RCE. Visualizing the results, part 4. Wealth and Human Rights

In this post, we are going to look at whether the correlation we saw between democracy and better protection of human rights is in part driven by wealth. (Remember correlation is not causation and all that). Several possibilities could be true: perhaps wealth makes both democracy and better protection of human rights possible (it is the cause of both); or perhaps it helps by promoting values that lead to democracy which in turn leads to better protection of human rights (so that democracy causes better human rights protection, but only at high levels of wealth); or perhaps democracy helps protect human rights independently of the level of wealth. We may not be able to tease out these possibilities (or others) without very sophisticated methods, but I can at least give you some ideas about what’s going on. (Please ask questions if you’re confused - this post may be a bit technical in places!).

The first thing to do, however, is to simply look at the correlation between a measure of wealth (real GDP per capita, as reported by the Penn World Table for international comparisons, version 9.0) and the Chris Fariss’ measure of human rights protection:

As we can see, there is a very strong relationship between wealth and the protection of human rights, though except at the extremes there’s still a wide range. What we want to do now is to check if there is a correlation between democracy and human rights protection even after we factor out the correlation with wealth. What I’m going to show you is a “partial correlation” plot, depicting the relationship between democracy and the residuals of a regression between wealth and your measure of democracy:

Using the POLS209 measure of democracy, it looks like there is still a relationship between democracy and better human rights even after “controlling” for the degree of wealth of a country. We can repeat the exercise by using more data (the V-Dem measure of democracy), just to be sure:

The same overall result obtains, though of course with a lot of variation. (For those familiar with the language of regression, if you do a regression of the human rights scores on the measure of democracy and on the measure of wealth, both coefficients are positive and significant).

Of course, those relationships vary by country. Here I show you partial correlation plots for all the countries for which there is data:

Countries with upward-sloping lines experienced increases in human rights protection from becoming more democratic “above and beyond” what could be expected simply by increasing wealth, whereas countries with flat lines experienced increases in human rights protection that are only commensurate with changes in their wealth (so democracy does not appear to explain much there), and countries with downward-sloping lines experienced less improvement in human rights from becoming more democratic than any increases in wealth would suggest (e.g., Venezuela).

(Note that some countries were highly democratic and had very high human rights scores for the entire period, so the plots don’t show much relationship with anything - e.g., Norway, Iceland, New Zealand).