The Logic of Life: The Rational Economics of an Irrational World - Tim Harford

~~Moved from GR~~

The Logic of Life

By Tim Harford

 

Recommended to Carly by: the author's radio programme and other book--that'll teach me to theorize ahead of the evidence.
Recommended for: Readers who can identify the wild jumps in logic, yet still enjoy the book

 

Tim Harford, you're breaking my heart, and more importantly, you're undermining my faith in quantitative economics. I am a passionate fan of your BBC radio show, "More or Less." What could possibly be more entertaining than a topical radio programme that uses statistics to fact-check the politicians, especially if it occasionally measures things in whales and/or Wales? Sure, I don't have the same faith in rationality that you do--well, not without completely bending the meaning of "rationality" out of joint to allow for irrational influences such as spite-motivation--but I accept the premise. I greatly enjoyed your previous book, The Undercover Economist, and that book was just as insistent on human rationality. But what I can't cope with is this book's betrayal of the primary credo that separates idle speculation from science.

Mr. Harford, what, oh what, happened to your belief in the rallying cry of "More or Less?" What happened to "correlation does not imply causation?"
 


The Logic of Life is, of course, well-written and entertaining. The premise, that humans are rational agents, so apparently irrational actions must have hidden rational motivations, is a fascinating approach to studying and understanding social systems. The issues that Harford covers include the effects of marriage and divorce ("Is Divorce Underrated"?), how monetary reward is used and misused to encourage labour ("Why Your Boss is Overpaid"?), how the structure of a neighborhood can encourage or discourage crime ("In the Neighborhood"), how racism can create a "rational" vicious cycle ("The Dangers of Rational Racism"), and more. And yet I think this book betrays the very foundations of quantitative analysis.

For example, he takes a study about speed-dating, in which it was demonstrated that relative differences in attractiveness of partners mattered far more than absolute differences, to conclude that we as humans will all "settle" for what we can get. It's a nice theory and it does have the empirical backing that Hartford mentions, and he even cautiously notes that speed-date behavior may not precisely correlate to marriage behavior. However, he does not point out the most important potential issue with the study: that the set of individuals who choose to speed-date are emphatically not a uniform sample of society. One might imagine that individuals who choose to speed-date are far more desperate to form a relationship than the general population and therefore are predisposed to "settle." He claims, "it is official: rich men are a turn-on and rich women are a turnoff," using as evidence information from online dating data and the "fact" that we
"Find lots of women in places where there are lots of rich men: that is, in the cities....areas with high male salaries are areas where a lot of women live, especially young women. Consciously or not, plenty of women seem to have decided they would rather compete for scarce wealthy males."
(This last is laughably insubstantial "proof;" consider the sheer number of contributing factors that are ignored. Correlation confused with causality? Never.) And just in case these massive jumps in logic haven't yet managed to raise your blood pressure, he also asserts that young black women tend to be single mothers because (a)they tend to only marry within racial lines, (b)black men in prison means less black men available in the marriage market, and (c)black men out of prison use their additional bargaining power to avoid marriage. Ye gods.

According to Harford, the increase in women pursuing higher education and higher-powered jobs are motivated (solely, apparently) by marriage markets: "I'll show that supply and demand in the dating market motivates people to work, to study...we'll see that in places where men are scarce, women respond by staying in school longer." First, women seek to "increase their attractiveness in marriage prospects" by getting jobs and education. He then invokes the garbage gender-role theories promoted by evolutionary biologists: that
"In the African savannah, then, our rational male forbears wanted young and beautiful forebears while our rational ancestors down the maternal would have preferred high-status males."
and that Homo sapiens beat out the Neanderthal via division of labour into gender-based roles. In this division of labour, men specialized as breadwinners (hunters) and women as homemakers (gatherers). A woman discarded by a man could no longer benefit by specializing in these tasks. Divorce, therefore, galvanized women to pursue education and job opportunities, and this in turn gave them more bargaining power. I found this theory tantalizing, but it involved far too much soft reasoning and jumps in logic to be presented as "fact." By simplifying all motivations to a single utility (marriage prospects), it implies that all women pursue higher education to gain bargaining power and additional security in marriage.

If you haven't yet begun to foam at the mouth at Harford's absolute faith in his breezy simplifications, wait until you get to the chapters on racism, where he discusses Schelling's Segregation model, "acting white," and racism in the workplace. His writing here is even less tactful than in the marriage chapters (in which home-makers are repeatedly denigrated as staying home as "baking cookies"). Before making such sweeping generalizations on delicate and emotional issues, I want to see a lot less simplification and a lot more evidence. The end result was a book that I might describe as "Freakonomics-lite," or maybe as a cross between Malcolm Gladwell and Freakonomics. I actually think Gladwell is an engaging writer able to package and sell neat little concepts in intriguing examples and "sticky" catch-phrases. As Gladwell is not a scientist, I don't hold him as responsible for his tendency to single out a qualitative example, tell a story, and then present an idea as "truth"--all while completely failing to separate correlation from causality, and sometimes not even showing the correlation. But how can Tim Harford have published a book that does the same thing? How can he assert, with a straight face, the "truth" of those evolutionary biology theories that take a single fact and work backwards to find a self-serving explanation?

It comes down to this: I believe that cavalierly confusing fact with appealingly simplistic theory is both dangerous and unethical. I think that it is important to use mathematical techniques about the world and I fully understand the necessity of constructing simplified models. However, I think it is immoral to conclude with such absolute certainty that these theories reflect the "truth." In Harford's case, most of these theories simplify incredibly delicate, controversial, and politically sensitive issues, and they involve so many variables that it is inherently impossible to reason about them without invoking one's own biases. Much of the "evidence" Harford presents are correlations at best, and in others, he utilizes the incredible logical fallacy of arguing backwards from fact to theory without acknowledging the inherent biases in this method. To be clear, I have very little issue with the theories presented here--as long as they are presented as theories. However, Harford tends to present his points in assertive, definite language. The book is engaging and thought-provoking, but much of the content is insubstantial and lacks scientific rigour. I recommend first reading Duncan Watts' Everything is Obvious: Once You Know the Answer to fully understand why Harford's methodology is so problematic. The book completely failed to answer my most urgent question: how can a book celebrating human rationality be so inherently irrational?

So, Mr. Harford, even if you have gone over to the other side, I'm sticking to my guns. I'm going out with the battle-cry, "correlation does not imply causation!"