# Freakonomics by Steven D. Levitt and Stephen J. Dubner Source: https://thegisthub.com/books/freakonomics-by-steven-d-levitt-and-stephen-j-dubner Last updated: 2026-10-24 Published 2005 · 320 pages · ISBN 9780061234002 · Categories: Money and investing, Society and history Economist Steven Levitt and journalist Stephen Dubner use data to answer odd questions about cheating teachers, drug dealers, crime and parenting. For anyone who likes to see conventional wisdom tested. ## The gist Freakonomics applies the tools of economics to questions that economists usually ignore. Its method rests on a few ideas: **incentives** drive behavior, **conventional wisdom** is often wrong, big effects can have distant causes, experts use their **information advantage** for themselves, and good data can expose all of this. Levitt and Dubner show teachers and sumo wrestlers cheating, real-estate agents keeping their own homes on the market longer, drug dealers earning less than minimum wage, and the 1990s crime drop tracing back to legalized abortion. On parenting, they find that who parents are predicts a child's test scores more than what parents do. ## Key ideas ### 1. Incentives are the cornerstone of modern life The authors argue that most puzzles about behavior dissolve once you find the incentive. They name three kinds: economic, social and moral. An Israeli day-care center introduced a small fine for late pickups and late pickups rose, because a cheap price replaced the guilt that had kept parents on time. Chicago teachers facing high-stakes tests filled in answer sheets. Sumo wrestlers with nothing at stake lost to opponents who needed a win. Change the incentive and you change the behavior, often in ways the designer did not intend. ### 2. Data can catch cheating Levitt built an algorithm to find classrooms where students shared long strings of identical answers on hard questions while missing easy ones, and where scores jumped one year and fell the next. The district retested the suspect classrooms under supervision, scores dropped, and teachers were dismissed. In sumo, 7-7 wrestlers on the final day beat 8-6 opponents far more often than their records predict, then lose the return match. The authors' point is that cheating leaves patterns, and the right question reveals them. ### 3. Experts exploit information asymmetry When one side of a deal knows more, it uses the gap. Real-estate agents earn a small share of the sale price, so a quick sale matters more to them than a higher price for the client. Chicago data show agents keep their own homes on the market about ten days longer and sell for about 3 percent more. Stetson Kennedy destroyed the Ku Klux Klan's mystique by leaking its secrets to the Superman radio show. The internet has cut expert advantages in markets like term life insurance. ### 4. Conventional wisdom is often manufactured The authors borrow the term from John Kenneth Galbraith and argue that received opinion is often shaped by people with reasons to shape it. Advocate Mitch Snyder claimed three million Americans were homeless, a number he later admitted making up for reporters, yet it circulated for years. Experts predicted a 1990s wave of young superpredators just before crime collapsed. The lesson is to ask who made a widely repeated number and what they gained from it. ### 5. Drug dealing is a tournament with bad odds Sociologist Sudhir Venkatesh obtained the financial records of a Chicago crack gang. The structure looked like a franchise. The leader, J.T., earned about $100,000 a year, and a small board above him earned far more. The foot soldiers on the street, who faced the highest risk of arrest and death, earned about $3.30 an hour and lived with their mothers. The authors explain this as a tournament: like aspiring actors, foot soldiers accept bad pay for a small chance at the top. ### 6. The 1990s crime drop traces back to Roe v. Wade Crime fell sharply from the early 1990s after experts predicted a rise. More police, more prisons and the end of the crack boom each helped. The authors' main explanation, from Levitt's work with John Donohue, is that legalized abortion in 1973 reduced the number of children born into the conditions most associated with later crime. The first post-Roe cohort reached peak crime age around 1991. States that legalized earlier saw crime fall earlier. A 2005 critique found a coding error and a weaker effect, and the dispute continues. ### 7. Who parents are matters more than what parents do Using the Early Childhood Longitudinal Study, the authors sort factors by whether they correlate with children's test scores. Correlated: educated parents, high socioeconomic status, a mother over thirty at first birth, many books in the home. Not correlated: an intact family, a mother who stayed home, Head Start, museum visits, television, daily reading. The correlated factors describe the parents' identity. The uncorrelated ones describe their actions. The authors stress that this is correlation and concerns test scores only. ### 8. A name reflects a family, it does not shape a child Roland Fryer's study of California birth records shows that distinctively Black names rose sharply in the early 1970s and now signal a child's socioeconomic background. But once background is controlled for, the name itself predicts nothing. The authors also show how names flow through society: high-status parents adopt a name first, and it spreads downward over a decade or two. The brothers Winner and Loser Lane, one a detective and one a repeat offender, illustrate how little a name decides. ## Questions this book answers ### What is Freakonomics about? Freakonomics by Steven D. Levitt and Stephen J. Dubner applies economic analysis to unusual questions. It shows how incentives explain cheating by teachers and sumo wrestlers, how information asymmetry lets real-estate agents and experts profit, why most drug dealers earn less than minimum wage, why American crime fell in the 1990s, and which parenting factors actually predict children's test scores. The book's method is to find the data and follow them even when they contradict conventional wisdom. ### What does Freakonomics say about abortion and crime? The book presents work by Levitt and John Donohue arguing that the legalization of abortion in 1973 contributed to the sharp fall in American crime in the 1990s. Children who would have been born into the conditions most associated with later crime were not born, and the first post-Roe cohort reached peak crime age around 1991. In 2005, Foote and Goetz found a coding error and a weaker effect. Donohue and Levitt replied that the result held, and economists still debate it. ### What are the main ideas in Freakonomics? Five ideas hold the book together. Incentives are the cornerstone of modern life. Conventional wisdom is often wrong because it is shaped by people with reasons to shape it. Dramatic effects often have distant causes. Experts use their information advantage to serve themselves. And knowing what to measure makes a complicated world simpler. Each chapter applies these ideas to a case, from cheating teachers to baby names. ### What does Freakonomics say about parenting? Using data from the Early Childhood Longitudinal Study, the authors find that factors describing who parents are (education, income, the mother's age at first birth, books in the home) correlate with children's test scores, while factors describing what parents do (daily reading, museum visits, staying home, Head Start) do not. They argue that parents' identity is largely set before the child is born, and they stress that this is correlation and concerns test scores only. ### Why do drug dealers still live with their moms? Sociologist Sudhir Venkatesh obtained the financial records of a Chicago crack gang. The structure resembled a franchise. The leader earned about $100,000 a year and a small board above him earned much more, but the foot soldiers who sold on the street earned about $3.30 an hour while facing the highest risk of arrest and death. They lived with their mothers because they could not afford to live elsewhere. The authors call it a tournament: bad pay now for a small chance at the top. ## Quotes > Incentives are the cornerstone of modern life. (Introduction, stating the book's first premise) > Morality, it could be argued, represents the way that people would like the world to work, whereas economics represents how it actually does work. (Introduction, on the book's approach) ## Full summary ### The hidden side of everything Steven Levitt is an economist at the University of Chicago who won the John Bates Clark Medal in 2003. Stephen Dubner is a journalist who profiled him for The New York Times Magazine. Freakonomics grew out of that profile. The book has no single theme. Its unity comes from a method: ask an odd question, find the data that bear on it, and follow the numbers even when they contradict what everyone believes. The introduction lays out the ideas that hold the chapters together. **Incentives** are the cornerstone of modern life, and most puzzles dissolve once you find the incentive at work. **Conventional wisdom** is often wrong, because it is made by people with reasons to shape it. Dramatic effects often have distant, even subtle, causes. **Experts** use their information advantage to serve their own interests. And knowing what to measure, and how, makes a complicated world much simpler. The authors preview two of the book's stories here. One is the sharp fall in American crime in the 1990s, which experts had predicted would rise. The other is a real-estate agent who, when selling her own house, waits longer and gets a higher price than she does for her clients. Both stories, they argue, make sense once you look at incentives and data instead of received opinion. ### What do schoolteachers and sumo wrestlers have in common? The first chapter is about cheating, and it opens with a day-care center in Israel. When the center introduced a small fine for parents who picked up their children late, late pickups increased. The fine replaced a moral incentive (guilt) with an economic one (a cheap price), and parents bought the extra time. The authors use this to introduce three kinds of incentive: economic, social and moral. The main case is the Chicago public schools. After high-stakes testing arrived, some teachers had an incentive to raise their students' scores by any means. Levitt and a colleague built an algorithm to find classrooms where strings of identical answers appeared on hard questions while students missed easy ones, and where scores jumped one year and collapsed the next. The pattern pointed to teachers filling in answer sheets. The district retested suspect classrooms under controlled conditions. Scores fell, and several teachers were dismissed. The second case is sumo wrestling in Japan. A wrestler needs eight wins in a fifteen-bout tournament to keep his rank. The authors look at bouts where one wrestler is 7-7 on the final day and the other is 8-6, with nothing at stake. The 7-7 wrestler wins far more often than his record predicts, and he then loses to the same opponent more often than expected the next time they meet. The data fit a quiet exchange of favors. The chapter ends with **Paul Feldman**, who left a job to deliver bagels to offices on the honor system. His payment records show that about 87 percent of people paid. Smaller offices paid more reliably than large ones, executives paid less reliably than lower-level staff, and payment fell around Christmas and rose after the September 11 attacks. The authors read this as evidence that most people are honest most of the time, and that the conditions around them change the rate. ### How is the Ku Klux Klan like a group of real-estate agents? The second chapter is about **information asymmetry**: one party to a deal knows more than the other and uses the gap. The historical case is **Stetson Kennedy**, who infiltrated the Ku Klux Klan in the 1940s and passed its passwords, rituals and internal gossip to the producers of the Superman radio show. Once the Klan's secrets became material for a children's program, its mystique collapsed and recruitment fell. Information, the authors argue, was the weapon. The modern case is real-estate agents. Their commission is a small share of the sale price, so an extra few thousand dollars for the seller means little to the agent, while a quick sale means a lot. Levitt's data from Chicago home sales show that agents selling their own homes keep them on the market about ten days longer and sell them for about 3 percent more than they get for clients. The authors also decode listing language: words like "fantastic" and "charming" signal a weak house, while specific words like "granite" and "maple" signal a strong one. The chapter adds other examples. Term life insurance prices fell sharply once price comparison sites made them public. Car mechanics, doctors and funeral directors all benefit from clients who cannot judge their advice. Online daters inflate their attributes in predictable ways. The internet, the authors argue, has cut the value of expert information advantages across many markets. ### Why do drug dealers still live with their moms? The third chapter opens with a warning about **conventional wisdom** and its makers. Advocates and experts have reasons to produce numbers that help their cause. The homeless advocate Mitch Snyder once claimed three million Americans were homeless, a figure that spread for years before he admitted it was made up for the press. The heart of the chapter is the work of sociologist **Sudhir Venkatesh**, who spent years with a Chicago crack gang, a branch of the Black Gangster Disciple Nation. The gang's leader, **J.T.**, kept detailed financial records, and Venkatesh obtained them after the gang broke up. The books showed a structure much like a franchise. A small board of directors at the top earned a great deal. J.T. earned about $100,000 a year. Below him, three officers earned decent wages. The foot soldiers, who sold the drugs on the street and faced the highest risk of arrest and death, earned about $3.30 an hour. Most lived with their mothers because they could not afford anything else. The authors explain the structure as a **tournament**. Like aspiring actors or athletes, foot soldiers accept bad pay for a small chance at the top position. The chapter also traces the crack boom of the 1980s, which turned gangs from social clubs into businesses and drove the violence that followed. ### Where have all the criminals gone? The fourth chapter is the book's most debated. Violent crime in the United States rose through the 1980s, and criminologists predicted a coming wave of young "superpredators." Instead, crime fell sharply from the early 1990s onward. The authors test the common explanations. Innovative policing in New York, they argue, gets more credit than the data support, since crime fell in cities that changed nothing. Rising prison populations and more police officers both contributed. The end of the crack boom helped. Gun control laws, a strong economy and an aging population explain little. They then present the explanation from Levitt's work with John Donohue. The Supreme Court legalized abortion nationwide in 1973 with Roe v. Wade. The children most likely to be aborted after that date were those who would have been born to young, poor, unmarried mothers, and those are the children with the highest statistical risk of becoming criminals. Around 1991, the first cohort born after Roe reached the age at which crime peaks, and crime fell. States that had legalized abortion earlier saw crime fall earlier, and states with high abortion rates saw larger declines. The authors stress that this is a statistical finding, not a policy proposal. They also describe the opposite case: Romania under Nicolae Ceausescu banned abortion in 1966, and the cohort born after the ban did worse in school and in work and committed more crime than those born just before. ### What makes a perfect parent? The fifth chapter turns to parenting, and it opens with risk. A child is roughly a hundred times more likely to die in a swimming pool than from a gun in the home, yet parents fear the gun far more. The authors use this to argue that people misjudge risks that are unfamiliar or outside their control. The main data come from the **Early Childhood Longitudinal Study**, a federal survey of about 20,000 children. Using regression analysis, the authors sort parenting factors into those that correlate with test scores and those that do not. Correlated: the parents are highly educated, have high socioeconomic status, the mother was thirty or older at her first child, the child had low birth weight (negatively), the parents speak English at home, the child is adopted (negatively), the parents are involved in the PTA, and the home has many books. Not correlated: the family is intact, the parents recently moved to a better neighborhood, the mother did not work between birth and kindergarten, the child attended Head Start, the parents take the child to museums, the child is spanked, the child watches television, and the parents read to the child nearly every day. The pattern, they argue, is that the correlated factors describe who the parents are, while the uncorrelated factors describe what the parents do. Books in the home predict scores not because reading them raises scores but because the kind of parents who own books also do many other things that matter. The authors are careful to say this is correlation and that it concerns test scores, not happiness or character. ### Would a Roseanne by any other name smell as sweet? The final chapter asks whether a child's name affects their life. It opens with a New York father who named one son Winner and another Loser. Loser Lane became a police detective. Winner Lane built a long criminal record. The data come from economist **Roland Fryer**'s study of California birth records. Distinctively Black names rose sharply in the early 1970s and are now strong indicators of a child's socioeconomic background. But when the authors control for that background, the name itself does not predict outcomes. A child's name reflects the parents' circumstances rather than shaping the child's future. The chapter also tracks how names move through society: high-income, high-education parents adopt a name first, and it spreads downward over a decade or two until the original adopters move on. An epilogue returns to two boys introduced earlier, one white and raised in a stable home, one Black and abandoned by his mother. The white boy became the Unabomber, Ted Kaczynski. The Black boy became the economist Roland Fryer. The authors offer the pair as a caution against confident predictions. ### What the critics say The abortion and crime chapter drew the most scrutiny. In 2005, economists Christopher Foote and Christopher Goetz of the Federal Reserve Bank of Boston reported a coding error in the Donohue and Levitt paper and found that, after correcting it and adjusting for population, the effect was much weaker. Donohue and Levitt replied that the finding held with the corrections, and they published updated work in 2019 with more years of data. The dispute remains open among economists. More broadly, some critics argue that the book presents clever correlations as causes and that its title oversells a loose collection of essays. ### Who should read it Read it if you want to see how an economist tests a claim with data, and if you enjoy having common beliefs overturned. It is short, fast and easy to argue with. It is not a guide to economics or to policy. ## Do this week - Pick one rule at work or home that people keep breaking, and write down what the current incentive rewards. - Find one statistic you have repeated this month, trace it to its original source and note who produced it. - Before your next purchase from an expert (mechanic, agent, advisor), write down how their pay depends on your choice. - Take one belief you hold about parenting or health and look for data that separates correlation from cause. - Read the Foote and Goetz critique summary and the Donohue and Levitt reply, then write three sentences on which you find stronger. Summary by The Gist Hub (https://thegisthub.com). In our own words, not the book's text.