or provides the additional income needed to self-insure. Finally, in Panel IV, we
see no effect on women who were working full-time in the previous year. They
neither leave the labor market nor change their labor force participation. These
results are highly consistent with the idea that married women value health
insurance and seek out their own employer-provided health insurance or seek
greater income in order to pay for potential medical expenses once their husbands
are offered publicly provided health insurance.
Table 7 provides various robustness checks. In our base specification,
we included all wives over the age of 18 in our sample. However, women of
younger ages may have different labor market attachment than older women.
Panel I provides results limiting to middle-aged and older wives. When wife ages
are limited to 45-64, the percentage point magnitude of the results for not working
and hours worked outcomes is very similar to that in the main specification in
Table 3. Limiting women to the same ages as the men in the sample, age 55-64
provides larger magnitude results than our earlier sample. In this case, women
decrease not working by 3.4 ppt and increase hours worked by 1.12, about twice
the magnitude as our base regression.
In Panel II of Table 7, we explore outcomes for wives whose husbands
are working part time or are not working. Our main results in Table 3 include all
wives regardless of their husbands’ labor force attachment because the timing of
22