The Incidence of Child Health Improvements

• Published date: 01 May 2017
• DOI: http://doi.org/10.1111/rode.12262
• Author: Stephen D. Younger,David E. Sahn
• Date: 01 May 2017

The Incidence of Child Health Improvements

David E. Sahn and Stephen D. Younger*

Abstract

Economic growth accelerated during the first decade of the 2000s in many poor countries, especially in

Africa. This welcome news is widely discussed, even in the popular press. Yet as economies grow, some

analysts express concerns that such growth is not having as significant an impact on poverty as one would

hope. Analysts and policymakers ask whether this growth is adequately pro-poor, shared, or inclusive.

Less noticed is that improvements in children’s health are accelerating, too. This paper examines the

extent to which these health improvements are equitably shared or “inclusive.” We use a descriptive

method, which is analogous to growth incidence curves, and apply it to eight countries from Africa, Asia

and Latin America. We draw two principal conclusions. First, within countries, health improvements

often have a different distribution than income/expenditure growth, and that distribution is usually more

hopeful in the sense that it is more likely to be strongly pro-poor than the distribution of income growth.

Second, we have yet to see clear patterns in terms of the within-country relationship between growth

incidence curves and health improvement incidence curves. Thus, one cannot rely on the information in

the growth incidence curve to infer the inclusiveness of health improvements.

1. Introduction

This paper sits at the intersection of two important literatures about improving

living standards and poverty reduction. The first literature addresses the distribution

of income growth. Many developing countries enjoyed an acceleration of economic

growth in the first decade of this century. Of the 104 developing countries from the

regions in Table 1, 69 had faster growth in the 2000s than they did in the 1990s; but

even as economies grow faster, some analysts have expressed concerns that such

growth is not having as significant an impact on poverty as one would hope because

it is disproportionately concentrated among the better off. Such concerns for “pro-

poor growth” or “shared growth” or “inclusive growth” lead to a closer

examination of the distributional consequences of economic growth in the first

decade of the century (African Development Bank, 2012; Kimenyi, 2006; Younger,

2013).

The second literature concerns poverty or welfare as a multidimensional

phenomenon. In recent years, Sen’s widely accepted theoretical argument that

poverty must be understood as deprivations in multiple dimensions of well-being

has found a variety of empirical approaches and applications (Sen, 1976; Alkire and

Foster, 2007, 2011; Duclos et al., 2006a,b). While economic growth is welcome and

pro-poor growth even more so, improvements in incomes are not synonymous with

poverty reduction or improved living standards. Broad measures of improvements

in living standards must consider welfare in multiple dimensions.

Our purpose here is to consider the extent to which improvements in children’s

health are distributionally progressive, or pro-poor. We choose this particular

*Sahn (Corresponding author): Cornell University, B16 MVR Hall, Ithaca, NY, 14853, USA. Tel: +1-

607-255-8931; Fax: +1-607-255-0178; E-mail: David.Sahn@cornell.edu. : Forschungsinstitut zur Zukunft

der Arbeit (IZA), Bonn, Germany. Younger: Ithaca College, 953 Danby Road, Ithaca, NY, 14850, USA.

Review of Development Economics, 21(2), 304–320, 2017

DOI:10.1111/rode.12262

©2016 John Wiley & Sons Ltd

dimension of well-being because, like incomes, there has been a marked

acceleration of improvements in children’s health in the past 20–30 years. Table 1

gives trends in infant mortality and child stunting, as well as information on gross

domestic product (GDP). In Africa, infant mortality has improved throughout the

decades, though the reductions decelerated slightly in the 1990s and accelerated in

the 2000s.

1

The share of countries showing improvements in the infant mortality

rate (IMR) is 90% or above for all periods in Africa, while in the other regions,

this improvement is seen in all countries in all periods. Stunting actually worsened

in the 1990s in Africa before returning to improvements in the 2000s,

2

while in

other regions there is a steady decline in stunting, like IMR, across the decades.

As with the positive GDP growth picture, shown in Table 1, which accelerated in

all the periods, these improvements in health are encouraging, though they have

Table 1. Improvements in Children’s Health, 1971–2011

Decade IMR

Change

in IMR

(%)

Share of

countries

reducing

IMR Stunting

Change in

stunting

(%)

Share of

countries

reducing

stunting

GDP p.c.

(2005 US$

at PPP)

Change in

GDP p.c.

(%)

Share of

countries

increasing

GDP p.c.

Africa

1970s 117 2,368

1980s 100 16 0.94 0.39 2,380 1 0.38

1990s 89 12 0.90 0.41 5 0.42 2,575 8 0.43

2000s 72 21 0.96 0.38 8 0.68 3,028 16 0.73

South Asia

1970s 121 962

1980s 93 26 1.00 0.58 1,156 18 1.00

1990s 68 31 1.00 0.48 19 1.00 1,608 33 1.00

2000s 47 37 1.00 0.38 23 1.00 2,422 40 1.00

South America

1970s 65 6,361

1980s 46 34 1.00 0.29 5,798 9 0.17

1990s 32 36 1.00 0.21 32 0.88 6,512 12 0.83

2000s 22 37 1.00 0.18 15 0.88 7,716 17 0.92

Southeast Asia

1970s 68 15,581

1980s 53 25 1.00 0.47 13,072 18 0.57

1990s 40 28 1.00 0.38 21 1.00 14,872 13 0.71

2000s 28 35 1.00 0.30 24 1.00 17,291 15 0.71

Middle East and North Africa

1970s 92 5,039

1980s 57 47 1.00 0.26 5,611 11 0.78

1990s 37 43 1.00 0.23 12 0.80 6,371 13 0.78

2000s 25 39 1.00 0.17 30 1.00 8,319 27 1.00

Central American and the Caribbean

1970s 58 7,212

1980s 40 37 1.00 0.32 7,545 5 0.61

1990s 28 35 0.95 0.27 17 0.78 8,971 17 0.89

2000s 20 33 0.89 0.21 25 0.89 11,197 22 1.00

Source: World Development Indicators (World Bank, 2013).

Notes: All statistics are unweighted averages across countries. Only countries with data in all three

decades (four for IMR) are included in the calculations. Data are decadal averages for each country.

2000s include up to 2011. Height and weight data are sparse before 1990 and may include only one

observation per decade. IMR is infant deaths per 1000 live births. Stunting measures the share of

children under 5 years old who are more than two standard deviations below the median of the WHO

reference population.

CHILD HEALTH IMPROVEMENTS 305

©2016 John Wiley & Sons Ltd