Back to Parity Homepage.
mounted by the Louis de la Parte Florida Mental Health Institute
University of South Florida
Top of page
CBO'S ESTIMATES OF THE IMPACT ON EMPLOYERS
OF THE MENTAL HEALTH PARITY AMENDMENT IN H.R. 3103
CBO estimated that the mental health parity amendment in H.R. 3103, as passed by the Senate, would impose direct costs on the private sector of 4 percent of private health insurance premiums. Most employers currently provide for mental health coverage, but with significantly less generous benefits than for other medical services. The provision intends to place mental health services on an equal footing with these other medical services.
Employers could respond to the mandate in a variety of ways to reduce the impact of that increase. Possible responses include dropping mental health coverage, reducing the generosity of their health insurance benefits, or ceasing to offer health insurance altogether. Any remaining increases in premiums would most likely be passed on to workers in the form of lower wages or reductions in other fringe benefits. This memorandum describes the definitions of covered mental health services and mental health parity that CBO assumed in its estimates, and discusses the estimated increase in private sector costs, potential responses by employers to higher premiums, and the implications of state benefit mandates.
Covered Mental Health Services
The Congressional Research Service (CRS) provided CBO with an estimate of the increase in private health insurance premiums that would result from the mental health parity requirement Consistent with the provisions of HR 3103, CRS did not restrict coverage to severe mental illness only. The estimate assumed that treatment for all mental health conditions, including alcoholism and substance abuse, would be covered under the parity amendment Additional information on the cost of treating severe mental illness is provided in the last section of this memorandum.
Definition of Parity
CRS estimated the premium increase for a typical indemnity plan, incorporating the following assumptions about the definition of mental health parity:
The estimate defines parity in terms of service and spending limits, rather than in how care is provided. CRS assumed that insurers would continue to use current techniques, including "carve-outs", to control utilization of mental health services.
Increase in Private Sector Costs
Under those assumptions, CRS estimated that the premium increase for an indemnity plan that incorporated mental health parity would be 5.3 percent Because the CRS estimate was confined to the impact of parity on indemnity plans, CBO lowered the increase to 4.0 percent to reflect the prevalence of managed care plans in the employer-sponsored health insurance market and the significant savings in mental health spending that managed care plans can produce relative to indemnity plans. Thus, in aggregate, CBO estimated that the panty amendment would impose direct costs on the private sector of 4 percent of private health insurance premiums.
Responses by Employers
Most employers offering health insurance coverage to their workers could face additional costs as a result of the mental health parity amendment Data on employment-based plans from the Bureau of Labor Statistics (cited by Coopers and Lybrand, LLP, in a study of the effects of the parity amendment) indicate that almost all employment-based plans offered some mental health benefits in 1991, but less than 2 percent of them had parity for outpatient coverage of mental health services.
[p] 2
Employers could respond to those additional costs in a variety of ways. They could lessen the impact on health insurance premium, by reducing the overall generosity of their health insurance benefits or by dropping mental health coverage. Some employers might cease to offer health insurance to their workers altogether. Any remaining increase in premiums, would most likely be passed on to workers in the form of lower wages or reductions in other fringe benefits.
Projections of the relative magnitude of the possible responses are, inevitably, speculative. The best studies of the effects of mandates on health insurance coverage have large margins of error associated with their estimates. Some empirical questions, such as the degree to which other components of health benefits would be dropped in response to a mandate about a specific component of coverage, have simply not been addressed by academic studies. Nonetheless, we can provide estimates of the plausible magnitudes of various responses, albeit with a high degree of uncertainty.
For calendar year 1998, CBO estimates that the panty amendment would impose direct costs of $11.6 billion on the employment-based insurance market-costs that we assume would ultimately be borne by workers. Based on our assumptions, we estimate that the additional costs would be allocated in the following ways:
Most of the $6.7 billion in additional premium payments would come from increased contributions by employers with the remainder being employee contributions. Employers would shift dm additional costs onto workers in the form of lower non-health compensation (cash wages and other fringe benefits.) Non-health compensation for workers who continued to have employment-based coverage would be reduced by almost $6 billion, or by about 0. 15 percent of what they otherwise would have received.
Employers would also reduce the generosity of their health insurance benefits by about $4.8 billion in 1998. Reductions in coverage generosity could encompass moving toward more restrictive forms of managed care, increasing cost-sharing requirements, restricting or dropping other benefits, or dropping mental health coverage entirely (if permitted under applicable state laws).
If employers took no action to reduce the generosity of benefits, annual premiums would rise, on
average, by approximately $190 per policy or $95 per covered person. The reductions in benefit
generosity would offset those increases, however, so that the net increase in premiums would
average about $110 per policy or about $55 per covered person.
CBO estimates that the parity requirement could result in 400,000 fewer workers (800,000 fewer workers and dependents) having employment-based coverage than otherwise. But those estimates are highly uncertain because of the large margins of error in the study on which they are based. (Indeed, the possibility that the parity amendment would have no effects at all on the number of covered workers is within the margin of error.) Had those workers continued to have coverage, their premiums would have risen by $80 million, reflecting the additional costs attributable to mental health parity. By dropping coverage, the $80 million in mental health parity costs would be saved. CBO assumes that the affected workers would receive the equivalent of the premium contributions formerly made by their employers as additional non-health compensation.
Effects of Existing State Benefit Mandates
Many states currently have mandates requiring insurers; to provide or offer mental health coverage. According to the most recent data from GAO, 15 states require insurers to cover mental health benefits 23 require insurers to cover alcoholism treatment, and 13 require insurers to cover drug abuse treatment. In addition, 16 states require insurers to offer at least one plan with mental health coverage, 16 require insurers to offer alcoholism treatment, and 10 require insurers to offer substance abuse treatment Specific requirements under those mandates vary considerably from state to state. Thus, the effects of mental health parity on insurers and employers could vary significantly by state.
In general, employers with fully-insured plans could not drop mental health, alcoholism, or substance-abuse treatment in states where the corresponding mandates apply. Those employers would have fewer options open to them than self-insured plans, which are not covered by state benefit mandates, and employers in states with no mental-health-related mandates. Employers not affected by state mandates could choose to drop all mental health coverage in order to avoid the parity requirements, although it seems unlikely that many employers, other than small firms, would choose that option.
In states that require insurers to offer a plan that provides mental health (or alcoholism or substance abuse) coverage, insurers may already be experiencing adverse-selection problems between Plans that do and do not have mental health
[p] 4
coverage. That is, people who believe that they will need to use mental health services are more likely to purchase policies offering mental health coverage, thereby pushing up the premiums for such plans. If more employers sought policies without mental health coverage to avoid premium increases induced by the parity requirement, the result could be somewhat greater adverse selection among health plans in the future.
The behavioral response of employers, regardless of whether they were affected by state benefit mandates, would depend critically on whether use of managed care "carve-outs" and other managed care techniques could continue and expand under the parity requirement If those options were precluded by legislation, CBO's cost estimates would be substantially higher.
Severe Mental Illness
Information is limited on the differences in the costs of coverage for the severely mentally ill as compared to those with less serious mental health conditions. Research conducted by Richard Frank at Harvard University suggests that patients with severe mental illness, defined as schizophrenia, manic-depressive disorder, and depression with a history of hospitalization, may account for 40-60 percent of all mental health care costs. The high end of that range reflects experience in Medicaid programs, whereas the low end of the range may be more likely among the privately insured population
To our knowledge, the only estimates of the costs of the mental health parity provision that attempt to distinguish between different levels of severity are those conducted by Milliman and Robertson, Inc. Using a somewhat broader definition of severe mental illness than did Professor Frank, that firm concluded that premiums would increase by an average of 3.9 percent if the parity amendment governed all mental health conditions, compared with 2.5 percent if severe mental illness only were covered. We are unable to evaluate those estimates because the Milliman and Robertson report provides little information about its data sources or estimation techniques. We do, however, have concerns about the effectiveness of a policy to limit the parity provisions to treatment of certain specified diagnoses, because of the incentives providers would face to "upcode" less serious diagnoses to meet the parity provisions.
Items referenced in the CBO report:
Back to Parity Homepage.