Saturday, January 30, 2016

Book: Hive Mind by Garett Jones

Jones, Garett. 2015. Hive Mind: How Much Your Nation's IQ Matters So Much More than Your Own. Stanford, CA: Stanford Economics and Finance.

I read this because the folks at Marginal Revolution were pushing it. The title makes the book's thesis clear and the author does a good job of making the case for his views by marshaling evidence from a variety of disciplines. I was surprised at how much of the material was familiar to me from other things I had read, though I still appreciated the thoughtful arrangement of the familiar bits into a clear and fairly compelling argument.

Four things bothered me about the book. First, even conditional on being aimed at a general audience of intelligent readers, I found the writing annoyingly simple at times. The book is, say, a 3.5 / 10.0 on a scale from "not academic at all" to "Varian's graduate micro text". I would have preferred a 5.0 or 6.0 book. Maybe the author will write that one next. Second, in trying to make the case that national average IQ matters, the author minimizes the value of own IQ conditional on national IQ, but does so without putting much effort into justifying that point. Perhaps I am overly influenced by the Murray and Herrnstein Bell Curve book which I read as showing that individual IQ is a non-trivial predictor of nearly every adult behavior. Plus there is no real need to make this case: it can be true that both individual IQ and national average IQ have meaningful explanatory power. One does not preclude the other. Third, I would have liked a chapter on within-country sorting by IQ and why it does not function to partly undo the effects of national IQ. Certainly in countries like the US with a lot of residential and social segregation by education and income (and thus implicitly by IQ) this would seem relevant. Finally, the part of the book that was newest for me was a discussion near the end of a model of the labor market that combined a sector with an o-ring technology, where the average skill matters, with a standard Cobb-Douglas sector where it does not. Trying to explain this model without the math and at the 3.5 level of academic sophistication made things really frustrating for this reader, particularly as it seemed to me that some of what were offered up as general points may well have depended on specific functional form assumptions.

Recommended with the above caveats.

No comments: