Random Variable

Lijphart’s basic argument is that there are two major types of democracies – majoritarian democracies, based upon the Westminster model, and consensus democracies, to be found in Europe. He claims that consensus democracies provide better outcomes for people on a variety of indicators. I decided to try our one indicator, not mentioned in Patterns of Democracy [1] – poverty reduction.

Using data from the Luxembourg Income Study [2], and plotting the executive-parties dimension of democracies from Lijphart’s Patterns of Democracy produces the following graph:

My poor grasp of statistical tests tells me that scoring higher on Lijphart’s first dimension is positively correlated with effective poverty reduction at the 5% level, and that’s with the United States as a huge outlier.

* Relative effectiveness of income redistribution systems in poverty reduction around 1990, where poverty threshold = 60 per cent of median equivalent disposable income.

Make of that what you will.

[1]  Arend Lijphart, Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries (New Haven [Conn.]: Yale University Press, 1999).

[2]  J. Fritzell and V. M. Ritakallio, “Societal shifts and changed patterns of poverty,” Luxembourg Income Study working paper 393 (2004).

Update: Replaced Excel graphs with plots from R.

The Guardian reports that The Gallup Coexist Index 2009 finds that exactly 0% of British Muslims tolerate homosexuals, compared with 19% of German Muslims and 35% of French Muslims.

Now, before everyone goes all Geert Wilders, some problems.

Given that migrants to France, Germany and the UK come from specific ethnic groups, are we missing some correlation between ethnicity and religion? Have we got some true control variables, rather than just asking non-Muslims.

Also, seems like British Muslims are doing pretty badly economically, with the second lowest unemployment, and very low expectations of social mobility:

I wonder how well non-Muslim groups in a similarly dire economic position would fare on moral questions.

In addition, perhaps British immigrant groups are secularising faster than their continental counterparts, especially if they are actually LGBT.

If I had the data, it’d be interesting to see if you could run it through some sort of multivariate analysis and see if other variables correlate higher on moral issues than others.

Over at Liberal Conspiracy, Martin Robbins has a post on a more smeary opinion poll about Muslims eeking out of the ironically titled Centre for Social Cohesion.