Estimating the Number of Individuals in the U.S. Without Health Insurance. The CPS Data in Context


Any discussion of the uninsured should be clear about the multiple methods and data sources available for measuring the number of individuals who lack health insurance. There are three basic ways to count the number of uninsured: 1) those uninsured for a full year,3 2) those ever uninsured during a year, and 3) those uninsured at a specific point in time. The estimates obtained from the CPS, and noted above, are technically full-year uninsured estimates.4 While there is some debate on what the CPS actually measures, the structure of the CPS questionnaire asks respondents about their insurance coverage at any time during the prior calendar year, which implies those who indicate no coverage therefore must have been uninsured for the entire year.

There are three other major government surveys that also measure the number of uninsured (in addition to measuring many other topics). These are the Medical Expenditure Panel Survey (MEPS), the National Health Interview Survey (NHIS), and the Survey of Income and Program Participation (SIPP). While the CPS provides the most widely-quoted estimate for the number of uninsured, the design of the survey only allows an estimate of the number of people uninsured for a full year. The MEPS, NHIS and SIPP each provide three estimates for the number of uninsured: 1) full year uninsured, 2) point in time uninsured, and 3) ever uninsured in year.

The following chart lays out estimates of the uninsured from these four surveys.

Table 2. Uninsured Estimates from Various National Surveys

Survey Most Recent Year Method of Estimate
Uninsured For Full Year Point in Time Uninsured Ever Uninsured During the Year
Current Population Survey (CPS) 2003 45.0 million 15.6% of total pop N/A N/A
Medical Expenditure Panel Survey (MEPS) 2001 31.7 million 11.2% of total pop 45.9 million 16.7% of total pop 64.4 million 22.7% of total pop
Survey of Income and Program Participation (SIPP) 1998 21.1 million 8.0% of total pop 40.5 million 14.6% of total pop 56.8 million 20.8% of total pop
National Health Interview Survey (NHIS) 2002 25.8 million 9.4% of total pop 39.7 million 14.3% of total pop 49.9 million 18.4% of total pop

Note: “N/A” = Survey does not capture this dimension

The estimates of the uninsured differ widely based on the time frame used to measure the lack of coverage and based on the survey used. The CPS full-year uninsured estimate appears more consistent with the point-in-time measures of the NHIS, MEPS and SIPP, and is far higher than those surveys’ full-year measure. For this reason, some analysts have suggested, as far back as 1986, that the CPS represents a measure of the uninsured at a point in time, rather than for a full year.5 For reasons to be discussed later in this memo, ASPE believes that it is more likely that the CPS does represent a measure of the full-year uninsured, but is an estimate that is inflated due to poor reporting of Medicaid coverage, and perhaps other coverage types as well. The three other surveys have a more modest undercount of the Medicaid population, and are also thought to have a more accurate estimate of the uninsured.

More detailed data from the 2001 MEPS provide an example of the dynamic nature of insurance coverage. According to MEPS, 64 million people, or 23% of the population, experienced at least one month without coverage. Of those, 51% were uninsured for all of 2001, and the remaining 49% were uninsured for varying lengths of time:

  • 20% were uninsured for 3 months or less;
  • 14% were uninsured for 4 to 6 months;
  • 12% were uninsured for 7 to 9 months;
  • 4% were uninsured for 10 or 11 months.

Those who are uninsured for shorter time periods may represent different policy challenges and solutions. In addition, both the health and economic consequences of spells of uninsurance of different lengths are likely to vary for different subgroups of the uninsured. Policy makers who seek to expand coverage to the uninsured need to consider the dynamic nature of insurance coverage.

While persons uninsured for all of CY 2003 could be uninsured for longer than 12 months, the reference period of the CPS is limited to a single year.

There has been considerable debate on the interpretation of the CPS in the academic and policy communities. As will be discussed, many believe the CPS is actually a better measure of the number of uninsured at the time of the survey, not the number of uninsured for the previous calendar year.

Katherine Swartz, "Interpreting the Estimates from Four National Surveys of the Number of People Without Health Insurance," Journal of Economic and Social Measurement, vol. 14 (1986), pp. 233-242.

View full report


"report.pdf" (pdf, 121.72Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®

View full report


"TechAppendix.pdf" (pdf, 366.49Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®