Growing Apart in Marriage, or Coming Together? The Structure and Dynamics of Differences in Risk Aversion between Partners

Abstract

Households are often faced with financial choices that have a large and lasting impact on the well-being of individual family members. Typical examples include the amount of borrowing for purchasing a home, a car or other durable goods, whether or not to take out various types of insurance, and so on. One significant decision is whether or not to invest in the stock market and additionally, for those willing to invest, the fraction of wealth to be invested stocks. Finance theory suggests that risk aversion is an important determinant in each of these cases.

In most societies, the majority of households consist of couples in cohabitation or marriage, with or without children. In these households, decision making is likely to be more complex than suggested by a unitary actor model, particularly when preference and opinions differ. While one partner may be highly risk averse, the other may not be. This creates a dilemma. To see its magnitude, it is worthwhile to study the structure and dynamics of differences in risk aversion between partners.

We use panel data from the Netherlands, in particular, the Dutch Household Survey (DHS) which is maintained by the Center for Economic Research (CentER) at Tilburg University. The annual data start in 1993 and extend through 2016. We measure risk aversion and discrepancies in risk aversion within couples by drawing on assorted methods, subsamples and subperiods. The samples are large compared to previous literature. Much of our analysis is based on couples who are married for longer than a decade, and that we observe for (minimally) seven years.

For instance, we compare differences in risk aversion between truly married partners vs. what hypothetically would be observed with male and female individuals who are randomly matched (but individually similar in age, income etc., to the ones that make up married couples). The actual gaps in risk aversion between truly married or co-habiting partners are substantially smaller. The resemblance in risk aversion may be due to (i) partner selection; (ii) to a convergence of values or beliefs over time, or (iii) to both factors. After all, couples that stay together for some years share similar life experiences.

We examine various measures of convergence over time. Our findings suggest, however, that the similarity between partners is largely driven by partner selection. At the same time, there is evidence for a small, marginally significant convergence effect.

Our study is relevant for financial advisers. Evidently, differences in risk perceptions and preferences between married partners may create decision impasse, but it is often far from evident what advisers can or should do about persistent disagreements of this kind.

Keywords: Risk aversion, Household Finance

JEL Classification: G02, D14