October 25, 2005

Corruption's Correlates

What contributes to corruption? My criminologist reader has calculated correlation coefficients between Transparency International's 2005 Corruption Perception Index and various national demographic factors.

One confusing aspect is that the CPI gives the highest scores to the least corrupt countries, so it's actually an honesty index. Thus, the following coefficients are for correlations with honesty. Positive correlations are good, negative correlations are bad.

- IQ (from Lynn and Vanhanen's list) correlates at an impressive .708 with the honesty index.

Interesting how the IQ correlation is stronger than any of the other factors. Perhaps smarter societies rationalize their systems. National IQ is turning out to be a powerful shaper of society. And my brilliant professors all told me that individual traits do not matter at aggregate levels!

- predominantly Muslim country, -.257

- predominantly Protestant country, .345;

- predominantly white county, .521.

- ethnic diversity, -.364 for 121 countries (that offers some supports Lee Kuan Yew's contention that multiculturalism combined with democracy equals corruption).

- first and second cousin marriages (percent consanguineous), -0.160.

One difficulty is that most of the samples for cousin marriages were of subpopulations so they did not match up with the CPI data which is national. I found comparable data for 35 countries. Sub-Saharan Africa and southern Asia were not represented. The 7 Muslim countries fit your model perfectly, but the heavy sampling of Latin American countries (13) weakened the correlation greatly--it ended up being -.16 (remember that high CPI is low corruption). Also, consanguinity is not normally distributed--the Muslim countries averaged around 30% while no other country in the sample was higher than 6.3%.

So, my cousin marriage theory -- that the Arabs and Muslims have trouble forming effective nation-states because their extended families are too strong for patriotism to flourish due to high rates of cousin marriage -- is neither supported nor debunked by this data. The methodological problem is that, as I observed in 2003, the Muslim countries have lots of cousin marriage and lots of nepotism. But the rest of the world has so little cousin marriage that there isn't much relationship to corruption. It could well be that cousin marriage helps makes the Muslim countries the way they are, but this methodology won't nail it down.

My published articles are archived at iSteve.com -- Steve Sailer

No comments: