Super Freakonomics

Steven D. Levitt and Stephen J. Dubner continue right where they left off in Freakonomics. This book is a very quick read that tackles a number of unconventional economics topics. It does quite well when it focuses on the data. However, at times it ventures off in to the realm of pet projects that don't really have the data.
It presents a great chapter on car seats. The data shows that a child above the age of two is no more likely to die in a car crash when wearing a seatbelt than when in a car seat. Yet we have increasing mandates for carseats. Why is this? Car seat manufacturers have a vested interest in selling more car seats. They often support this interest by showing car seats make children much more safer than leaving them unrestrained. There is also a possibility that they may be less likely to suffer a serious injury. However, the safety would be much greater if we just had child-sized seat-belts. This would also be a much cheaper alternative. (Since the majority of the people riding in the back are children, why not optimize for them? Adjustable seat belts would also be beneficial for smaller sized people. Unfortunately, there is little lobbying effort for this, while there is plenty for car seats. Even adjustable seatbelt makers can be ambivalent. (They could sell the belts with or without a carseat mandate.)

Parents can also be irrational when it comes to safety measures. They would be adamant that they do the "safe thing" for their kid and keep them in the car seat, even if that $50 could help the child's safety in many other ways. (Flame retardant regulations also come to mind. Children's pajamas and furniture are required to be filled with carcinogens to help provide a small delay in the remote chance that they catch on fire. A significant amount of "sure harm" is exchange for a remote possibility of reduced harm. Eliminating cigarettes in the house has proven much more effective. However, flame retardent chemicals let both the chemical industry and the tobacco industry increase sales - while providing a further revenue stream for the health care industry down the road.)

Other sections, however, are much less convincing. A discussion on global warming is especially poor. At the start of the chapter the authors discuss the "dangers of global cooling" that had worried scientists 40 years ago. Today, of course, the worry has turned to global warming. They present a little data to say that "its happening, but its not the end of the world." Then they proceed to speak in favor of piping chemicals to the stratosphere to induce global cooling.

From that discussion, it would seem to follow that we'd want a couple of pipes, one to spew out greenhouse gases and the other cooling gases. Then we could regulate the earth's temperatures. Only, we don't even have models that can accurately predict the weather for one location a week in advance. How do we expect to predict and manage the world for years? The "solution", unfortunately, could become a big part of the problem. (And could lead to many geopolitical issues, depending on who controls it.)

Agriculture discussions are also somewhat week. Ammonium Nitrate is praised as allowing great population booms. However, the externalities are not mentioned. The localvore movement is criticized because the small scale. However, no data is given to back this up or discuss other differences. (You could easily critique inappropriate local food as being wasteful [corn in the desert]. However, moving beyond commodities, local produce will likely be differet, and could even be grown on a greater scale than the industrial produce.)

The pitty of these shortcomings is that the book acknowledges the same shortcomings in other sections. The problems of "unintended consequences", externalities and detailed data analysis are all given ample discussion in the book. Unfortunately, many of the cases fail to take these in to account. This leaves the book bouncing around from rigourous pieces to fluff, and even investigative journalism. It's all entertaining, just not entirely convincing.