For the first few days it worked well with sweets changing hands in return for timely toilet visits. On the third or fourth day Levitt, 38, took Amanda to the toilet. She passed just a dribble and took the sweets. A few minutes later she was back on the toilet, passing yet another dribble and putting her hand out for yet more sweets. She'd rumbled the system. Levitt smiles.
"I never thought my three-year-old daughter could outwit my incentives in just three or four days," he says. "But it's a great example of how incentives can have unpredictable effects."
Welcome to Levitt's world - the unintended and unexpected outcomes arising from various initiatives and incentives; the peculiar relationships between things one would generally not relate. His methods are intriguing - comparing the behaviour of some Chicago school teachers to sumo wrestlers (both have incentives to cheat) and the Ku Klux Klan to estate agents (they both derive their power from secret information). His subjects are original - examining a study of a crack-running gang on Chicago's South Side to see why drug dealers live with their mothers.
His conclusions are, to some, offensive. In one chapter in Freakonomics he argues that the crime rate in the United States plummeted in the 90s because more liberal abortion laws in the 70s meant fewer potential criminals were being born. In another, he claims that children born to uneducated poor parents are more likely to be unsuccessful, regardless of how they are brought up. "It isn't so much a matter of what you do as a parent; it's who you are." Both are arguments that could be used to justify eugenics.
And his popularity is skyrocketing. Not only has the book, which was co-written by New York Times journalist Steven Dubner, become a bestseller since its publication earlier this year; but he has a monthly column in the New York Times and a regular stint on NBC's Today Show. The CIA invited him to talk to them about whether his method could be used to track down terrorists. Even before Freakonomics, the Clinton administration invited him to work as an adviser: "The Monica Lewinsky thing had just happened and it didn't seem like they were going to be doing much economic policy anyway." And Team Bush called him to ask him to advise them on crime in the run up to the 2000 election. "I told them they should have a look at what I said about abortion first ... They never called me back."
"I don't have strong political views," says Levitt. "I don't find the idea of politics itself very interesting. I try to ask questions that are of personal interest to me. Questions where conventional wisdom will probably be wrong and problems that are really easy but look really hard."
Levitt, an economics lecturer at the University of Chicago, refers to himself as a data detective. "I'm not a great economist," he says. "I'm not an intellectual. But everybody has got their framework for the way they think about the world. Mine just happens to be a micro-economic framework. It's the story of how people respond to incentives. The scripts that run through my head tend to be those kinds of scripts."
Levitt's scripts reveal both tragedy and comedy strung together with an apparently random stream of consciousness. In his research on state school teachers in Chicago he discovered that 5% of the city's teachers inflated their test scores after the city introduced high-stakes testing in 1996. His algorithmic evidence was sufficiently compelling that the city retested a number of children and then fired teachers who they believed were cheating.
He used a similar approach to prove that sumo wrestling in Japan was also rigged with superior wrestlers throwing high-stakes bouts, when a loss would inflict far more damage on their opponent's prospects than a victory could help their own. What links the wrestlers and the teachers, he says, is their incentive to cheat.
"So if sumo wrestlers, schoolteachers, and day-care parents all cheat, are we to assume that mankind is innately and universally corrupt? And if so how corrupt? The answer may lie in ... bagels," he says in his book. Levitt then studies the experience of Paul Feldman, who delivers bagels to companies and lets people pay on an honour system and collects data on how often they pay. His figures show that people are more honest when they work in a smaller office, in pleasant weather and closer to national holidays (though not all - holidays that carry attendant stress, such as Christmas and Thanksgiving, can have adverse effects).
To the extent that Levitt has a method this is it - searching for numerical data that will answer questions and make or break connections that have become embedded "conventional wisdom".
His work shows no overarching theme or underlying philosophy that draws it altogether. Recently awarded the John Bates Clark medal which is given every two years to the best American artist under 40, his work has been embraced by economists. "I thought when I went into sumo wrestling that I'd finally crossed the line," he says. "But it was published by the American Economic Review," one of the most prestigious academic journals in the field.
Levitt just follows the facts, setting out with a question in mind rather than an answer and hoping the data will lead him there. If some holy cows get slaughtered in the process then so be it. "Facts are amoral," says Levitt. "I don't think we should shy away from facts that people don't want to hear."
But he does have a worldview, spelt out at the beginning of the book in a few basic tenets: "Incentives are the cornerstone of modern life; the conventional wisdom is often wrong; dramatic effects often have distant, even subtle causes; experts use their informational advantage to serve their own agenda; knowing what to measure and how to measure it makes a complicated world much less so."
"Facts speak only when the historian calls on them," wrote the historian EH Carr in his landmark work, What Is History? "It is he who decides which facts to give the floor to and in what order or context. It is the historian who has decided for his own reasons that Caesar's crossing of that petty stream, the Rubicon, is a fact of history, whereas the crossing of the Rubicon by millions of other people before or since interests nobody at all."
For historian here read data detective. Levitt may be right when he says facts have no intrinsic morality but they do have an inherent power. That power is shaped by context. It depends on who collects the facts, why they collect them and about whom they are being collected. The most glaring contemporary example is the war in Iraq - we know how many US soldiers have died because they have been counted. We can only guess how many Iraqi civilians or insurgents have died because they have not been counted. Our conclusion on how the Iraq war is going based purely on the available data regarding casualties will inevitably be skewed.
None of this makes Levitt any less of a liberating thinker; he is meticulous in his analysis, rigorous in his conclusions and both inventive and accessible in his style. It does, however, put a question mark over his claim to have liberated himself from conventional wisdom since the facts themselves are the product of conventional wisdom. At one point in his chapter about crime he concedes that increased imprisonment "is one of the key answers ... to explain the drop in crime in the 1990s". Different facts tell a different story. As a deliberate result in policy, between the 70s and the 90s Finland went from having one of the highest prison populations to one of the lowest; meanwhile both the patterns and rates of crime remained consistent with its Nordic neighbours.
In person Levitt is extremely self-deprecating: "I've always loved the opportunity to be the dumbest guy at the table and to learn from people who know much more than I do," he says. But in print he can appear bullish. "Morality," he writes, "represents the way that people would like the world to work - whereas economics represents how it actually does work."
He admits that his method demands that the reader abandons their ethical judgment for the span of his argument. "If people can suspend their moral feeling for just 30 seconds then they can see that while they might not like a conclusion it still makes sense," he says. "You're not supposed to suspend your feelings for ever."
Nowhere is this approach more evident than in his claim that more liberal abortion laws are primarily responsible for the precipitous drop in the crime rate. "One study has shown that the typical child who went unborn in the earliest years of legalised abortion would have been 50% more likely than average to live in poverty [and] 60% more likely to grow up with just one parent. These two factors - childhood poverty and a single-parent household - are among the strongest predictors that a child will have a criminal future ... The crime drop was, in the language of economics, an 'unintended benefit' of legalised abortion."
When it was first reported, Levitt was denounced from both the left and the right. "I just didn't answer my phone," he says. "People called and left angry messages but I didn't answer. I didn't do any media ... Politicians use my work how they want. They decide what they want to do and then they go in search of the facts to do it. There's nothing I can do about that. What's great about this book is we get to tell our own story."
Levitt's facts may speak for themselves and in so doing they have certainly sparked both lively and heated conversations throughout the United States. But, like his trick with the M&M's, there may be a considerable incentive for others to use them cynically with consequences that were neither intended nor desirable.
