On net, these changes would reduce large employers' health expenditures roughly $7 billion. Taking into account higher tax payments as health expenditures drop, their spending would be reduced $5.6 billion, or 5.8 percent.

Since smaller firms are less likely than larger employers to now provide health insurance, their costs would increase. Table C-2 illustrates the impact on small employers under the incentives the Commission recommends and under requirements, if incentives prove insufficient.

If all smaller employers were to offer insurance in response to incentives, costs would increase $18.8 billion, after taxes, or 1.8 percent of payroll. For smaller employers, the cost of covering previously uninsured persons ($34.7 billion) and expanding benefits to meet the minimum standard ($5.9 billion) would total $40.6 billion. This increase would be offset for five years by a 40 percent refundable tax credit for employers with fewer than 25 workers and an average annual payroll of $18,000 per worker amounting to $5.8 billion; by $7.6 billion in lower taxes (including use of the new opportunity to deduct full premium costs); and by $8.4 billion in reduced responsibility for covering workers of other employers and for uncompensated

care.

The estimated 28.6 million workers in small firms who are not covered through their own jobs-some 3.7 million of whom work part time-would receive job-based coverage under these provisions. Another 11.7 million dependents would also get job-based coverage. In addition, the Commission's plan would affect some existing small employer plans that fall short of the plan's minimum benefit and premium sharing standards. About 25 percent of all small firms that now offer insurance would fail to meet the minimum benefits and premium sharing provisions of the Commission's plan.

If small employers were to become subject to the requirement to provide private or public health insurance, their after tax health expenditures would increase by about $20.6 billion, a 2.1 percent increase in their cost of labor. The cost of requiring employers to contribute to insuring all workers and dependents not now insured would be about $30.5 billion (see Table C-2). (This cost assumes a cap on employer payments for health insurance at 7 percent of payroll, an example of the contribution rate to obtain public coverage.) The cost of the minimum benefits standard and premium sharing requirements would be about $5.9 billion. Estimated total new costs are therefore $36.4 billion.

These costs would be offset by savings of about $15.8 billion. Firms would save $5.9 billion as working spouses and dependents obtain coverage through their own employers, and $2.5 billion in reduced. payments for uncompensated care. In addition, some employers could reduce their health spending by covering their employees under the public plan through

the payroll contribution, lowering their costs by about $0.5 billion. Finally, higher health expenditures are offset by lower tax payments ($6.9 billion), including reductions due to the new opportunity to deduct full premium costs. The net increase in small employer benefit payments would be $20.6 billion (see Table C-2).

Appendix D

The primary data source used in the model is the 1980 NMCUES data, which have been "aged" to be representative of the population and health expenditures in 1990. NMCUES is a representative sample of persons in the United States. It provides information on family income; employment; family structure; sources of health insurance; utilization and expenditures by type of health care service (inpatient, outpatient, and so forth); and sources of payment (insurance, out-of-pocket, and so forth). These data were aged so that they are representative of the levels of

income, population, insurance, and employment as reported in the 1988 CPS data; utilization of health services as reported in the NHIS data; and health expenditures as estimated in the NHA data.

The model was used to estimate the number of persons potentially affected by various components of the Commission's proposal, as well as the number of eligible persons who will enroll based upon observed patterns of enrollment behavior in existing public and private insurance programs. In addition, the model was used to estimate the increase in utilization that will occur as coverage is extended to previously uninsured persons. The model was then used to estimate program costs based on the household health expenditures data using the reimbursement and cost sharing rules of the Commission. The model was also used to estimate savings realized by current payers of health care benefits (such as family out-of-pocket, county hospitals, and charity care) as services previously paid by these sources become covered under the Commission's proposal.

The impact of the Commission's plan on employers and the likely employer responses to various tax incentive proposals were incorporated in the model. The analysis was based largely on a Lewin/ICF survey of employers in 1986, which provides information on (1) the number of persons offering health insurance; (2) characteristics of those persons' health plans; and (3) firm size, payroll, and revenues. These data were used to analyze the impact of (1) the number of employees in firms that fall short of various minimum benefits standards, (2) the number of workers eligible for various tax credits, and (3) the number of workers in firms that face an incentive to participate in the public plan rather than offering insurance to their employees. Each worker in the aged NMCUES data was assigned to one of 850 employers in the employer survey to estimate the impact of employer behavior on worker and dependent coverage.

The microsimulation approach permits aggregated estimates of program impacts while also providing information on the distribution of these impacts across