The National Institute for Health Care Management has issued two readable and extremely useful papers on the uninsured.

The first NIHCM paper, released in August 2006, is a primer on the alternative ways the uninsured are measured.

Most newspapers stories on the uninsured cite the measure put out by the Census Bureau every August, which is based on sample data collected as part of the Current Population Survey (CPS).

But, as the NIHCM primer notes, there are two important problems with this data.

First, the sample is collected every March, and respondents are asked essentially whether or not they were uninsured during the prior year. The question is supposed to elicit a “yes” only when someone went without coverage for the entire calendar year in question. But many people apparently answer “yes” based on an imperfect understanding of the question. Hence, many people who are uninsured at the time they are asked the question seem to answer yes, as well as those who were uninsured during a portion of the calendar year.

Other surveys, like the Survey of Income and Program Participation (SIPP), follow the same people for a number of years and collect data from them more frequently. This sampling approach allows better “point in time” versus “all year” determinations. And it is noteworthy that, based on SIPP data, the number of Americans uninsured all year in 2002 was only 23 million (compared to about 42 million using CPS data). The SIPP survey also found that about 43 million Americans were uninsured at any given point in time during 2002—fairly close to the CPS estimate of the full-year uninsured count.

The second NIHCM paper, issued in April of this year, looks at different subpopulations of the uninsured and the appropriate policy remedies to expand coverage to them. The paper sorts the uninsured into those eligible for public program coverage (about 12 million) and those who are not. Of those not eligible, the paper further estimates that about 40 percent live in households with incomes exceeding 400 percent of poverty.

The paper also notes that large numbers of the uninsured are in “transition” age groups: children leaving their parents’ coverage but not settled in stable job-based plans, and the near elderly who have left the workplace but are not yet eligible for Medicare. Public policy may need to target coverage options to the specific needs of these age groups.

The primer of CPS, SIPP, and other uninsured data is available here. The paper on uninsured subpopulations and possible policy responses is available here.


September 3, 2008