The right panel of Table 5 summarizes the evidence. As before, non-citizens who live in
the less generous states experienced a decline in Medicaid participation, with the decline
being particularly steep for children. Despite the decline in Medicaid coverage, however,
the non-citizens most affected by these cutbacks did not experience a sizable drop in health
insurance coverage, partly because of an increase in their rate of ESI coverage. The thrust of the evidence on health insurance coverage rates, therefore, is not sensitive to the definition of
the generosity index. As a result, the remainder of the analysis will use my initial definition
of the generosity index, which is based on the programmatic information summarized in
the first two columns of Table 2.Regardless of the definition of the state’s generosity index, any comparison between
naturalized citizens and non-citizens may be contaminated by the potential endogeneity
of the naturalization decision. After all, the non-citizens most affected by welfare reform
could neutralize many of the restrictions in the legislation by becoming naturalized.22 In
fact, there was a rapid rise in the number of naturalization applications during the period
(Wasem, 1998). This increase in the number of naturalization applications generated a huge
backlog at the INS, further delaying the time it takes to become a naturalized citizen One solution to the endogeneity problem would be to compare persons who differ in
terms of how long they have resided in the United States, rather than in terms of their
citizenship status. Immigrants have to live in the United States for 5 years before they
can apply for naturalization, but the lags in the application process imply that it may take
8 years or more before an immigrant can become a naturalized citizen. I estimated the
triple-difference regression model using an immigrant vector defined in terms of whether
the person was native-born, was an immigrant who had been in the United States for
fewer than 10 years, or was an immigrant who had been in the United States for more
than 10 years. These regressions (not shown) indicated that although the most recent immigrants
suffered the greatest declines in Medicaid participation rates, their health insurance
coverage rates remained relatively constant because of a concurrent increase in
the rate of ESI coverage. Alternatively, the endogeneity of the naturalization decision can
be avoided by simply comparing the immigrant and native populations, so that the vector
I in Eq. (1) would contain a single variable indicating if the household is headed
by a foreign-born person. The evidence (not shown) suggested that Medicaid participation
fell for immigrants, while health insurance coverage rates remained constant because
of a corresponding increase in the probability of being covered by employer-sponsored
insurance In sum, the results presented in this section strongly suggest that the state-funded assistance
programs helped to attenuate the decline in Medicaid participation in the immigrant
population. At the same time, however, these state-funded programs (or their absence)
had important “unintended” consequences. Non-citizens who did not have access to the
state-funded programs found ways of replacing the cutbacks in publicly provided health
insurance by increasing their probability of coverage with employer-sponsored insurance.
In the end, the state-funded programs did not seem to substantially alter the probability that
the immigrants had some type of health insurance coverage The evidence, therefore, implies the existence of a strong crowdout effect of publicly
provided health insurance. The results effectively offer a “mirror-image” perspective to the
crowdout findings first reported in Cutler and Gruber’s (1996) influential study. Cutler and
Gruber document that an expansion of Medicaid eligibility substantially reduced the number of persons covered by private health insurance. My study reveals that a cutback in public
assistance induces many immigrants to replace the lost benefits with employer-sponsored
insurance.
As noted earlier, however, the evidence presented in this paper differs in an important
way from the results in the crowdout literature. The welfare reform legislation affected
immigrant eligibility and participation in all public assistance programs. As a result, the
crowdout effects estimated in this section capture the behavioral response to the changing
value of the entire package of public benefits, rather than the behavioral response to a shift
in the parameters of the Medicaid program.