20070711MRavallion

New Evidence on the Urbanization of Global Poverty : 

New Evidence on the Urbanization of Global Poverty Martin Ravallion, Shaohua Chen and Prem Sangraula Development Research Group, World Bank

Development theory: Population urbanization is the main dynamic: 

Development theory: Population urbanization is the main dynamic Long tradition of theoretical models in which population urbanization drives both growth and distribution (Lewis, Kuznets, Harris-Todaro, Robinson, Fei-Ranis, Fields, Anand-Kanbur) “Kuznets process”: Key assumptions: The economy comprises a low-inequality and poor (low-mean) rural sector, and a richer urban sector with higher inequality. Migration process is such that a representative slice of the rural distribution is transformed into a representative slice of the urban distribution. => Absolute poverty will fall with urbanization; income inequality will rise up to some point then fall (inverted U)

Slide3: 

Inequality Urban population share Total poverty Rural poverty rate Urban poverty rate Poverty and inequality under the Kuznets process

How quickly is poverty in the world urbanizing? Is this a positive or negative force for overall poverty reduction?: 

How quickly is poverty in the world urbanizing? Is this a positive or negative force for overall poverty reduction? “The world’s poor once huddled largely in rural areas. In the modern world they have gravitated to the cities.” (Gerard Piel, 1997) Is that right? Is it good or bad news? A positive force in poverty reduction as rural workers take up more remunerative urban jobs? Or the unwelcome forbearer of new poverty problems as economic disadvantages become more geographically concentrated? => Five claims, four of which we test

Claim1: Most people live in rural areas, but this will soon change with urbanization : 

Claim1: Most people live in rural areas, but this will soon change with urbanization Data from the UN’s World Urbanization Prospects Differences in how “urban” is defined; technical differences + administrative/political influences “Urban-rural” distinction is becoming blurred. Both “migration” and “reclassification”; but how much? However, for this study we have little choice but to take the demographic data underlying Claim 1 as given.

Claim 2: The incidence of absolute poverty is lower in urban areas : 

Claim 2: The incidence of absolute poverty is lower in urban areas Past evidence is entirely from country-specific studies; e.g., World Bank’s Poverty Assessments However, different countries have different definitions of what “poverty” means and the differences are not statistically ignorable. Higher real poverty lines tend to prevail in richer countries, which tend also to be more urbanized. And the urban composition of the poor probably varies with the level of economic development and urbanization. => Biases in estimates of the global trends in the urbanization of absolute poverty => Biases in the x-country relationship with population urbanization and economic growth.

What does Claim 1 imply for the future validity of Claim 2? : 

What does Claim 1 imply for the future validity of Claim 2? Does urbanization of the population as a whole come with lower overall poverty? What about within sectors? Does population urbanization mean that the urban poverty problem has overtaken the rural problem in scale? Or will it overtake some time soon?

Claim 3: The urban share of the poor is rising over time Claim 4: The poor are urbanizing faster than the population as a whole : 

Claim 3: The urban share of the poor is rising over time Claim 4: The poor are urbanizing faster than the population as a whole Urban share of the poor: Pu=SuHu/H Claim 3: Pu rises over time Claim 4: Hu/H rises over time => growth rate of Pu > growth rate of Su

Why might Claim 4 hold?: 

Why might Claim 4 hold? This is what one expects when urbanization entails gains to the poor, ..but the gains are not large enough for all previously poor new urban residents to escape poverty. Thus the urbanization process puts a brake on the decline in urban poverty incidence, even when total poverty is falling. => urbanization is not a pure Kuznets process.

Simple model: rising urban poverty <=> falling overall poverty: 

Simple model: rising urban poverty <=> falling overall poverty

Claim 5: Urbanization is a positive force in overall poverty reduction: 

Claim 5: Urbanization is a positive force in overall poverty reduction Direct gains to migrants Indirect gains to rural poor Remittances Tighter rural labor markets And Claim 5 can be valid even when urban poverty is rising.

Data and methods: 

Data and methods

Key assumptions: 

Key assumptions 1. Consumption + additivity (“sub-group monotonicity”): Poverty is defined in terms of household consumption per capita. Additively separable poverty measures for which the aggregate measure is the (population-weighted) sum of individual measures. Relocation invariance: Simply moving people between urban and rural areas (or countries), with no absolute loss in their real consumption, cannot increase the aggregate measure of poverty. => Absolute poverty measures Poverty lines aim to have constant real value both between urban and rural areas and across countries

Cost-of-living (COL) differences: 

Cost-of-living (COL) differences Relatively weak internal market integration and the existence of geographically non-traded goods =>substantial cost-of-living differences between urban and rural areas in developing countries. Any assessment of the urbanization of poverty that ignored these COL differences would not be credible. Yet existing Purchasing Power Parity (PPP) exchange rates used to convert the international line into local currencies do not distinguish rural from urban areas. How can we construct an international urban-rural poverty profile?

Our approach: 

Our approach We use country-specific urban and rural poverty lines from the World Bank’s Poverty Assessments in setting the urban-rural differential in the international poverty lines. The fact that PA’s have now been completed for most developing countries makes this feasible. Otherwise, our methods closely follow the Bank’s past methods.* * Chen, Shaohua and Martin Ravallion, “How Have the World’s Poorest Fared Since the Early 1980s?” World Bank Research Observer, Fall 2004.

“$1 a day”: 

“$1 a day” For country-level analysis one should clearly use a poverty lines(s) appropriate to each country. However, for global poverty monitoring, the Bank has taken the position that to measure absolute consumption poverty on a consistent basis across countries one must use a common poverty line. But whose poverty line should it be? In the 1990 WDR, the Bank chose to measure global poverty by the standards of what poverty means in the poorest countries. Using rural lines when available => the “$1/day” line.

Slide17: 

Log poverty line at PPP Log consumption per capita at PPP $1 a day

Purchasing Power Parities: 

Purchasing Power Parities International poverty line is converted back to local currency at the base date (1985 originally; 1993 now) using PPP rates for consumption. EKS method of setting PPPs: a multilateral extension of standard bilateral Fisher price index Geary-Khamis (GK) method (Penn World Tables) uses quantity weights to compute the international price indices; too high a weight to consumption patterns in richer countries when measuring poverty globally. Since 2000 the World Bank’s global poverty and inequality measures have been based on the Bank’s PPPs, which use the EKS method.

Doubts about PPPs: 

Doubts about PPPs Continuing concerns about quality of PPP’s Incomplete ICP participation Differences in quality of goods Relevance to poverty Major new PPP efforts underway at the Bank to construct PPPs for the poor (Deaton). Preliminary results suggest little difference to standard consumption PPP using EKS method

Latest “$1 a day” poverty lines: 

Latest “$1 a day” poverty lines We chose the median poverty line of the lowest 10 lines => $32.74/month ($1.08 a day) at 1993 PPP for consumption. Regression based method gives $1.05 (95% CI: $0.88,$1.24) for poorest country. All numbers revised back in time to assure consistency. Note: Not valid to compare different sets of PPPs. For example, adjusting “$1” at 1985 PPP only for inflation in the US yields a poverty line in 1993 that is well above that found in low-income countries

Taking “$1 a day” to the surveys: 

Taking “$1 a day” to the surveys Poverty line in 1993 local currency is updated using local CPI (urban/rural for major countries) All estimates are our own, from primary data (unit record/specially designed tabulations). we do not rely on any secondary sources for the poverty and inequality measures Consistent methods across countries and time. Obvious comparability problems are eliminated However, there are comparability problems galore that can’t be readily fixed Income vs. consumption; recall periods; valuation This study: 208 surveys, 87 countries (95% of pop.)

Some remaining data problems : 

Some remaining data problems Urban-rural poverty line differential may vary by level of consumption/income National differential in middle-income countries is not then valid for the international $1 a day line Different countries have used different methods of setting their urban-rural poverty lines Cost-of-basic needs method is very common But differences in implementation Different countries have different definitions of “urban” Not clear that this is a bad thing. But there have been reports of political redefinition Can’t distinguish migration from re-classification

Results: 

Results

New urban poverty lines: 

New urban poverty lines Urban poverty line is about 30% higher than rural line, on average Regional differences Highest for LAC Lowest for EECA Poorer countries tend to have higher ratios of urban to rural poverty lines => Fig. Suggests weaker internal market integration in poorer countries.

Slide25: 

r=0.58

Global poverty measures for $1 a day: 

Global poverty measures for $1 a day

Rural poverty incidence is more than double that in urban areas: 

Rural poverty incidence is more than double that in urban areas The “$1 a day” rural poverty rate in 2002 of 30% is more than double the urban rate. Similarly, while we find that 70% of the rural population lives below $2 a day, the proportion in urban areas is less than half that figure.

Global poverty measures for $1 a day: 

Global poverty measures for $1 a day

Faster progress against rural poverty: 

Faster progress against rural poverty $1 a day

Rising number of urban poor; falling number of rural poor: 

Rising number of urban poor; falling number of rural poor 100 million fewer “$1 a day” poor = + 150 million fewer rural poor - 50 million more urban poor.

China is an unusual case: 

China is an unusual case China has an unusually large rural-urban poverty gap H($1)=22% in rural, 1% in urban Well-known urban-rural disparities in living standards But data problems (sample frame; “floating population”) Excluding China: urban share of the poor rises to 28% (from 24%).

Decomposition: 

Decomposition rural poverty reduction urban poverty reduction population urbanization

Are the poor urbanizing faster?: 

Are the poor urbanizing faster? “Yes,” for $1 a day but “no” for $2. For $1 a day, the share of the poor in urban areas rose by 0.3% points per year However, if we drop China from the calculation then Claim 4 is supported for both poverty lines. Country level estimates also support Claim 4. Let be the urban share of the poor = 0.30 (s.e.=0.07)

Is urbanization a positive force in poverty reduction?: 

Is urbanization a positive force in poverty reduction? Both the regional aggregates and the underlying country-level estimates are consistent with Claim 5. Polled data; Figures=> However, confounding latent regional/country effects Regression of poverty measures on urban population share with regional/country fixed effects: Possible biases due to time varying effects (correlated with urbanization) cloud the causal interpretation.

Slide35: 

$1 a day $2 a day

Slide36: 

Test for a Kuznets curve in distributional component of poverty: Controlling for a quadratic function of mean, no sign of a distributional effect of urbanization. Positive interaction effect; urbanization reduces the (absolute) growth elasticity of poverty reduction The main channel linking population urbanization to poverty is economic growth. Distributional effect?

Regional differences: 

Regional differences LAC has the most urbanized poverty problem and the steepest rise in urban share of the poor. East Asia has the least urbanized poverty. “Ruralization of poverty” in EECA; also China.

Regional differences cont.,: 

Regional differences cont., South Asia: Rural H=1.21 x Urban H East Asia: Rural H = 8.7 x Urban H (!) Due to China Urban poverty rate in China in 2002 that is barely 4% of the rural rate, while it is 86% for India. LAC and SSA: rising urban poverty + falling rural SSA: population urbanization (including the poor) has come with little reduction in aggregate poverty.

Some implications: 

Some implications

Slide40: 

For many countries, rising or stagnant urban poverty is the “other side of the coin” to what is in large part a poverty-reducing process of urbanization. by providing new opportunities to rural out-migrants some, but not all, of whom escape poverty in the process and through the second-round impacts of urbanization on the living standards of those who remain in rural areas (remittances + tighter rural labor markets) Africa stands out as an exception; further research needed on why. Rising urban poverty is consistent with pro-poor economic development

Slide41: 

Compared to Chen-Ravallion (2004): $1 a day poverty rate 2% points higher And falling at a slightly slower rate Chen-Ravallion underestimated poverty in a segment of the economy with below average rate of poverty reduction. Slower pace of poverty reduction than has been thought

On the future urbanization of poverty: 

On the future urbanization of poverty The latest WUP predicts that the urban share of the population of the developing world will reach 60% by 2030. If the urban share of the poor evolves consistently with our data then the share of the $1 a day poor in urban areas will reach 39% by 2030. Using Bocquier’s forecasts for urbanization, the urban share of the poor will reach 31% by 2030.* * Bocquier, Philippe, 2005, “World Population Prospects: An Alternative to the UN Model of Projection Compatible with Urban Transition Theory,” Demographic Research 12(9): 197 236.

Policy implications: 

Policy implications Too much or too little urbanization? None of this implies that governments should actively promote urbanization. Causality remains unclear. Consistency with the economic growth process is key Externalities: urban governments are typically answerable only to their urban constituents. A city government will probably devote too few resources to actions that yield external benefits to its rural hinterland. Indeed, some incumbent urban residents may expect to be worse off from policies that help rural migrants. It is not surprising that past urban policies have often ignored the needs of migrants and even burdened them with extra costs.

Take aways: 

Take aways Poverty appears to be even more rural than we thought. Urban poverty is falling more slowly than rural poverty, But that is not a bad thing, in general; it typically comes hand-in-hand with overall poverty reduction through urbanization. Economic growth is the main link between population urbanization and poverty reduction. No sign of a systematic distributional effect. Yes, the poor are gravitating to towns and cities, but more rapid poverty reduction will probably require a faster pace of urbanization, not a slower one. Development policymaking will need to facilitate this process, not hinder it.