Pun Salad

A bit of explanation first:

I saw this book by Christopher Snowdon favorably mentioned out there in one of the right-wing fever swamps that I routinely visit. Since I knew that the library of the University Near Here owned The Spirit Level (TSL from here on), I suggested via their online form that they pick this up as well. In order, primarily, to give our local scholars a shot at seeing both sides of the inequality debate.

Somewhat surprisingly, the library purchased it at my suggestion. So I felt obligated to also read TSL (which I would not ordinarily have bothered to do); if you missed them, my TSL comments are here. Summary: I wasn't impressed. Although I read Snowdon's book in parallel with TSL, I tried to restrict myself to criticisms I came up with independently.

Snowdon's book deals primarily with fact-checking (and mostly refuting) many of TSL's arguments, although other works in the same genre with similar theses are mentioned. Snowdon accuses TSL authors, Wilkinson and Pickett, of assuming their conclusion (inequality causes all sorts of bad stuff), then cherry-picking data that seem to bear that out.

For example: when doing comparisons and correlations between "rich" nations, Wilkinson and Pickett include Portugal (which isn't particularly rich), but exclude Slovenia, Hong Kong, and Singapore (which are). Their justification seems weak, and it just so happens that different selections of countries can weaken or eliminate a number of TSL's strong correlations between inequality and various dysfunctions.

Similarly, in some cases, so-called "outliers" cause TSL to conclude cause-and-effect. They graph homicide rate vs. inequality and (no surprise), they spy a strong correlation. But this conclusion relies heavily on the inclusion of Portugal and (unfortunately) the USA. If you remove these two countries from the mix, the correlation goes away, as does TSL's conclusion. It's not robust.

Some of the refutations don't require any heavy statistical lifting whatsoever. For example, TSL correlated inequality against the percentage of waste recycled; they use the resulting regression line to "demonstrate" that more-equal societies are more civic-minded.

But Snowdon argues (convincingly) that this just shows there are two kinds of countries: those whose governments have set up mandatory recycling laws, and those who haven't. People aren't recycling more because they look around and don't see a lot of income disparities; they recycle more because they get fined if they get caught doing otherwise.

So I had a higher opinion of Snowdon's book than TSL, not surprising given my general ideological slant. Readers should feel free to make up their own minds, not that readers need me to tell them to do that. If you don't want to shell out the bucks for one or both books, you can get the flavor of the (ongoing) argument from duelling websites: The Equality Trust from Wilkinson/Pickett, and The Spirit Level Delusion from Snowdown. Particularly interesting is Snowdon's "Chapter 10", a freely-available PDF addon to this book, a discussion of Wilkinson and Pickett's response to criticisms of TSL.