Panel I, columns (1) and (2), we find that wives are between 1 and 2 ppt less
likely to be not working once their husbands receive VA insurance. This implies a
3-4% increase in the probability of working relative to the pre-period average of
0.473, although once full controls are added this effect is only marginally
significant. These results appear to be driven by wives with high school or less
education, as demonstrated in Table 3, Panel II, columns (1) and (2). Women with
high school education or less are 3 ppt more likely (a 5.6% increase off a base of
0.53) to work when their husbands are offered VA insurance. Coefficients are
smaller, insignificant, and change signs when full controls are added, for women
with at least some college education as shown in Tables 3, Panel III, columns (1)
and (2). Similarly, we find in Table 3, Panel I, columns (3) and (4) that average
hours worked per week for all women increases by approximately half an hour
(between 0.44 and 0.63 hours). When hours worked are examined by women’s
education level, women with a high school education or less in Table 3, Panel II,
columns (3) and (4), work 0.6 to 0.8 hours more per week when their husbands
are offered health insurance, while results in Table 3, Panel III, columns (3) and
(4) are positive but insignificant for those with more education. We proxy for
work on the intensive margin with weekly hours worked and weekly earnings
conditional on working any hours or having non-zero earnings respectively
However, conditional on working any hours, hours worked do not increase
18