P a k i s t a n P o l i c y N o t e — R e c o v e r i n g s t r o n g p o s i t i v e t r e n d s i n p o v e r t y a n d o pp o r t u n i t y 79572 Pakistan Policy Note 8 John Newman 1 Recovering Strong J une 2 0 1 3 Positive Trends in Poverty and Opportunity Household survey data point to declining poverty data from Punjab are mostly free of anoma- over the past decade, despite some difficulties with lous results, so this note provides more detailed data from three provinces. Indeed, Pakistan has analysis for that province. turned economic growth into poverty reduction. In Punjab (the largest province, where the data appear In addition to providing information on poverty head- more reliable and are thus analyzed in more detail), count rates, the note introduces a possible measure of poverty has fallen considerably, from 33.5 percent in shared prosperity applied to the consumption growth 2001/02 to 16.4 percent in 2007/08, after adjusting of the bottom 40  percent of the population. This for higher food prices. This improvement was driven measure—an indicator the World Bank is test- largely by increasing returns in the nonfarm sector, ing for application worldwide—recognizes that in both urban and rural areas. Over the period, consumption growth in the bottom 40 percent the growth of per capita consumption of the bottom will matter most in building shared prosperity. 40  percent of Punjab’s population exceeded GDP (It differs from the Gini coefficient, which pays per capita growth. Subsequently, over 2007/08– attention to the equitable distribution of wealth 2010/11, per capita real consumption growth in but not growth at the bottom of the curve.) THE WORLD BANK GROUP SOUTH ASIA REGION Punjab was stagnant, and the equality of opportu- nity for primary education completion rates seemed As changes in social indicators are also important, to improve but alongside a slowdown in the rate of this note updates the Human Opportunity Index for improvement in indicators for water and sanitation 2010/11. It builds on earlier work that explored and for primary enrollment. Nationwide, the rate the trend in equality of human opportu- of improvement for gas and electricity deteriorated. nity between 1998/99 and 2007/08. Histori- Institutionally, there is a need for an independent cally, Pakistan has done better in reducing expert committee to work with the Pakistan Bureau of consumption-­ based poverty measures than in Statistics to strengthen poverty monitoring. improving social indicators. This note sheds some light on the last decade’s pov- Trends erty trends in Pakistan. Based on the Household Income and Expenditure Surveys in 2001/02, National 2004/05, 2007/08, and 2010/11, the broad trend appears to be one of substantial poverty Poverty reduction in Pakistan, as in most countries, reduction. However, given data concerns, it is is aligned closely with its growth in per capita income. premature to draw conclusions beyond this While there are data concerns, the overall pov- broad trend (see Box  1 below). Nevertheless, erty rate appears to have fallen steeply over the P a k i s t a n P o l i c y N o t e — R e c o v e r i n g s t r o n g p o s i t i v e t r e n d s i n p o v e r t y a n d o pp o r t u n i t y past decade (Figure 1). China and Indonesia Pakistan’s GDP per capita in 2005 international cut poverty substantially while their GDPs per $ PPP grown at least 3 percent a year for four capita increased sharply. Poverty reduction is consecutive years (2004–07)—and never has it strongest at the initial phase of growth, and as for five consecutive. Unsurprisingly, countries a country moves up the growth scale over time, that have been better at reducing poverty tend it takes a greater increase in GDP per capita to to be better at generating sustained growth achieve a given percentage point reduction in (Figure 3). poverty, with an even bigger impact in moving 2 down at the lower levels of poverty. The line for This observation—that Pakistan is successful in Pakistan between 2001/02 and 2007/08 is quite reducing poverty when GDP grows but cannot sus- steep, indicating a much stronger proportional tain that growth—has two important policy impli- reduction in poverty for the improvement in cations. First, with more growth interruptions, GDP per capita—an experience similar to an adequate social protection system becomes those of other countries. Many other countries, more important. The poor are vulnerable to either over their entire period or over a limited shocks—be they of natural disasters, health, period, also experienced significant reductions or macro policy. An adequate system would in poverty relative to their gains in GDP per ensure that when shocks hit, the poor and capita. Further progress will require not only vulnerable can still maintain the investments sustained growth but also stronger growth. they need to increase their incomes and their children’s welfare. With the advances in the This sharp reduction in poverty over a short period Benazir Income Support Programme, Pakistan also suggests that poverty in Pakistan is highly elas- has made solid progress in developing such a tic to growth. Figure 2 compares the elasticity of system in recent years. The second policy impli- poverty reduction with respect to growth across cation is that a renewed effort to address the countries. From 2001/02 to 2007/08, Pakistan problems that work against sustained growth had high elasticity—well above the mean and would be well justified for faster poverty reduc- toward the upper end of the distribution of tion. This effort should lead to policy priorities developing countries. for poverty reduction (such as more focus on the macro environment) different from those However, while Pakistan is doing well converting in countries better able to sustain growth but growth into poverty reduction, it is struggling to unable to convert that growth into rapid pov- sustain that growth. Since 1985, only once has erty reduction. Time path of the relation between poverty and GDP per capita, Pakistan and Figure selected countries 1 100 Poverty rate at PPP $1.25 a day (percent) China, 1980–2004 Vietnam, 1992–2007 75 Indonesia, 1983–2009 India, 1993/94–2009/10 50 Pakistan, 2001/02–2007/08 25 Sri Lanka, 1984–2006 0 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 GDP per capita (2005 international PPP $) Note: Each line represents the history of a country. For China, India, Indonesia, Sri Lanka, and Vietnam, the time paths are shown over a longer period; for other countries performance is shown only since 2000. Source: Author’s calculations. Poverty data are from PovcalNet, and GDP per capita data are from the World Development Indicators database. Distribution of elasticities of poverty reduction with respect to growth, Figure selected countries 2 Fiji Nepal China Swaziland Mali Pakistan, 1987–2008 Guatemala Elasticity = –2.163 Niger Mozambique Namibia Malawi 3 Cameroon Central African Rep. Azerbaijan Moldova India Panama Thailand Bangladesh Costa Rica Sri Lanka Nicaragua El Salvador Tunisia Guinea-Bissau Tanzania Brazil Jamaica Egypt, Arab Rep. Iran, Islamic Rep. Malaysia Algeria Turkey Belarus Rwanda Croatia Macedonia, FYR Serbia Slovenia Lithuania Czech Rep. Seychelles Argentina Mean = –0.716 Honduras Paraguay Belize Bolivia Yemen, Rep. –5 –4 –3 –2 –1 0 1 Elasticity Note: The elasticities are calculated over the entire range for which data are available (which varies for each country). Thus only one elasticity is presented per country. The figure only shows countries with at least six years between the first and last observation. The distribution is trimmed by dropping observations where the elasticities were less than –21 and greater than 20. Source: Author’s calculations. Poverty in the poverty elasticity calculations is the $1.25-a-day poverty headcount in international 2005 $ PPP from PovcalNet. Growth is calculated for GDP per capita (international 2005 $ PPP) from the World Development Indicators database. In Punjab, overall consumption base, but starting in 2006/07 the index for food and other essential items began Given other provinces’ data concerns, the focus to grow faster than the general consumer price is on Punjab, the only province free of anomalous index. Table 1 illustrates how the poverty line, results (Box 1). As the largest province, Punjab in both urban and rural areas, would change can provide some insight into the country as once such an adjustment is made. The over- a whole. But even here it is necessary to adjust all trend remains strongly toward poverty for increasing relative prices of food on the reduction, from 33.5  percent to 14.6  percent poverty line. The official poverty line uses the rather than to 11.8  percent. And rural pov- general consumer price index to adjust the erty remains much higher than urban poverty, P a k i s t a n P o l i c y N o t e — R e c o v e r i n g s t r o n g p o s i t i v e t r e n d s i n p o v e r t y a n d o pp o r t u n i t y Number of growth episodes when GDP per capita grew at least 3 percent a year for Figure five consecutive years 3 Countries with average poverty reduction of more than 1 percent a year or end-period poverty of less than 5 percent 20 15 Growth episodes 4 10 5 0 Costa Rica Nepal Guatemala Fiji Namibia Senegal The Gambia Cameroon Pakistan Burkina Faso Uganda Timor-Leste Ghana Tajikistan Ethiopia Cambodia Bhutan Thailand Indonesia China Vietnam Lao PDR Countries with average poverty reduction of less than 1 percent a year 20 15 Growth episodes 10 5 0 Lesotho Kenya Guinea-Bissau Mauritania Philippines Zambia South A rca India Panama Nigeria Rwanda Tanzania Botswana Bangladesh Madagascar Note: The top panel shows the distribution of episodes of five consecutive years of growth above 3 percent for countries that started with poverty rates of more than 20 percent and that have averaged at least a 1 percentage point reduction a year or that had poverty rates below 5 percent at the end of the period. The bottom panel shows the distribution for countries that had initial poverty rates of more than 20 percent and averaged less than a 1 percentage point reduction a year. Source: Author’s calculations based on World Development Indicators database. Box Concerns over poverty data in three provinces 1 Pakistan’s household survey data has the potential to provide useful evidence for guiding policy decisions on reducing poverty. Yet examining the poverty data from four recent Household Income and Expenditure Surveys at three-year intervals (2001/02, 2004/05, 2007/08, and 2010/11) reveals some data concerns that need to be resolved before a more detailed assessment can be made. These concerns include large reductions in rural poverty in Sindh combined with a decrease in food expenditure as a proportion of the total; large swings in poverty in Balochistan and Khyber Pakhtunkhwa (they are conflict-prone, so systematic data capture from all regions needs to be confirmed); the exclusion of many durable consumer goods from the estimate of total expenditure (which can understate improvements in poverty reduction as income grows, as well as the impact of an economic downturn); and the need to update the poverty line to account for changes in relative food and nonfood prices (as when food prices increase, higher expenditure will be required to generate the same minimum calorific intake). Until these problems are resolved, it will remain difficult to drill down to more detail using the national poverty estimates— hence the spotlight on Punjab, the only province free of anomalous results. Clearing up the concerns on national poverty data requires a review of the data collection practices in each province for each survey year. This can be done only with staff from the Pakistan Bureau of Statistics, who would have detailed information on field practices. Illustrative example of the effect of adjustment of poverty line to the increase in Table food prices relative to the general consumer price index, Punjab 1 Adjustment of original poverty line using Adjustment of original poverty line with adjustment general consumer price index for increased relative food prices in poverty line Nominal All Nominal All poverty line Punjab Urban Rural poverty line Punjab Urban Rural 2001/02 723.4 33.5 24.4 37.1 723.4 33.5 24.4 37.1 2004/05 878.6 24.3 17.3 27.5 888.2 25.0 18.2 28.1 2007/08 1,141.5 14.1 9.0 16.5 1,179.3 16.4 10.8 19.0 2010/11 1,671.7 11.8 5.3 14.9 1,771.4 14.6 6.5 18.4 5 Source: Author’s calculations based on Household Income and Expenditure Surveys 2001/02, 2004/05, 2007/08, and 2010/11. but some rates of poverty reduction would reduction. These growth incidence curves be somewhat slower, particularly in the most should be viewed with an appreciation of pop- recent period. This adjustment is only illustra- ulation movement to urban areas. Indeed, tive, however. although the share of the rural poor stayed roughly constant at 85  percent, overall pov- Ideally, the poverty line should be recalculated, using erty rates would have been higher without the either the 2007/08 or 2010/11 Household Income migration from higher poverty rural areas to and Expenditure Survey, at which time the govern- lower poverty urban areas. ment should also discuss the choice of price index to adjust the poverty line for inflation. It would be The bottom 40  percent can fully benefit from useful to have an expert committee pursue growth. Another indicator of whether the these proposed changes and resolving some of lower income population is benefiting fully the anomalous results, work with the Pakistan from growth and poverty reduction is to look Bureau of Statistics, and recommend ways that at the growth rate of the lower 40 percent of more reliable, regular, and timely poverty data the population—a proxy measure of shared can be produced and made available to the prosperity. Table 2 presents information on public. the growth of consumption of that 40 percent relative to the entire distribution and to the In Punjab, by income group growth rate of real GDP per capita. During the first two periods, growth of the bottom Overall poverty reduction is important, but it is 40  percent exceeded the growth of GDP per equally important to disaggregate the data to see capita. Only in the last period did growth of the impacts at different incomes. Figure 4 pres- that 40  percent fall below the growth of per ents the growth in consumption at each capita GDP—yet it still stayed above that of the income group for the province as a whole entire distribution. Thus for Punjab at least, it and for urban and rural populations for would appear that the bottom 40 percent can the three growth periods. During the first fully benefit from growth. growth period (2001/02–2004/05), poverty reduction was spurred by good consumption The end of growth tends to signify an end to pov- growth among all income groups, though the erty reduction. The problem seems to be more wealthy tended to benefit more. In the sec- the history of overall inconsistent growth with ond (2004/05–2007/08), overall growth in short periods of positive (but not terribly high) consumption remained strong but was most growth not being sustained over long periods pronounced in the lowest income groups. (see Table 1 and Figure 2). One final observa- In the third (2007/08 –2010/11), growth tion: the bottom 40 percent tends to be more was far lower among all percentiles, as posi- rural than the top 60  percent. However, this tive growth in urban areas was outweighed does not mean that the bottom 40 percent are by negative growth across all percentiles in all in rural areas and the top 60 percent are all rural areas, hence the reduced rate of poverty in urban areas (Table 3). P a k i s t a n P o l i c y N o t e — R e c o v e r i n g s t r o n g p o s i t i v e t r e n d s i n p o v e r t y a n d o pp o r t u n i t y Figure Growth incidence curves (for three-year periods), Punjab 4 Growth incidence Growth at median Growth in mean Mean growth rate 99 percent con dence bounds 2001/02–2004/05 Total Urban Rural 40 40 40 30 30 30 20 20 20 10 10 10 0 0 0 0 25 50 75 100 0 25 50 75 100 0 25 50 75 100 Expenditure percentile Expenditure percentile Expenditure percentile 2004/05–2007/08 Total Urban Rural 30 30 30 20 20 20 10 10 10 0 0 0 –10 –10 –10 0 25 50 75 100 0 25 50 75 100 0 25 50 75 100 Expenditure percentile Expenditure percentile Expenditure percentile 2007/08–2010/11 Total Urban Rural 20 20 20 10 10 10 0 0 0 –10 –10 –10 –20 –20 –20 0 25 50 75 100 0 25 50 75 100 0 25 50 75 100 Expenditure percentile Expenditure percentile Expenditure percentile Note: The growth rates are over the entire three-year period (not annualized). The time periods are the same so that the growth can be compared directly. The scales of the various growth incidence curves are different to show more detail. Thus, in comparing the figures, please note the scale of the y axis. Source: Author’s calculations based on Household Income and Expenditure Surveys 2001/02, 2004/05, 2007/08, and 2010/11. Human Opportunity Index education, health, and infrastructure services are considered opportunities. An opportunity An opportunity can be defined as a good or ser- is said to be distributed according to a princi- vice important enough to be made available to all, ple of equality of opportunity if circumstances regardless of background. In most societies, basic exogenous to the individual—such as birth Mean real consumption and growth in mean real consumption per adult equivalent Table of bottom 40 percent of population and total population 2 Punjab National Growth of mean real per capita consumption expenditure Growth of GDP per capita (constant local currency units; percent) All households (percent) Bottom 40 percent (percent) 2001/02–2004/05 3.2 5.0 3.9 2004/05–2007/08 4.5 4.8 5.2 2007/08–2010/11 1.1 –0.6 0.1 Source: GDP per capita from the World Development Indicators database; growth rate of mean real per capita consumption expenditure—author’s calculations based 7 on Household Income and Expenditure Surveys 2001/02, 2004/05, 2007/08, and 2010/11. Distribution of population in bottom 40 percent and top 60 percent of real monthly Table consumption per capita 3 (percent) Bottom 40 percent Top 60 percent Urban Rural Urban Rural 2001/02 22 78 34 66 2004/05 23 77 37 63 2007/08 23 77 37 63 2010/11 20 80 40 60 Source: Author’s calculations based on Household Income and Expenditure Surveys 2001/02, 2004/05, 2007/08, and 2010/11. place, gender, ethnicity, income, and educa- while Balochistan and Khyber Pakhtunkhwa tion of parents—have no bearing on how the were behind Punjab and Sindh, the pace of opportunity is distributed in the population. progress in the two provinces has picked up. In The Human Opportunity Index (HOI) com- sanitation, there are striking differences in the bines the overall coverage rate of the opportu- pace of improvement, with good recent prog- nity with a “penalty� for the share of access to ress in Balochistan and Sindh but decreases in opportunities distributed unequally.1 HOI in Khyber Pakhtunkhwa and Punjab. Opportunity is picking up. Figures 5 and 6 measure Improvement is slowing some. Table 4 provides a the education and sanitation opportunities for summary of all the indicators that could be children in Pakistan. Opportunity is growing in readily analyzed using the Social and Living both urban and rural areas, a very positive sign. Standards Measurement Survey. In general, the Urban children have more absolute opportunity annual pace of improvement in the indicators has than rural children, but the rate of growth in slowed over the latest period relative to 1998/99– rural areas is growing faster. The slowing rate of 2007/08, apart from the primary completion opportunity in sanitation is worrying, however, rate. The slippage is seldom much, but in sanita- as it can have implications for stunting. tion, gas, electricity, and primary enrollment the reduction is ­noteworthy—especially as Pakistan There is a mixed picture of opportunity by province. has not reached full coverage in these areas. Figures 7 and 8 provide the same information Rather than repeating similar figures for but for the other provinces. It is important to all the indicators, the interested reader can look at both the level of coverage and HOI as go directly to a website and explore interactive well as their growth to judge whether there may “dashboards.�2 be cause for concern. In primary completion, there appears to be a problem in Sindh. The Conclusions level is still low, and the rate of growth is low and showed little improvement in 1998/99– Pakistan has a rich set of data that suggest a decline 2007/08 and 2007/08–2010/11. By contrast, in poverty over the past decade. The data contain P a k i s t a n P o l i c y N o t e — R e c o v e r i n g s t r o n g p o s i t i v e t r e n d s i n p o v e r t y a n d o pp o r t u n i t y Figure Primary completion rates of children ages 15–19 5 1998/99 Human Opportunity Index National 2007/08 2010/11 1998/99 Urban 2007/08 2010/11 8 1998/99 Rural 2007/08 2010/11 0 25 50 75 100 Coverage rate (percent) 1998/99–2007/08 2007/08–2010/11 National Urban Rural 0 1 2 3 4 5 0 1 2 3 4 5 Annualized growth rate of Annualized growth rate of Human Opportunity Index (percent) Human Opportunity Index (percent) Note: The top panel shows the coverage rate for each year (shown by the bar) and the value of the inequality-adjusted coverage rate, the Human Opportunity Index (shown by the line in the bar). The closer the line to the end of the bar, the smaller the penalty. The two bottom graphs show the growth rate of the Human Opportunity Index, with the line in the bar showing the division between growth in coverage and growth in equality. Source: Pakistan Social and Living Standards Management Surveys 1998/99, 2007/08, and 2010/11. some concerns and need to be considered International comparisons suggest that Pakistan has with caution, but the data for Punjab are free been a good performer in turning growth into pov- of some of the other provinces’ anomalous erty reduction. Pakistan has demonstrated that results. The Household Income and Expendi- it can reduce poverty even at the relatively low ture Survey data for Punjab tell us that between rates of growth of 3.2–4.5 percent—but not at 2001/02 and 2007/08 poverty declined rapidly, growth of GDP per capita of 1 percent. Coun- especially for the rate of growth of real GDP tries that are more successful in reducing pov- per capita. Much of this decline was driven erty tend to be better at generating sustained by increases in returns in the nonfarm sector growth. The issue for Pakistan will thus be sus- in urban and rural areas, which reflects an taining growth. It has had only one period of increase in the relative price of labor. The rate four—and unlike many other developing coun- of poverty reduction, however, appears to have tries, never five—consecutive years of growth slowed since, and without population move- of GDP per capita in international $ PPP above ment from rural to urban areas, poverty would 3  percent. Because it is subject to stop-go have been worse. growth and to many natural disasters, it has to ensure a strong safety net program as part of Over 2001/02–2007/08, growth of mean per capita an overall poverty reduction strategy—like the consumption of the lowest 40 percent of the popula- Benazir Income Support Programme—though tion increased faster than real per capita GDP. When this is no substitute for sustained growth. growth in per capita GDP declined, growth rates in per capita mean consumption expendi- Results from the Household Income and Expenditure ture and poverty reduction also declined. Surveys data on trends in social indicators can be Figure Improved sanitation for children ages 0 –16 6 1998/99 Human Opportunity Index National 2007/08 2010/11 1998/99 Urban 2007/08 2010/11 9 1998/99 Rural 2007/08 2010/11 0 25 50 75 100 Coverage rate (percent) 1998/99–2007/08 2007/08–2010/11 National Urban Rural 0 2 4 6 8 10 0 2 4 6 8 10 Annualized growth rate of Annualized growth rate of Human Opportunity Index (percent) Human Opportunity Index (percent) Note: The top panel shows the coverage rate for each year (shown by the bar) and the value of the inequality-adjusted coverage rate, the Human Opportunity Index (shown by the line in the bar). The closer the line to the end of the bar, the smaller the penalty. The two bottom graphs show the growth rate of the Human Opportunity Index, with the line in the bar showing the division between growth in coverage and growth in equality. Source: Pakistan Social and Living Standards Management Surveys 1998/99, 2007/08, and 2010/11. used immediately to guide policy, though more work important implications for the pace of improve- is required to increase confidence in the poverty esti- ment in nutrition outcomes, because there is mates. The analysis for this note could be car- increasing evidence that sanitation plays a very ried out readily outside Punjab, as soon as some important role in affecting stunting—particu- potential problems are resolved. Until then, larly in South Asia. There has been some good some caution should be exercised in generat- news with improvements in gender equality— ing national figures on poverty. The problems and the HOI analysis provides a useful metric would be best addressed by an independent to gauge whether inequality in social indicators task force of experts consisting of representa- is narrowing. Finally, the last three years have tives from the Pakistan Bureau of Statistics and seen some sizable differences in the improv- from the community of poverty researchers. ing social indicators. Sindh has been lagging in its primary completion rates, and Khyber The HOI can be useful in identifying whether there Pakhtunkhwa has been lagging in coverage has been progress in coverage and equality of service of improved sanitation. Policy makers and provision. Analysis of level and growth together interested readers can explore the details of provides an indication of where increased performance of the social indicators through efforts are needed. The pace of improvement interactive dashboards that have been pre- in sanitation has declined. This is likely to have pared to complement this note (see References). P a k i s t a n P o l i c y N o t e — R e c o v e r i n g s t r o n g p o s i t i v e t r e n d s i n p o v e r t y a n d o pp o r t u n i t y Figure Primary completion rates of children ages 15–19, by province 7 1998/99 Punjab 2007/08 Human Opportunity Index 2010/11 1998/99 Sindh 2007/08 2010/11 1998/99 Balochistan Pakhtunkhwa Khyber 2007/08 10 2010/11 1998/99 2007/08 2010/11 0 25 50 75 100 Coverage rate (percent) 1998/99–2007/08 2007/08–2010/11 Punjab Sindh Khyber Pakhtunkhwa Balochistan 0 2 4 6 0 2 4 6 Annualized growth rate of Annualized growth rate of Human Opportunity Index (percent) Human Opportunity Index (percent) Note: The top panel shows the coverage rate for each year (shown by the bar) and the value of the inequality-adjusted coverage rate, the Human Opportunity Index (shown by the line in the bar). The closer the line to the end of the bar, the smaller the penalty. The two bottom graphs show the growth rate of the Human Opportunity Index, with the line in the bar showing the division between growth in coverage and growth in equality. Source: Pakistan Social and Living Standards Management Surveys 1998/99, 2007/08, and 2010/11. Figure Improved sanitation for children ages 0 –16, by province 8 1998/99 Punjab 2007/08 2010/11 1998/99 Sindh 2007/08 2010/11 1998/99 Human Opportunity Index Balochistan Pakhtunkhwa 2007/08 Khyber 11 2010/11 1998/99 2007/08 2010/11 0 25 50 75 100 Coverage rate (percent) 1998/99–2007/08 2007/08–2010/11 Punjab Sindh Khyber Pakhtunkhwa Balochistan –5 0 5 10 15 –5 0 5 10 15 Annualized growth rate of Annualized growth rate of Human Opportunity Index (percent) Human Opportunity Index (percent) Note: The top panel shows the coverage rate for each year (shown by the bar) and the value of the inequality-adjusted coverage rate, the Human Opportunity Index (shown by the line in the bar). The closer the line to the end of the bar, the smaller the penalty. The two bottom graphs show the growth rate of the Human Opportunity Index, with the line in the bar showing the division between growth in coverage and growth in equality. Source: Pakistan Social and Living Standards Management Surveys 1998/99, 2007/08, and 2010/11. Table Trends and average annual change in Human Opportunity Index 4 Average annual change in Human Opportunity Index (percent) 1998/99 2007/08 2010/11 1998/99–2007/08 2007/08–2010/11 Education Primary enrollment 47.8 66.6 68.5 3.7 1.0 Secondary enrollment 41.8 55.0 59.0 3.1 2.3 Primary completion 45.6 55.0 59.7 2.1 2.7 Secondary completion 18.6 24.5 26.6 3.1 2.6 Health Percentage of births assisted by traditional and formal birth attendants 71.1 80.2 80.6 1.3 0.2 Percentage of births assisted by formal birth attendants 12.7 30.4 37.4 9.7 6.9 Institutional birth 10.8 27.7 34.9 10.5 7.7 Any postnatal care 6.0 15.7 21.1 10.7 9.8 Ever immunized 76.6 94.1 95.3 2.3 0.4 Full immunization (card and self-reported) 43.2 68.9 71.0 5.2 1.0 Full immunization (card) 25.4 45.1 25.1 6.4 –19.6 Free of diarrhea in last 30 days 86.8 88.6 87.6 0.2 –0.4 Infrastructure Gas 7.0 14.2 15.9 7.9 3.8 Electricity 61.7 84.7 87.6 3.5 1.1 Sanitation 27.3 46.0 48.0 5.8 1.5 Water 71.5 84.6 82.5 1.9 –0.9 Source: Author’s calculations based on Household Income and Expenditure Surveys 1998/99, 2007/08, and 2010/11. P a k i s t a n P o l i c y N o t e — R e c o v e r i n g s t r o n g p o s i t i v e t r e n d s i n p o v e r t y a n d o pp o r t u n i t y Notes ———. n.d. Contribution National. Washington, The author would like to thank Olivier Dupriez DC. http://dataviz.worldbank.org/t/SASEP/ and Tefera Bekele for their help in preparing views/PakistanContributionNational/ the aggregate total consumption expendi- DshContributionNational?:embed=y. ture estimates, Hernan Winkler and Gabriela ———. n.d. Contribution Provinces. Washington, Inchauste for useful conversations on the micro DC. http://dataviz.worldbank.org/t/SASEP/ decompositions, and Minh Cong Nguyen for views/PakistanContributionProvinces/ his help in preparing this note. DshContributionProvinces?:embed=y. 12 1. Recently developed at the World Bank, the ———. n.d. Dashboard HOI National. Wash- HOI has been estimated for more than ington, DC. http://dataviz.worldbank. 20 countries in Latin America and Africa. org/t/SASEP/views/PakistanHOINational/ 2. Links to the dashboards can be found in DshHOINational?:embed=y; the References section of this note. ———. n.d. Dashboard HOI Provinces. Wash- ington, DC. http://dataviz.worldbank. References org/t/SASEP/views/PakistanHOIProvince/ Government of Pakistan. Various years. House- DshHOIProvince?:embed=y; hold Income and Expenditure Survey. Islamabad. ———. n.d. World Development Indicators. ———. Various years. Pakistan Social and Living Washington, DC. http://data.worldbank.org/ Standards Management Survey. Islamabad. data-catalog/world-development-indicators. World Bank. n.d. PovcalNet. Washington, DC. http://iresearch.worldbank.org/PovcalNet/ index.htm. © 2013 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street NW Washington, DC 20433 USA All rights reserved This report was prepared by the staff of the South Asia Region. The findings, interpretations, and conclusions expressed herein are those of the authors and do not necessarily reflect the views of the World Bank’s Board of Executive Directors or the countries they represent. The report was designed, edited, and typeset by Communications Development Incorporated, Washington, DC.