WDR ON POVERTY AND DEVELOPMENT 2000/01 Stiglitz Summer Research Workshop on Poverty, Washington DC ARE THE INTERNATIONAL DEVELOPMENT TARGETS ATTAINABLE? Lucia Hanmer and Felix Naschold July 6 July 8, 1999
Work in progress (May 1999) (Please do not cite without the prior consent of the authors) Comments welcome. Are the International Development Targets Attainable? Lucia Hanmer and Felix Naschold 1 l.hanmer@odi.org.uk, f.naschold@odi.org.uk Overseas Development Institute Portland House Stag Place London SW1E 5DP 1 The authors would like to acknowledge the support of the Department for International Development, who commissioned the research on which this paper is based. We would like to thank Paul Mosley, John Healey and Simon Maxwell for their comments. The usual disclaimer applies. 1
1. Introduction In 1996 the Development Assistance Committee (DAC) of the OECD announced a "global development partnership effort" to reduce poverty and increase human development by 2015 (OECD, 1996). The success of this effort will be judged with reference to the attainment of a set of goals selected from the series of UN conferences held in the 1990s. These International Development Targets (IDTs) are shown in Box 1. This article assesses the likelihood that these targets will be met. Section 2 outlines the methods used to make this assessment. Section 3 presents our results for the income poverty targets and section 4 our results for the human development targets. We discuss some caveats to our findings in section 5 and present our conclusions in section 6. Box 1 The International Development Targets 1. The proportion of people living in extreme poverty in developing countries should be reduced by at least one-half by 2015. (Copenhagen 1995) 2. There should be universal primary education in all countries by 2015. (Jomtien 1990, Beijing 1995, Copenhagen 1995) 3. Progress towards gender equality and the empowerment of women should be demonstrated by eliminating gender disparity in primary and secondary education by 2005. (Cairo 1994, Beijing 1995, Copenhagen 1995) 4. The death rates for infants and children under five should be reduced in each developing country by two-thirds the 1990 level by 2015. (Cairo, 1994) 5. The rate of maternal mortality should be reduced by three-fourths by 2015. (Cairo, 1994, Beijing 1995) 6. Access should be available through the primary health-care system to reproductive health services for all individuals of appropriate ages, no later than the year 2015. (Cairo 1994) 7. There should be a national strategy for sustainable development, in the process of implementation, in every country by 2005, so as to ensure that current trends in the loss of environmental resources are effectively reversed at both global and national levels by 2015. (Rio 1992) 2
2. Estimating poverty and human development in 2015 Our method here is the same as Hanmer et al (1999) which estimated future levels of income poverty and human development (HD) indicators using forecasts of economic growth for the major developing regions and elasticities of poverty and human development with respect to real per capita GDP growth. The poverty (HD) elasticity shows the percentage decrease in the poverty headcount ratio (improvement in the HD indicator) resulting from a one per cent increase in per capita GDP. The poverty elasticities used to make the projections are calculated econometrically using cross section data measuring the percentage of population living on less that $1 per day at 1985 PPP and the level of real per capita GDP. The data are taken from Chen Datt and Ravallion (1994) and World Bank (1998a). The HD elasticities are estimated econometrically using cross section data from the early 1970s and 1993, taken from the Human Development Report (UNDP, various years) and the World Bank (1998a). Hanmer et al (1999) found that whether the income poverty targets can be met globally or in particular developing regions depends critically on the pattern as well as the rate of growth. Targets are met in all regions, except sub-saharan Africa, when high or medium levels of income equality accompany growth. The prospects for achieving the human development targets are good for most regions. This paper updates and extends the analysis and results presented in Hanmer et al (1999) by estimating the effects of the following factors on the likelihood of meeting the IDTs: the effect of the financial crisis in S.E Asia and E. Europe on poverty and growth; the impact of the HIV/AIDS epidemic on mortality; the effect of policy variables and growth strategies on poverty reduction; and, The effect of social sector development on the human development indicators. It uses a larger, updated data set on poverty, income inequality and GDP per capita that has recently become available from the World Bank (World Bank, 1998a) 3
3. Reaching the income poverty targets We extend our analysis of the implications of the characteristics of the growth path for poverty reduction by examining the proposition that broad-based (labour intensive) growth reduces poverty more effectively than other types of growth. Strategies that favour employment creating growth have the potential for being propoor as long as labour markets clear at an equilibrium real wage that is higher than the poverty line. Adopting an employment creating growth strategy has, therefore two aspects: (a) labour intensity and (b) labour productivity (because in the long run real incomes/ wages can only increase as fast as productivity). A labour intensive growth strategy implies creating maximum scope for substitution of labour for other factors of production. This means: encouragement of activities in sectors that have flexible technical coefficients of production; maximising backward and forward linkages between the large scale and the small scale, the formal and informal sectors; Fostering a price system and an investment strategy that gives maximum scope to these sectors, including creating a financial sector that can capitalise them. Flexible coefficient sectors can (but may not necessarily) include agriculture, rural non-farm activities, construction, export processing/ manufacturing and services. Whether these activities are carried out in a labour intensive way will depend on the institutional setting (e.g. the land tenure system, the financial sector) and the structure of economic incentives in any particular country setting. A labour productivity increasing strategy implies a key role for investment growth. There are limits to the extent that labour can be substituted for other factors of production. Growth of the capital stock and other assets that are complementary to labour need to keep up with labour force growth. 4
Research has also established the importance of the government s policy stance in creating the economic environment for growth to be pro-poor. (See for example Collier and Dollar (1998) or Wade (1990)). Maintaining competitive exchange rates and an open trade regime, fostering domestic price stability and controlling deficits on the balance of payments and the government s budget are generally considered to be important preconditions for economic growth. We use a dummy variable developed by Sachs and Warner (1995) to capture these effects. The Government Policy 2 dummy indicates the degree of interventionism in the economy by the government. The dummy is equal to one if the black market exchange rate premium is larger than 20% for at least ten years, and/ or there is a public sector export monopoly for crops, and/ or the country is classified as socialist, and/ or more than 40% of customs import code lines are affected by some sort of quantitative restrictions. Otherwise the dummy equals zero otherwise. Hence government policy is good if the dummy equals zero and bad if the dummy equals one. Income Inequality The updated and expanded poverty data set confirmed that the poverty elasticities vary systematically with income inequality. Pooling observations from all years for all countries we found that: when inequality is low (gini< 43) : the poverty elasticity = -0.93 when inequality is high (gini>43) : the poverty elasticity= -0.34. If real per capita GDP increases by one per cent the poverty headcount ratio will fall by 0.93% if income inequality is low. It will only fall by 0.34% if income inequality is high. The systematic difference in poverty elasticities holds at the regional level for Sub-Saharan Africa, Latin America and the Caribbean. 3 2 Sachs and Warner (1995) call it the Openness dummy. 3 Numbers of observations limited testing for systematic differences in poverty elasticities in other regions. 5
Growth strategy We were not able to directly test whether countries that use more labour intensive production techniques have greater poverty reduction for a given growth rate than those that use less labour intensive ones. Our regression results did, however, show that for a given level of real per capita GDP: poverty is lower when investment grows faster than the labour force; poverty is lower the more efficiently capital is used (the incremental capital output ratio is positively correlated with the poverty headcount); and in sub-saharan Africa poverty is lower when value added per worker in agriculture is higher than modern sector value added per worker. The first two results confirm the importance of investment for poverty reduction and are consistent with a growth strategy that results in growing labour productivity and increasing real wages. We interpret the third result as support for the poverty reducing capacity of a broad-based growth strategy. Agriculture provides the bulk of employment, output and export revenue in the majority of sub-saharan African countries. It can be a flexible coefficient sector, capable of absorbing labour in productive activities and, as most of the poor live in rural areas in sub-saharan Africa, it has high poverty reducing potential. The ratio of value added per worker in the modern sector to value added per worker in agriculture will reflect the productivity ratio of these two sectors. So higher agricultural than modern sector value added per worker fits the preconditions for structural change observed in the advanced capitalist economies at the beginning of the century and the South East Asian economies in the 1960s (c.f. Maddison, 1982). Higher real incomes in the agricultural sector stimulate demand for the products of other sectors and labour within the sector. Higher agricultural outputs stimulate the creation of non-farm rural and urban employment opportunities, through backward and forward linkages to service and manufacturing sector activities. During the next stage of development the supply side effects of increasing agricultural productivity lead to a shift in sectoral employment out of agriculture into industry and services. 6
Government policy Our regression results show that, for a given level of GDP, poverty reduction is improved if government policies fit the Sachs and Warner definition of good policies. So: good government policies reduce poverty. Poverty in 2015 We project income poverty in 2015 using two growth scenarios and two policy scenarios. The growth rates are shown in Figure1. Figure 1: Growth Scenarios - Real GDP per capita Developing Countries Eastern Europe and Central Asia Latin America and Caribbean South Asia East Asia & Pacific Middle East and North Africa Sub Saharan Africa 0.0 1.0 2.0 3.0 4.0 5.0 6.0 average annual % growth Global Economic Prospects December 1998 Projections for 2001-2015 Historic Growth Rates 1965-1997 The High Growth Scenario: is optimistic and uses high growth rates produced by the World Bank (1998b). The Low Growth Scenario: projects economic growth at historic average, achieved between 1965 and 1997 (World Bank 1999). The Low Inequality policy scenario assumes good government policy and a broad-based growth strategy. The High Inequality policy scenario assumes no additional beneficial effects from government policy or the growth strategy. Table 1 shows the projected level of poverty in the world developing regions in 2015. 7
Table 1. Poverty incidence (%) in 2015 in the world s developing regions ($1 a day at 1985 PPP). Growth Scenario Policy Scenario High Low Poverty incidence in 2015 High Inequality 19 20 Low Inequality 9 11 Poverty incidence in 2015 as % of 1990 poverty High Inequality 64 66 Low Inequality 29 (yes) 36 (yes) Note: (yes) indicates target has been hit. Table 1 shows that the target of halving extreme poverty by 2015 is met at the global level if, and only if, low levels of income inequality accompany growth. In the low inequality scenario poverty falls to between 29 and 36 per cent of its 1990 level, depending on whether growth is high or low. The projections of future poverty rates are more sensitive to the assumptions about policy scenarios than growth rates, as Figure 2 shows. Figure 2 Global Poverty in 2015 the effect of income inequality 30 Percentage of population under $1/day 25 20 15 10 5 1990 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Table 2 gives a regional break down of the prospects for meeting the IDTs. 8 2008 2009 2010 2011 2012 2013 2014 2015 high inequality low growth high inequality high growth low inequality low growth good policies low inequality high growth good policies
The best case scenario is the high growth rate, low inequality with good government policy and a broad based growth strategy. The worst case scenario is the low growth rate high inequality scenario. Table 2 Poverty incidence (%) by region in 2015: best and worst case scenarios; ($1 a day, at 1985 PPP). 1990 2015 Best Case Worst Case Sub Saharan Africa 39.3 23.1 38.8 Middle East and North Africa 4.3 2.7 4.1 East Asia & Pacific 28.5 6.0 (yes) 14.8 (yes) South Asia 43 14.6 (yes) 33.2 Latin America and Caribbean 23 10.6 (yes) 20.9 Eastern Europe and Central Asia 2.1 1.4 3.4 Developing Countries 29.6 9.5 (yes) 21.1 Note: yes indicates the target has been attained. What, then, has been the impact of the financial crisis on achieving the international development income poverty target? Figure 3 shows that it has set back the achievement of the target (approximately 14% poverty incidence) by about 4 years. Figure 3 Poverty incidence in East Asia Pacific - Comparison of pre and post Asian crisis projections 30 25 % of population under $1/day 20 15 10 1999 2003 5 0 1990 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 9 Post Crisis Pre Crisis
4. Reaching the human development targets Hanmer et al (1999) established that improvements in human development indicators are correlated to the rate of growth of real per capita GDP and to technological progress, which is independent of national rates of per capita GDP growth. 4 We extend the analysis by including the effect of: HIV/AIDS on infant, under five and maternal mortality (prevalence rates of women attending ante-natal clinics (UNAIDS 1998)), And the effect of the following policy variables: expenditure on health services (physicians per thousand population) access to maternal health services (births attended by trained personnel) school expenditure per student We use physician per capita to illustrate the effect that greater health sector expenditures can have on infant, under-five and maternal mortality. Health sector expenditures are not necessarily representative of the quantity and availability of health services. Money can be inefficiently spent and some health care systems are more expensive than others are and so the cost of the same health services varies widely between countries. An output rather than an input indicator of health services can better capture the effect of increasing health expenditures. A similar caveat holds for education, but there are no internationally comparable data on relevant output indicators (e.g. pupil teacher ratios). Mortality targets We found that HIV/AIDS is strongly and positively correlated with levels of infant child and maternal mortality. The quantity and availability of health services (as proxied by the number of physicians per thousand population and the number of births attended by trained personnel) is negatively correlated with infant and child mortality. 4 We found that between 1970 and 1990 the intercept of the regressions of real GDP per capita on life expectancy and adult literacy respectively shifted upwards. Hence at 10
We project future infant mortality rates using the high growth rates and two scenarios: The Better Health Scenario keeps HIV/AIDS infection rates at their present level and assumes that the availability of health services per capita increases and that technological progress continues to have effects in the 21st century. The AIDS Pandemic Scenario projects HIV/AIDS infection rates spreading throughout all developing regions at the rate predicted by epidemiological models. All regions have adult HIV/AIDS prevalence of 22.5% by 2015. The benefits of technological progress are negated by the spread of AIDS in this scenario. Table 3 shows that infant mortality remains at 76% of its 1990 level in the AIDS Pandemic Scenario. The target of reducing infant mortality by two thirds is, however, met in the Better Health Scenario. The same conclusion holds for the under-five mortality target. Table 3. Prospect for meeting the infant mortality rate target. Infant mortality for all developing countries, population weighted average. 2015 1990 Better Health AIDS Pandemic Infant Mortality (per thousand live births ) 56 18 42 Infant Mortality as a percentage of its 1990 level 33 (yes) 76 Note: yes indicates target is hit. Figure 4 shows the breakdown by regions. Even with the spread of AIDS infant mortality falls in all regions apart from the Middle East and North Africa. However if the Better Health Scenario prevails the infant mortality target is met in all regions except for sub Saharan Africa where it remains at 56% of its 1990 level. (In S. Asia the infant mortality rate falls to 34% of its 1990 level in the Better Health Scenario). the same level of per capita GDP life expectancy (literacy) was higher in 1990 than in it was in 1970. 11
Figure 4 Infant Mortality Rates Better Health 140 120 Infant Mortality Index (1990=100) 100 80 60 40 SSA MENA EAP SA LAC EEUCIS 20 0 1990/3 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 The Universal Primary Education Target Achieving universal primary school enrolment by 2015 is the most easily attainable human development target as most regions had primary school enrolment rates of over 90% in 1990. Table 4 shows the 1990 levels and 2015 projections of the primary school enrolment rate for the high and low growth scenarios by region. Table 4 Primary school enrolment rates in 2015, by region 1990 2015 High growth Low growth Sub Saharan Africa 78 92 89 Middle East and North Africa 103 118 115 East Asia & Pacific 129 158 157 South Asia 94 116 113 Latin America and Caribbean 110 129 126 Eastern Europe and Central Asia 94 116 113 Developing Countries 107 131 129 12
Universal primary school enrolment is reached in 2015 in both growth scenarios and all regions except for sub-saharan Africa. An important caveat must be kept in mind. Enrolment does not ensure education. Our country case study of the Universal Primary Education (UPE) policy in Uganda shows that quality concerns can be critical. Teacher pupil ratios have soared to 70:1 in government schools with UPE and the challenge facing the Government of Uganda is to ensure that all pupils leave primary school literate and numerate and that the overall standard of education does not fall (McGee et al, 1999). Our case study from Tanzania shows that deteriorating standards of education lead to education becoming highly price elastic. The recent imposition of a T Sh 1000 charge per year (about 1) led to falling enrolments as parents would not pay something for nothing (Eele et al, 1999) Gender equality targets Using the ratio of female to male primary and secondary enrolment, we found that there is only a very weak correlation between gender equality in school enrolment on the one hand and per capita income and income inequality on the other. The model explains only about 20% of the variation in the data for primary and secondary school enrolment. We do not consider our results strong enough to use as the basis for projections, especially as the correlation with GDP per capita is so weak. Intuitively this makes sense, as the role of women, and hence gender equality, very much depends on local culture and customs. Evidence from the country studies on the attainment of the IDTs shows that policies to promote women s economic and political empowerment, rather than economic growth are the critical determinants of advances in gender equality. Such polices need to address the overall position of women in society if the incentive system is to be changed so that parents want to send their girl children to school and that girls want to remain there and have the opportunity to succeed when they do. 5. Some Caveats 13
Our projections are based on sustained growth rates, yet this is rarely the case in many low-income countries. Our case studies of Tanzania and Uganda (Eele, 1999 and McGee, 1999) revealed that it is the extreme variability in growth rates, often due to exogenous shocks such as drought and civil strife, that policy makers find most threatening to the achievement of the IDTs. The capacity of the state to tax and spend efficiently is also an important component of a broad-based growth strategy. The case studies of Uganda and Tanzania point to the importance of public investment in infrastructure to enable rural communities to be linked to markets. In Pakistan poverty reduction and growth in the 1970s and 1980s was associated with high levels of public development expenditure that benefited the agricultural sector (Zaidi, 1999). The size and allocation of public investment can thus be an important component of a broad-based growth strategy. Gini coefficients tend to remain roughly similar between time periods in stable economies. However significant structural changes can remove redistributive mechanisms and produce rapid changes in income distribution, which are difficult to reverse. The case study of Pakistan gives a recent example of exogenous shocks leading to large changes in income equality. By the end of the 1980s remittances from Pakistani workers had reduced substantially as changes in the price of oil reduced the demand for expatriate labour in the Middle East. Remittances had a strong poverty reducing effect as they were used for investment in family based rural and urban small-scale enterprises. Hence an important redistributive mechanism was removed with the contraction of demand for labour in the Middle East. Another recent example of rapid changes of income distribution due to the removal of redistributive mechanisms is the experience of the transition economies. For example, gini coefficients in Russia, Bulgaria and Ukraine rose from 0.23-0.26 to 0.40. In ten years these economies have gone from being among the world s most equal to high inequality economies and new and effective mechanisms to redistribute to the poor are proving difficult to find (Mosley, 1999). Finally the recent stress on the multidimensional aspects of poverty should be borne in mind. Income poverty is one manifestation of being disadvantaged. Recent research has shown that non-economic dimensions are as important. Powerlessness, isolation, 14
lack of political and economic freedoms and lack of autonomy and the gendered nature of poverty experiences are dimensions of poverty that are frequently stressed by the poor. There is no reason why achieving the IDTs should not create an environment where poverty in its broadest sense is reduced. But targets should be seen as a means to an end rather than an end in themselves. Too narrow a focus on targets could mean that in the process of achieving them the voice of the poor remains ignored. 6. Conclusions The attainment of IDTs is conditional. Halving income poverty by 2015 is conditional on low-income inequality growth paths. Reducing infant and under-five mortality rates by two-thirds is conditional on halting the spread of HIV/AIDS, increasing the capacity of developing countries health sectors to deliver more health services and ensuring that technological progress spills over to benefit the developing world. Attaining universal primary education is conditional on maintaining the quality of teaching in primary schools as enrolment becomes universal. Attaining gender equality in primary and secondary education depends critically on policy measures that promote the political and economic empowerment of women. The caveats that: growth in poor countries often fluctuates wildly; the government s capacity to invest is important; income inequality can rise quickly and rapidly in unstable economies; and, Not all aspects of poverty reduction can be quantified are further conditionalities that effect the attainment of the IDTs and their consequences. 15
References Chen, S., G. Datt and M. Ravallion (1994) Is Poverty increasing in the developing world? Review of Income and Wealth, 40 (4):359-376. Collier, P. and D. Dollar (1998) Aid allocation and poverty reduction, World Bank, unpublished (summary presented by P. Mosley at Reading DSA conference) Eeele, Graham et al (1999) Meeting the International Development Targets in Uganda. Paper presented to the meeting the International Development Targets Workshop, ODI, London; 26-27 th April, mimeo. Hanmer, L., N. de Jong, R. Kurian and J. Mooij (1999) Are the DAC targets achievable? Poverty and Human Development in the Year 2015. Journal of International Development, forthcoming. Hanmer, L. and H. White (1999) The impact of HIV/AIDS on under five mortality in Zambia and Zimbabwe in Human Development in Sub-Saharan Africa: The Determinants of Under-Five Mortality, ISSAS: The Hague, January. Maddison, Angus (1982) Phases of Capitalist Development Oxford:OUP McGee, Rosemary et al (1999) Meeting the International Development Targets in Uganda Paper presented to the Meeting the International Development Targets Workshop, ODI, London; 26-27 th April, mimeo. Mosley, Paul (1999) Are Poverty and social goals attainable in the Transition region? Paper presented to the Meeting the International Development Targets Workshop, ODI, London; 26-27 th April, mimeo OECD Development Assistance Committee (1996) Shaping the 21 st Century: The Contribution of Development Cooperation. OECD/DAC, Paris, May. Sachs, J. and A.M. Warner (1995) Economic Convergence and Economic Policies, National Bureau of Economic Research, Inc. Working Paper Series 5039. UNDP (various years) Human Development Report, New York: UN. Wade R (1990) Governing the Market World Bank (1998a) World Development Indicators CD ROM. World Bank (1998b) Global Economic Prospects for Developing Countries 1998/99: Beyond Financial Crisis, Washington D.C.: World Bank, December. World Bank (1999)World Development Indicators. Zaidi, Akbar (1999) Is poverty now a permanent phenomenon in Pakistan? Paper presented to the Meeting the International Development Targets Workshop, ODI, London; 26-27 th April, mimeo 16