, it is important to include a) Income versus consumption food consumed by the household from all In most industrialized countries, living possible sources. In particular, this measure standards and poverty are assessed with should include not just (i) food purchased reference to income, not consumption. in the market place, including meals pur- The empirical literature on the relationship chased away from home for consumption at between income and consumption has or away from home, but also (ii) food that is established, for both rich and poor coun- home-produced, (iii) food items received as tries, that consumption is smoother and gifts or remittances from other households, less-variable than income. Observing con- as well as (iv) food received from employers sumption over a relatively short period, even as payment in-kind for services rendered. a week or two, will tell us a great deal more BLSS 2017 food consumption mod- about annual–or even longer period–living ule questionnaire contains separate sets of standards than will a similar observation on questions on: (a) purchased imported; (b) income. Although consumption has seasonal purchased domestic; and (c) non-purchased components they are of smaller amplitude food items. BLSS 2017 food purchases than seasonal fluctuations in income in agri- module contains questions on purchases cultural societies. of a fairly comprehensive list of food items There are several other reasons why during a relatively short reference period, it is more practical to gather consumption i.e., last seven days, last 30 days, and last rather than income data. Where self-em- 12 months. Data are collected on the total ployment, including small business and amount spent on purchasing each food agriculture, is common, it is notoriously item, and also on the quantities purchased, difficult to gather accurate income data, during the specified recall period. 45 Bhutan Poverty Analysis Report 2017 Calculating the food purchases of consumption of each home-produced sub-aggregate involved converting all food item. The home-production food reported expenditures on food items to a sub-aggregate can thus be calculated by uniform reference period—one month— adding the reported value of consumption and then aggregating these expenditures of each of the home-produced food items across all food items purchased by the in a manner analogous to that followed in household. the case of food purchases. The ‘last 30 days’ data measure over Consumption of food derived from the ‘last 7 days’ or the ‘last 12 months’ payment in-kind, as well as in the form has the advantage of being closer to the of gifts, remittances, etc., is added to the concept that we want—usual consump- overall food aggregate. All quantities are tion—over what actually happened in the reported in standard units. Analysis is per- last 7 days, which could have been unusual formed on the quantities and unit prices to for any number of reasons—and reduces treat missing data and identify inconsistent problems of seasonality, but suffers from data. Cases are noted where a household measurement error if respondents find it had declared consuming a non-zero quan- difficult to calculate a reasonable answer. tity of a particular item, or households The last ‘12 months’ may be too long a reported consumption values, but no recall period to reveal accurate data. Thus, corresponding information on quantities. we prefer the ‘last 30 days’ data. If there Others had inconsistent data on quan- are no available ‘30 days’ data, we use the tities, or values (yielding outliers of unit ‘last 7 days’ data and rescale the results. If prices). In such instances, median regional there are no available ‘30 days’ or ‘last 7 unit prices are used to make imputations. days,’ we use the ‘last 12 months’ data and Median prices are preferred to mean prices, rescale the results. as they are less sensitive to outliers. When BLSS 2017 questionnaire also asks median price is not available at the lowest explicitly about the total value of meals geographic level, we use prices reported by taken outside the home by all household other households in the same Dzongkhag, members; this amount is included in the depending on whichever is the next higher food consumption aggregate as part of level of aggregation for which price infor- purchased consumption. mation is available. Medians of unit price The questionnaire contains a sep- are computed and used separately for pur- arate set of questions on consumption chased and home-produced items. of home-produced food items. Data are collected on both the value and quantity 46 Annex II: Technical Notes c) Non-food consumption households rent their dwellings, the rent Unlike many homogeneous food items, paid is the obvious choice to include in the most non-food goods are too hetero- consumption aggregate. Whenever such geneous to permit the collection of rental data are available, they were used information on quantities consumed, so for constructing the housing sub-aggregate BLSS 2017 collected data only on the value and the consumption total. of non-foods purchased over the reference In most cases, however, households period. Data on purchases of non-food own the dwelling in which they reside and items are collected for two different recall do not pay rent as such. Others are pro- periods, i.e., over the 12 months, or the last vided with housing free of charge (or at 1month, depending on how frequently the subsidized rates) by their employer, a friend, items concerned are typically purchased. a relative, government, or other such enti- Constructing the non-food aggregate ties. Non-renter households are asked how thus entails converting all these reported much it would cost them if they had to rent amounts to a uniform reference period, the dwelling in which they reside, and this one year, and then aggregating across the ‘implicit rental value’ is used in place of various items. actual rent. Not all non-food expenditures 2) Taxes are included in the consumption aggre- Expenditures on taxes and levies are not gates. Also, some ‘expenditures’ require part of consumption, and are not included imputations. in the consumption total. 1) Housing 3) Repayment of debt and interest payments What is required is a measure in monetary All purchases of financial assets, as well terms of the flow of services that the house- as repayments of debt, and interest pay- hold receives from occupying its dwelling. ments are excluded from the consumption Because house purchase is such a large aggregate. and relatively rare expenditure, under no 4) Education circumstances should expenditures for a Education expenditure paid by the housing purchase be included in the con- households is included in households’ sumption aggregate. consumption. Expenditure on house repairs and 5) Health improvements were also excluded from the Expenditure on health is to a large extent consumption aggregates. a lumpy expenditure. One argument for In the hypothetical case where exclusion is that such expenditure reflects rental markets function perfectly and all a regrettable necessity that does nothing 47 Bhutan Poverty Analysis Report 2017 to increase welfare. By including health 6) Remittances expenditures for someone who has fallen Another group of expenditures are sick, we register an increase in welfare charitable contributions, and remittances when, in fact, the opposite has occurred. to other households. Their inclusion in the The fundamental problem here is our consumption aggregate would involve dou- inability to measure the loss of welfare ble-counting if, as one would expect, the associated with being sick, and which is transfers show up in the consumption of (presumably) ameliorated to some extent other households. We therefore excluded by health expenditures. them from household consumption. Including the latter without allowing 7) Other lumpy expenditures for the former is clearly incorrect, though While almost all households incur excluding health expenditures altogether relatively large expenditures on relatively means that we miss the difference between infrequent expenditures such as marriages two people, both of whom are sick, but and dowries, births, and funerals at some only one of which pays for treatment. It is stage, only a relatively small proportion also true that some health expenditures— of households are likely to make such for example cosmetic expenditures—are expenditures during the reference period discretionary and welfare enhancing, and typically covered by the survey. Ideally, that it is difficult to separate ‘necessary’ we would want to “smooth” these lumpy from ‘unnecessary’ expenditures, even if expenditures, spreading them over several we could agree on which is which. It is also years, but lacking the information to do difficult without special health question- so—which might come, for example, by naires to get at the whole picture of health incorporating multi-year reference periods financing. Some people have insurance, so for such items— we left them out of the that expenditures are only ‘out of pocket’ consumption aggregate. expenditures which may be only a small 8) Durable Goods fraction of the total, while others have Another important group of items none, and may bear the whole cost. Simply to consider are items such as consumer adding up expenditures will not give the durables whose useful life typically spans right answer. a time-period greater than the interval for Expenditure on hospitalizations, con- which the consumption aggregate is being sultations, and analyses are excluded from constructed. From the point of view of the household consumption. Purchase of household welfare, rather than using expen- medicine is, however, included. diture on the purchase of durable goods during the recall period, the appropriate 48 Annex II: Technical Notes measure of consumption of durable goods itself and, therefore, differ from one house- is the value of services that the household hold to another. In other words, these receives from all the durable goods in its indexes involve, not only the prices faced possession over the relevant time period. by household h in relation to the reference prices, but also household h’s expenditure d) Computing regional price deflators pattern, something that is not true of a Before our measure of consumption could Laspeyres index. 0 be used to compare standards of living The reference price vector p was of individuals residing in different parts inevitably selected as a matter of conve- of the country, it is necessary to take into nience. To ensure that the vector is not very account differences in cost of living. To different from prices actually observed, we convert total expenditure into money met- chose to take the median of the prices ric utility, the price index must be tailored observed from individual households as to the household’s own demand pattern, a reference. The use of the national median demand pattern that varies with the house- price vector ensures that the money metric hold’s income, demographic composition, measures conform as closely as possible location, and other characteristics. The to national income accounting practice, calculation of money metric utility thus as well as eliminating results that might requires that the nominal values be deflated depend on a price relative that occurs only by a Paasche price index, in which the rarely or in some particular area. weights vary from household to household. Quantities and unit values were avail- Data collected by the BLSS 2007 able at the household level only for foods were used to construct the regional price items. For non-foods, data is not available deflators. at the household level. The Paachse price The Paasche price index for house- indices were thus computed for food items hold h is given by: only. Pp h = (�wkh ( pk 0 / pkh  ))–1 Technical Note 2 (Food Poverty Line) The Food Poverty line for 2017 is updated 0 where p k isthe reference unit price from 2012 using the food inflation between h for good k, p k isthe unit price paid for 2012 and 2017. BLSS 2007 collected data h good k by household h, and wk is the share on 118 different food items. Consumption of household h’s budget devoted to good data is available in standard quantity units k. The weights used for the price index are for all these items. For 94 of them, calories the quantities consumed by the household intake data is available, and of these items, 49 Bhutan Poverty Analysis Report 2017 53 items are used to create a reference of each item in the food basket are estab- food basket. These items are used to com- lished by considering the consumption pute the food poverty line since the most pattern of the reference population. The frequently consumed food items by the quantities are scaled up in such a way that reference population (i.e., the second to the the resulting basket provides a total of fourth deciles of the nominal per capita 2,124 Kcal. The cost of the basket is calcu- consumption distribution). These 53 goods lated using the national median unit prices accounted for 80% of the food consumed for each item. by the reference population. The quantities Table A.17  Food Bundle and Costs of Nutritionally Adequate Food Bundle Per Person Per Day, 2007 Calories per Daily quantity Daily calories Items Unit units (kcals) consumed (units) provided (kcals) Price per unit Cost Cereals and Pulses 101 Rice Bhutanese Gram 3.5 92.3 319.3 0.0 2.3 102 Rice fine Gram 3.5 59.8 208.8 0.0 0.8 103 Rice FCB Gram 3.5 110.2 381.4 0.0 1.5 104 Processed rice Gram 3.3 9.6 31.2 0.0 0.3 (zaw, sip) 105 Maize (kharang) Gram 3.4 93.0 318.0 0.0 1.0 106 Ata, Maida Gram 3.4 9.8 33.2 0.0 0.2 107 Noodles Gram 3.5 12.1 42.1 0.0 0.5 108 Confectionery Gram 2.5 0.2 0.5 0.3 0.1 109 Biscuits Gram 3.6 4.7 17.0 0.1 0.4 110 Pulses Gram 3.4 11.5 39.3 0.0 0.3 Dairy Products 201 Liquid milk Ml 0.7 19.1 12.8 0.0 0.5 202 Milk powder Gram 5.0 6.5 32.3 0.2 1.1 203 Local butter Gram 7.3 10.4 76.1 0.2 1.6 204 Local cheese Gram 4.7 12.4 58.5 0.1 1.4 205 Egg Gram 1.7 3.7 6.4 0.1 0.3 Meat 301 Fresh fish Gram 0.1 2.3 2.2 0.1 0.2 302 Dried fish Gram 2.6 11.2 28.6 0.1 0.8 303 Fresh beef Gram 1.1 7.2 8.2 0.1 0.4 304 Dried beef Gram 2 1.8 3.5 0.2 0.4 305 Fresh pork Gram 1.1 4.1 4.7 0.1 0.4 306 Chicken Gram 1.1 2.9 3.2 0.1 0.3 50 Annex II: Technical Notes Calories per Daily quantity Daily calories Items Unit units (kcals) consumed (units) provided (kcals) Price per unit Cost Fruits 401 Apple Gram 0.6 0.7 0.4 0.0 0.0 402 Orange Gram 0.5 21.2 10.2 0.0 0.3 403 Mango Gram 0.7 0.5 0.4 0.0 0.0 404 Banana Gram 1.2 18.1 21.0 0.0 0.1 405 Cucumber Gram 0.1 6.0 0.8 0.0 0.1 406 Sugarcane Gram 4.0 2.7 10.7 0.0 0.1 407 Guava Gram 0.5 2.4 1.3 0.0 0.0 408 Walnut Gram 6.9 3.9 26.9 0.0 0.0 409 Other fruits Gram 0.5 0.6 0.3 0.0 0.0 Vegetables 501 Fresh beans Gram 1.6 17.4 27.4 0.0 0.4 502 Tomato Gram 0.2 17.8 4.1 0.0 0.4 503 Spinach Gram 0.3 32.9 8.6 0.0 0.4 504 Cabbage Gram 0.3 20.4 5.5 0.0 0.2 505 Potato Gram 0.1 60.6 58.8 0.0 0.7 506 Pumpkin Gram 0.3 4.4 1.1 0.0 0.0 507 Radish Gram 0.2 26.5 4.5 0.0 0.3 508 Cauliflower Gram 0.3 8.1 2.4 0.0 0.2 509 Brinjal Gram 0.2 5.5 1.3 0.0 0.1 510 Gourd Gram 0.1 2.7 0.3 0.0 0.0 511 Fresh Gram 0.3 2.0 0.5 0.2 0.4 mushroom 512 Fern (damru) Gram 0.3 6.3 2.1 0.0 0.1 513 Mustard oil Ml 9 14.1 127.0 0.1 0.9 514 Dalda oil Ml 9 3.1 27.6 0.1 0.2 515 Refined oil Ml 9 6.6 59.5 0.1 0.4 Spices, Seasonings and Pastes 601 Fresh chili Gram 0.3 21.3 6.2 0.0 0.6 602 Dried chili Gram 2.5 6.2 15.2 0.1 0.6 603 Haldi, Jeera Gram 3.5 0.8 2.9 0.1 0.1 Coriander 604 leaves Gram 0.4 6.2 2.7 0.0 0.2 605 Salt Gram 0 8.8 0 0.0 0.1 607 Sugar/gur Gram 4.0 16 63.7 0.0 0.5 Beverages 701 Beer Ml 0.4 3.9 1.4 0.1 0.2 702 Juice Ml 0.5 4.4 2.0 0.1 0.2 TOTAL PER DAY 2,124 Kcal Nu 22.49 51 Bhutan Poverty Analysis Report 2017 Technical Note 3 (Non Food whose food consumption is nearest the food Adjustment to the Poverty Line) poverty line. We increase the bandwidth to Having set the food poverty line, a non- 2% and recalculate the average non-food food component must be added to obtain per capita expenses, and keep iterating up an overall poverty line that incorporates to a plus or minus 10% bandwidth. Then overall needs. As M. Ravallion and Bidani we take an average of all these mean per (1992, 1999) suggest that the total poverty capita non-food expenditures and use this line is obtained by scaling up the food pov- as our non-food adjustment. In effect, the erty line to allow for the purchase of some resulting non-food adjustment is a weighted essential non-food items to reach a final average of non-food expenses of house- poverty line. The non-food needs must be holds whose food expenses are near the consistent with the consumption behavior food poverty line, with the highest weight of households who can just afford basic on the households whose food spending are food needs. closest to the food poverty line (and with A number of methodologies have weights that decline as the food spending been proposed for making this non-food goes farther from the food poverty line). adjustment, including the use of another Similar to the Food Poverty Line, the basket of non-food items. The best solution Non-food Poverty Line for 2017 is updated is to measure what is the typical value of from 2012 using the non-food inflation non-food spending by a household that is between 2012 and 2017. just able to reach its food requirements. This will equal the lowest level of non-food Technical Note 4 (Poverty Measures) spending for households that are able to Incidence of Poverty ( P0 ) acquire the basic food bundle. It can thus The incidence of poverty is the proportion be considered a minimal allowance for of the population that is poor (percentage non-food goods. of the total population below the poverty What we use here is a non-paramet- line). The percentage of households below ric estimate of the non-food consumption the poverty line may also be computed of households in the reference population (since poor households usually have a whose food consumption is close to the smaller size, the proportion of poor house- food poverty line. First, we calculate the holds is usually lower than the proportion mean per capita non-food expenditures of poor population). of households in the reference population P0 = q / n whose food spending lies within a plus or minus 1% bandwidth of the household wh e re P0 i s t h e p ro p o r t i o n o f 52 Annex II: Technical Notes population deemed to be poor (poverty index multiplied by total population may headcount), q is the number of poor peo- be thought of representing the total cost of ple (below the poverty line), and n is the poverty reduction assuming perfect poverty total population. targeting. The fact that poverty calculations are The poverty gap index can also be based on a sample of households, or a sub- written as set of the population, carries implications. P1   = H * ( z – yp  ) / z Samples are designed to reproduce the whole population, but they can never be as where ( z – yp ) / z is referred to as the exact as information that covers everybody ‘income gap ratio’ (mean depth of poverty in the country. They carry a margin of as a proportion of the poverty line). error, as do poverty rates calculated from The income gap ratio is not a good these sample surveys. When monitoring the poverty measure. To see why, suppose that incidence of poverty over time, it is crucial someone just below the poverty line is made to remember that the figures are based on sufficiently better off to escape poverty. The samples. Instead of considering one figure, mean of the remaining poor will fall, and it is better to use confidence intervals. so the income gap ratio will increase. And yet one of the poor has become better Poverty Gap Index ( P1 ) and Income off, and none are worse off; one would be Gap Ratio loathe to say that there is not less poverty, The poverty incidence alone will not pro- and yet that is what the income gap ratio vide a complete picture of poverty. It is also would suggests. This problem doesn’t arise important to look into the depth of poverty. if the income gap ratio is multiplied by the For one individual, the depth of poverty is head count index to yield P1 . the proportion by which that individual is The poverty gap index doesn’t tell us below the poverty line (it has a value of 0 how the poverty is distributed among indi- for all individuals above the poverty line). viduals; it may not convincingly capture The poverty gap index is the average differences in the severity of poverty. The depth of poverty for the population. This poverty gap will be unaffected by a transfer is the sum of the depth of poverty of each from a poor person to someone who is less individual, divided by the total number of poor. However, when the poverty gap index individuals in the population. This gives a is multiplied by the total population and good indication of the depth of poverty, the result further multiplied to the poverty in that it depends on the distances of the line, we obtain the aggregate gap. This poor below the poverty line. Also, this represents the cost of eliminating poverty 53 Bhutan Poverty Analysis Report 2017 assuming perfecting targeting and no tar- and q is the number of individuals (or geting costs. households) with per capita consumptions below the poverty line. Poverty Squared Gap Index ( P2 ) The Poverty Severity Index ( P2 ) gives a Technical Note 5 (Inequality weight to the poverty gap (more weight to Measures) very poor than to less poor). It is the average a) Gini value of the square of depth of poverty for Graphically, the Gini coefficient can be each individual. Poorest people contribute easily represented by different areas of relatively more to the index. the Lorenz curve, a cumulative frequency While this measure has clear advan- curve that compares the distribution of a tages for some purposes, such as comparing specific variable such as per capita expen- policies, which are aiming to reach the diture with the uniform distribution that poorest, it is not easy to interpret. For pov- represents equality. To construct the Gini erty comparisons, however, the key point is coefficient, graph the cumulative percent- that a ranking of dates, places or policies in terms of P2 should reflect well their ranking in terms of the severity of poverty. It is the ability of the measure to order distributions in a better way than the alternatives that makes it useful, not the precise numbers obtained. The poverty incidence, poverty gap and poverty squared gap measures all belong to a family of measures proposed by Foster, Greer, and Thorbecke (1984). q α Pα   = (1/n) � z – yi     i=1 z age of households (from poor to rich) on where αis some non-negative param- the horizontal axis and the cumulative eter, z is the poverty line, y denotes per percentage of consumption-expenditure capita consumption, i represents indi- on the vertical axis. This gives the Lorenz viduals (or households), n is the total curve as shown below. The diagonal line number of individuals (or households) in represents perfect equality. The Gini coef- the population (or household population), ficient is calculated as the area A divided 54 Annex II: Technical Notes by the sum of areas A and B, where A and quintile dispersion ratio is readily interpre- B are as shown on the graph. If A=0 the table, by expressing the consumption of Gini coefficient becomes 0 which means the top 20% as a multiple of that of those perfect equality, whereas if B=0 the Gini in the poorest quintile. However, it ignores coefficient becomes 1 which means com- information about consumptions in the plete inequality. middle of the consumption distribution, Formally, let xi be a point on the and does not even use information about X-axis, and yi a point on the Y-axis. Then the distribution of consumption within the top and bottom quintiles. Gini  = 1 – �( xi – xi–1 ) ( yi + yi–1). N i=1 c) Palma Ratio It is also a measure of inequality. It is the When there are N equal intervals on the ratio of the average consumption of the X-axis this simplifies to richest 10% of the population divided by the average consumption of the poorest 40%. N Gini  = 1 – 1 �( yi + yi–1). N i=1 The Gini coefficient of inequality varies between 0, or complete equality of expen- ditures, and 1, or complete inequality (one person has all the expenditure, all others have none). b) Quintile Dispersion Ratio A simple measure of inequality is the quintile dispersion ratio, which represents the ratio of the average consumption of the richest 20% of the population divided by the average consumption of the bottom 20%. This ratio can also be calculated for other percentiles (for instance, dividing the average consumption of the richest 5%– the 95th percentile– by that of the poorest 5%– the 5th percentile). The 55