Poverty In Pakistan - Trends Over Time

Transcription

1. POVERTY IN PAKISTAN : Trends over time 1. Poverty in Pakistan, 2001-2014 In 2001, some 14 percent of the population had a real per capita expenditure level of Rs. 550 per person. By 2014, this was down to just 2 percent. The poverty rate in Pakistan fell from a high of 34.7 percent in 2001-02 to about 9.3 percent in 2013-14. 2. Corroborating the Poverty Decline The Trend in Poverty between 2001 and 2014 40% The credibility of the poverty decline has been verified using a number of criteria, including looking at independent surveys done in the same time period, as well as improvements in other measures of household welfare. 35% 30% 25% 20% 15% 10% 5% The Pakistan Panel Household Survey (PPHS), which tracked the same households in rural Punjab and Sindh from 2001-2010 shows a decline in poverty in rural Punjab and Sindh which is close to the HIES 0% 2001-02 2004-05 2005-06 2007-08 2007-08 base CPI from 2011-12 2010-11 2011-12 2013-14 2007-08 base CPI from 2010-11 Source: HIES and World Bank Staff calculations. The gains were shared with everyone—even the poorest 5 percent of the population saw real improvements. Poverty Headcount in Rural Punjab and SindhHIES and PRHS (percent) 60 49.5 45.5 50 36.0 40 27.5 23.2 30 20 30.2 24.9 18.9 2001 27.4 22.6 14.4 16.6 2004 2010 10 0 Punja b PRHS Sindh PRHS Punja b HIES Sindh HIES Source: HIES and World Bank Staff calculations. Note: The vertical line represents the poverty line set in 2001-02, equal to Rs. 723.4 per month per adult equivalent. Source: HIES and World Bank Staff calculations. When poverty declines, it usually goes hand in hand with other gains in household welfare. Pakistan saw substantial gains in a number of areas. There was a substantial increase in the ownership of relatively expensive assets like motorcycles and refrigerators among the poorest 20 percent. Looking at the distribution of households below the poverty line clearly shows this gain. National .1 .2 40% 35% 30% 25% 20% 15% 10% 5% 0% 0 Cumulative Density .3 Asset Ownership in the Bottom Quintile (Percent Who Own) 100 150 200 250 300 350 400 450 500 550 600 Real Per Adult Equivalent Food Expenditure 2001-02 650 700 2001-02 2004-05 2007-08 2010-11 2011-12 2013-14 750 2013-14 Refrigera tor Stove TV Motorcycle Source: HIES and Survey-Weighted CPI (base change-2011-12) Source: HIES and World Bank Staff calculations 1
2. Housing quality improved . Homes with some type of latrine (pit or flush connected to a drain) increased considerably Housheold Toilet Type in the Bottom Quintile In Pakistan, the reduction in poverty led to an increase in dietary diversity for everyone. For the poorest, the share of expenses devoted to milk and milk products, chicken, eggs and fish, as well as vegetables and fruits rose. In contrast, the share of cereals, which provide the cheapest calories, declined steadily between 2001-02 and 2013-14. 60% 50% Engel Curve for 2001 and 2014 Real per adult equivalent Food Expenditure (2001-02 rupees) 40% 30% 20% 10% 0% 2001-02 2004-05 2007-08 2010-11 2011-12 2013-14 Flush La trine Other Source: HIES and World Bank Staff calculations. And, school enrollment rates, particularly among girls increased considerably during the period of the sharpest poverty decline, from 2001-02 to 2007-08. School Enrollment Rates (Percent) 2001-02 2007-08 2013-14 Elementary school (age 5-9) Gross enrollment ratio Boys 83 97 98 Girls 61 83 81 Total 72 91 90 Net enrollment ratio Boys 46 59 60 Girls 38 52 53 42 55 57 Total Source: Pakistan Bureau of Statistics PSLM reports Changes in consumption patterns over time are also consistent with the poverty decline: Increases in income are strongly associated with households spending less of their budget on food, and more on non-food items. 1600 1400 1200 1000 800 2001-02 600 2013-14 400 200 0 0 1000 2000 3000 4000 5000 6000 7000 Source: HIES and World Bank Staff calculations. Since foods like chicken, eggs, vegetables, fruits, and milk and milk products are more expensive than cereals and pulses, and have lower caloric content, this shift in consumption increased the amount that people spend per calorie over time. For the poorest quintile, expenditure per calorie increased by over 18 percent between the years 2001-02 and 2013-14. 2. What caused the Poverty Decline? Growth in GDP While there is some question about the extent to which measured GDP captures economic activity in the country, GDP growth was quite high in the initial years of the 2000‘s, when poverty declined the fastest. Growth Rate of GDP and Poverty Headcount (%) 40 9 8 7 6 5 4 3 2 1 0 35 30 In Pakistan, the 25 percentage point decline in poverty between 2001-02 and 2013-14 was associated with a 10 percentage point reduction in the share of expenditure devoted to food. 8000 Real per adult equivalent Total Expenditure (2001-02 rupees) 25 20 15 10 5 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Increases in income are also associated with households moving towards more balanced and diverse dietary patterns. Poverty Headcount (Primary Axis) Annual GDP growth (Secondary Axis) Source: World Bank data (data.worldbank.org/country/pakistan) However, measured GDP growth is unlikely to have produced this kind of a continuous decline. 2
3. These issues need to be carefully examined in order to understand both the key determinants of the decline in poverty thus far , and the prospects for a continued robust decline in poverty looking ahead. Remittance flows There has been a dramatic increase in remittances in Pakistan over the past 10 to 15 years. Measured in current US dollars, remittance inflows increased from 1.5 billion dollars in 2001 to about 17.1 billion dollars in 2014. Remittance Inflow (billion dollars-current USD) 18 17.1 14.6 16 14.0 14 12.3 3. Critiques of the Poverty Decline Two sets of concerns regarding the declining trend in poverty which have been raised by various analysts require a response: Does the declining food share represent a possible ‗food squeeze‘ rather than an increase in wellbeing? And could the decline be due to the use of the CPI to inflate the poverty line? 12 9.7 10 8.7 7.0 8 i) Does the declining food share represents a possible ‗food squeeze‘ rather than an increase in wellbeing? 6.0 5.1 6 4.3 3.6 4.0 3.9 4 2 1.7 1.3 1.7 1.2 1.0 1.1 1.5 0 This argument stems from the lack of improvement in nutritional outcomes in Pakistan. The question often raised in this context is: If poverty has declined why has there been so little improvement in child stunting and wasting indicators? Source: http://www.migrationpolicy.org/programs/data-hub/global-remittances-guide Pakistan became the seventh largest remittance receiving country in the world in 2014, with remittances accounting for some 7 percent of GDP. As remittances tend to go toward poorer families, they help move households up the income distribution. ‘Hidden’ urbanization and growth of the Informal economy Many secondary and tertiary cities have sprung up in the rural periphery and, with them, the informal economy has burgeoned. This needs to be better captured in national data, including the GDP-and is likely an important source of the reduction in poverty. An important indicator is the lack of change in the share of the employed in the rural economy combined with the reduction in male participation in agricultural work. Percentage Distribution of Employed persons age 10 and above Percent Employed Total Male Female 2014-15 Agriculture 42.47 25.38 16.88 Non-Agriculture 57.73 51.38 6.35 Total 100 76.76 23.24 2001-02 Agriculture 42.09 32.63 9.46 Non-Agriculture 57.91 52.73 5.18 Total 100 85.36 14.64 Source: LFS various years This line of argumentation assumes that nutritional deficits are due to a lack of adequate food, both in terms of quantity and quality. If this were the case then any improvement in income would immediately translate into improved nutrition and would show up in a reduction in the proportion of young children that are stunted or wasted. Before turning to the data to look at evidence for a ―food squeeze‖ it is important to note that the link between food availability and nutritional status is weak at best, and is mediated by the ambient disease environment and the quality of water and sanitation. There is now a considerable body of research showing this. While nationally representative data on water quality is not yet available, data from select rural districts in Punjab, Sindh and KP, shows that between 1/3rd - 2/3rd of drinking water in rural areas is contaminated with Ecoli and unfit for drinking. Sanitation and associated health behaviors also remain quite poor, and are much worse in relatively poorer communities. Without a substantial improvement in these critical public services and much greater awareness about detrimental heath behaviors at the household and community level, it is unclear that a rise in the income of the poorest will substantially improve nutritional outcomes. 3
4. Turning now to the issue of food availability , several check are possible. The evidence on the increase in dietary diversity already indicates that food shortages are an unlikely cause of stunting in children. Important to note that the same pattern of decline in measured calories is also evident in other developing countries which have seen a decline in poverty, including, in particular, neighboring India. Another check could be see if the basket of food consumed in 2001 became more or less affordable by 2014. It is clear in the table below, that the share of the expenditure required to purchase the 2001 food basket actually declined in real terms. For the bottom 20 percent, it fell from 55 percent to 45 percent, while for the top quintile it fell from 42 to 34 percent. ii) Could the decline be due to the use of the Consumer Price Index to inflate the poverty line? Percentage of Total Expenditure Required to Purchase the Average Food Basket Consumed in 2001-02 2001-02 2007-08 2013-14 National Bottom Quintile 55.32 48.87 45.45 Top Quintile 41.98 33.84 31.12 Urban Bottom Quintile 50.56 42.03 38.78 Top Quintile 32.57 26.08 24.38 Rural Bottom Quintile 57.04 49.78 47.04 Top Quintile 48.46 42.63 39.54 Source: HIES and World Bank staff calculations. What is an apparent puzzle, however, is that measured calories in the HIES decrease over time—but only in rural areas and across all income quintiles-with the wealthiest rural quintile decreasing their calories the most. Further, the only items for which calories declined were cereals, mainly wheat and pulses, which are also the cheapest foods. Rural households have apparently chosen to move away from cheap calorie dense foods towards more nutritious and lower calorie foods—and their consumption patterns are also becoming more closely aligned to those of urban households. What is more likely is that the decline in measured calories in rural areas is due to the increasing importance of food consumed away from home – a category of expenditure that is currently poorly captured in the HIES. This is also consistent with the shift of male labor out of agriculture as seen above. Non-agricultural wage workers tend to purchase their lunch time meal, and roti and daal is the most common lunchtime meal among workers! In short, there is no credible evidence of a ‗food squeeze‘ or a decline in total calorie consumption. The argument is that the CPI underestimates rural inflation and thus overall inflation- thereby exaggerating the poverty decline. In some instances, the quality of price data collected in the CPI has also been questioned. Both can be checked to a degree. The quality of price data can be checked, but only for food items, using unit values (the best available price proxy) reported in the HIES or other surveys, such as the PPHS. A comparison of these for urban areas, with reported CPI prices in the inflation monitor for the year 2010 show a close match with CPI price data, indicating that the quality of CPI price data is adequate. However, the CPI has a clear ―urban bias‖ due to the omission of rural markets. Due to this, the CPI will underestimate (overestimate) inflation in periods when rural inflation exceeds (is lower than) urban inflation. In the period under review, rural food inflation exceeded urban food inflation – particularly between 2007-08 and 2010-11—so using the CPI to inflate the poverty line overestimated the decline in poverty in this period. Correcting for urban bias in the CPI dampens the pace of poverty reduction by between 6 to 8 percentage points by 2013-14. While this is significant, it does not overturn the trend in poverty. Pakistan still more than halved its poverty rate between 2001 and 2014. 4. The impact of improvements in the CPI and in the HIES to account for food consumed away from home on the trend in poverty going forward Capturing food consumed away from home more comprehensively will likely result in higher measured expenditure and higher calorie consumption. This will contribute to a faster than currently measured decline in the poverty rate. Including rural markets in the CPI will dampen the impact of higher expenditure on the poverty rate, to the extent that the rural inflation rate exceeds the urban inflation rate. For these reasons, and for the sake of continuity, the government will continue to track and report trends for both the old and the newly-adopted poverty line, and the 4
5. poverty rate with and without the changes in the HIES and the CPI for at least the next 5-6 years . 5