If a family has a baby in December, it is eligible for child-related tax benefits for the whole year. If the baby arrives just one month later, in January, the benefits kick in a full year later. So economists analyzed what happened in low-income families whose first child was born in either December or January during the 1980s and ’90s. They found that the extra tax benefits that families received as a result of having a child before Jan. 1 significantly boosted what those children earned as adults, especially for males. Much of the effect was transmitted through increased childhood test scores, higher rates of high school graduation, and reduced suspensions. The economists also estimated that these higher lifetime earnings generated tax revenue for the government that exceeded the cost of the initial benefit.
Barr, A. et al., “Investing in Infants: The Lasting Effects of Cash Transfers to New Families,” Quarterly Journal of Economics (forthcoming).
Patients who had been referred to a medical specialist were more likely to seek a second opinion if the first specialist was a woman, according to researchers who analyzed data from the Massachusetts All-Payer Claims Database, which contains medical claims data for virtually every resident of the state. This was especially true for male patients, and the dynamic held up even when researchers adjusted the data to control for the patient’s age, diagnosis, and insurance type and the specialist’s level of experience. In turn, this preference resulted in significantly lower billing revenue for female specialists.
De Vaan, M. & Stuart, T., “Gender in the Markets for Expertise,” American Sociological Review (forthcoming).
Be fair to the algorithm
One of the criticisms of the use of artificial-intelligence algorithms in areas like judicial, medical, or human-resources decision-making is that there’s little transparency in how an algorithm arrives at its decision. But in a new study, researchers argue that human decision-making is just as much of a black box. In experiments, people were asked to consider how a judge evaluates the risk of recidivism, how a radiologist evaluates an MRI, or how a recruiter evaluates a video interview. After being asked to explain one of the human evaluation processes, study subjects indicated that they understood it less. Subsequent experiments showed similar results if the study subjects were reminded how different they are from the human decision makers described in the exercises or how difficult it is in general to make decisions.
Bonezzi, A. et al., “The Human Black-Box: The Illusion of Understanding Human Better Than Algorithmic Decision-Making,” Journal of Experimental Psychology: General (forthcoming).
Looking the other way
A study found that the careers of NFL players didn’t suffer after they were arrested for violence against women. In fact, by comparing players who had been arrested with those who hadn’t but were similar in race, position, age, and career statistics, the researchers found that before the mid-2000s, being arrested for violence against women was actually associated with a longer career. Since then, there is evidence of an increasing negative effect, but it has been limited to low-value players.
Sailofsky, D., “More Talent, More Leeway: Do Violence Against Women Arrests Really Hurt NFL Player Careers?” Violence Against Women (forthcoming).