Life Expectancy and GDP (a complicated relationship)

Life Expectancy and Gross Domestic Product (GDP) are two fairly common statistics that are published and can be easily compared for countries around the world. For this project I pulled the data from the the World Trade Organization and World Bank websites for six countries (Chile, China, Germany, Mexico, the US, and Zimbabwe) and looked at annual data compiled for the years 2000–2015. Based on this data I started mining the information and making a review.

In an initial comparison of GDP to Life Expectancy for all six countries there seemed to be a very clear correlation. In each case there was an apparent correlation between increased GDP and increased Life Expectancy.

Zeroing in on each country individually we seem to observe a similar trend.

Chile — overall trends
China — overall trends
Germany — overall trends
Mexico — overall trends
US — overall trends
Zimbabwe — overall trends

While these individual, and even side-by-side, reviews of national trends point towards a strong correlation of Life Expectancy to GDP it is a poor reflection of scale. GDP growth within the US, China, Germany and Mexico are in the range of trillions and even tens of trillions of dollars. Meanwhile Chile and Zimbabwe GDP is scaled in the billions. When viewed together for a proper sense of proportion the correlation between GDP and Life Expectancy becomes less clear.

Chile and Germany have Life Expectancy rates that are comparable to or surpass the US even though their GDPs are a fraction of the size. The same is true when comparing Mexico and China.

Let’s take a moment to zoom in on the data.

If we review the GDP for scale and growth it is obvious that the US outpaces all of the other countries in our dataset. China meanwhile has seen the greatest change in GDP over the course of the observed 15 years. The other four countries have seen relatively little change in their GDP and are nowhere near the pace of China or the US.

Let’s take a look at how that compares to Life Expectancy.

At a glace the difference is shocking. Zimbabwe, while seeing comparatively little GDP growth has had immense growth in Life Expectancy. Meanwhile the other five countries, even those with very large GDP increases have seen comparatively little change in Life Expectancy. To get a better look at this we can remove the outliers. First the GDP without the economic giants China and the US.

Here it is quite apparent that the economic growth of Germany far surpasses the others, and Mexico’s growth being more than double that of Chile.

However when we look at Life Expectancy among the top five countries (removing Zimbabwe as an outlier) we see the following:

Germany and Chile are neck and neck with one another both surpassing the US in terms of Life Expectancy even though Germany and Chile are in the bottom four economically. Even Mexico which has a lower life expectancy than the US, has a life expectancy higher than China. This is again worth noting because China is in the top 2 economically, while Mexico is in the bottom 3.

While these results are compelling on their face it is worth running the data through some hypothesis testing models to review whether the difference between countries is significant. At the end of the day is a difference of 1–2 years Life Expectancy statistically significant? Is the difference of GDP growth between 1 trillion and 10 trillion between two countries statistically significant? Either way it’s still a lot of zeros, right?

Statisticians use a model referred to as the Tukey Test that helps determine this very question. While I won’t go into all of the testing methods here I think it is worth reviewing the results. First lets look at GDP.

The far right column of the table notes whether the comparison between group1 and group2 (far left columns) has a difference that is statistically significant. If the final column returns “True” then there is a statistically significant difference. If the final column returns “False” then the difference between the two groups is significantly insignificant. The data shows that there is a significant difference in the GDP between nearly all of the pairings with the exception of three: Chile, Mexico and Zimbabwe. The GDPs of these three countries are statistically comparable to one another. The rest o the countries show significant differences. Now lets run the Life Expectancy data through the same test.

The results here are far more mixed. The test reveals that there is no statistically significant difference between the Life Expectancy in Chile, the US, Mexico, Germany, and China. These five countries are linked in a range as was revealed earlier in our plot of the top five countries by life expectancy. There is no significant difference in the live expectancy of these five countries even though there are significant differences in their GDPs.

After a review of the data the connection between GDP and Life Expectancy is shaky at best. While there may be some correlation on a nation by nation basis, when compared internationally the wheels seem to come off the cart. In the end the search for an indicator of Life Expectancy must be looked for elsewhere.

Areas of future research

While the data provided by the World Bank and World Health Organization does not cover this information it is interesting to look at what else might contribute to increased Life Expectancy. One note that would be worth more exploration is that Chile, Germany, Mexico and Zimbabwe each have systems of nationalized healthcare services. The US and China do not. This lack of a nationalized health service may explain why the US and China with all of their economic advantages do not see a significant difference in Life Expectancy when compared with countries with significantly smaller GDPs. We do not have the data at this time to test the hypothesis of whether nationalized medicine is a contributing factor, but it is certainly worth exploring at a future date.