Hi Experts:

I have a panel data contains information of cancer indicator and wage. I want to know "what is the wage would be if cancer individuals do not get cancer". I want to know the difference between actual wage and counterfactual wage. The aim of my research is to use this difference to predict other outcomes, say, probability of partner's wage. I hope my data should looks like following:

ID Time Cancer or not Actuarial wage Counterfactual wage Difference
1 2000 Yes 150 159 159-150
1 2001 Yes 151 140 140-151
1 2002 Yes 145 162 162-145
2 2000 No 160
2 2001 No 158 162-158
2 2002 Yes 154 140 140-154

1. I first regress cancer on other covariates, and get predicted probability of cancer (p).
2. Then I use p to derive weight, pw, following the DLF paper, "Labor Market Institutions and the Distribution of Wages, 1973-1992: A Semi-Parametric Approach"

However, I do not know how to use this weight to create counterfactual wage. May be I use the wrong method from the beginning?

I would appreciate if you have any comments, thanks.

Connie