To: jlallen who wrote (554694 ) 3/12/2010 9:55:01 AM From: Alighieri Read Replies (1) | Respond to of 1572102 You don't know what the F you are talking about...go haunt some other thread...imbecile... Al ================================================================== Construction of the Base Case The core data set in CBO’s model is taken from the SIPP, a longitudinal survey of roughly 70,000 individuals interviewed at 4-month intervals over a 40-month period. Although the Census Bureau’s Current Population Survey has a larger sample and includes an annual health insurance module, CBO relied on the SIPP in part because of its iterative design: It asks the same questions of the same respondents over an extended period. Thus, it is a better tool for modeling policy proposals that predicate eligibility on a person’s having been uninsured for a specified period. The core data on demographics, family structure, income, employment, and insurance coverage are drawn from SIPP responses covering the month of May 2002, which included a special supplement (Topical Module 5) with detailed questions about health insurance, offer and eligibility status for ESI, and health status. Those data are further supplemented with several additional questions that were presented to SIPP respondents shortly before or after May 2002. Income data are included for a full year to give an accurate measure of income and earnings. Tax-related data are taken from a previous SIPP supplement to help inform estimates of marginal tax rates. Additional data, on health status and chronic conditions, are taken from a subsequent supplement. Health Insurance Coverage The health insurance unit (HIU) is a key unit of analysis in CBO’s model. Each HIU consists of all the members of a family who could be covered by one family health insurance policy. HIUs are created from indicators in the SIPP data that describe familial relationships of individuals in households (all individuals in a given household are surveyed in the SIPP). Three-generation families, families with children over age 22, and unmarried adults are usually grouped into separate HIUs. Only the income earned within the HIU is considered available to the unit’s members, and family coverage identified within the HIU is considered as covering all HIU members and no others. Several SIPP variables were used to ascertain the health insurance coverage status of each survey respondent. The use of SIPP indicator variables makes it relatively straightforward to determine whether a person is an ESI policyholder or a dependent. Those covered under insurance that is “privately purchased,” as opposed to sponsored by a current or former employer or by a government program, were considered as having “individual market,” or nongroup, coverage. 5 In CBO’s model, SIPP survey responses indicating multiple sources of coverage are generally assumed to be accurate, and the model is designed to handle that complexity. About 3 percent of the respondents reported simultaneous enrollment in ESI and Medicaid, 1 percent reported ESI single and dependent coverage, and 1 percent reported ESI and TRICARE coverage. (TRICARE is the Department of Defense’s health care program.) In some cases—for example, if it appeared that a transition from one form of coverage to another had occurred during the four-month survey period—multiple coverage was reduced to a single form of coverage. Offers of Employer-Sponsored Insurance The model identifies who is offered employer-sponsored insurance, as derived from participants’ responses to SIPP questions, mostly in Topical Module 5. The SIPP survey asks workers who are not ESI policyholders whether health insurance is offered to any employees where they work. (Respondents who are ESI policyholders are assumed to have an offer of insurance even if they are not working. They could have coverage under COBRA [the Consolidated Omnibus Budget Reconciliation Act of 1985], which extends ESI after employment ends, or they could be on temporary leave.) If the answer is “yes,” respondents are asked why they are not policyholders themselves. The available answers are “not eligible,” “denied coverage,” “chose not to be covered,” and “other.” Only people who give one of the final two responses are categorized as having been offered insurance. Some researchers distinguish between people who work at nonoffering companies and those who are not eligible to enroll in the health insurance plan offered by an employer (for example, because they work part time). This distinction is not critical to the model: Incentives that might cause some employers to begin to offer insurance could just as easily cause others to extend eligibility to previously ineligible employees. If a specific policy proposal might affect the “not offered” category differently from the “offered but not eligible” category, such a distinction could be introduced into the model. Where the responses to questions are missing or incomplete, offer status is imputed according to probability tables of offer, which are based on data from the 2004 Medical Expenditure Panel Survey Insurance Component (MEPS-IC). The probability tables are conditional on firm size, industry, uninsured status, and full- or part-time worker status.2 Nearly 80 percent of full-time workers are eligible for and are offered ESI in the base case, ranging from 51 percent of workers in firms with fewer than 25 employees to 88 percent of workers in firms with more than 100 employees. Overall, 71 percent of all nonelderly people (including part-time employees and nonworking people) have access to an ESI offer, meaning they are offered coverage themselves or they are the 2. MEPS-IC data are based on an extensive annual survey of U.S. employers. The data were supplied to CBO by the Agency for Healthcare Research and Quality in the Department of Health and Human Services. 6 coverable dependents of someone who has an offer. Of the 71 percent with access to ESI, 7 percent remain uninsured; of the 29 percent of people who do not have access to ESI, 41 percent are uninsured and 48 percent have public coverage.