Marriage and Wages

This work investigates the commonly observed relationship between marriage and wages among men in Britain using panel data covering the 1990s. We explicitly test several hypotheses developed in the literature to explain this relationship, including the household division of labour and specialisation, differential rates of human capital formation, employer favouritism, and self-selection. After accounting for individual-specific time-invariant effects, and a wide range of individual, household, job and employer related characteristics, we find a small but statistically significant premium remains that can be attributed to productivity differences. Our estimates provide evidence for the existence of a large selection effect into marriage based on both observable and unobservable characteristics that are positively correlated with wages (consistent with employers using marriage as a positive signal), and also evidence in support of the specialisation hypothesis. NON-TECHNICAL SUMMARY Much research in applied economics has commented on the advantages associated with marriage. Marriage has been found to have positive effects both on reported levels of happiness and health. In addition, a male marriage premium is a common finding in wage equations, indicating that marriage is associated with higher wages for men. However, there is the question of causality – does marriage itself make men more productive and therefore increase their earnings? Or alternatively do more productive, higher earning men get married? If marital status is genuinely productivity enhancing, then changes in the marital status composition of the workforce will affect productivity. If there are no productivity effects of marriage, then changes in the marital status composition of the workforce will have no impact on economic output. In this paper we investigate in detail the presence and causes of a marriage wage premium among men in Britain using panel data from the first eleven years of the British Household Panel Survey (BHPS), covering the period 1991–2001. We explicitly test several hypotheses developed in the literature to explain the relationship between marriage and wages. By using panel data we are able to allow for possible correlations between unobserved characteristics, marriage and wages. Failure to do so will bias the coefficient of interest – some of the returns attributed to marriage may actually be returns to some unobserved qualities correlated with marriage. If so, the observed wage premium associated with marriage largely reflects unobserved individual characteristics that are also valued by the employer. Our work provides new evidence on the existence and causes of the marriage wage premium among married men in Britain by using a long run of panel data. Cross-sectional analysis yields a wage premium for married men of about 15%, consistent with much of the previous literature. Using panel data and panel data methods, we find that this premium falls dramatically, indicating that about three-quarters of the observed premium in cross-sectional analysis is caused by unobserved individual heterogeneity and/or selection effects. Married men have unobserved characteristics that are also correlated with wages. Our results are consistent with the hypothesis that employers use marriage as a signal – a large proportion of the marriage premium is due to unobservable characteristics that are valued both by wives and by employers, such as motivation, loyalty, dependability and determination. Nevertheless, a relatively small but statistically significant marriage premium remains even when allowing for a wide range of individual, household, job and employer-related characteristics and time invariant individual specific unobservable effects. Our preferred panel estimates indicate the size of this premium increases with the number of domestic chores for which the spouse is mostly responsible, and falls with the wife’s working hours. The relative sizes of the coefficients suggest that a married man whose wife does not work but whose wife is mostly responsible for four domestic chores enjoys a wage premium of about 4% relative to a single never married man. However this premium almost disappears if the wife also works 40 hours per week in the labour market. We show that the effects of the hours worked and domestic chores carried out by the wife are genuine, and not due to the potential endogeneity of the wife’s decision to work. Our estimates therefore provide some evidence in favour of the specialisation explanation for the enhanced productivity of married men.