they might work for employers who drop or change the offer of coverage, corresponding to an
ongoing trend that may be unrelated to the availability of public coverage.
Evidence about crowd-out that may be related to changes in the employer offer of coverage
for employees is relatively consistent. Cutler and Gruber (1996) used variation arising from
differential timing of Medicaid expansions to compare employers’ offer decisions from the late
1980s (before expansions) to early 1990s (afterwards). They found no evidence that employers
were less likely to offer coverage when a higher proportion of employees or their dependents
were eligible for Medicaid. Using data in the same period and focusing on small firms with
fewer than 100 employees, Shore-Sheppard et al. (2000) drew the same conclusion about
again that there was no significant reduction of firms’ offer rates overall when a state-initiated
coverage program was expanded. However, they did estimate a 3.1 percent reduction in the rate
of offer specifically among small, low-wage firms.
With respect to private coverage overall, the evidence regarding of the potential effect of
expanded public coverage is very murky. Even within the same study, the range of estimates is
so large that no conclusions can be drawn.
A large number of recent studies have investigated consumers’ price responsiveness in
different segments of the insurance market, with most depending on observational data analysis
and a few using natural experimental or experimental designs. In general, the demand for health
insurance is price-inelastic. However, there is no further consensus about the magnitude of the
11 However, they found that a 10-percent increase in the proportion of employees eligible for Medicaid reduced by 6 percent the likelihood that employers would offer coverage for dependents. They also found weak evidence of a negative effect of expansions on take up of employer-offered coverage.