June 21, 2007 | WebMemo on Health Care
A path-breaking 1996 paper by economists David Cutler and Jonathan Gruber provided quantitative support for the idea that expansions of public health care programs often have little "bang for the buck" in terms of reducing the ranks of the uninsured. While some benefits will flow to those who are uninsured, program expansions also cause many individuals and families to lose their private coverage, often in favor of public coverage. The result is that many individuals are transitioned out of private coverage, a large portion of increased spending on public health flows to those who do not need it, and the ranks of the uninsured shrink by less than expected. Looking at Medicaid eligibility expansions in the 1980s and 1990s--mostly expansions to cover children from families with higher incomes--Cutler and Gruber found that nearly half of program expansions were offset by decreased private coverage. Recent research confirms this effect and its magnitude. As Congress considers expanding SCHIP up the income ladder, it should recognize that throwing more money into the program will increasingly "crowd out" private funding and coverage while doing less to expand overall coverage.
Cutler and Gruber describe three mechanisms by which crowd-out might occur: employers cutting back on private coverage, workers declining coverage for themselves and their families, and workers declining coverage for dependents eligible for public programs. They found that program expansions had no impact on employers' decisions to offer coverage but "a large and statistically significant effect" on workers' decisions to enroll in employer-based coverage. Cutler and Gruber also found that Medicaid expansions targeting women and children led many workers to drop coverage for their dependents while maintaining individual coverage.
Altogether, the crowd-out effect is especially strong in the expansion of public health programs that cover children. When programs targeting children are expanded, Cutler and Gruber conclude, "On net, the rate of uninsurance for children falls, but by only one-half as much as the increase in Medicaid coverage."
Cutler and Gruber conclude that, though a popular policy option, "expanding public insurance . . . may be quite expensive, due partially to private crowd-out." Among other alternatives to lessen the impact of crowd-out, they suggest "a sliding scale" of subsidies for lower-income people to purchase private coverage--in other words, premium support.
The Recent Research
Ten years after the publication of his initial crowd-out study, Gruber, along with economist Kosali Simon, revisited the issue with "improved methods and data."
An interesting statistic, based on data from 1984 through 2004, prompted the reevaluation: "[D]espite an enormous expansion in the public health insurance safety net in the U.S., the number of uninsured continues to grow." Could the culprit be, once again, crowd-out?
The new study's findings are even clearer, and more relevant to current policy debates, than those of the 1996 paper. First, the crowd-out effect is larger than previously estimated: "[T]he number of privately insured falls by about 60% as much as the number of publicly insured rises." Second, anti-crowd-out mechanisms such as waiting periods, which many states inserted into their SCHIP programs, failed to work--indeed, "if anything these provisions cause crowd-out to rise, not fall."
While Gruber has noted that his and Simon's model of the crowd-out effect of public insurance expansions may not be directly applicable to SCHIP and that their SCHIP estimates are unreliable, the paper is significant because it confirms the magnitude of crowd-out generally and suggests that anti-crowd-out mechanisms are, at best, ineffective. Gruber's two papers (with Cutler and with Simon), consistent with most studies on the subject published since Cutler and Gruber's initial study, demonstrate that crowd-out is a major drawback of expanding public insurance programs and present no evidence to suggest that this conclusion does not apply to SCHIP. In fact, Cutler and Gruber's study confirms an insight that applies directly to SCHIP expansions:
Since most of the people made eligible by the expansions are dependents, a natural reaction for workers may be to drop coverage of eligible dependents only, while maintaining insurance for themselves.
While Gruber and Simon, using different data and methodology, do not attempt to address this hypothesis directly, they do, as described above, confirm the general results on crowd-out of Cutler and Gruber.
To a large extent, a new report by the Congressional Budget Office draws on a wide body of recent research to fill in what Gruber has acknowledged to be the shortcomings in his and Simon's paper with respect to SCHIP. In addition to those suggested by Cutler and Gruber, the CBO paper describes two especially interesting mechanisms by which crowd-out might occur:
[P]reviously unemployed parents might be more likely to decline coverage at a new job if their children are enrolled in SCHIP. To the extent that SCHIP makes private coverage less important for some families, the program might also increase the likelihood that low-income parents take jobs that offer higher cash wages rather than health insurance.
In this latter way, SCHIP expansions to cover more children could actually further erode private coverage among adults. States' outreach efforts to boost SCHIP enrollments have also led to increased Medicaid enrollment, reducing private coverage still further. Studies cited in the CBO report also make the point that crowd-out encompasses more than just transitions from private to public coverage; rather, it represents a shift away from private coverage. For example, some individuals and families with public coverage may opt to stick with it rather than transition to private coverage when it becomes financially feasible.
The CBO concludes that reliable estimates of crowd-out in SCHIP range from 25 to 50 percent. As the report notes, however, the studies from which these numbers are drawn estimate the reduction only among children and so "probably understate the total extent to which SCHIP has reduced private coverage" because SCHIP crowds out adults as well. Another reason that crowd-out in SCHIP expansions is likely to be greater than the current literature's estimates is "because the children eligible for SCHIP are from families with higher income and greater access to private coverage." Thus, as SCHIP expands up the income ladder, more of the children enrolled in it will be transitioning from private coverage and fewer will come from the ranks of the uninsured.
This risk is real. Fully 61 percent of children who became eligible for public insurance due to the creation of SCHIP already had private coverage. As SCHIP grows to allow children from wealthier families, this figure will rise. According to a CBO analysis of Census data, current proposals in Congress to expand SCHIP eligibility would reach children in income groups in which 89 percent or more of children currently have private coverage. (See Chart 1.) Raising income eligibility limits for SCHIP will inevitably draw in more children who today have private coverage, increasing the problem of crowd-out. In this way, expanding SCHIP actually diverts the program from its original purpose--providing health coverage to uninsured children from needy families.
While precise estimates vary, study after study consistently shows that expansions of public insurance programs come at the expense of private coverage, especially when those expansions begin to encompass families with higher incomes. As more money is poured into expanding SCHIP, less of the new funds will go to providing coverage to children who currently go without. Rather, expansion will increasingly serve to coax individuals and families out of the private insurance market and into government coverage. Undermining private coverage, which is the linchpin of most approaches to improving coverage and care, with government dependence is not an effective way to address shortfalls in coverage.
Andrew M. Grossman is Senior Writer, and Greg D'Angelo is a Research Assistant in the Center for Health Policy Studies, at The Heritage Foundation.
 David M. Cutler and Jonathan Gruber, "Does Public Insurance Crowd Out Private Insurance?" The Quarterly Journal of Economics, Vol. 111, No. 2 (May 1996), pp. 391-430.
 Their "final estimate" is 49 percent. Ibid., p. 426.
 See, e.g., Jonathan Gruber and Kosali Simon, "Crowd-Out Ten Years Later: Have Recent Public Insurance Expansions Crowded Out Private Health Insurance?" NBER Working Paper No. w12858, January 2007, and Noelia Duchovny and Lyle Nelson, "The State Children's Health Insurance Program," Congressional Budget Office, May 2007.
 Cutler and Gruber, p. 411.
 Again, Cutler and Gruber are especially clear: "There is no effect of Medicaid eligibility on the decision to offer insurance…. Thus, the estimates suggest that all of the reduction in worker coverage is coming through lower take-up rates, not reduced employer offering of insurance." Cutler and Gruber, p. 417. Only one study has directly examined employers' responses to SCHIP. That study found "no evidence that employers responded to SCHIP by either dropping health insurance altogether or by dropping coverage for the dependents of employees." It did find, however, that "employers whose workers were likely to have been affected by the introduction of SCHIP did respond to the new program by raising family employee contributions relative to those for single coverage." It is not clear, however, the extent to which these increases independently affected employee take-up. The authors helpfully explain one point that policymakers should take from their research: SCHIP expansions mean that "some non-eligible families working in these firms also might face higher health insurance premiums for family coverage, as would families who were eligible for SCHIP based on income but who would not qualify because they had not been without insurance for a certain period of time." In other words, expanding SCHIP eligibility may actually increase the cost of private coverage for families that are not eligible for SCHIP coverage, including some families that fall within the SCHIP target population.
 They are particularly clear on this point, as concerns children: "For children, each ten-percentage-point increase in Medicaid eligibility results in a 0.74 percentage point decline in private insurance coverage. On net, the rate of uninsurance for children falls, but only by half as much as the increase in Medicaid eligibility." Cutler and Gruber, p. 410.
 Ibid., p. 410.
 Ibid., p. 428.
 Gruber and Simon, p. 27.
 Ibid., p. 1.
 Ibid., p. 2.
 Ibid., p. 3.
 Letter from Jonathan Gruber to Representative John Dingell, March 1, 2007 (on file with authors).
 For reviews of other literature, see Duchovny and Nelson, pp. 9-13, and Gruber and Simon, Table 1. Gruber and Simon find fault in the studies that they review finding low levels of crowd-out and seem to intend their study to grapple with the criticisms raised in those studies while not succumbing to the faults they have identified. See Gruber and Simon, pp. 9-13.
 Cutler and Gruber, p. 418.
 Again, they are especially clear on this point: "[C]rowd-out remains a pervasive phenomenon for recent public insurance expansions." Gruber and Simon, p. 2.
 See Letter from Jonathan Gruber to Representative John Dingell (D-MI) ("The estimates of crowd-out that are being cited from our paper are estimates of the total impact of expanding public insurance to families. The relevant estimate for interpreting the crowd-out effects of SCHIP differ…."), and Duchovny and Nelson, pp. 9-13.
 Duchovny and Nelson, p. 9.
 Ibid., p. 10.
 A Robert Wood Johnson Foundation issue brief calls this "'within enrollment' crowd-out." See Claudia Williams, Gestur Davidson, Ph.D., Lynn A. Blewitt, Ph.D., and Kathleen Thiede Call, Ph.D., "Public Program Crowd-Out of Private Coverage: What Are the Issues?" Robert Wood Johnson Foundation, Policy Brief no. 5, June 2004.
 Duchovny and Nelson, p. 11.
 Julie L. Hudson, Thomas M. Selden, and Jessica S. Banthin, "The Impact of SCHIP on Insurance Coverage for Children," Inquiry, vol. 42, no. 3 (Fall 2005), Table 1.
 Duchovny and Nelson, p. 12, n. 39.