Does marriage increase couples’ life satisfaction? Evidence using panel data and fixed-effects individual slopes

Abstract

Many contemporary studies find that married couples are more satisfied with life than unmarried people. However, whether marriage makes people more satisfied with life or whether more satisfied couples are more likely to marry remains a debated question. We reassess this relationship with panel data from the German Family Panel (pairfam) and extend previous analyses by adding individual trajectories (slopes) to standard fixed-effects regressions (FEIS). We are thereby able to distinguish – controlling for time-constant unobserved heterogeneity – whether there is in fact an effect of marriage on life satisfaction, whether people who are simply happier in their relationship are more likely to get married, or whether people whose development in life satisfaction is more positive are more likely to get married. We translate these different social mechanisms into different analytical strategies and find that OLS regression – due to its confounding effects between and within persons – overestimates the effect of marriage on life satisfaction. A fixed-effects estimator reveals a much lower effect of marriage on life satisfaction for couples who marry compared to those who continue to live apart together or cohabitate. Additionally, using a FEIS estimator and adjusting for – non-linear – development of individual life satisfaction over time, suggests that this effect is in fact causal.