Positive assortative mating and spouses as complementary factors of production: a theory of labour augmentation

Abstract

This article develops a model of intellectual labour augmentation to explain both the marriage wage premium and educational assortative mating. We suggest that husbands and wives are complementary factors of production where a spouse's education and skills augment their partner's productivity and earnings potential. We test this proposition using data from the 2000 US Census of Population and the 2003 Current Population Survey. Our results indicate that for working couples the marriage premium for husbands and wives is directly related to the education level of their spouses–suggesting that positive assortative mating may be attributable to the labour market effects of intellectual augmentation of married households.

Acknowledgements

We would like to thank Dan Black and Rich Hill for useful comments. We would also like to thank Jana Groothuis and Susan Schmitz for useful comments and for years of augmenting our productivity. All errors, however, remain ours.

Notes

¹ A fourth outcome is possible where the marginal product of labour augmentation is zero for both and no labour augmentation wage effect is found in a marriage.

² We use working couples for our samples in order to concentrate on testing the labour augmentation hypothesis for both male and female wage earners. In our sample we use all observations where both partners are either working part time or full time. We do the same analysis for a subset of workers where both spouses are full time married workers, because hours worked might be an endogenous decision in the household. The empirical results, however, are similar. We report these results in Appendix 1.

³ The census aggregate occupational and industry categories were used to construct the SAMEOCC and SAMEIND variables. The occupational categories are: managerial, professional, sales, clerical, operatives, production/craft, labourers, construction/extractive and service; the industry categories are: construction/extractive, manufacturing, wholesale/retail trade, information, finance, professional services, education and health services, recreation and hospitality, other services and public administration.

⁴ In addition, approximately 10% of the working couples in the Census and CPS samples are employed in the same industry and occupational categories. These percentages are significantly higher than what would occur under a random matching of couples by industry or occupation.

⁵ The regression results reported in Table 3 have been corrected for possible sample selection bias via Heckman's (1979) technique. The selection model is based on a probit estimation of a labour force participation equation (either in or out of the labour force) as a function of a respondent's experience, education, disability status, and the presence of young children. The full set of empirical estimates is available from the authors upon request.

⁶ For the CPS sample, the distribution of men and women with spouses in the same occupational category is: managerial (18%), professional (49%), service (9%), sales (7%), clerical (9%), operatives (1.5%), production/craft (5.3%), labourers (0.3%) and construction/extractive (0.5%). The corresponding percentages for the industry categories are: construction/extractive (1%), manufacturing (18%), wholesale/retail trade (16%), information (2%), finance (6%), professional services (7%), education and health services (34%), recreation and hospitality (5%), other services (3.5%) and public administration (5%). The corresponding percentages for the Census sample are similar.