Update: On May 15, 2018, CBO released a report revising how the agency defines and estimates health insurance coverage for people under 65.
In the United States, most people under age 65 are covered by employment-based (or group) private health insurance that they or their family members obtain. A smaller number of people buy private health insurance individually, known as a nongroup policy. Those nongroup policies are available through the health insurance marketplaces established under the Affordable Care Act (ACA) or outside them, through brokers or directly from insurers. Two of the major sources of public insurance coverage for people under 65 are Medicaid and the Children’s Health Insurance Program (CHIP).
The federal government subsidizes private and public insurance coverage through various tax preferences and federal programs. Because those subsidies affect the federal budget in many ways, defining what constitutes coverage and projecting health insurance coverage for people under 65 is an important aspect of CBO’s budget projections. The most recent year with actual coverage data serves as the starting point for CBO’s coverage projections. (More information about the methods that CBO uses to make its projections of coverage is available on CBO’s website.)
This blog post is the first in a series over the coming year to explain more about how CBO analyzes health care topics. Here we describe how CBO defines health insurance coverage (private and public) for people under 65 who are not institutionalized and who are not members of the active-duty military. We explain how the agency estimates the number of insured and uninsured people in that population for the most recent year for which data on actual coverage exist. We also describe where CBO obtains the data to estimate coverage, the limitations of those sources, and how CBO adjusts its estimates because of those limitations. (A bibliography detailing the studies that discuss the issues in estimating public and private insurance coverage from survey and administrative data accompanies this post.) CBO’s most recent detailed projections of health insurance coverage for people under age 65 are shown in Table 1 on page 4 of Federal Subsidies for Health Insurance Coverage for People Under Age 65: 2016 to 2026.
- How Does CBO Define Private Insurance Coverage?
- How Does CBO Define Public Insurance Coverage?
- What Data Sources Does CBO Use to Estimate the Number of People With and Without Health Insurance Coverage?
- What Are the Challenges in Using Survey Data to Estimate the Number of People With and Without Health Insurance Coverage?
- What Are the Challenges in Using Administrative Data to Estimate the Number of People With Health Insurance Coverage?
- How Does CBO Estimate the Number of People With Private Insurance Coverage?
- How Does CBO Estimate the Number of People With Public Insurance Coverage?
- How Does CBO Estimate the Number of People Without Health Insurance Coverage?
How Does CBO Define Private Insurance Coverage?
Health insurance policies vary widely, ranging from some that offer substantial coverage for a variety of health care services to some with a limited scope or amount of coverage. Therefore, in preparing any estimate of the number of people covered by health insurance, it is useful and important to identify where to draw the line to distinguish policies that provide some type of comprehensive coverage from those that do not.
An important function of insurance is to provide financial protection against high-cost, low-probability events. Consistent with that notion, CBO broadly defines private health insurance coverage as a comprehensive major medical policy that, at a minimum, covers high-cost medical events and various services, including those provided by physicians and hospitals. The agency grounds its coverage estimates on that widely accepted definition, which encompasses most private health insurance plans offered in the group and nongroup markets. The definition excludes policies with limited insurance benefits (known as “mini-med” plans); “dread disease” policies that cover only specific diseases; supplemental plans that pay for medical expenses that another policy does not cover; fixed-dollar indemnity plans that pay a certain amount per day for illness or hospitalization; and single-service plans, such as dental-only or vision-only policies.
When specific requirements are established in law, CBO relies on those definitions to further determine what policies count as private insurance coverage. To define coverage under the ACA, CBO relies on provisions in that law that established detailed requirements governing the benefits of private insurance coverage in the large-group market, which is generally defined as employers with more than 50 employees. (For more about the minimum value standard used to govern benefits in the large-group market and other aspects of the private health insurance market, see CBO’s report about private health insurance premiums.)
CBO also relies on separate provisions of the ACA that define the requirements governing plans offered in the small-group (generally defined as employers with up to 50 employees) and nongroup markets. Since 2014, new plans sold in those markets must cover 10 categories of health benefits that the ACA defines as essential. Other provisions require that the actuarial value (a summary measure of the depth of coverage) of those plans fall into specified categories (60 percent, 70 percent, 80 percent, and 90 percent actuarial value). Only in limited circumstances may plans with an actuarial value of less than 60 percent—known as catastrophic plans—be sold to certain individuals. (A plan with an actuarial value of 60 percent means that for a standard population, the plan will pay for 60 percent of covered health care expenses, while enrollees are responsible for 40 percent of health care expenses through some combination of deductibles, copayments, and coinsurance.) Some plans that existed before 2014 and have continued are exempt from those requirements or some of the ACA’s regulations. CBO counts those noncompliant plans and catastrophic plans as private insurance coverage because they typically provide major medical coverage and are permitted under the ACA in limited circumstances.
If the provisions of the ACA governing the definition of private insurance coverage were repealed, CBO would revert to the broader definition of private insurance coverage—a comprehensive major medical policy, as described above. Such a broad definition of private insurance coverage is in keeping with what the agency has used to estimate coverage in the past. For a discussion of how CBO would estimate coverage under alternative proposals, see CBO’s blog post about challenges in estimating health insurance coverage under proposals for refundable tax credits.
How Does CBO Define Public Insurance Coverage?
CBO defines as publicly insured people who receive full Medicaid or CHIP benefits. The agency’s definition of publicly insured does not include people who receive only partial Medicaid benefits—such as women who receive only family planning services or unauthorized immigrants who receive only emergency services. In addition, CBO defines as publicly insured adults under age 65 with disabilities who are covered by Medicare. (For more about those definitions and coverage provided through other sources, see CBO’s report about federal subsidies for health insurance coverage for people under age 65.)
What Data Sources Does CBO Use to Estimate the Number of People With and Without Health Insurance Coverage?
To estimate the number of people with and without health insurance coverage, CBO combines data from household and employer surveys with administrative data provided by government sources. CBO uses survey data as the basis for estimating employment-based private insurance coverage because no comprehensive administrative data exist. The agency also uses survey data to estimate the number of people without coverage because no administrative data on the uninsured are available. By contrast, CBO relies on newly available administrative data from the Centers for Medicare & Medicaid Services (CMS) to estimate the number of people with private insurance coverage in the health insurance marketplaces. Similarly, CBO relies primarily on data from administrative records to count people with public insurance coverage through Medicaid and CHIP.
Certain administrative records, such as those that record program participation, are generally more accurate than reports from household surveys because they are complete tabulations rather than samples and because they support program funding. Survey data, however, are used to determine the demographic and income characteristics of Medicaid and CHIP enrollees because that information is lacking in administrative data. Survey data can thus fill in the gaps by providing estimates for types of coverage that are not included in administrative data or by showing how coverage varies by demographic variables, such as income and age.
CBO uses data from several household surveys—including the National Health Interview Survey (NHIS), the Medical Expenditure Panel Survey (MEPS)—Household Component, the Current Population Survey (CPS), and one employer survey, the MEPS—Insurance Component—to estimate the number of people with private insurance coverage or without coverage. Those surveys are also used to estimate how public insurance coverage varies by demographic variables. In addition, CBO uses the Census Bureau’s Survey of Income and Program Participation (SIPP) as the base data in its Health Insurance Simulation Model (HISIM). For additional details, see CBO’s technical description of that model. In the next generation of its microsimulation model, now under development, CBO will use the CPS as its base data.
CBO uses the SIPP as the base data in HISIM because it includes detailed information on individuals and families, such as demographic characteristics, income, health status, employment status, insurance coverage, and employers’ offers of insurance. That detailed information allows CBO to make coverage projections for current and future years, and it supports the simulation of behavioral responses of individuals and families to changes in policy and the resulting changes in coverage.
The SIPP data alone, however, do not accurately indicate the extent of current insurance coverage. That is because the survey data are from an earlier year, and survey respondents typically misreport their sources of coverage to some degree (because they do not always accurately recall what those sources were). As a result, CBO adjusts the SIPP data to match coverage estimates developed from a combination of administrative data and other household and employer surveys (as discussed below). CBO takes that step so that its historical coverage estimates reflect the most recent year for which data on actual coverage distributions exist; the estimates serve as the starting point for the development of future projections.
What Are the Challenges in Using Survey Data to Estimate the Number of People With and Without Health Insurance Coverage?
The main challenge in using household survey data to estimate the number of people with and without health insurance coverage involves measurement. Important aspects include the following:
- Errors in the reporting of coverage status,
- Different reference periods (or reporting windows) across surveys, and
- A lack of information on the depth and extent of private insurance coverage.
The potential for error on the part of respondents in reporting their insurance coverage is always present in household surveys, but it can depend, in part, on how the survey questions are structured. For example, some surveys ask whether anyone in the household had coverage, whereas others ask whether each person in the household (by name) had coverage.
Data in the household surveys used by CBO and listed above substantially undercount the number of people with coverage through Medicaid and CHIP because of misreporting. Methodological research suggests that the reason for the undercount is that some respondents confuse those public insurance programs with other types of coverage, such as private insurance. Also, some people appear not to report having public insurance coverage because of the stigma associated with receiving public assistance. To correct for those measurement problems, CBO uses administrative data to count enrollees in Medicaid and CHIP.
The reference period also varies across the surveys listed above. Some surveys ask respondents about their coverage at a point in time, such as on the date of the interview or during the previous few months. Other surveys ask respondents about their coverage at any time during the previous calendar year. The length of the reference period and the time since that reference period occurred (the recall period) can affect the accuracy of respondents’ answers. The more time that has passed since the reference period, the more difficult it is for respondents to correctly recall their coverage status.
Furthermore, the different reference periods might affect the estimates of the number of people with and without coverage. For example, the number of people who are uninsured at any time during the year is generally higher than the number of people uninsured at a specific point during the year, which, in turn, is higher than the number of people uninsured for the entire year.
A related issue is that different reference periods might affect estimates of the number of people with specific types of coverage. In surveys that ask about coverage at any time during the year or over a certain period, respondents have the potential to report more than one type of coverage (employment based, Medicaid, or nongroup, for example). That approach can generate higher estimates for specific types of coverage because many people may have different sources of coverage or temporary lapses in coverage throughout the year, such as between jobs.
Another challenge with household survey data is that they provide very little information on the depth and extent of private insurance coverage—in terms of the scope of benefits, the level and structure of cost sharing, and the actuarial value of plans. Although that information is lacking for households, some total statistics on the depth and extent of private insurance coverage in the employment-based market are available. The Agency for Healthcare Research and Quality has recently begun to publish such data from the MEPS—Insurance Component (a survey of private and state and local government employers). For policies in the health insurance marketplaces, detailed information on the scope of benefits, the amount and structure of cost sharing, and the actuarial value of plans is publicly available.
In addition to those measurement challenges, there is often a delay between when survey data are collected and when they are made available. The delay can be even longer if respondents are asked to report on their insurance coverage for a time before the date of collection, such as the previous year.
What Are the Challenges in Using Administrative Data to Estimate the Number of People With Health Insurance Coverage?
Using administrative data to estimate the number of people with health insurance coverage presents three main challenges. The first is the delay between the measurement period and the availability of the data. The second is that most sources of administrative data lack detailed information on a person’s demographic characteristics, such as age, sex, income, and employment status. The third is that administrative data have the potential to misreport coverage. For example, they might double-count people who have more than one insurance policy within a state or who sign up for coverage in more than one state during a year.
How Does CBO Estimate the Number of People With Private Insurance Coverage?
CBO uses data from the MEPS—Insurance Component survey as a benchmark to estimate the number of people with private insurance coverage in the employment-based market. CBO then adjusts that benchmark to incorporate federal employees’ health coverage (because data from the MEPS—Insurance Component survey do not include federal agencies). CBO uses the MEPS—Insurance Component survey because it applies to employers and because no administrative data provide a complete count of the number of people with employment-based coverage.
Estimating private insurance coverage for the nongroup market has become much easier following enactment of the Affordable Care Act. Since the establishment of the health insurance marketplaces, CMS has collected administrative data that CBO use as a benchmark of total enrollment in the marketplaces. To estimate enrollment in the nongroup market outside the marketplaces, CBO uses administrative data from insurance filings.
How Does CBO Estimate the Number of People With Public Insurance Coverage?
CBO uses data from two sources to estimate public insurance coverage provided through Medicaid and CHIP. To count enrollees in those programs, CBO uses administrative data submitted by the states to CMS. Those data provide the most accurate counts of public insurance coverage because people often misreport that coverage in household surveys. To determine the demographic and income characteristics of those Medicaid and CHIP enrollees, CBO uses household survey data from the SIPP.
CBO then adjusts the administrative data to better match its definition of public insurance coverage. For example, CBO excludes people who receive only partial Medicaid benefits. But even though those enrollees are not considered covered by Medicaid as defined by CBO for purposes of determining public insurance coverage, they are included in CBO’s counts of total Medicaid enrollment and spending.
Furthermore, CBO counts only people actually enrolled in Medicaid and CHIP when estimating coverage in those programs. Some people argue that individuals who are eligible for public programs but have not enrolled should not be counted as uninsured because those people could enroll at any time. CBO does not count as covered people who are eligible but not enrolled in Medicaid and CHIP because they do not generate federal spending on those programs.
How Does CBO Estimate the Number of People Without Health Insurance Coverage?
CBO uses data from the MEPS—Household Component as a benchmark to estimate the number of people who are uninsured. No single source of data exists regarding people’s coverage status, and the only reliable information on uninsured people comes from federal surveys. Because of differences in design across surveys, CBO compares estimates from the MEPS—Household Component with estimates from the NHIS and the CPS. It considers the strength of each survey to arrive at a final estimate of the number of people who are uninsured.
Jared Lane Maeda and Susan Yeh Beyer are analysts in CBO’s Health, Retirement, and Long-Term Analysis Division. This blog post was prepared with guidance from Jessica Banthin, a deputy assistant director in that division.