Freakonomics, by Steven D. Levitt and Stephen J. Dubner

If there is no unifying theme to Freakonomics, there is at least a common thread running through the everyday application of Freakonomics. It has to do with thinking sensibly about how people behave in the real world. All it requires is a novel way of looking, of discerning, of measuring. Steven D. Levitt and Stephen J. Dubner

Steven D. Levitt (born in 1967) is an American economist and author, who won the John Bates Clark Medal for his work on crime, was co-editor of the Journal of Political Economy, and teaches economics at the University of Chicago. His co-author Stephen J. Dubner (born in 1963) is an American journalist (he published in the New York Times, The New Yorker, and Time), author, and podcast and radio host. Dubner initially wrote a profile on Levitt for The New York Times Magazine, and they eventually wrote a series of books together, including Freakonomics, SuperFreakonomics, Think Like a Freak, and When to Rob a Bank. I first read Freakonomics in 2010, and was immediately appealed by the authors’ focus on real-life topics, their uncanny ability to quantify things, and their unconventional findings. Below is a summary of the book:

Incentives:

• Economics is the study of incentives.
• An incentive is a means of urging people to do more of a good thing and less of a bad thing.
• There are 3 types of incentives: (1) Economic, (2) Social, and (3) Moral.

Trade-offs and unforeseen results:

Incentives are trade-offs: a slight tweak can produce drastic and unforeseen results.

• The enactment of a $3 fine for late pick-ups at several day-care centers resulted in more late pickups (the guilt from being late was substituted by an easier to bear economic penalty, and the small fine sent a signal that late pickups weren’t such a big problem).
• Paying blood donors a small stipend resulted in less donations (the stipend turned a noble act of charity into a painful way to make a few dollars).

Cheating:

Whatever the incentive, dishonest people will try to cheat:

• On April 15, 1987, seven million American children suddenly disappeared: the IRS had requested tax filers to provide a Social Security number for each dependent child.
• Studies and testimonies have shown extensive cheating in such unsuspecting groups as (1) schoolteachers (those whose students test poorly can indeed be censured or passed over for a raise or promotion, while those who do well may receive praise, promotions, or bonuses), (2) sumo wrestlers (some who secure enough victories in a tournament to get promoted agree to lose their last bout to opponents on the fence, in quid pro quo arrangements), or (3) white collar office workers (a bagel salesman who delivered boxes of bagels to offices in the morning and collected cash later in the day on a pay-per-use basis found that 13% of the bagels eaten were not paid for, with a higher than average propensity to cheat in bigger companies, at higher echelons, during bad weather periods, and during lengthy holidays).

To catch a cheater, think like a cheater (understand their incentives and the way they may go about to limit suspicions) and look for unusual patterns:

• e.g., to uncover schoolteacher cheating, look for blocks of identical answers among the harder questions, students getting the hard questions right while missing the easy ones, or students testing far better than past scores then falling significantly the following year.
• e.g., to uncover match-rigging, look for situations in which some players have far more to gain from a victory than an opponent has to lose, compare the predicted win percentage with the actual win percentage and likewise with subsequent rematches.

The power of fear:

• The Ku Klux Klan was founded by former Confederate soldiers after the Civil War.
• In the beginning, their actions were pranks, but the organization soon evolved into a terrorist organization designed to frighten and kill emancipated slaves and whites who supported their right to vote, acquire land, or gain an education.
• The lynchings were rare but worked, as few incentives are more powerful than the fear of random violence.
• The public was frightened and powerless, and opponents had little leverage or information about the Klan.

The power of information:

Sunlight is said to be the best of disinfectants Supreme Court Justice Louis D. Brandeis

A Floridian man named Stetson Kennedy, who set out to damage the Klan in the 1940s, initially wrote to the attorney general and the governor of Georgia with evidence to have their charter revoked, but to no avail. He then proceeded to turn the Klan’s secrecy against itself by making private information (obtained thanks to an infiltrated agent) public on a radio show, converting precious knowledge into ammunition for mockery. He understood that the KKK’s power was derived in large part from the fact that it hoarded information: once that information was leaked, the group’s advantage disappeared. The power of information is also illustrated in the following examples:

• In the late 1990s, the price of term life insurance fell dramatically, causing the industry’s revenues to reduce by $1 billion. The culprit was the Internet, specifically websites enabling customers to compare the prices of term life insurance sold by dozens of companies.
• A brand-new car that is driven off the lot instantly loses as much as a quarter of its value, as potential buyers assume that the seller has found it to be a lemon.

The power of experts:

Many experts use their expertise to your detriment: they depend on the fact (1) that you don’t have the information they do, (2) that you are confused by the complexity of their operation and wouldn’t know what to do with the information if you had it, or (3) that you are in awe of their expertise and wouldn’t dare challenge them. Armed with information, experts can exert fear:

• Fear of suffering from a heart attack if you do not have angioplasty surgery.
• Fear that a cheap casket will expose your dead grandmother to a terrible fate.
• Fear that a $25,000 car will crumple in an accident, whereas a $50,000 car will protect your loved ones.
• Fear that you will sell your house for far less than it is worth or not be able to sell it at all if you do not use a real-estate agent.

Question conventional wisdom:

Some theories or statements become articles of faith because they appeal to factors that, according to John Kenneth Galbraith, contribute to the formation of conventional wisdom: (1) the ease with which an idea may be understood, and (2) the degree to which it affects our personal well-being.

Experts

Experts must be bold if they hope to convert their theories into conventional wisdom (arguments with a lot of restraint or nuance often don’t get much attention). Their best chance is to engage the public’s emotion, notably fear, the most potent one. In the early 1980s, Mitch Snyder, an advocate for the homeless, took to saying that there were 3 million homeless Americans, and even testified before Congress. He later admitted that the figure was a fabrication: journalists had been pressing him for a specific number, and he did not want them to walk away empty-handed. Parenting experts contradict one another and even themselves, making conventional wisdom on parenting seem to shift by the hour. Crime experts in the mid-1990s forecast a “bloodbath” of youth violence, which never materialized. Note: this reminds me of unsourced conventional wisdom statements I had heard in my early private equity years (e.g., “Decorative paints is driven 75% by renovation and 25% by new build”, or “There is a deficit of 900,000 housings in France”).

Advertisers

Listerine did not make mouthwash as much as it made halitosis James B. Twitchell

Listerine was invented in the 19th century as a surgical antiseptic. It wasn’t a runaway success until the 1920s, when it was pitched as a solution for “chronic halitosis”, i.e. bad breath. This condition had not been considered to be catastrophic, until Listerine came out with ads picturing young women eager for marriage but turned off by their mate’s bad breath.

Rights advocates

Women’s rights advocates have hyped the incidence of sexual assault, claiming that 1 in 3 American women will be a victim of rape or attempted rate, when the actual figure is closer to 1 in 8. But they know it would take a callous person to publicly dispute their claims.

Police

The police in Atlanta was found to have been radically underreporting crime since the early 1990s: the city needed to shed its violent image to secure the 1996 Olympics.

Fear of death

The risks that scare people and the risks that kill people are very different. Peter Sandman

Contrary to conventional wisdom, it is safer to let your child attend a playdate at a house with a gun than at a house with a swimming pool. Annual child deaths from drowning in residential pools amount to 550 out of 6 million pools, so the odds are 1 in 11,000, while annual child deaths from gunfire amount to 175 out of 200 million guns, so the odds are 1 in 1 million. According to risk communications consultant Peter Sandman, several factors may distort our perception of risk, notably (1) the “control factor“, linked to how much control we believe we have over the risk (most people are more scared of flying in an airplane, over which they have no control, than driving a car, which they believe to control, although the per-hour death rate of driving versus flying is about equal), and (2) the “dread factor“, linked to the perceived immediacy and outrage of the risk (more people are afraid of terrorist attacks, which can be immediate and dreadful, than heart attacks from fatty food consumption, although the risk of death from the former is much lower than the risk of death from the latter). Note: the “dread factor” and the “control factor” bring to mind to the following cognitive biases: the “salience bias” (tendency to focus on items that are more emotionally striking), the “illusion of control” bias (tendency to overestimate one’s degree of influence over external events) and the “overconfidence effect” bias (tendency to overestimate one’s abilities or judgements).

Drug dealers

Examining the detailed financial accounts of a crack gang, Sudhir Venkatesh and the authors concluded that contrary to conventional wisdom, most drug dealers did not make much money ($3.30 an hour for foot soldiers), especially in light of their terrible working conditions (arrest, injuries, and 1-in-4 chance of being killed).

The laws of labor

The reason drug dealers took such a job is that they wanted to succeed in a very competitive field in which, like acting or sports, if you reach the top, you are paid a fortune ($100,000 for franchise leaders, $500,000 for top bosses), with the attendant glory and power. The problem with “glamour professions” (such as sports, movies, fashion, or to a lesser extent advertising, publishing, and media) is that a lot of people are competing for very few prizes, must be willing to start at the bottom, work long and hard at substandard wages, and in order to advance, prove themselves not merely above average but spectacular. The main factors that determine a wage are (1) the supply of people willing and able to do the job (2) specialized skills, (3) unpleasantness, and (4) demand for services that the job fulfills. The delicate balance between these factors explains why drug dealers do not earn much (a lot of people willing to do the job for the chance to gain big money, power, and glory at the top) or why some professions earn more than others despite their being less specialized (e.g., prostitutes typically earn more than architects because there is more demand for their services and the supply is lower). Note: I would suggest to reduce the factors to two, the supply of people willing and able to do the job (which is a factor of the skills required and the attractiveness of the job, relative to alternatives), and the demand for services that the job fulfills.

Conflicting incentives in gang wars

Gang wars are bad for business (sales plummet as customers are scared, foot soldiers want to be compensated for the extra risk), especially for the bosses, but serve a purpose for foot soldiers (one of the ways to make a name for themselves is to prove their mettle for violence).

The impact of crack on Black neighborhoods

For black Americans, the situation had improved drastically between WW2 and the 1980s, as evidenced by the civil rights legislation in the mid 1960s, and the shrinking gap with whites in terms of income, children’s test scores, and infant mortality. Then came crack, which was an ideal drug for low-income, street-level customers (low cost, powerful high that fades fast and sends the user back for more), and which turned black street gangs (whose leaders had made the connections to Mexican gang members and Colombian dealers in federal prison) into lucrative commercial enterprises. Crack impacted black neighborhoods much harder than most: beginning in the 1980s, infant mortality soared, as did the rate of low-weight babies and parent abandonment, the gap widened in terms children test scores, the homicide rate among young urban blacks quadrupled within 5 years, and the number of blacks sent to prison tripled.

Reasons for the rise and drop in crime

The detriment that the State would impose upon the pregnant woman by denying this choice altogether is apparent…. Maternity, or additional offspring, may force upon the woman a distressful life and future. Psychological harm may be imminent. Mental and physical health may be taxed by child care. There is also the distress, for all concerned, associated with the unwanted child, and there is the problem of bringing a child into a family already unable, psychologically and otherwise, to care for it. Justice Harry Blackmun, U.S. Supreme Court’s ruling in Roe v. Wade

From the mid 1970s to the late 1980s, violent crime had risen 80% in America. The violence associated with the crack boom coincided with a broader crime wave that had been building since the early 1960s, driven in part by a more lenient justice system (conviction rates declined and criminal served shorter sentences, partly as a result of the expansion of accused people’s rights, and politics becoming softer on crime for fear of sounding racist). This led experts to make apocalyptic forecasts. In the early 1990s, the crime rate began falling with such speed and suddenness that it surprised everyone. The most frequently mentioned crime-drop explanations were (1) innovative policing strategies, (2) increased reliance on prisons, (3) changes in crack and other drug markets, (4) aging of the population, (5) tougher gun-control laws, (6) strong economy, (7) increased number of police, (7) increased use of capital punishment, and (8) gun buybacks. The authors conclude that only three of these explanations can be shown to have contributed to the drop in crime (increased imprisonment, increased number of police, bursting of the crack bubble), along with an ignored yet potent fourth explanation (the legalization of abortion):

• Strong economy: no evidence. While a stronger job market may make certain financially motivated crimes less attractive (studies have shown that a 1% unemployment decline accounts for a 1% drop in nonviolent crime), reliable studies have shown no link between the economy and violent crime, which is precisely the segment which declined most in the 1990s (40% decline compared to a 2% decline in the unemployment rate). A look at the 1960s, which had both a wild growth in the economy and in violent crime, further weakens the “strong economy” argument.
• Increased punishment: c1/3 of the crime drop. Harsh prison terms have been shown to act as both deterrent for would-be criminals on the street and prophylactic for would-be criminals already locked up. Note: while the logical argument makes sense, the authors do not provide backup.
• Increased use of capital punishment: no evidence. Given the rarity with which executions are carried out (478 executions during the 1990s) and the long delays in doing so (2% annual execution rate on death row), the negative incentives for criminals aren’t serious enough to change their behavior. According to Isaac Ehrlich’s tentative estimate of the tradeoff between executions and murders, an additional execution per year may result in 7 fewer murders. Given the number of executions went from 14 in 1991 to 66 in 2001, these additional 52 executions would have accounted for 364 fewer homicides, less than 4% of the decrease that year.
• Increased number of police: c10% of the crime drop. The number of police officers per capita in the United States rose about 14 percent during the 1990s, a period during which violent crime fell drastically. To show causality and not merely correlation, the authors found scenarios in which more police were hired for reasons unrelated to rising crime: in the months leading to Election Day, mayors often try to lock up votes by hiring more police. By comparing the crime rate of cities that recently had an election with that of cities that had no election, the authors found that additional police does substantially lower the crime rate. From 1960 to 1985, the number of police officers had fallen more than 50% relative to the number of crimes, translating into a roughly equal decline in the probability that a given criminal would be caught. Coupled with leniency in the courtrooms, this decrease in policing created a strong positive incentive for criminals.
• Innovative policing strategies: no evidence. In New York City, mayor Rudolph Giuliani and his police commissioner William Bratton vowed to fix the city’s desperate crime situation. Bratton demanded accountability, introduced a computerized method of addressing crime hot spot, and applied the “broken window” theory, policing minor nuisances that would otherwise turn into major nuisances. New York City was a clear innovator in police strategies during the 1990s crime drop, and also enjoyed the greatest decline in crime (homicide rates fell from by 74%, from 30.7 per 100,000 people in 1990 to 8.4 per 100,000 people in 2000) among large cities. But innovative policing strategies probably had little effect: (1) crime was well on its way down before Giuliani and Bratton arrived in 1994 (by the end of 1993, the rate of property crime and violent crime had fallen by 20%), (2) the new police strategies were accompanied by a hiring binge, much of it started by Giuliani’s predecessor David Dinkins (between 1991 and 2001, the NYPD grew by 45%, three times the national average, which would be expected to reduce crime by 18%, putting New York back in the average of large cities), (3) crime went down everywhere in the 1990s; even in Los Angeles, a city notorious for bad policing, crime fell at about the same rate as in New York once the growth in police force is accounted for.
• Tougher gun laws: no evidence. The Brady Act, passed in 1993, requires a criminal check and waiting period before a person can buy a handgun. While appealing to politicians, this solution doesn’t make sense for an economist: regulation of a legal market is bound to fail when a healthy black market exists for the same product. With guns cheap and easy to get, the standard criminal has no incentive to fill out a firearms application at his local gun shop. A study of imprisoned felons showed that even before the Brady Act, only 1/5 of them had bought their guns through a licensed dealer. Washington DC and Chicago both instituted gun-control bans before crime began to fall, and those two cities were laggards in the national reduction of crime. One deterrent that has proven moderately effective is a stiff increase in prison time for anyone caught with an illegal gun.
• Gun buyback: no evidence. (1) The payoff to the gun seller ($50, $100, or free hours of psychotherapy) isn’t an adequate incentive for criminals planning to use their gun, (2) the number of surrendered guns is much lower than the number of new guns coming to market, (3) the likelihood that a given gun is used to kill someone is 1 in 10,000; given the typical gun buyback programs yields less than 1,000 guns, this would translate into an expectation of 0,1 less homicide per buyback program.
• Bursting of the crack bubble: c15% of the crime drop. (1) Crack cocaine was so potent and addictive that it created a hugely profitable market overnight, making leaders of crack gangs rich, but also contributing to a huge increase in violent crime (more than 25% of homicides in New York City in 1998 were crack-related) as street-level dealers were willing to kill their rivals to advance or to secure lucrative drug-selling corners, (2) The violence associated to crack began to fall in 1991, not because crack smoking stopped (it remains as popular) but rather because the huge profits for selling crack vanished (the price of cocaine fell and dealers began to underprice one another), thus making young dealers realize that the smaller profits didn’t justify the risk. From 1991 to 2001, the homicide rate among young black men (disproportionately represented among crack dealers) fell 48%, compared to 30% for older black men and older white men.
• Aging of the population: no evidence. The theory goes that since people mellow out as they get older, more older people must lead to less crime. But demographic change is too slow to explain the suddenness of the crime decline.
• Legalization of abortion: strong evidence. In several European countries that declared abortion illegal (Romania in 1966) or required permission from a judge in order to obtain one, studies found that women who were denied an abortion often resented their baby, and failed to provide them with a good home, making the children more likely to become criminals. When a woman does not want to have a child, she usually has good reason (e.g., unmarried or in a bad marriage, too poor to raise a child, too unstable or unhappy, drinking or drug use could damage the baby’s health, too young or without enough education). In the late 1960s, several states allowed abortion under extreme circumstances. By 1970, five states had made abortion entirely legal. In January 1973, legalized abortion was extended to the entire country (Supreme Court ruling in Roe v. Wade). The number of abortions went from 750,000 the year after to 1.6 million by 1980, where it leveled off. Before Roe v. Wade, it was mainly middle or upper class women that could arrange a safe illegal abortion (which cost $500). After Roe v. Wade, any woman could easily obtain an abortion, often for less than $100. The women most likely to take advantage of Roe v. Wade were unmarried, in their teens, poor, and with a low education, all factors among the strongest predictors that a child will have a criminal future. In other words, legalized abortion leads to less unwantedness, which leads to less crime. In the early 1990s, as the first cohort of children born after Roe v. Wade was hitting its late teen years (during which young men enter their criminal prime), the rate of crime began to fell: missing from the cohort were the boys who stood the greatest chance of becoming criminals, and the girls most likely to become teenage mothers. To prove the impact of legalized abortion on crime, the authors point that (1) the five states which legalized abortion before Roe v. Wade saw crime rates fall earlier than the other 45 states, (2) the states with the highest abortion rates in the 1970s experienced the greatest crime drops in the 1990s, (3) in states with high abortion rates, the entire decline in crime was among the post-Roe cohort, (4) studies in Australia and Canada have established similar links.

The art of parenting

The authors point out that obsessive parenting or race do not matter as much as we think:

• Bad parenting matters a great deal: unwanted children, who are disproportionately subject to neglect and abuse, have worse outcomes than children who were eagerly welcomed by their parents.
• Genes alone are responsible for about 50% of a child’s personality and abilities, according to studies including research into twins who were separated at birth.
• Good test scores are correlated to what parents are rather than to what parents do: well educated, successful, healthy parents who gave birth to their first child after 30 tend to have children who test well in school; whether they read to their children every day, enroll them in preschool, let them watch TV frequently, spank them, have a computer at home, take them to museums, divorce or stay married has no bearing on their test scores.
• Black and white children with the same socioeconomic background have the same abilities in math and reading upon entering kindergarten. The issue is that (1) the average black child is more likely to come from a low-income, low-education household, and thus to score worse on average, and (2) the black-white gap reappears within two years of entering school, because the typical black child tends to go to a school whose environment is not conducive to learning (e.g., gang problems, nonstudents loitering in front of the school, bullying if they are perceived to “act white”). However, black students in “good” schools do not lose ground to their white counterparts, and outperform whites in “bad” schools.