AN UPDATE ON POVERTY AND INEQUALITY

Transcription

1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized AN UPDATE ON POVERTY AND INEQUALITY IN NICARAGUA: 9 STYLIZED FACTS ( ) María Eugenia Dávalos and Renos Vakis 1 March 2011 Summary: Between 2005 and 2009, general poverty declined in Nicaragua, driven by welfare improvements in rural areas. At the same time, inequality fell across the board. Still, one in three rural Nicaraguans lives in extreme poverty. A preliminary analysis suggests that muted economic growth, weak labor market dynamism and low coverage of social programs cannot explain the results. Instead, external shocks such as the global economic crisis likely had a bigger role in shifting employment and income opportunities, while the food crisis seems to be explaining part of the rural gains through improvements of farm incomes and deterioration of urban household welfare. A more in-depth analysis of poverty and inequality in Nicaragua from 2005 and 2009 is urgently needed by expanding this work to uncover the determinants of the poverty and inequality changes, the constraints that the poor still face to escape poverty, the vulnerability to external shocks and the role of social programs in overall poverty reduction. 1 Poverty and Gender Unit, Poverty Reduction and Economic Management Sector Department, Latin America and the Caribbean Region, The World Bank. 1

2 THE EVOLUTION OF POVERTY AND INEQUALITY IN NICARAGUA: Nicaragua has not had official poverty estimates since 2005 due to the lack of national representative household data. As such, little has been known about the evolution of welfare in the country. In 2009, the Instituto Nacional de Información de Desarrollo (INIDE) collected a new national representative survey which allows the estimate of poverty and other welfare measures 2. This short note presents 9 stylized facts that emerge related to the evolution of poverty and inequality since It also includes a brief discussion on methodology and comparability of poverty estimates across years (Annex 1). STYLIZED FACT 1: POVERTY REDUCTION, DRIVEN BY IMPROVEMENTS IN RURAL AREAS Poverty declined in Nicaragua between 2005 and Around 230,000 people escaped poverty, with the poverty rate significantly changing from 48.3 to 42.5 (5.8 percentage points, Figure 1). The decrease in poverty was mainly driven by rural areas (Figure 2), where poverty declined 7 percentage points. By contrast, urban areas experienced no significant changes in any of the poverty measures during the same period (see Annex 2 for more detail). 3 On the other hand, extreme poverty over the period persists. There is no significant change at the national, urban or rural levels (Figure 1 and 2, Annex 2). In fact, as of 2009, 1 in three rural Nicaraguan s still lives in extreme poverty. 2 See Appendix for a discussion on comparability of poverty estimates with previous surveys. 3 In this note, consistent with INIDE, estimates are considered significant at the 5% level or lower, as signaled with a circle in the figures presented. 2

3 Figure 1. Headcount ratio in Nicaragua, 2005 and General Poverty Extreme Poverty Source: EMNV 2009 Urban Areas Figure 2. Headcount ratio by Area, 2005 and 2009 Rural Areas General Poverty Extreme Poverty 0 General Poverty Extreme Poverty Source: EMNV 2009 Mirroring these results, rural areas experienced reductions in the other measures of poverty, namely the poverty gap and poverty severity. For example, the general poverty gap in rural areas decreased by 5 percentage points to 23.2 (Figure 3). Again, no significant changes exist in urban areas nor among the extreme poor (Annex 2). 3

4 Figure 3. Depth and Severity of Poverty by Area, Urban Rural Pov. Gap Pov. Gap Squared Pov. Gap Pov. Gap Squared Pov. Gap Pov. Gap Squared 8 7 Pov. Gap 3 3 Pov. Gap Squared General Poverty Extreme Poverty General Poverty Extreme Poverty Source: EMNV 2009 With respect to regional heterogeneity, the results also suggest that the decline in poverty between 2005 and 2009 has been concentrated. In fact, only in two out of the seven regions in Nicaragua (the Urban Pacific and Rural Central) general poverty declined significantly (Figure 4). In particular, Rural Central, one of the poorest rural areas in the country, experienced declines in both general and extreme poverty with a decrease in the general poverty rate by 8 percentage points to 69 percent. Similarly, the Urban Pacific also experienced a significant reduction in general poverty from 38 to 29 percent. Interestingly, while not significant, Managua is the only of the seven regions where poverty experienced a slight increase. Figure 4. Headcount ratio by Region, 2005 and 2009 Source: EMNV

5 Gini STYLIZED FACT 2: INEQUALITY IMPROVES Inequality in Nicaragua declined across the board during between 2005 and Using the Gini coefficient (based on consumption), overall inequality decreased from 40.5 in 2005 to 37.1 in 2009 (Figure 5 and Annex 3). Inequality also declined within urban and rural areas. These trends are also consistent with a declining trend in inequality in Latin America over the last few years. In fact, inequality in Nicaragua is below the regional average. Figure 5. Inequality trends - Gini Coefficient (2005 and 2009) National Urban Rural Source: EMNV 2009 A related inequality measure (the Theil index) also shows a decline in inequality from 2005 to 2009 (0.30 to 0.24, Figure 6). The index is useful as the inequality changes can be further decomposed into changes between and within rural and urban areas. The decomposition shows that the largest component of the inequality reduction over the period is explained by inequality within each area (Figure 6) in both years. In addition, the changes seem to be driven by a decline in the within component, suggesting that the welfare distribution in rural and urban areas (within each area) are becoming more equal. Still, although explaining a smaller fraction of total inequality, inequality between urban and rural areas remains. 5

6 Figure 6. Decomposing the Theil Index: Inequality between and within rural and urban areas Within Between Source: EMNV 2009 STYLIZED FACT 3: FEW CHANGES IN ACCESS TO BASIC SERVICES The improvements in poverty and inequality from 2005 to 2009 are reflected in muted changes in the access to basic services. In fact with the exception of access to cellular telephones (both urban and rural) and access to electricity (urban areas only), there has been no change in the provision of or access to basic services in neither urban nor rural areas in the last 5 years (Figure 7)! This includes services like electricity (for rural areas), water, fixed telephones, paved roads and trash collection. The lack of changes also reflects a persistent sharp divide in access between rural and urban areas in Nicaragua. Urban areas already had high access in the majority of the services. For example, by 2009, 90 percent of the urban population had access to water, 98 percent to electricity, 77 to paved roads and 72 disposed had acc1ess to trash disposal services (Figure 7). 4 By contrast, only one in four rural households had a water connection, less than one in two to electricity, 18 percent to paved roads and 3 percent to trash disposal services. Interestingly, in both areas, access to mobile phones was the exception, with sharp increases in this period going from 37 to 79 percent in urban areas, and from 5 to 39 percent in rural areas. As in other countries in the region and the world, the fast expansion of private cellular telephone 4 Although very few cases, this also includes households that pay to have the trash disposed of and those that convert it into organic fertilizers. Conversely, it excludes households that burn or bury their trash, or those that throw it into the field or bodies of water. 6

7 companies in both urban and rural areas has allowed bridging an existing gap in access to telecommunication (and as can been seen by the low availability and expansion of landline phones) Figure 7. Access to Services by Area, Urban Areas Water Electricity Telephone Mobile phone Paved road access Trash disposal Rural Areas Water Electricity Telephone Mobile phone Paved road access Trash disposal Source: EMNV

8 UNDERSTANDING THE TRENDS The overall results between 2005 and 2009 in Nicaragua can summarized as follows: (i) poverty reduction, driven by rural areas; (ii) no change in extreme poverty; (iii) a decline in inequality across the board, driven by improvements within areas; and (iv) access to basic services has remained unchanged, highlighting stark difference in accessibility between urban and rural areas. This section aims to provide some preliminary ideas on the context in Nicaragua during this period to further understand the poverty and inequality trends between 2005 and 2009 presented in the previous section. STYLIZED FACT 4 WEAK GROWTH, IN QUANTITY AND QUALITY Nicaragua experienced a moderate-to-low growth rate between 2005 and 2009, with an average annual growth in GDP of 1.7 percent (Figure 8). Growth rates by sector were equally unimpressive, with a yearly sharp decline of 3.8 percent in construction, a labor-intensive and mostly urban sector (Figure 8). Nonetheless, there was a positive though moderate growth in other traditionally laborintensive activities such as commerce (1.8 percent) and manufacturing (2.3 percent), sectors in which almost half of the urban poor are employed. These sectors, however, suffered a decline during the latest years of the period due to the global economic crisis (see Stylized Fact 8). The two most dynamic sectors in the period, financial services (5 percent) and transport and communications (4 percent) employed very little of the total workforce in Nicaragua (around 5 percent of country s workforce, and around 1.5 percent of the working poor). On the other extreme, the mining sector experienced the largest drop from 2005 and 2009 (4.1 percent); this economic sector is mostly capital-intensive, thus also employing a very small fraction of the total Nicaraguan workforce (0.5 percent in 2009). 8

9 Annual growth rate % Annual growth rate % GDP Financial services Transport and communications Livestock, fishing, silviculture Manufacturing Electricity and water Commerce, hotels and restaurants Housing Ownership Other Services Agriculture Construction Mining Figure 8. Annualized GDP and sectoral growth, % 4% 5.0% 4.0% 3.0% 2% 1.7% 2.3% 1.9% 1.8% 1.8% 1.6% 0% -0.6% -3.8% -4.1% -2% -4% -6% Source: Own calculations with Central Bank data. Growth incidence curves (GIC) of the annual growth rate in both consumption and income suggest that both consumption and income increased in Nicaragua between 2005 and 2009 across the whole welfare distribution. The larger gains happened to the lower half of the distributions (Figure 9). A similar phenomenon (not-shown) is observed in the distribution of growth in both labor and nonlabor income in this period. Figure 9. Growth Incidence Curves (GIC), Consumption GIC Income GIC 11 Total (years 2005 and 2009) 10 Total (years 2005 and 2009) Growth-incidence 95% confidence bounds Growth-incidence 95% confidence bounds Growth in mean Mean growth rate Growth in mean Mean growth rate Expenditure percentiles Expenditure percentiles Source: EMNV

10 Persons in thousands STYLIZED FACT 5 - LOW DYNAMISM IN LABOR MARKETS Changes in the labor market during the period were far from dynamic 5. For example, there were little or no changes in formality and unemployment rates, although the number of those considered economically inactive increased by 7 percent from 2007 to 2009 (Figure 10). Figure 10. Labor Markets: Economically Active Population (EAP) in Nicaragua, Share (%) of the EAP by Employment Status Persons in the EAP 80% 60% 40% 20% % Formal Informal Unemployment 1800 Economically active population (EAP) Economically inactive population Source: Central Bank of Nicaragua. Related to this low dynamism, the real average wage index between 2005 to 2009 remained at best the same (manufacturing, agriculture and livestock); at worst it declined (commerce, construction) (Figure 11). Figure 11. Real average wage index, Agriculture and livestock Manufacturing Construction Commerce Transport and communication General average Source: Central Bank of Nicaragua. 5 With the available data from 2007 and 2009 of the Central Bank of Nicaragua. 10

11 STYLIZED FACT 6 AGRICULTURE INCOMES IMPROVE FOR THE RURAL POOR The majority of poor and non-poor rural households are employed in agriculture, livestock and fishery (80 and 59 percent respectively, Figure 12). Employment in public sector, commerce and manufacturing consist the majority of the remaining labor sources. Figure 12. Rural employment, by poverty status (2009) Agriculture, Livestock, Fishery Government & community services Commerce, Hotels and Restaurants Manufacturing Transport Construction Finance Electricity and Water Mining Non Poor Poor Source: EMNV 2009 By 2009, poor households increased their participation in agricultural activities (by one percentage point) and in the non-agricultural sector of commerce, hotels and restaurants (Figure 13). In addition, the evidence suggests that the poor participated less in the manufacturing and transport sector. On the other hand, non-poor households experienced a larger increase in their participation (7 percentage points) towards agricultural employment, accompanied by a proportionate decline in their participation in manufacturing and commerce, hotels and restaurants. While indicative, increases in agricultural activities and a reduction in activities such as manufacturing are likely to be related to the global food and economic crises that occurred during the period (see stylized fact 8). 11

12 Figure 13. Percentage Point Change in Activity of Rural Employed Population, Poor Agriculture, Livestock, Fishery Commerce, Hotels and Restaurants Government & community services Finance Electricity and Water Construction Mining Transport Manufacturing Non Poor Agriculture, Livestock, Fishery Transport Finance Mining Electricity and Water Construction Government & community services Manufacturing Commerce, Hotels and Restaurants Source: EMNV 2009 Additional exploration of changes in rural incomes reveals a number of interesting trends. First, there was an average increase in income of rural households, of which around 53% came from nonagricultural labor and around 20% from agricultural sources. Focusing on the rural poor vs the nonpoor shows that income changes between 2005 and 2009 behaved very different across groups. Among non-poor rural households, the net income change was close to zero (Figure 14). While they 12

13 experienced increases in income from non-agricultural activities, the gains were offset by declines in incomes from agriculture and to a lesser extent remittances. By sharp contrast, incomes across all categories (with the exception of remittances) resulted in large income changes among the rural poor. More than half of the income gains among poor rural households can be attributed to increases in either wage or self-employment in agriculture (Figure 14). An additional 25 percent of the income increases are related to non-agricultural income increases (mainly driven by self-employment income). Figure 14. Absolute Income Changes by Source, Rural non-poor Absolute Income Changes TOTAL INCOME CHANGE -33 Non-Agricultural Wages Non-Agricultural Self-Employed Agricultural Wages Remittances Agricultural Self-Employed Other income 13

14 Rural poor Absolute Income Changes TOTAL INCOME CHANGE 776 Agricultural Self-Employed Agricultural Wages Non-Agricultural Self-Employed Non-Agricultural Wages Remittances Other income Source: EMNV Note: Other income includes income from own home value, food (donations/gifts), pensions, institutional donations, rent, interest, dividends, educational transfers and others. An income increase in rural areas, particularly for the poor, is likely to be related with the decline in inequality. In fact, decomposing the contribution of each source of income to total inequality (Shorrocks, 1982) suggests that while in 2005 the element explaining the highest proportion of inequality was agricultural self-employment (explains almost 60 percent), this share fell considerably by 2009 and it actually points to a small equalizing effect of agricultural wages (Figure 15). Figure 15. Contribution to total inequality - by income source Non-agricultural self-employed Agricultural self-employed Non labor income Non-agricultural wages Agricultural wages Source: EMNV

15 STYLIZED FACT 7 SOME DYNAMISM OF NON-AGRICULTURAL SECTORS IN URBAN AREAS, BUT NOT ENOUGH In contrast to rural areas, labor employment in urban areas is more diversified for both the poor and the non-poor. For example, about a third of urban households are employed in the public sector and another third in commerce (this share is larger among non-poor individuals relative to poor, Figure 16). In addition, around 15 percent of urban individuals work in manufacturing, and 8 percent in agriculture (this is more than 15 percent among the urban poor). Figure 16. Urban employment, by poverty status (2009) Commerce, Hotels and Restaurants Government & community services Manufacturing Transport Agriculture, Livestock, Fishery Construction Finance Mining Electricity and Water Non Poor Poor During the period, overall changes in activities of employment were modest. The largest changes were among the non-poor. For example, the non-poor reduced their participation in manufacturing and commerce, hotels and restaurants (4.2 and 1.2 percentage points, respectively, Figure 17). On the other side, the non-poor shifted employment towards public jobs (3.6 percentage points). Smaller shifts were experienced among the poor notably an increase in commerce and a decline in participation in manufacturing. As in rural areas, the observed generalized move away from manufacturing by 2009 might be a result of the global economic crisis. 15

16 Figure 17. Percentage Point Change in Activity of Urban Employed Population, Poor Commerce, Hotels and Restaurants Construction Transport Government & community services Mining Finance Electricity and Water Agriculture, Livestock, Fishery Manufacturing Non-Poor Government & community services Transport Agriculture, Livestock, Fishery Mining Finance Construction Electricity and Water Commerce, Hotels and Restaurants Manufacturing Urban non-poor households had, on average, a positive income change during the period. Urban income increases among the non-poor is mainly attributable to increases in non-agricultural income (Figure 18). A large fraction of the income gains from these non-agricultural sources was, however, offset by a large decline in agricultural income and to a lesser extent decrease in remittances. Among poor urban households, income changes are driven by non-agricultural gains in self employment. 16

17 Figure 18. Absolute Income Changes by Source, Urban non-poor Absolute Income Changes TOTAL INCOME CHANGE 408 Non-Agricultural Self-Employed Non-Agricultural Wages Remittances Agricultural Wages Agricultural Self-Employed Other income Urban poor Absolute Income Changes TOTAL INCOME CHANGE 748 Non-Agricultural Self-Employed Agricultural Self-Employed Agricultural Wages Non-Agricultural Wages Remittances Other income Source: EMNV Note: Other income includes income from own home value, food (donations/gifts), pensions, institutional donations, rent, interest, dividends, educational transfers and others. 17

18 STYLIZED FACT 8 CAN THE CRISES EXPLAIN THE RURAL POVERTY REDUCTION? The decrease in poverty and inequality occurred amidst the food crisis of 2007/2008 -in which the prices of staples in many countries around the world, including Nicaragua, skyrocketed- and the unprecedented global economic crisis of which reduced the world output by 0.5 percent and the regional output by 1.7 percent in To some extent, this was an important context (and partial cause) of the non-impressive economic growth in Nicaragua during this period. In the pre-crisis period of , the economy grew by around 2.8 percent annually; from 2007 to 2009, this growth rate was less than one half of one percent (Figure 19), with a fall of 1.5 percent in GDP in the post-crisis year of During this time, more traditionally urban sectors like construction and manufacturing (employing close to 30% of the urban poor) suffered the most, while more traditionally rural sectors like agriculture, livestock, fishing and silviculture (employing around 80% of the rural workface) remained closer to their pre-crisis growth rates. Commerce, hotels and restaurants stagnated in the period from 2007 and Figure 19. Annualized GDP growth in % by Sector National Transport and communications Commerce Construction Manufacturing Livestock, fishing, silviculture Agriculture Source: Central Bank of Nicaragua. 6 World Economic Outlook (IMF) data. 18

19 Percentage point change in $1/day poverty rates Moreover, inflation data suggests that the food crisis resulted in a much higher overall inflation (around 17 percent in 2007 and 14 in 2008) 7, primarily driven by the higher cost of food. Inflation of food and beverages reached over 20 percent in 2007 and 2008, and that of agricultural products climbed to nearly 40 percent in Although the Latin American region seems to have been more resilient to these external shocks than in the past, poverty simulations of the effect of the food crisis in Nicaragua show increases in poverty and that urban areas were potentially more affected by the shock than rural areas. Ivanic and Martin (2008) 8 estimate two scenarios, one simulating a 10 percent increase in the prices of beef, dairy, maize, poultry, rice, sugar, wheat, and the other simulating the actual price increase in these products. In both cases, urban poverty changes were close to doubling those of rural areas (Figure 20). It is plausible then, that rural households might have been better able to cope with the crisis, either from production for self-consumption or as net-sellers of agricultural products. Figure 20. Simulations: Effect of the Food Crisis on Poverty in Nicaragua Rural Urban Rural Urban Simulation 1: 10% increase Simulation 2: global food price increase Source: Ivanic and Martin (2008). A final exercise to better understand the potential impact of the food crisis is to explore the income sources among those that exited poverty in rural areas. In the absence of panel data, the exercise explores those households that are above (but near) the poverty line in both years. The results are striking: incomes for this group increased by 26 percent between the two years (around 5% per year) and it was driven entirely by a huge increase in agricultural self employment income (Figure 21). To 7 Central Bank of Nicaragua. 8 Ivanic, Maros and Will Martin (2008) Implications of Higher Global Food Prices for Poverty in Low-Income Countries. Policy Research Working Paper 4594, The World Bank. 19

20 the extent this is capturing those households that may have exited poverty, the exercise is suggestive of the positive impact of the food crisis among households in agricultural self-employment. It is also indicative of the potential channel by which rural poverty reduction may have benefited by the food crisis. Figure 21. Absolute Income Changes of Rural Population Near and Above the Poverty Line, Absolute Income Changes Total Income 2058 Agricultural Self-Employed Other Income Non-Agricultural Self-Employed Non-Agricultural Wages Agricultural Wages Remittances Source: EMNV Note: Other income includes income from own home value, food (donations/gifts), pensions, institutional donations, rent, interest, dividends, educational transfers and others. The exercise considers households with income up to 10 percent above the poverty line in each year. 20

21 Proportion of population in each group receiving transfer STYLIZED FACT 9 DO SOCIAL POLICIES REACH THE POOR? COVERAGE AND TARGETING The household data include a module of program participation in a range of social programs ranging from programs in education, health, productive interventions and property titling. While different programs will by design have objectives beyond poverty, in exploring the potential channels of the welfare improvements discussed above, understanding to what extent social policies are reaching the poor is a useful starting point. This can be done in a number of ways. Focusing on the three flagship social programs in Nicaragua Usura Cero, Hambre Cero and Amorreveals three trends. First, the overall coverage of these flagship programs is low in 2009 (Figure 22). Program coverage was highest for Hambre Cero (a program that focuses on training and transfer of productive goods to poor rural families). Still, it only reached less than 4 percent of the total population and 8.3 percent of the total rural population. At the household level, this suggests that less than 40,000 households were beneficiaries of the program by 2009, half the government target of 75,000 households. For Usura Cero and Amor, overall national coverage was also very low at 3.2 percent and 0.8 percent, respectively. Note that for the case of these programs (especially Amor), the data presented here was collected as the program was expanding and as such, the coverage trends are likely to be underestimated. The point though does remain that these trends do suggest that the low coverage of these programs during the period makes it difficult to associate them as drivers of the observed poverty reduction. Figure 22. Social Programs Coverage by area Usura Cero Hambre Cero Programa AMOR Total Urban Rural Source: EMNV

22 In terms of targeting towards the poor, these programs can be largely improved, especially if one of the underlying objective is to target poor households. A number of indicators suggest large exclusion and inclusion errors. For example, four in ten beneficiaries of Hambre Cero were non-poor compared to 37% and 25% of moderate and extreme poor, respectively (Figure 24). In the case of Usura Cero, a program that aims to provide small loans to the urban poor, more than 80 percent of the beneficiaries are non-poor. Again, while these programs do include additional considerations in defining eligibility beyond poverty, and as such, these results do not suggest inefficient targeting per se. They do indicate though that the extent by which these programs may have affected the observed poverty and inequality reduction in Nicaragua is limited towards the poor who received them. Figure 23. Total beneficiaries of social programs: share by poverty status (%) Hambre Cero Usura Cero AMOR Extreme Poor Moderate Poor Non-Poor Source: EMNV 2009 The targeting effectiveness towards of social programs the poor is also reflected by the concentration index on the progressivity of program targeting vis-à-vis the poor. The concentration index takes values between -1 and 1, with negative values indicating a progressive distribution (most of the beneficiaries are poor), positive values indicating a regressive distribution (most of the beneficiaries are non-poor), and a zero value indicating an equal distribution among population groups. This exercise is done for 10 social programs ranging from programs in education, health, productive interventions and property titling. As discussed above, the exercise does not aim to imply that these programs are badly targeted (since their objectives go beyond reducing poverty), but instead they serve as an input in understanding how much the poor have benefited from them. 22

23 For half of these programs (Hambre Cero, Amor, Merienda and Mochila escolar, Campana de Alfabetización), the concentration index is negative suggestive an overall progressive trend in terms of these programs targeted/reaching the poor. Still, the indices are far from -1 (perfect targeting towards the poorest, Figure 24). This implies that, if poverty is one of the criteria for program eligibility, these programs have some space to be improved. The results also show that a number of other programs (Credito rural, titling, vaccination campaigns and the grains program) are at best neutral. Finally, Usura Cero is the program with the least progressive targeting (consistent with Figure 23 and the fact that it targets urban households linked to entrepreneurship). While these results are to be used as a starting point for further analysis, they do indicate that the low overall coverage and low number of poor benefiting from these programs cannot fully explain the poverty and inequality reduction in Nicaragua. Figure 24. Targeting of Social Programs: Concentration Index Usura cero Jornada anti epidémica Crédito rural Venta de granos básicos Titulación de la propiedad Merienda escolar Atención Integral de la Niñez (AMOR) Campaña Nac. De Alfabetización Mochila escolar Hambre cero Source: EMNV

24 SUMMARY During 2005 and 2009, poverty declined in Nicaragua, driven by poverty reduction in rural areas. The extreme poverty remained unchanged. Inequality also fell during this period. Preliminary exploration suggests that muted economic growth, labor market dynamism, low coverage and low reach of social programs to the poor cannot fully explain some of these welfare improvements. Instead, external shocks (such as the global food and economic crises during the period) likely played a larger role in shifting employment and income opportunities. For example, the food crisis can potentially explain part of the rural gains (by improving farm incomes and potentially hurting urban households). Despite the observed progress in Nicaragua many challenges remain. Poverty clearly remains a rural phenomenon with 6 out of 10 rural Nicaraguans still poor and one in three in extreme poverty. The stylized facts presented here aim to provide some initial hypothesis for a more in depth analysis of poverty and inequality in Nicaragua from 2005 and A much closer look could expand this analysis and uncover the determinants of the poverty and inequality changes, the constraints that the poor still face to escape poverty, the vulnerability of the poor and non-poor to external shocks, and the role of social programs in poverty reduction. In particular, given the current food prices crisis, it is relevant to further explore the role of the 2007/08 food crisis coupled with the global economic crisis of 2008, and the channels through which they impacted poverty and inequality changes (e.g. labor markets, prices, remittances). Specifically, it would be interesting to identify the households that benefitted from the price increase (net buyers vs. net sellers) and their profile, from which important lessons can be drawn to face current challenges. Similarly, the data presented only looks at the role of social programs vis-à-vis reaching the poor. More analysis should focus on the potential impact of these programs to improve welfare and better place them in their design with respect to their specific objectives which in many cases are beyond poverty per se. Finally, as Annex 1 specifies in more detail, future work should focus on harmonizing the poverty methodology in Nicaragua so that comparisons across years, beyond that of 2005 and 2009, can be made. 24

25 ANNEX 1. ON POVERTY MEASUREMENT IN NICARAGUA Since 1993 Nicaragua has been collecting living standards household survey data with subsequent rounds in 1998, 2001 and The Encuesta Nacional de Hogares Sobre Medición de Nivel de Vida of 2009 (EMNV 2009), collected by the Instituto Nacional de Información de Desarrollo (INIDE) of Nicaragua, is the more recent round and has allowed for the calculation of updated poverty figures for the country. Still, even if Nicaragua has five survey rounds from 1993 to 2009, technical and methodological differences across years have resulted in poverty estimates that are not fully comparable with each other. First, the 1993 round differs with subsequent rounds in its sampling and the calculation of welfare measures rendering poverty comparisons difficult. On the other hand, the implementation of the 1998 and 2001 rounds were similar, allowing for comparability between these two years. The 2005 round introduced some additional challenges in comparability. While the survey itself was identical with previous rounds, a new census that was completed in 2005 provided with a number of options for poverty estimation, each with its advantages and disadvantages. Specifically, for the two rounds of 1998 and 2001, the caloric value for calculating the extreme poverty line was estimated using population structures based on projections from the 1995 census. Given the new census, INIDE was faced with 2 options: (i) use the new census (with actual populations) but lose comparability with prior rounds; or (ii) keep using the old projections (gaining comparability with previous rounds) at the cost of estimating results based on outdated population projections. A third option to use the new census to update the 1998 and 2001 results was not explored. At the end, INIDE decided to use the new census population to estimate the poverty line and as such, comparability (of poverty estimates) with previous years (1998 and 2001) is invalid. The per capita annual poverty line for 2009, in nominal terms, is C$6,903 for extreme poverty (US$335) and C$11,725 for general poverty (US$569). Finally, in the 2009 round, while several changes were introduced in the survey and data collection protocols (mainly due to budget constraints), e.g., smaller number of survey questions, a meta analysis suggests that these changes did not alter the comparability with 2005 data. Given that the methodology for poverty calculation in 2009 replicates that of 2005 and uses the 2005 census for population projections, these two last rounds provide comparable poverty estimates. 25

26 There are two pending tasks for INIDE. First the construction of a full comparable poverty series, dating as back as 1998, can be done by re-estimating poverty lines in a consistent way using all available information. Second the 1998 poverty methodology needs to be updated as it needs to reflect changes in, for example, consumption patterns in Nicaragua. To achieve these, the creation of a Poverty Committee of national and international experts has proven to be a successful initiative in other countries in the region (e.g., Peru), and could be one way to achieve consensus on the methodological approach for poverty calculation in the country. 26

27 ANNEX 2. POVERTY STATISTICS, Difference P-value General Poverty Urban General Poverty Rural General Poverty Extreme Poverty Urban Extreme Poverty Rural Extreme Poverty General Poverty Gap Urban General Poverty Gap Rural General Poverty Gap Extreme Poverty Gap Urban Extreme Poverty Gap Rural Extreme Poverty Gap General Poverty Gap Squared Urban General Poverty Gap Squared Rural General Poverty Gap Squared Extreme Poverty Gap Squared Urban Extreme Poverty Gap Squared Rural Extreme Poverty Gap Squared

28 ANNEX 3. INEQUALITY STATISTICS Confidence Intervals (95%) Gini Urban Gini Rural Gini