Testing Beckers Prediction on Assortative Mating on Spouses Wages

It is often believed that spousesʹ wages are positively related even when other traits such as age and education are controlled. This is mainly based on the observation of two-earner couples. This paper uses the standard sample selection technique to correct for the sample censoring and to compute potential wages for nonworking women. Using data from Taiwan, it is found that after accounting for sample censoring and cross-productivity effects, there is weak evidence that the partial correlation between spousesʹ wages can be negative. This lends first weak but direct support for Beckerʹs prediction of negative assortative mating on spousesʹ wages.