If you live in a country with a despotic dictator you don’t really trust, and you’re afraid your life is just too safe and easy, good news: you’re also probably more likely to die in a car wreck. At least, that’s the conclusion James O’Malley came to when he correlated data about road deaths and people’s trust in their governments.

To get to these conclusions, O’Malley first took data about traffic deaths and the results of a survey about how well various governments rated on qualities like corruption, justice, etc. He gives more detail:

To find out, I took the World Health Organisation’s 2013 data on road deaths per 100,000 people in different countries (which seems a sensible proxy for quality of driving), and compared it to the scores given by the World Justice Project on Rule of Law in 2015. This latter score is generated by surveying 100,000 people and 2,400 experts on 44 indicators like the openness of government, corruption, fundamental rights and justice.

Pairing up the 94 countries that are included in both datasets, it reveals that – amazingly – there appears to be a correlation between the two. For the stats nerds out there, that’s a Pearson correlation coefficient of -0.68.

Let’s just take a moment to explain these Pearson correlation numbers. The Pearson scale is between 1 and -1, where 1 is a total positive correlation (that is, as one factor goes up, so does the other) and -1 is a negative correlation (as one factor goes up, the other goes down), and 0 is no correlation at all.

So, when he says that these World Justice ratings of governments and road deaths have a -0.68 correlation, that means that as the scores of the governments go up, traffic deaths tend to go down, at least 68% of the time (assuming that a Pearson score of -1 would mean 100% of the time). That’s a pretty significant correlation.


Then, he takes data about how much people seem to trust their governments, and compares that to the traffic deaths:

Comparing traffic data to how much citizens trust their governments creates an even more striking correlation. Using data collected by the OECD, it reveals a correlation of -0.81.


Holy crap – a Pearson score of -0.81! That’s a really strong negative correlation. So, the more you trust your government, the less statistically likely you are to die in a car crash.

Of course, like all statistical things, you sort of have to take this with a grain of salt:

This isn’t to imply that correlation means causation: there is also a correlation with average income, for instance, but both of these things are essentially measures of how successful a country is. And of course, we shouldn’t rule out that it could just be one massive coincidence.

But in any case, at least on this initial glance, it certainly appears that there is a link between the two, whether they are causally related or whether they are merely outcomes of the same parent phenomenon. If I’m right, and this isn’t a coincidence, then it suggests that the state of a country’s roads could also act as a neat heuristic for understanding the quality of a country’s government.


This is still a fascinating bit of numbers-relating here, and it seems to make sense. This certainly doesn’t mean that countries with low traffic deaths are utopias; look at North Korea, which has relatively minimal automotive traffic period, and one of the most completely oppressive and dictatorial governments around. Or, at the opposite end, you could consider India, a massive democracy with plenty of traffic deaths, but that seems to come more from the sheer chaos of driving there than any fundamental mistrust of government.

It would be interesting to follow a country with increasing levels of trust in their government to see if their roads get safer as well. Get on that, James.