How Good Is GDP Growth At Reducing Child Mortality?

Last weekend I wrote about UNICEF’s report that child mortality had declined by large amounts across the globe from 1990 to 2010 (UNICEF defines child mortality rates by the number of children per 1000 live births who do not make it to age 5). Of course, the report did not speculate about how much economic growth may have contributed to this decline, but the capitalist in me wanted to know. Wouldn’t it be good to know if it was better to save children’s lives by encouraging the growth of government programs or by encouraging the growth of the marketplace? So I spent some free time in the last few days gathering data from Wikipedia to see how well GDP numbers correlated with child mortality numbers.

The quick summary is, yes, countries that had stronger GDP growth tended to see greater reductions in child mortality, although the correlation wasn’t as strong as a pure capitalist might like to see. There is perhaps nothing surprising or remarkable about this, but I’m trying to stimulate more discussion of the great news about world reduction in child mortality, which I don’t feel has gotten nearly as much attention as other recent reports like the one on US poverty. A few disclaimers before I present the data:

1. I am assuming that the child mortality numbers from UNICEF are reliable. Gathering comparable data from every single country in the world is quite difficult, and there was much surveying and modeling.
2. I am assuming that Wikipedia’s GDP numbers are reliable. (The third Wikipedia link has three different sources for 2010 GDP; I went with the CIA World Factbook source because it listed the most countries.)
3. I am assuming that I copied and pasted all of those numbers inerrantly.
4. I decided to compare a percentage of child mortality reduction with a percentage of GDP growth per capita, as opposed to with a percentage of GDP growth, or comparing absolute numbers for any of them. I also did not account for inflation or changes in the value of the dollar. By using the same basic reported numbers for every country I thought this was the simplest and best way to analyze the average rise in living standards per person, but you may disagree.
5. Because I decided to use percentages, I did not analyze every country in the entire UNICEF report. I arbitrarily excluded countries that had a reported 1990 child mortality rate below 30 deaths per 1,000 live births because I don’t think that a 60% reduction from 10 to 4 says as much about the living standards of a country as does a 60% reduction from 100 to 40. In other words, once your country gets to a relatively decent child mortality number, the marginal cost of reducing that number by a given percent is much higher than reducing that number by the same percent for a country with a much more horrid child mortality number. I’m just looking at the “bad” countries and seeing how much their economy grew on their way to getting better, but you may argue that I should have chosen a different number than 30 as my cut-off. Additionally, you may argue that I should have chosen high or low cutoffs for GDP per capita to exclude outliers from that dimension.
6. I also excluded a few ad hoc countries for which I could not find GDP data for both 1990 and 2010, but this should not greatly affect the overall averages and trends based on the 134 countries that I did analyze.

So this is not a report on “How Child Mortality Reduction Correlates With GDP Growth By Country.” It’s a report on “How Child Mortality Reduction Correlates With GDP Growth From High Child Mortality Countries For Which I Could Find Decent Data Under A Variety Of Hopefully Reasonable Assumptions.” Here we go…

First, let’s look at how child mortality compares to GDP per capita in 2010 for the 134 analyzed countries, before we look at how much those numbers have changed since 1990.

We see that countries with higher GDP per capita tend to have lower mortality rates, but there is also a lot of variance and a long tail. Countries with low GDP have child mortality rates ranging from pretty good to very bad. Only three countries with a GDP per capita of over $5,000 have child mortality rates around 50 or greater, while the vast majority of countries under $5,000 have rates over 50. This graph suggests that it does not take much economic growth to significantly reduce the child mortality rate of very bad countries, but it also suggests that there are ways to reduce the rate without economic growth.

If countries with higher GDP per capita tend to have lower mortality rates, then we could expect the countries that had the most growth in GDP per capita to have the greatest reduction in mortality rates. So what do we have? Let’s start with the averages:

First, I notice that the world averaged a reduction of 30% or more regardless of GDP growth. Second, I notice that countries with larger growth experienced more reduction in child mortality. Third, I notice that once you get beyond a doubling or tripling or so, the reduction gains peter into noise.

Now let’s break this out by individual countries, where we see another definite but also fairly weak trend:

There is wide variety; some countries reduced their mortality rates by 45% with hardly any growth in GDP per capita, and one country (Lebanon) reduced its rate by the same amount with a ten-fold growth in GDP per capita. However, we also see that all of the countries which failed to reduce their rates by more than 20% also had less than a ~350% increase in GDP per capita, while all of the countries with more than a ~350% increase in GDP per capita saw more than a 20% reduction in child mortality rates.

This is even more starkly seen by comparing the countries with the ten greatest reductions in child mortality to the countries with the ten lowest:

Every single one of the bottom 10 countries had less economic growth than every single one of the top 10 – although South Africa did not have much less growth than Timor-Leste. Other factors are clearly at work, but the trend is clear. Zimbabwe and Haiti were the only countries that actually regressed in child mortality, and Zimbabwe was one of the few countries that actually had a loss of GDP per capita over the last 20 years.

Correlation is not causation, and there are many unique factors at work in each country. But I believe the data supports a theory that general economic growth leads to a reduction in child mortality, even though the specifics of each country’s situation will determine if they can achieve a greater reduction than economic progress alone can bring, or if they can achieve reductions in spite of a lack of economic progress. It is also unlikely that the causation is reversed. While a reduction in child mortality should lead to greater economic growth in the long run as the children grow up and contribute to the economy, this study covers a span of only twenty years. Any child who is alive today because of a reduction in child mortality is probably not contributing to his or her country’s economy yet (if anything, he or she is probably a net cost, as any parent, with all love, would probably agree).

So I think this analysis contains nuggets for both the capitalist and the charitable enthusiast. Economic growth may be the “rising tide” that lifts more children to life, but there are undoubtedly other ways to save lives more directly through governments and organizations and the like. However, I think this data shows that we should consider a country’s disposition toward economic growth (by which I mean enforcement of property rights, accountability for public officials, freedom to create businesses and jobs, etc, etc) as a very important part of the equation. I don’t think there’s anything especially extraordinary or surprising about my findings, but feel free to support or dispute them anyway.