Political Diversity and Human Development: a one plot check.

This is a very simple post about the association of political diversity with good or bad outcomes.

Elections took place recently in Portugal and Spain. In these elections the public voted for a more diverse parliament than usual. A frequent scare tactic that political parties at the center use is to claim that disperse voting will lead to lots of small parties in parliament and as a result the country will be ungovernable. I decided to have a very simple and quick look to see if the claim holds up.

For measures of political diversity I looked up the most recent electoral results (vote percentage) of the majority of EU countries (by looking at wikipedia reports of the elections). Then I calculated the diversity as the “Inverse Simpson Index” – the more diverse the political voter turnout is the higher the index will be.  For measures of good political outcomes I looked at the Human Development Index (HDI) data from the United Nations. The test is simple: if higher diversity is bad for governance, then the diversity measure and the HDI will have a negative correlation. Do they?

So here is the first order result:


Electoral Diversity (horizontal axis) and Human Development (vertical axis). Correlation is 0.36 with p-value 0.06.



The plot shows only a weak association – but the opposite of the one predicted by the claim. In this data, the more electoral diversity, the higher the human development index score. It’s obvious this proves nothing rigorously, but if anything more electoral diversity is associated with more human development in a set of reasonable comparable countries (i.e. the EU). This may be because political concentration leads to corruption, it may be because a culture for diversity is also associated with a  culture that leads to development, or the data may be confounded for many other variables that matter. However at the very least then, the claim that “votes should be concentrated to avoid problematic governance and bad outcomes” can be discarded if no better evidence is produced.






3 thoughts on “Political Diversity and Human Development: a one plot check.

  1. In my opinion, the plot does show very little (but it’s interesting that, if you ignore most outliers, there’s even a stronger correlation in the “core”, from Romania to Denmark). I think it’s too simplistic an argument and, quite possibly, plotting Political Diversity vs HDI would let you take other conclusions. Yes, I know it shows exactly the same thing, but you can say precisely the associative opposite of «if anything more electoral diversity is associated with more human development» and argue that political diversity tends to arise from and be more successful when HDI is high.

    Your plot says that, on average, political diversity coexists well with higher developed societies. However it does not necessarily invalidate the original claim, in an amended form, in which one could state that “a lot of small parties in parliament IN A LOW HDI COUNTRY result in the country being less governable”. I’m exaggerating, for the sake of arguing, but I don’t think you’re necessarily making the strongest of cases.

    Also, even if would still not prove much, it would be interesting to see the time-dependency of both these variables and see which arose first: electoral diversity, high HDI, or if they both increase at the same time.

    • I think I agree with everything.

      My point was *not* that one thing leads to another – in any direction. My point was (and is) that «the claim that “votes should be concentrated to avoid problematic governance and bad outcomes” can be discarded if no better evidence is produced.». If anything the association is positive (across countries), so regardless of any causal claim, the association starts already in the opposite direction. Of course, as you say, inside a country the association may be the other way around – and if someone wanted to claim that, you and me would both say “very well then, show us the evidence”.

      Doing it across time would be very interesting, I agree. I considered it. But realized I’d have to spend many days scraping election results off wikipedia. If you create a crawler to scan the pages, I’ll do the stats!

      • Adorava poder escrever-te isso, mas infelizmente também não posso agora. Tenho um projecto desses de scraping em mente e acho que já lá vão dois anos que não sai da mente!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s