Assortative mating: The overlooked factor in America’s rising income gap

Despite the saying that opposites attract, marriages are increasingly being formed between individuals who are similar in many respects. Scientists call this phenomenon "assortative mating."

Economists have known for at least a decade that such assortative mating fosters income inequality between married couples. It's easy to understand why this happens. The difference in total income between a household consisting of two well-earning individuals and a household of two low-wage earners is more significant than the difference between families in which one person earns a lot and the other earns little.

According to one hypothesis, the intensification of assortative mating is due to the proliferation of dating services. In 1998, only 2% of marriages met online. A decade later, this percentage was 20%; in 2017, it reached 50%. This means that dating services have become the dominant way to initiate relationships.

In a research paper published in mid-September, three American economists argue that the data do not support this hypothesis. According to Anton Cheremukhin from the Federal Reserve Bank of Dallas, Paulina Restrepo-Echavarria from the Federal Reserve Bank of St. Louis, and Antonella Tutino from Haverford College, assortative mating has indeed intensified in recent decades. It contributed to income inequality, but this largely occurred before the advent of dating services.

The authors of the new study calculate that in the US, income inequality between households formed by married couples (traditional) has increased by nearly 25% since the 1960s. The Gini coefficient, a popular measure of inequality, increased from around 25 points to 31 points (the extreme value is 100 points). Initially, however, as economists' simulations suggest, inequalities would have grown similarly if couples had randomly paired off. It was only later, after 1980, that inequality increased due to the more selective (assortative) nature of marriage. What were the sources of this phenomenon?

Economists, based on data from censuses and extensive surveys (covering even 5% of the US population), noticed that over the past decades, the preferences of those seeking a partner have clearly changed. However, this happened before 2008, which is the period of rapid development of online dating services.

Racial preferences are horizontal, meaning that participants in the marriage market look for people who are similar to them in this respect, preferences regarding education level and age are moderately horizontal, while preferences regarding income level and skill level are decidedly vertical (participants look for partners with higher incomes and qualifications than their own – editor), the article states. From 2008 to 2021, these preferences changed minimally, but compared to the 1960s and 1980s, horizontal preferences have increased, particularly concerning income and skills.

In short, marriages are now more often formed between individuals who are similar in many areas. This change, according to the three economists, raised the Gini coefficient by about 3 points, accounting for roughly half of the mentioned increase in income inequality in the US. The most significant contribution to this was the change in potential partners' preferences concerning education and skill level (measured in the study based on the job held), and then concerning income and age.

The findings of Cheremukhin and his co-authors suggest that preferences have largely changed due to women's emancipation. In the past, women preferred men with higher education levels, but today they seek men with higher income and skills but similar education levels. Traditionally, men sought "housewives," partners with lower status regarding education, skills, and income. However, the increase in the percentage of educated and working women means there are now fewer such partners in the marriage market. As a result, the way couples match has changed: couples are more often formed by individuals similar in education and other long-term income-affecting traits.

Why didn't dating services have at least an equally significant impact on this market?

According to economists, these services reduced the costs of searching for partners, which theoretically could make it easier for similar people to pair off. In practice, however, human cognitive abilities are the limiting factor. While technological advancements have increased the choice of potential partners, people's ability to process this information has not increased.