POVERTY PROFILE OF THE FEDERATED STATES OF MICRONESIA BASED ON THE 2013-14 HOUSEHOLD INCOME AND EXPENDITURE SURVEY BASED ON THE 2013-14 HOUSEHOLD INCOME AND EXPENDITURE SURVEY OCTOBER 2017 TECHNICAL REPORT TABLE OF CONTENTS Introduction 6 HIES 2013/14 9 Sample Design 9 Survey Instruments 10 POVERTY MEASUREMENT METHODOLOGY 12 ESTIMATES OF POVERTY INCIDENCE 13 National-Level Outcomes 13 State-wise poverty lines and poverty incidence 14 CHARACTERISTICS OF THE POOR 19 MULTIDIMENSIONAL ASPECTS OF POVERTY 26 ANNEX I: WELFARE INDICATOR 28 Total Household Expenditure 28 Treatment of Errors and Outliers 28 Calorie Standard 30 ANNEX II: FOOD BASKET 31 02 ANNEX III: MEAN CONSUMPTION 32 03 TABLES AND FIGURES Table 1: FSM HIES 2013/14 sample used for 9 Figure 1: Average number of transactions 10 poverty assessment per day per household by state Table 2: Survey components by module 11 Figure 2: FSM and international poverty lines 14 Table 3: National poverty line 13 Figure 3: Cumulative distribution of welfare 15 (deflated to Pohnpei prices) Table 4: National poverty measures 14 Figure 4: State-wise poverty rates 16 Table 5: State-wise poverty lines 15 (with 95% confidence interval bands) Table 6: State-wise poverty measures 15 Figure 5: Distribution of the poor population 16 Table 7: Inequality indicators 17 Figure 6: Food share by quintile across States 19 Table 8: Average household size by State and 22 Figure 7: Non-food share by welfare level 20 poverty status Figure 8: Expenditure share by category 21 Table 9: Poverty and gender 23 and welfare level Table 10: Multidimensional deprivation 27 Figure 9: Household size and expenditure levels 21 indicators, HIES 2005 and 2013 (% of population in households with the following deprivations) Figure 10: Poverty rate by household group 22 with different numbers of children Figure 11: Poverty rates by education 23 Table A1: Expenditure components for the 29 level (individual) FSM HIES 2013/14 poverty assessment Figure 12: Poverty rate by employment type 24 04 Table A2: Estimated rental regression equation 29 05 Figure 13: Education level by working sector 24 Table A3: Reference national food basket 31 Figure 14: Poverty rate by industry 25 Table A4: Mean real annual expenditure 32 per adult-equivalent by category and state ($US) Figure 15: Education level by industry 25 INTRODUCTION This report presents the poverty profile of the Federated States of Micronesia based on the 2013/14 Household Income and Expenditure Survey (HIES) conducted by the government of the Federated States of Micronesia (FSM) over July, 2013 - July, 2014. Two key questions are addressed in this report: • Who are the poor and how are they distributed across the states of FSM? 06 • What are the general characteristics of people 07 living in poverty in FSM? HIES 2013/14 The 2013/14 HIES collected a wide array of was allocated to these strata in proportion to the information on the living conditions of the FSM population in each stratum. Within each stratum, population including household consumption Primary Sampling Units (PSUs) were chosen based expenditures, household demographics, household on Probability Proportional to Size (PPS), and 10 assets, and education and health status of households were randomly chosen from each household members. Technical assistance in the selected PSU. Further details on sampling design design and implementation of the 2013/14 HIES are available in the methodological report.2 was provided by the South Pacific Community. Details on survey design can be found in the HIES While data was collected from a total of 1,664 Report. 1 households, 16 households had to be removed from this assessment due to the absence of SAMPLE DESIGN records on food consumption. The sample size The 2013/14 HIES used the sample frame from used in this assessment is therefore 1,648 the 2010 census and was designed to generate households and sampling weights used in statistically valid estimates at the State level. generating estimates were adjusted accordingly. Each state, except Kosrae, was stratified into three strata based on accessibility to services and facilities, and the sample size for the state Table 1: FSM HIES 2013/14 sample used for poverty assessment State Households Persons 08 Yap 353 1,735 09 Chuuk 572 4,101 Pohnpei 524 3,049 Kosrae 199 1,050 FSM 1,648 9,935 1 Government of FSM Statistics Division (2014). Household Income and Expenditure Survey 2013/14 Main Analysis Report, FSM government. 2 Statistics Division (2014). Household Income and Expenditure Survey 2013/14 Methodological Report, FSM government. SURVEY INSTRUMENTS consumption was captured only in the diaries, while Table 2: Survey components by module Interviews for the household survey were conducted non-food purchases and payments for services were from July 11, 2013 to July 10, 2014. Data was elicited also recorded in the recall modules. NO MODULE LABEL using two types of survey instruments: diaries self- The figure below shows the number of transactions 1 DEMOGRAPHICS Age, sex, and relationship profile of household members recorded by households, and recall-based information recorded in the diaries over the two week period. The collected by interviewers. number of reported transactions typically fell from the Labor-force status of household members start to the end of each week. Educational and health status of household members Diaries Information household members who left the household Over a period of 14 consecutive days, sampled Recall Modules household members directly recorded daily acquisitions The survey contained four recall modules: (i) 2 HOUSEHOLD CHARACTERISTICS Housing characteristics and rental information of goods and services from the market, foods consumed demography, (ii) household expenditures, (iii) individual AND EXPENDITURES Utilities and communication from home production, and gifts given and received. expenditures, and (iv) income earnings. A summary of Household assets Households were required to record quantity, unit, topics included in each module is provided in Table 2. and total monetary amount of each transaction. Food Vehicles and accessories Travel expenses Household services Contributions to special occasions Figure 1: Average number of transactions per day per household by state Provisions of financial support Loans 8 Insurance and taxes Average Daily Diary Transactions 7 6 3 INDIVIDUAL EXPENDITURES Education expenses 5 Health expenses 4 Clothing expenses 3 2 Communication expenses 1 4 INCOME Wages and salaries 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Non-subsistence business Day 10 Agricultural and forestry activities 11 Total Kosrae Handicraft and home-processed foods Yap Chuuk Pohnpei Livestock and aquaculture Source: FSM HIES 2013/14 dataset. Property income, transfer income and other receipts Remittances and other cash gifts POVERTY MEASUREMENT METHODOLOGY ESTIMATES OF POVERTY INCIDENCE NATIONAL-LEVEL OUTCOMES The principal welfare indicator used for this 03 assessment is total consumption expenditure per To compute the initial food poverty line for each adult-equivalent3. A household is considered poor state, the typical food basket of households In FSM, meeting essential caloric needs requires an average of $US1.84 if its total consumption expenditure per adult- belonging to the bottom 40% when ranked by per adult per day; meeting both food and non-food basic needs requires equivalent is below the poverty line. nominal per adult-equivalent total expenditures on average $US4.34 per day. nationally is considered as the reference food Two poverty lines are specified using the Cost of basket. The basket is then scaled up to yield 2,565 At the national level, the average daily cost of food needed to obtain recommended calories for an adult is Basic Needs (CBN) methodology. The food poverty Kcal a day per adult equivalent and priced out for line is the minimum expenditure needed to acquire $US1.84 (Table 3). The national value denotes the weighted average of food poverty lines in the four states. each state using the state’s median unit prices for recommended calorie intake using the food basket the various foods included in the food basket. actually consumed by the poorer groups (bottom Likewise, the total poverty line which includes a non-food allowance is $US4.34 per day at the national level. 40%) in the FSM. The total poverty line is computed 04 A non-food allowance is then added to the food by adding to the food poverty line an allowance for poverty line to obtain the total poverty line. Table 3: National poverty line essential non-food expenditures. This allowance is equal to the average non- Food Poverty Line Total Poverty line Ratio of Food to Food Poverty Line Total Poverty Line food expenditures of households who have food (Daily per adult (Daily per adult Non-food cost in the (Annual per adult (Annual per adult Key steps used in determining poverty status are as expenditures equal to or near the food poverty line. equivalent) equivalent) Total Poverty Line equivalent) equivalent) follows: 01 05 National $1.84 $4.34 42:58 $670.6 $1,583.9 Average With the computed poverty lines, three main sets Household consumption is defined as the total of poverty measures are calculated: (i) poverty annualized value of all food and non-food items headcount index denoting the percentage of As is shown in Figure 2, FSM’s national total poverty line is significantly higher than the World Bank’s $1.9 (2011 consumed, including estimated values of non- population below the poverty lines; (ii) poverty purchased items such as those produced by the PPP) extreme international poverty line, equivalent to about $2.00 in 2013 prices. While the food poverty line is 12 household or received in kind as gifts. gap index to gauge the depth of poverty; and (iii) relatively close to the international extreme poverty line, the national total poverty line is almost 80% higher.5 13 squared poverty gap index to measure the severity 02 of poverty. The food poverty line is anchored to a daily intake of While only one out of 10 people in the FSM lives below the food poverty 2,565 Kcal per adult equivalent.4 line, more than 40 people out of 100 live below the total poverty line. 5 Since the international line is specified on a per capita basis, the adult-equivalent-based FSM poverty lines are adjusted to per capita terms using the 3 A child below the age of 15 is counted as half (0.5) an adult. ratio between the average number of persons per household and average adult-equivalents per household, both at the national level. Thus poverty line 4 The minimum daily requirement per capita is 2,100Kcal. This was adjusted to per adult equivalent basis. See Annex 1 for more details. per capita = poverty line per adult equivalent ($4.34) * (Average adult-equivalent per household (7.03) ÷ Average number of persons per household (8.58). Figure 2: FSM and international poverty lines Table 5: State-wise poverty lines USD (FSM local currency) 4 Food Poverty Line Share of Non-food Cost Total Poverty Line Food Poverty Line Total Poverty Line 3 State (Daily per adult equivalent) in Total Poverty Line (Daily per adult equivalent) (Annual per adult equivalent) (Annual per adult equivalent) 2 2 3.55 1 Yap $2.46 51.95% $5.11 $896.3 $1,865.3 1.5 0 Chuuk $1.72 48.37% $3.33 $628.1 $1,216.6 Food Poverty Line Total Poverty Line Pohnpei $1.80 66.81% $5.41 $655.2 $1,974.0 FSM Poverty Lines $1.9 2011 PPPin LCU Kosrae $1.80 58.92% $4.39 $657.6 $1,600.8 At the national level, about 10% of the people followed by $5.11 in Yap. Both are substantially in FSM spend below what is needed to secure a higher than in Chuuk ($3.33). Table 6: State-wise poverty measures minimal health diet. When both food and non- Food Poverty Line Total Poverty Line Poverty incidence is higher and foods are considered, 41.2% of the population live more severe in Pohnpei and Chuuk Headcount Poverty Gap Squared Poverty Gap Headcount Poverty Gap Squared Poverty Gap below the total poverty line. The poverty gap index, than in Yap and Kosrae. INDEX (Incidence) (Depth) (Severity) (Incidence) (Depth) (Severity) which indicates the extent to which average adult- Yap 10.0% 2.4% 0.9% 39.4% 14.3% 6.9% equivalent expenditures fall short of the poverty Generally, poverty is most severe in Chuuk: With lines, is estimated at 3.6% at the food poverty line more than 16% living below the food poverty Chuuk 16.6% 6.8% 4.1% 45.5% 17.7% 10.2% line, Chuuk primarily drives FSM’s overall food and 15.1% at the total poverty line. Pohnpei 2.6% 0.5% 0.1% 39.2% 13.4% 6.2% poverty rate. Further, Chuuk’s food poverty gap Table 4: National poverty measures Kosrae 0.0% 0.0% 0.0% 21.0% 5.2% 1.6% index (6.8%), and its food poverty severity index Headcount Poverty Gap Squared Poverty (Incidence) (Depth) Gap (Severity) (4.1%) are much higher than elsewhere. No Kosrae household lives below the food poverty line, and Food Poverty Line 9.9% 3.6% 2.1% Figure 3: Cumulative distribution of welfare (deflated to Pohnpei prices) food poverty in Yap and Pohnpei are 10% and 2.6% Total Poverty Line 41.2% 15.1% 7.9% respectively. The fact that poverty is relatively higher STATE-WISE POVERTY LINES AND POVERTY INCIDENCE in Chuuk even when the poverty line is the lowest 100% there implies that incomes are lower in Chuuk 90% Food poverty lines are quite similar than elsewhere. STATE-WISE POVERTY LINES AND 80% across Chuuk, Pohnpei, and Kosrae, Cumulative Percentage of Total POVERTY INCIDENCE Figure 3 shows the overall distribution of welfare 70% but higher in Yap. Pohnpei’s and Yap’s total poverty lines are higher levels by state. The cumulative density curve is 60% 14 than in other states. bounded on the top by Chuuk and at the bottom 50% 15 40% by Kosrae, confirming generally lower levels of 30% Except for Yap, all FSM states have similar food consumption in Chuuk and generally higher levels 20% poverty lines (Table 5), meaning that the expenditure of consumption in Kosrae. In between are Yap and 10% needed to meet the minimum calorie intake is Pohnpei whose density curves overlap substantially. 0% Pohnpei Total Poverty Line ($5.41) almost the same despite differences in food prices. The 95% confidence bands for the state-level 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 That is however not the case when the total poverty poverty estimates (Figure 4) indicate significant Per Adult Equivalent Expenditure (Real, Deflated to Pohnpei Prices) line is considered, as the cost of meeting non-food overlaps between Yap, Chuuk and Pohnpei, but not State goods/services is higher in more urbanized states. Yap Chuuk Pohnpei Kosrae with Kosrae. Poverty rate in Kosrae is significantly The total poverty line is highest in Pohnpei at $ 5.41 lower than the other states (p<0.01). Figure 4: State-wise poverty rates (with 95% confidence interval bands) Table 7 presents estimates of Gini coefficients and other distributional indicators. The Gini coefficient at the national level 60% is 38.6%, and these are similar across states. Also, at the national level, consumption expenditures of the top 10% of 50% the population ac count for nearly 29% of total consumption expenditure in the country. Shares of both the top 10% and 40% the bottom 40% account are similar across states. 30% Table 7: Inequality indicators 20% 10% Gini Share of top 10% of Share of bottom 40% of State Coefficient population in total expenditure population in total expenditure 0% Yap Chuuk Pohnpei Kosrae National Yap 38.8% 27.6% 16.5% Chuuk 38.8% 28.0% 16.6% Pohnpei 37.7% 28.2% 17.4% In terms of actual numbers, the majority of the poor live in Chuuk and Pohnpei. Kosrae 37.1% 29.0% 18.9% FSM 38.6% 28.5% 16.9% The pie charts in Figure 5 illustrate the distribution of the actual number of the poor across states. Chuuk contains nearly half of all the poor in FSM, and the two states of Chuuk and Pohnpei account for around 86% of the total population below the total poverty line. By itself, Chuuk accounts for 79% of the nation’s extreme poor that live below the food poverty line. By contrast, only about 3% of the country’s poor live in Kosrae. Figure 5: Distribution of the poor population Population below Total Poverty Line Population below Food Poverty Line 1,210 | 3% 4,728 | 11% 956 | 9% 1,199 | 12% 16 17 8,103 | 79% 14,463 | 34% 22,169 | 52% Yap Chuuk Pohnpei Kosrae The pattern of inequality is quite similar across states. CHARACTERISTICS OF THE POOR Low-income households mostly spend their resources on food, especially in Yap and Chuuk. As welfare levels rise, food share falls in all states. Except in Pohnpei, the poorest group (ranked by expenditure per adult equivalent) allocates more than half of its total expenditures on food. Food share of the poorest is especially high in Yap, in excess of 70% (Figure 6). However, food shares in total expenditures decline steadily with increases in total expenditures, except in Kosrae. In Yap, Kosrae and Chuuk, food share declines to close to 40% for households in the top expenditure quintile. Food share is consistently lower in all quintiles in Pohnpei than elsewhere, most likely because of the higher cost of basic non-food goods. Figure 6: Food share by quintile across States 80% 70% 60% 50% 40% 18 30% 19 20% 1 2 3 4 5 Yap Chuuk Pohnpei Kosrae There are some differences in expenditure patterns across states and welfare level. Regardless of poverty status, non-food expenditure account for almost 60% of total consumption in Figure 8: Expenditure share by category and welfare level Pohnpei, much higher than in other states (Figure 7) which is consistent with higher non-food poverty line. In contrast, share of non-food expenditure among the poor in Yap is very low at around 30%. The most important non-food expenditure item is rent in all states, and its share is the highest in Pohnpei (Figure ALL-FSM POOR 3.2% 29.3% 4.8% 3.5%2% 8). The second most important item is utility, followed by transportation and cash contribution to other NON-POOR 5.1% 25.0% 6.5% 5.6% 4.7% 5.2% households or school/church/community. Alcohol and tobacco are also important items of expenditures and are higher in Yap and Pohnpei than elsewhere. The non-poor across all states consistently have greater POOR 7.4% 11.8% 2.5% 4% 3.4% YAP shares for transportation and cash contribution compared to the poor. Education and health expenditures NON-POOR 9.4% 13.4% 6.8% 7.4% 2.2% 7.2% account for very small shares, mainly because they are heavily subsidized. POOR 1.5% 31.0% 3.7% 2.9% CHUUK NON-POOR 1.6% 26.4% 6.2% 4.0% 3.2% 1% Figure 7: Non-food share by poverty status POHNPEI POOR 4.5% 32.7% 7.2% 2.9% 5.9% 3.4% NON-POOR 8.4% 28.3% 6.6% 6.9% 7.1% 5.2% 66.3% 70% POOR 1.9% 27.2% 6.5% 4.4% 4.4% 5.4% KOSRAE 59.5% 60% 55.8% NON-POOR 2.1% 19.8% 8.0% 5.5% 5.5% 8.6% 52.4% 52.3% 50.2% 49.0% 47.9% 50% 44.0% Food Alcohol & Tobaco Rent Utility Cash Contribution 40% Communication Transport Education Health Other Non-Food 30.1% 30% 20% 10% Figure 9: Household size and expenditure levels 0% Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor Poor Non-Poor All-FSM Yap Chuuk Pohnpei Kosrae Household size is correlated with poverty. 20 21 The distribution of expenditure by household size clearly indicates that richer households have fewer Total annual exp. per adult-equivalent members (Figure 9). Overall, poor households have four more members than non-poor households (Table 8). Household Size Predicted hhsize 95% CI Table 8: Average household size by State and poverty status Table 9: Poverty and gender Poverty Incidence (%) Percent of Population Percent of Poor Average Household Size Yap Chuuk Pohnpei Kosrae FSM Yap Chuuk Pohnpei Kosrae Total Yap Chuuk Pohnpei Kosrae Total Yap Chuuk Pohnpei Kosrae Total Poor 9.2 12.6 9.6 9.7 11.1 Male headed 37.3 43.3 36.7 18.8 38.9 79.3 82.2 78.7 81.7 80.6 75.1 78.1 73.9 72.9 76.2 households Non-Poor 5.9 7.2 6.7 6.4 6.8 Female headed households 47.3 55.9 48.2 31.3 50.5 20.7 17.8 21.3 18.3 19.4 24.9 21.9 26.1 27.1 23.8 Difference 3.3 5.3 5.3 5.3 4.3 All households 39.4 45.5 39.2 21.0 41.2 100 100 100 100 100 100 100 100 100 100 Average Household Size Average No. of Children under 15 Children in family (%) Yap Chuuk Pohnpei Kosrae Total Yap Chuuk Pohnpei Kosrae Total Yap Chuuk Pohnpei Kosrae Total Poverty rates are higher in households with more children. Male headed 7.4 9.5 7.7 6.7 8.5 2.6 3.5 2.7 2.7 3.1 35.5 37.0 34.5 40.5 36.2 households Female headed 6.3 2.1 33.7 10.3 8.5 8.8 9.0 4.0 2.9 3.3 3.3 38.8 34.0 37.7 36.5 households Average household size in FSM is 8.6 members with about 3.1 members under age 15. While only 19.7% of the population in households with only adults are poor (Figure 10), households with three to four children—a All households 7.2 9.7 7.9 7.1 8.6 2.5 3.6 2.7 2.8 3.1 35.2 37.3 34.3 39.9 36.3 range that includes the national average—have a poverty incidence that is higher than 41%. It is even higher in households that have more children: 68 % of the households with five or more children are poor. Poverty is closely related with education levels. Figure 10: Poverty rate by household group with different numbers of children A strong and inverse relationship between the level of education and poverty is observed. Poverty rates are below 80% 68.0% the national average among those with at least secondary education (Figure 11) while almost half the population with 70% education of just up to the primary level is poor. In the FSM, primary education (8 years) is mandatory and free in public 60% schools, and public secondary schools (4 years), while not mandatory, are still free. 50% 42.8% 41.2% Figure 11: Poverty rates by education level (individual) 40% 29.3% 30% 19.7% 20% 50% 46.9% 10% 45% 41.2% 0% 40% 36.0% No Child 1-2 Children 3-4 Children 5 or more Children Total 35% 30% 25% 22.3% 22 20% 23 15% Female-headed households are poorer than male-headed households 10% in all states. 5% 0% Primary or less Secondary Tertiary Total About one-fifth of the FSM population lives in female-headed households and the incidence of poverty is higher in female-headed households than in male-headed households in all states. Female headed households tend to Note: Includes only population aged 20 and over. be both larger and have more children, except in Yap (Table 9). Poverty rate among workers in the public sector is lower than among workers elsewhere. As can be seen in Figure 12, poverty rate among workers in the public sector (20.9%) is substantially less Figure 14: Poverty rate by industry than workers in other types of jobs, including in the private sector (30.8%). Poverty incidence is the highest for workers involved in subsistence production (46.6%). This pattern is also quite well linked with education levels: more than 60% of workers in the public sector have tertiary level education compared to only 12.6% among workers producing for their own-consumption. (Figure 13). 50% 45.5% 45% 41.2% Figure 12: Poverty rate by employment type 40% 50% 46.6% 44.2% 35% 45% 29.3% 40% 30% 26.2% 35% 30.8% 25% 30% 20% 25% 20.9% 15% 20% 10% 15% 5% 10% 0% 5% 0% Primary Industry Secondary Industry Tertiary Industry Total Public sector Private sector Producing goods for Other own consumption Figure 13: Education level by working sector Figure 15: Education level by industry 12.6% 32.8% 12.9% 61.7% 19.8% 35.3% 31.9% 47.1% 42.4% 10.0% 43.8% 32.5% 41.5% 27.6% 44.8% 38.4% 29.6% 35.2% 36.1% 21.4% 25.3% 24.5% 24 7.6% 36.4% 24.1% 25 17.8% 14.6% Note: Only the population aged between 15 and 65 are considered for Figure 12 and 13. Related, poverty rates are significantly lower in the tertiary and secondary sectors than in the primary sector Note: Only the population aged between 15 and 65 considered for figure 14 and 15. where subsistence agriculture is predominant (Figure 14). Education levels are also lower in the primary sector with only 54% of the workers with at least secondary schooling or higher, compared to 80% or more in both the secondary (construction or utilities) and tertiary sectors (services and professionals) (Figure 15). (Unit %) Households Households Households Households Households Households Households Households Households primary completed working MULTIDIMENSIONAL with one 18.5 42.4 1.9 26.4 41.5 30.2 1.4 0.4 3.9 Chuuk 64.1 46.2 44.6 6.3 61.4 48.9 12.4 0.9 4.3 Kosrae 3.6 1.1 1.3 3.3 5.4 8.4 0.0 0.0 3.2 Kosrae 13.9 0.0 3.4 20.3 20.1 12.5 9.7 0.1 3.3 2005 National 39.1 45.9 25.7 12.2 44.5 32.9 12.6 0.8 2.7 Pohnpei 9.3 5.9 24.4 6.4 33.7 15.2 0.6 0.2 1.0 35.0 79.3 18.4 38.6 52.9 35.7 9.0 0.7 1.0 Chuuk 38.1 29.0 27.9 4.8 41.1 46.6 4.4 0.4 1.6 2013 National 23.6 20.8 22.2 7.8 36.5 31.4 2.4 0.3 1.7 Pohnpei 14.4 43.8 9.4 8.6 25.5 16.1 14.9 0.8 1.1 % % % % % % % % % ASPECTS OF POVERTY with few with no child with no one education that has The cost of basic needs approach captures a access to improved water source declined only Table 10: Multidimensional deprivation indicators, HIES 2005 and 2013 (% of population in households with the following deprivations) attending household’s living conditions by aggregating by 3.5 percentage points, and the proportion of school different components of consumption using their households with few assets remained almost monetary value. However, dimensions of poverty, unchanged. Progress also varies across states for like those related to health, education, shelter, and some indicators. In Yap, 42.4% of the households school assets other basic amenities are not adequately captured lack sanitary facilities in their dwellings, higher in this measure, especially when some of the even than in Chuuk which has a higher poverty services are provided publicly and therefore are not rate. Further, the proportion of households that fully reflected in private consumption expenditure. live in dwellings not made from solid materials with poor cooking Hence, directly monitoring these non-monetary is also higher in Yap than in the other states. fuel indicators offer a dual check on poverty reduction. On the other hand, dwelling quality, including access to electricity, is much better in Kosrae and Table 10 below presents a number of non- with poor sanitation water sources housing Pohnpei than elsewhere. While the proportion quality monetary poverty indicators for which there of households without access to electricity has is comparable data in the 2005 HIES such that declined significantly between 2005 and 2013 progress made in certain dimensions of poverty improved in Chuuk, more than one-third of households with no between 2005 and 2013 can be assessed at the there still lack access to electricity. Enrollment national and state levels. Significant achievement at the primary school stage is generally high in is seen in dimensions such as access to electricity, FSM education almost universal except in Chuuk improvement in sanitation, quality of housing, and improved with no and Pohnpei. access to education. Between 2005 and 2013, the proportion of households with no access to Definitions: -No electricity: Household has no access to electricity (public grid or electricity fell by 15.5 percentage points, from 26 27 electricity generators) with no 31.9% in 2005 to 23.6% in 2013. Likewise, there -No improved sanitation: Household has no flush toilet -No improved water sources: Household has no access to the improved was considerable reduction in the proportion of water system - piped water, cistern, community water supply Yap households living in dwellings without sanitary -Poor quality housing: Household lives in dwelling where building Yap materials (roof, walls, floor) are not solid (concrete, metal or wood) facilities: from 45.9% in 2005 to 20.8% in 2013. -Poor cooking fuel: Household does not use or have access to electricity, State gas, or kerosene for cooking Another area of significant improvement was in -Few assets: Household does not have any radio, television, telephone, school enrollment of children: the percentage of cell phone or automobile -No child attending school: Household with at least one child aged 6 to households with a school-age child not attending 11 currently not attending school school fell from 12.6% to 2.4%. -No one completed primary education: Household does not have any member that has completed primary or higher education FSM -No one working: None of the household members are either employed But there are other areas where improvements (salary/wage) or self-employed are not as significant. Households with no Table A1: Expenditure components for the FSM HIES 2013/14 poverty assessment Classification Category9.2 Description Consumption Food Purchased food items and eat-outs Food produced and consumed by households ANNEX I: WELFARE INDICATOR Alcohol & Tobacco Food received and consumed by households Alcohol/tobacco/narcotics purchased, produced or received Rent Actual rent paid self-estimated rental values Household Operation Household services (gardening, maid services, babysitting, etc.) TOTAL HOUSEHOLD EXPENDITURE obtained from a (hedonic) rental model estimating Non-durable/consumable goods the relationship between reported rental values and House maintenance/repair Expenditure components a number of observable dwelling characteristics Purchase of small tools Based on categorization typically used in FSM, (Table A2). Recreational goods household expenditures were first grouped into 13 Clothing Clothing and footwear mutually exclusive categories (Table A1). The following are not included in aggregate welfare Utilities Electricity/water bills indicator: i) production related expenditures and Fuel for cooking/lighting Total household expenditure is the sum all gifts given to others; ii) highly irregular spending Sewage/garbage disposal consumption expenditures reported in Table A1. and investments such as expenses for house Communication Internet, phone cards, phone bills, cable, postal/shipping services Expenditures are annualized, either based on the construction or for major alterations; and iii) Transport Travel, vehicle repair/services, fuel, parts, related fees 2-week diaries for food items, or the length of recall spending on jewelry. Finally, as sufficient information Education Tuition and education-related expenses used in recording non-food expenditures. For food was not collected to enable the computation of the Health Treatment, hospitalization, medication, health equipment items in the diary, the annualized value is obtained annual use value of durable goods, expenditures Miscellaneous Personal care goods/services, legal/administrative services by multiplying the diary expenditures by 26. For on durables are not included in the final welfare Non-Consumption Financial Insurance, tax, financial fees, fines, loan interest payment several non-food items recorded only in the diary aggregate. Total adult equivalent expenditures are Expenditure Cash contribution Cash contribution for functions (e.g., brooms), annualization factors used were computed by dividing household expenditures by Cash contribution to school/community/church based on consultations with staff from the National the adult-equivalent household size. Statistics Division. Table A2: Estimated rental regression equation TREATMENT OF ERRORS AND OUTLIERS Imputation of rent Indicator Estimated Coefficient P value The rental housing market is vastly underdeveloped There were 16 households without diary records Floor material is concrete 0.3485 0.000 in FSM and only 36 households in HIES 2013/14 and had to be excluded from the sample. Roof material is concrete 0.1495 0.001 28 sample reported actual rental payments while the Number of rooms (increase of one room) 0.1904 0.000 29 rest (1,612) provided their own estimates of rent Most of the other errors in the diary were due Electricity is accessible 0.2004 0.000 for the dwelling that they owned and lived in. Given to confusion in recording quantity units versus Constant 6.4653 0.000 the very thin rental market in the country, self- actual quantities in appropriate columns of the R-squared 0.5737 reported imputed rents can be subject to uncertain questionnaire. Identification of errors and their measurement errors. Hence, the information on corrections were made on the basis of comparing estimated rents is not used directly. Instead, when unit values of the same item within the same state. Note: i) In addition to the above variables, municipality dummy variables (not reported) were used to account for municipality-level differences. ii) Logged values are used for the regression analysis. constructing the rental component, actual rents Removing households without diaries decreased are used whenever available, but predicted imputed the final sample from 1,664 to 1,648 observations. rents are used otherwise. These predictions are Sampling weights were accordingly adjusted. CALORIE STANDARD Standard daily per capita calorie intake of 2,100Kcal is used ANNEX II: FOOD BASKET based on recommendation of the Humanitarian Charter and Minimum Standards in Humanitarian Response, collectively authored by a group of NGOs and international organizations.6 However, the analysis is based on adult- equivalent using the adjustment factor suggested by Table A3: Reference national food basket Deaton and Zaidi (2002) : 7 Item Share of the Total Rank based on Item consumption expenditure expenditure share Kcal per 100g A0+ C0 Rice (short grain) 12.92% 1 358 ADJ factor = (A0+0.5C0 ) Reef Fish 11.14% 2 110 Breadfruit 9.30% 3 107 Where A0 and C0 are the country’s average number of Taro 7.77% 4 97.5 adults and children (ages 0-14) respectively. The minimum Rice (medium grain) 5.98% 5 362 calorie requirement expressed in per adult equivalent is, 5.44% 6 Chicken 196 thus, set at 2,565Kcal. Food quantities are converted to Banana 5.00% 7 103 caloric values using conversion factors from the Pacific Tuna 4.58% 8 193 Islands Food Composition Tables8. Other Fish 3.91% 9 160 Mackerel 3.29% 10 182 Turkey Tail 2.28% 11 234 Ramen 2.14% 12 525 Coconut 1.94% 13 184 Bread 1.69% 14 242 Sugar 1.47% 15 394 Coffee 0.94% 16 132 Eat-out 0.82% 17 137 Spam 0.72% 18 112 30 Flour 0.51% 19 349 31 Corned Beef 0.51% 20 192 Hor Dog 0.46% 21 254 Soy Sauce 0.44% 22 33 Onion 0.28% 23 26 Soda 0.28% 24 43 6 The Sphere Project (2011). Humanitarian Charter and Minimum Standards in Humanitarian Response, the Sphere Project, Southampton, Oil 0.23% 25 880 United Kingdom. 7 Deaton, Angus and Salman Zaidi (2002). Guidelines for Constructing Consumption Aggregates for Welfare Analysis, the World Bank, LSMS Working Paper 135. 8 FAO (2004), Pacific Islands Food Composition Table, Second edition, Food and Agriculture Organization of the United Nations, Rome. ANNEX III: MEAN CONSUMPTION Table A4: Mean real annual expenditure per adult-equivalent by category and state ($US) TOTAL POPULATION POOR POPULATION Yap Chuuk Pohnpei Kosrae FSM Yap Chuuk Pohnpei Kosrae FSM Food 1544.7 1344.7 1034.0 1710.2 1277.2 Food 859.7 691.4 521.3 720.7 653.1 Alcohol & Tobacco 287.4 42.5 238.6 100.0 144.2 Alcohol & Tobacco 102.3 19.6 64.6 28.2 44.3 Rent 415.9 694.5 819.1 719.4 708.1 Rent 143.2 338.3 418.6 397.4 345.6 Household Operation 44.3 30.2 36.3 76.0 36.6 Household Operation 10.1 12.2 13.8 27.7 13.0 Clothing 31.8 43.7 38.0 65.8 41.5 Clothing 18.4 13.6 12.7 20.3 14.0 Utilities 204.1 175.2 232.5 297.5 205.9 Utilities 35.5 48.8 90.5 97.1 62.9 Communication 73.9 69.7 66.2 106.5 71.0 Communication 12.0 18.1 21.1 29.1 18.7 Transport 229.1 119.6 210.9 320.8 176.2 Transport 32.6 11.1 42.7 63.6 25.7 Education 26.1 25.0 27.5 9.9 25.2 Education 5.1 5.0 11.8 3.8 7.3 Health 12.6 9.4 17.7 9.7 12.8 Health 5.5 0.3 6.9 4.4 3.2 32 Miscellaneous 24.0 15.3 20.2 30.2 18.9 Miscellaneous 2.1 5.8 8.1 10.4 6.3 33 Financial 138.3 29.8 64.7 228.4 65.9 Financial 13.7 6.2 7.9 20.8 8.0 Cash Contribution 78.2 94.2 215.9 246.2 144.3 Cash Contribution 15.8 35.6 79.1 66.9 49.1 Total 3110.5 2694.1 3021.6 3920.6 2927.6 Total 1256.0 1205.9 1299.3 1490.4 1251.3 * Nominal values are deflated using the weighted average of regional poverty lines to derive real values 34