Freakonomics: A Rogue Economist Explores the Hidden Side of Everything and The Wisdom of Crowds
Published September 1, 2005 | September 2005 issue
What do schoolteachers and sumo wrestlers have in common? Where have all the criminals gone? Why can you walk into a convenience store at 2 a.m. and find the kind of orange juice you want? And why in the world don't NFL coaches "go for it" more often on fourth down?
These are among the intriguing questions addressed in two entertaining and engaging books, Freakonomics: A Rogue Economist Explores the Hidden Side of Everything by Steven D. Levitt and Stephen J. Dubner and The Wisdom of Crowds by James Surowiecki.
While not immediately apparent from their titles, these books share certain similarities. Both are relatively short, simple reads that offer interesting insights and captivating stories. Both ask unusual questions, seek to uncover hidden truths and are described on their dust jackets as "dazzling" by author Malcolm Gladwell (Blink and The Tipping Point)—leading one to wonder whether Gladwell might be dazzled a bit too easily. More substantively, the underlying themes in each are surprisingly in sync: Experts should be viewed with skepticism, conventional wisdom shouldn't be accepted without question and information is a powerful tool in decision-making.
But the authors come to these conclusions via very different routes. Levitt and Dubner (economist and writer, respectively) rely heavily on clever data analysis coupled with the creative force of the book's empirical superhero, Levitt himself, to look at a number of interesting but rather quirky issues. In contrast, Surowiecki, a staff writer for The New Yorker, uses a classic case-study approach to pursue a broader theme.
Let's take a closer look at each book:
Freakonomics: A Rogue Economist Explores the Hidden Side of EverythingBy Steven D. Levitt and Stephen J. Dubner
HarperCollins, 256 pages
Somewhere along the way, the authors (or editors) of Freakonomics decided to market economist Levitt as a lone-wolf superhero. Levitt is described (by Dubner) as a "noetic butterfly" who "unlike most academics, is unafraid of using personal observations and curiosities," and though he is "not good at math" has managed to "redefine the way we view the modern world." His "freakish" brand of economics is so revolutionary that "many people—including a fair number of his peers—might not recognize Levitt's work as economics at all."
Based on the fantastic sales of the book—it's number two on The New York Times bestseller list as this review is written—it might seem imprudent to criticize the marketing decision. Still, the portrayal is distracting and not especially accurate. Levitt is a chaired professor of economics at the University of Chicago. The members of the American Economic Association were able to recognize enough of his work to award him the John Bates Clark Medal, their biannual prize to the best American economist under the age of 40. Levitt may well be a fearless "noetic butterfly," but he is a butterfly firmly established in the highest pantheons of the economics establishment. Hardly a rogue. Furthermore, this clever marketing approach only serves to underscore an often-voiced criticism of Levitt's research and this book—that it is too clever, too cute and lacks lasting substance. More on this later.
But regardless of what one thinks of the marketing approach, underneath it is a book full of engaging stories that bring Levitt's professional research to life for a broad audience. Levitt's forte is to ask interesting and unusual questions, and then to combine his astute empirical skills with mountains of raw data to come up with answers, often answers that take the reader by surprise. Dubner's evident contribution is a compelling writing style that turns economic inquiry into page—turning who-done-it-and what-exactly-did-they-do?
Each of the book's six chapters is motivated by a tantalizing question. But behind such whimsical titles as "What Do Schoolteachers and Sumo Wrestlers Have in Common?" or "How Is the Ku Klux Klan Like a Group of Real Estate Agents?" lie more serious issues. The book does an excellent job of bringing the reader along for the ride, explaining the solution strategy and how data analysis is used to illuminate the way.
So, what do they have in common?
Consider the opening chapter. In two separate academic studies, Levitt analyzes schoolteachers and sumo wrestlers. Freakonomics points out that they share a common trait: When there is an incentive to cheat, some do.
In the case of schoolteachers—whose incentive to cheat derives from possible pay raises, or at least job security, if their students score well on standardized tests—Levitt uses 700,000 sets of test answers from Chicago public school students. With the aid of his trusty computer, Levitt searches for patterns in the answers to identify teachers who were changing their students' answers. The authors even include a few sample answer sheets and invite the reader to detect the cheating. Similarly, the book describes Levitt's clever strategy for examining the win-loss records of sumo wrestlers—an examination that points to especially important matches being "thrown," with the "winners" later reciprocating in less important matches, so that top wrestlers can maintain their status.
Another two chapters cover the Ku Klux Klan, real estate agents and why drug dealers live with their moms. The authors argue here that conventional wisdom is often wrong (most drug dealers make very little money), information is power (the secrets of the Ku Klux Klan) and experts should not be wholly trusted (real estate agents sell their own houses for 5 percent more than they sell their clients' houses).
Where have all the criminals gone?
A point repeated throughout the book is that experts should be viewed with skepticism, often because they have an incentive to use their information for personal advantage. Skewered experts include real estate agents, parenting experts and homeless advocate Mitch Snyder. But the authors save their sharpest barbs for criminologist James Alan Fox, who in the midst of the crack cocaine epidemic in the late 1980s "warned of a coming 'bloodbath' of youth violence." Instead, crime fell dramatically in the 1990s. How could this expert have been so wrong? What accounts for the decline in crime?
The surprising story of the falling crime rate is a hallmark of the book. First the authors describe the conventional wisdom—the eight most commonly cited explanations for the decline in crime. According to Levitt's analysis (the academic research, done with Stanford University's John Donohue, is published in the May 2001 Quarterly Journal of Economics), most of these factors had little or no impact [strong economy, capital punishment (see "A Punishing Debate," The Region, June 2002), innovative policing strategies, tougher gun laws, aging of the population] while others accounted for only part of the decline (increased reliance on prisons, increased number of police).
So what do Levitt and Dubner say drove the decline in crime? Legalized abortion and the Roe v. Wade Supreme Court ruling in 1973. The explanation for this association is surprisingly straightforward. Women who have abortions are disproportionately poor, uneducated and young, unprepared or unwilling to raise a child. Children born into such adverse family environments are more likely to commit crimes. There were over 1 million abortions per year during the mid and late 1970s, so by the early 1990s there were essentially millions of crime-prone teenagers missing from the population. Roughly 18 years after the passage of Roe v. Wade, crime began to decline.
Obviously, this explanation is emotionally and politically charged, and the authors make it clear that they are not advocating any policy. They are only answering the question of why crime went down.
But a plausible story does not suffice as an answer; it needs to be tested against the data. And here we see one of the hidden strengths of Freakonomics: the powerful use of data analysis with cautionary emphasis on the important difference between correlation and causation. In the example of abortion and crime, the authors clearly and carefully explain the empirical support for causation. Crime declined sooner in the five states that legalized abortion prior to Roe v. Wade, states with the higher abortion rates experienced the larger declines in crime, and there is no link between abortion rates and crime prior to the late 1980s when the "missing" teenagers would have first started arriving. Not conclusive proof, perhaps, but compelling facts to support the premise.
And the kids who were born? The final chapters focus on whether parenting advice from experts has much wisdom. One chapter addresses the fascinating topic—to parents at least—of how much parenting actions influence outcomes. Does reading to your child, limiting television time and having Mom stay at home raise school performance? By now you can probably guess the answer. Does the name you choose impact your child's life outcome? Not according to Levitt and Dubner, who make their point with amusing biographies of two boys named Winner and Loser. Such critique of expert wisdom is also the topic of the second book, The Wisdom of Crowds.
The Wisdom of Crowds
By James Surowiecki
Doubleday, 296 pages
Writer James Surowiecki, sans the swashbuckling style of Levitt and Dubner, pursues a simple but powerful theme: Groups can be surprisingly smart. He argues that by properly coalescing the individual knowledge of their members, groups can, in fact, be even wiser than experts. He refers to this as "collective wisdom."
More specifically, the author writes, "[U]nder the right circumstances [italics added], groups are remarkably intelligent, and are often smarter than the smartest people in them ... even if most of the people within a group are not especiallywell-informed or rational." Or, put differently, "The argument of this book is that chasing the expert is a mistake, and a costly one at that. We should stop hunting and ask the crowd."
In pursuing this thesis, Surowiecki is bucking a long-held conviction that groups bring out the worst, rather than the best, in people. This tradition is perhaps best summarized by Henry David Thoreau, quoted as stating, "The mass never comes up to the standard of its best member, but on the contrary degrades itself to a level with the lowest." More entertaining is the H. L. Mencken quotation: "No one in this world, as far as I know, has ever lost money by underestimating the intelligence of the great masses of the plain people."
A simple experiment exemplifies the basic reasoning. Place a large quantity of jellybeans in a jar. Have a group of people individually guess the number of jellybeans. The result: The average of all the guesses will be closer to the actual number than the guess of almost every individual. Surowiecki cites, for example, a finance class where only one of the 56 people was closer to the true count, 850 jellybeans, than the average guess of 871.
Early morning OJ
Of course, jellybean estimates are one thing; life decisions, quite another. But Surowiecki finds collective wisdom in many places the reader may not think about, but benefits from nonetheless. Here are some examples:
- Ever been impressed with how quickly Google can locate just the page you want? The PageRank software underlying Google relies on a methodology akin to a popularity poll of relevant Web pages.
- Know any techies that swear by the Linux operating system? This free software is created and continually improved by a decentralized crowd of individuals worldwide.
- Financial markets provide an example similar to the jellybean experiment; a recent study found that 90 percent of mutual fund managers (that is, experts) underperformed the unmanaged Wilshire 5000 Index.
- The identification of the SARS virus was a remarkable feat that was accomplished not by one or two experts having a eureka moment, but by the collective effort of a decentralized group of labs worldwide approaching the problem from different directions and sharing their results.
In fact, Surowiecki argues markets themselves are a powerful tribute to the wisdom of crowds. Each day millions of unorganized individual buyers and sellers determine the appropriate price and location for thousands of goods and services, "getting resources to the right places at the right cost" without any overall coordination of their activities. How else, asks Surowiecki, could you explain why you can find the carton of OJ you're looking for at a nearby convenience store at 2 a.m.?
The central theme of Wisdom is that groups can be collectively wise. But much of the book focuses on the obvious concern: Sometimes groups get it wrong. Surowiecki discusses at length the characteristics that are necessary for groups to achieve collective wisdom: diversity, independence, decentralization and a mechanism for effectively aggregating individual input.
Surowiecki also argues that groups are more conducive to providing good answers to some types of questions than to others. Questions that have, or will have, a definitive answer are more amenable to collective wisdom than are various coordination and cooperation problems such as traffic congestion and taxation. But given the proper environment, he suggests, groups can effectively address a wide range of problems. The first half of the book provides this analytical description of what's required for collective wisdom. Then Surowiecki turns to the meat and potatoes: case studies of groups making wise decisions and cases where groups went wrong.
Diversity and independence
Surowiecki argues that cognitive diversity is critical to collective intelligence because it adds perspectives that might otherwise be absent. Here he points to the failure of the Kennedy administration in the Bay of Pigs debacle and the current information failures of the CIA in Iraq. In both cases, the author argues, "groupthink" was allowed to take over, and dissenting views were unwelcomed or ignored.
Independence is another important group property because it keeps individual mistakes from cumulating across people; with independent thought, the group is more likely to acquire new information. The author contends that when people stop acting independently and start relying on conventional wisdom or imitation to make decisions (fittingly referred to as "herding behavior"), poor results can quickly follow. Here he points to the telecommunications stock bubble of the late 1990s. Surowiecki also summarizes entertaining research indicating that National Football League coaches tend to follow conventional wisdom and play conservatively on fourth down, though economic models indicate that going for the first down more often would be a better choice.
Diversity and independence are also critical in determining what new technologies and products should be pursued, according to Surowiecki. The early 20th century saw hundreds of independent automobile makers, for example, producing a wide variety of automobiles based on electricity, steam and gasoline. While such a process may seem wasteful, "no system seems all that good at picking winners in advance," he writes. "What makes a system successful is its ability to recognize losers and kill them off quickly." Counting on "experts" to pick winning technologies is almost always a mistake, he argues. Markets—representing the collective wisdom—will do the job far better.
Decentralize and aggregate
Decentralization and an effective aggregation mechanism are the final essential qualities for groups to reach good decisions, in Surowiecki's opinion. Decentralization allows for specialization and individualized knowledge about a problem, properties that are essential for collective intelligence. By "aggregation mechanism," he means a procedure for gathering input from every member of the decentralized group.
But as essential as these characteristics are for achieving collective wisdom, Surowiecki concedes that they can be hard to achieve. Voting and markets provide effective aggregation, but they're not always available. And decentralization can lead to disorganization. The recent failures of the CIA, says Surowiecki, are due largely to the lack of an effective strategy for collecting data from the agency's vast network of very knowledgeable decentralized specialists; centralizing the agency, though, would not result in better decisions but simply to greater groupthink.
The book goes on to explore other thought-provoking case studies of group failures and successes: traffic congestion, Long-Term Capital Management, CNBC, the Challenger and Columbia shuttle disasters, plank roads, taxes, restaurant tipping, the bowling craze of the 1950s and even the beneficial effect of capitalism on trust and cooperation.
The author also argues that modern businesses would do well to consider employing the wisdom of crowds. For example, he proposes tapping into the collective intelligence of employees by setting up internal decision markets whereby employees bet small amounts on the likely success of individual products, new marketing strategies and other relevant events.
So how convincing is Surowiecki's case for the wisdom of crowds? And for that matter, how well did Levitt and Dubner enlighten us about "the hidden side of everything"?
The wisdom of two books
The primary attraction of Freakonomics is the entertaining tales and insights, which make for great stories to share with family, friends and co-workers. Based on book sales and the abundance of talk-show appearances by the authors, the book has been a smashing success on that dimension. But if the engaging and offbeat stories are the strength of the book, they also present its biggest shortcoming. Levitt's research—as well as this book—is sometimes criticized for focusing on attention-grabbing curiosities rather than on fundamental issues of lasting importance. While each of his stories is entertaining and often thought-provoking, they don't add up to a substantive unifying theme or insight, a fact that Freakonomics' authors themselves acknowledge.
Still, even if the book doesn't provide profound economic insights, it does succeed in entertainingly conveying some of economics' basic lessons, such as the importance of incentives and information in analyzing behavior and outcomes. And if Freakonomics increases public awareness of economics and motivates students to learn more about the field, all the better.
Wisdom, for its part, has somewhat the opposite strengths and weaknesses. The book proposes an interesting, broad theme and then uses a wide array of real world examples and related research findings to make its case. But whereas Freakonomics is able to bring specific data to bear on its disparate list of questions, Surowiecki's task is in many respects more challenging due to the sweeping nature of his thesis.
Surowiecki writes artfully, and his theme is skillfully woven through a series of intriguing examples. But in the end his book does not present a completely convincing argument for its thesis. The case study methodology leaves the reader to wonder occasionally whether the examples have been judiciously chosen, or whether another explanation may do as well. Are we really sure, for instance, that a couple of economists with a model know more than NFL coaches—doesn't the collective wisdom and experience of the coaches, staff and players count for something, as long as they're not biased by perverse incentives? (Incidentally, Gladwell's Blink favors these informed hunches. Believe me when I tell you it's a dazzling book.)
Beware of experts
Whatever their shortcomings may be, both books make the same key point: Don't rely too heavily on the advice of experts when making important decisions. And why not? Freakonomics argues that experts, like everyone else, have self-interest and incentives, and they're wont to exploit their informational advantage for personal gain. Wisdom emphasizes that no matter how knowledgeable and virtuous experts are, groups can often do better, if appropriately organized.
Given this dual assault on experts, I'm reluctant to provide a personal recommendation. Instead, following the wisdom gained from reading Surowiecki's book, I'll conclude by providing the advice of a diverse, independent, decentralized crowd of readers. According to online reviewers at Amazon.com, each book merits four out of five stars. You decide the wisdom of that crowd.