This working paper explores the effect of joint labor decisions on the study of wage regression models. The estimation of Mincer equations suffers from numerous sources of bias, including the sample selection problem generated by the fact that the agent decision to work is not independent of the wage. Most of the papers corrects this bias using a model of individual labor participation. However, recent trends in the labor market show greater participation of women in the labor force and seem to indicate that the joint decision of the spouses is increasingly relevant in determining the selection mechanism. A bivariate version of Heckman's method appears as an interesting alternative to solve this problem. Although the estimates are in line with the previous literature, the results indicate that the joint decision of the couple is a relevant factor in the selection bias.