A sudden financial “injection” from lottery winnings increases men's likelihood of marrying and having children, while temporarily raising the risk of divorce among already married women, according to a new study published in the Journal of Human Resources. The researchers analysed data from nearly 77,000 Swedish lottery players, comparing people who played exactly the same games but won different amounts by chance.
The team of David Cesarini, Erik Lindqvist, Robert Östling and Anastasia Terskaya used the fact that the size of a lottery win is random and unrelated to the player's background or behaviour. This makes the comparison similar to a randomised experiment, they say. They linked player records to Swedish administrative registers of marriages, divorces and births, and followed adults aged 18 to 44 over periods of two, five and ten years. The results are expressed per 1 million Swedish kronor after tax, an amount equivalent to about $140,000 in 2010.
For unmarried men, a win of 1 million kronor increased the likelihood of marriage within five years by 4.7 percentage points. With a baseline marriage rate of about 16% in this group, the change amounts to an increase of around 30%. The effect was found mainly among men whose incomes were below the median before they won. Among unmarried women, the effect on marriage was small and statistically insignificant, while the difference between the sexes was not strong enough to rule out chance.
By contrast, married women who won showed a sharp increase in their short-term risk of divorce. A win of 1 million kronor nearly doubled the risk within two years, raising it by about 4 percentage points above a baseline rate of 4%. The effect faded by the ten-year mark, which the researchers interpret as evidence that the money accelerated divorces already under consideration. “The effect is sharp and short-lived,” Terskaya says, stressing that women probably still face financial barriers when they want to leave a marriage.
Among married men, winnings appeared to reduce the risk of divorce, but the authors consider this the least reliable finding. When they expanded the sample to include adults up to 64, the stabilising effect on men's marriages became much smaller and less consistent. Terskaya acknowledges that the effect weakens or disappears depending on the age range, definitions and statistical checks.
In terms of fertility, children act as what economists call a normal good: people tend to have more when their resources increase. Ten years after winning, men had an average of 0.056 more children per 1 million kronor, an increase in fertility of about 14%. The researchers estimate that 20% to 40% of this increase is due to more marriages and fewer divorces. For women, fertility showed a slightly positive trend, but it was not statistically significant; the gender difference was more pronounced among people over 35, which may reflect age-related limits on fertility.
The issue has policy implications, as many countries have fertility rates below the replacement level of around 2.1 children per woman, with consequences including population ageing and pressure on pension systems. However, the absolute effects are modest: across the full sample, each 1 million kronor increases the number of children over a decade by about 0.033. The researchers calculate that about 30 million kronor, roughly $4 million, is needed for one additional child. “Large sums of money would be an expensive way to change marriage or birth rates,” Terskaya says.
The study has limitations. It is based exclusively on Sweden, a country with a strong welfare state, low legal barriers to divorce and relatively high acceptance of non-traditional family structures. The researchers expect the results to apply more readily to other wealthy welfare states, especially the Scandinavian countries, and much less to developing countries. Jeong Jin Yu, a professor at Xi’an Jiaotong-Liverpool University who was not involved in the study, notes that confidence in the main findings is relatively high because of the large administrative dataset and the lottery-based design, but confidence in interpretations of gender differences is lower because the data do not directly measure relationship quality or decision-making power. Also, if wealthier men marry more because their position relative to other men improves, then a general increase in everyone's income — for example through universal basic income — may have a smaller effect than these estimates suggest.
Comments