Comparison of Expenditures in Nongroup and Employer-Sponsored Insurance
November 2006
Revised: February 2007 (see Errata)

Data from the insurance industry and reviews of premiums offered through on-line sellers show that premiums for nongroup health insurance are much lower than premiums reported on national surveys for employer-sponsored health insurance (ESI).  This paper uses data from the Medical Expenditure Panel Survey to compare the insurance payments for, and out-of-pocket payments by, people with nongroup health insurance and people with ESI.  The comparison finds that average payments by private insurance for enrollees with nongroup insurance are much lower than average payments for ESI enrollees, consistent with the observed lower premiums for nongroup coverage.  A potential reason for this difference is the much higher share of health expenses that nongroup enrollees pay out-of-pocket.  The higher out-of-pocket shares paid by nongroup enrollees suggests that nongroup policies have higher cost sharing and/or cover fewer health expenditures than employer-sponsored insurance.  Nongroup enrollees also are more likely than ESI enrollees to report that their health status and mental health status are excellent.  The better health of nongroup enrollees also helps explain why nongroup premiums are lower than premiums for ESI.

The nongroup health insurance market has received an increasing amount of attention from the health policy community and the media in recent times.  Encouraging people to purchase insurance on their own, rather than relying on benefit plans selected by their employers, is consistent with recent policy emphases on consumerism, individual responsibility for health, and individual ownership of financial benefits.  Interest in health savings accounts (HSAs) also has focused attention on the nongroup market, where the prevalence of HSA compatible plans has grown rapidly.¹  A recently enacted Massachusetts law, which aims to achieve near universal coverage in the state by requiring people who do not get insurance through their work to directly purchase coverage, should draw even greater attention to this market.

Understanding the costs for and benefits provided by nongroup coverage is important for those considering policies that would build on or expand the nongroup market. This is particularly important for policies that would extend nongroup coverage to the uninsured, who disproportionately are lower income. Data from convenience samples of insurers and from on-line brokers suggest that nongroup premiums are much lower on average than premiums for employer-sponsored insurance – maybe as much as 50% lower.²  They also suggest that cost sharing may be higher, on average, in nongroup policies than in employer-sponsored plans.³ For policy makers, the attractiveness of the apparent low price for nongroup policies needs to be balanced against questions about how much coverage these policies provide. Policies with high cost-sharing may not be attractive to lower income people with few resources to meet high out-of-pocket demands.

This brief looks at the amount of coverage provided in the nongroup market by comparing the spending of nonelderly people with nongroup coverage (“nongroup enrollees”) to nonelderly people with employer-sponsored health insurance (“ESI enrollees”). In particular we look for each group of enrollees at the relative amounts paid by private insurance and at the relative percentages of spending that are out-of-pocket. The average expenditures paid by private insurance tell us how much insurance pays for each enrollee; a significant difference between the average for nongroup and the average for ESI enrollees would suggest a difference in the level of coverage between the two markets. In addition, the percentage of expenditures that people pay out-of-pocket also provides information about the level of coverage that they have. Once people have insurance, it is fair to assume that they would prefer to pay for health care that they use with their insurance rather than out-of-pocket. If we see that one group of enrollees pays for a much higher share of their health expenditures out-of-pocket than the other group of enrollees, it would suggest that the first group has a lower level of coverage.

We also look at the self-reported health status of nongroup and ESI enrollees. Another possible reason why premiums for nongroup coverage may be lower than premiums for employer-sponsored insurance is that, on average, nongroup enrollees are in better health than ESI enrollees. People in better health should use less health care, which would result in lower payments by private insurance on their behalf. The next section of the paper describes the source of data for the analysis. The following sections compare for nongroup and ESI enrollees: (1) amounts paid by private insurance; (2) share of expenditures that are out-of-pocket; and (3) self-reported health status and mental health status. The final section concludes with some observations for policy.

Source of Data

For the analysis, we use data from the Medical Expenditure Panel Survey Household Component (MEPS HC), which is a national survey conducted annually by the Agency for Healthcare Research and Quality that provides annual estimates relating to the health status, expenditures, coverage and use of the U.S. civilian non-institutionalized population.⁴  To increase sample size within insurance groups, we pooled data from the 2000-2003 MEPS HC. Expenditure amounts were adjusted to account for differences in spending between years.⁵  The amounts shown are expenditures for health care services and do not include premium payments.

The method for identifying nongroup and ESI enrollees is discussed in Appendix A. In brief, the enrollees in each coverage group were insured by private health insurance for a full 12 months, although in some cases they may have had different coverage types during the year.⁶  On average, both nongroup enrollees and ESI enrollees had just less than 12 months of coverage during the year. People who had public coverage or were uninsured in any month were excluded from the analysis. Average ages were 35 for nongroup enrollees and 33 for ESI enrollees.

Comparison of Expenditures Paid by Private Insurance

Table 1 compares payments by private insurance for nongroup and ESI enrollees. Because health care spending, on average, increases with age, enrollees are divided into four age groups. Average and median values are shown for each age category.

TABLE 1

Average and Median Annual Expenditures Paid By Private Health Insurance,
Nonelderly Nongroup Enrollees and ESI Enrollees, By Age

The average and median annual amounts paid by private health insurance are higher for ESI enrollees than for nongroup enrollees for all age groups.  All differences for all age groups are statistically significant.  The results are presented two ways: the first column for each enrollee group shows total annual expenditures by private health insurance and the second column for each group shows annual expenditures for private health insurance adjusted to remove payments for dental and vision services.  Since dental and vision services, when covered, are often covered by separate or supplemental insurance policies, the adjusted amounts should provide a more accurate picture of expenditures for basic benefits usually covered by insurance.  ESI enrollees also are much more likely than nongroup enrollees to have any health insurance payments for vision and dental services, suggesting that coverage for these services is more prevalent for people who get coverage through work than for those who must buy it directly. 

Average and median amounts paid by private insurance are statistically significantly higher for ESI enrollees than nongroup enrollees in each age group.  Relative to the values for nongroup enrollees, the average expenditures paid by private insurance for ESI enrollees are 245% higher for people ages 0 to 17, 55% higher for people ages 18 to 34, 73% higher for people ages 35 to 49, and 87% higher for people ages 50 to 64.  The percentage differences within age groups do not change appreciably when expenditures for vision and dental services are removed.⁷ 

These very large differences in payments by private insurance for nongroup and ESI enrollees suggest that nongroup policies are providing less coverage than employer-sponsored insurance.  The results are consistent with the information from insurance industry sources that suggest that nongroup policies have higher up-front cost sharing than employer-sponsored insurance,⁸ although health status or other differences between nongroup and ESI enrollees also could help explain the differences in the amount paid by private insurance.  The next section looks at another way to measure level of coverage – the proportion of expenses paid out-of-pocket by people who have health expenditures. 

Out-Of-Pocket Spending as a Share of Total Health Expenditures

This part of the paper looks at the share of health expenditures that people pay out-of-pocket as a proxy for the level of coverage that people have – in other words, would the coverage be expected to pay for a relatively smaller or larger share of the health care expenditures that a person might have.  We assume that people who have insurance would rather use that insurance to pay for their health care expenses than to pay for them out-of-pocket.  Therefore, people who pay for a relatively large share of their spending out-of-pocket do so because a relatively large share of their health expenditures is not covered by their insurance.  This may be due to coverage restrictions or relatively high deductibles and other cost sharing.

For the people in our enrollee groups, payments from private insurance and out-of-pocket payments constitute the vast majority of their spending (98% for nongroup enrollees and 97% for ESI enrollees). The small remainder is made up of spending from other sources such as workers compensation, Veterans Administration or public programs.  For this analysis, we exclude the payments made by these other sources from the calculations because they may be paying for services that would not normally be covered by health insurance, such as work-related injuries.  Excluding expenditures from these sources does not materially affect the results.  Therefore, “total expenditures” refers to the sum of private insurance expenditures and individual out-of-pocket expenditures.

We look at out-of-pocket shares three ways.  In Table 2, for nongroup and ESI enrollees, we show the average total out-of-pocket expenditures as a proportion of the average total expenditures.  Table 3 shows the average and median out-of-pocket shares for nongroup and ESI enrollees who have health expenditures.  For this table, we calculate the out-of-pocket share for each enrollee and summarize the results.  The percentages in Table 3 are higher than in Table 2 because enrollees with lower spending, who make up the majority of enrollees, have relatively high out-of-pocket shares.  Enrollees with high total spending make up a small share of enrollees but a large share of total spending.  These enrollees have relatively low out-of-pocket shares.  Table 4 demonstrates this difference more directly by showing the average and median out-of-pocket shares for nongroup and ESI enrollees who were among the top 20% and the top 5% in total health expenditures in their respective enrollee groups. 

The story is similar for each way of looking at out-of-pocket shares.  Looking at aggregate spending across all nongroup and ESI enrollees, 43% of total heath expenditures by nongroup enrollees are paid out-of-pocket, as compared to 22% for ESI enrollees (Table 2).  The percentages fall somewhat when expenditures for vision and dental are removed from the calculation, but the large difference between nongroup and ESI enrollees remains.

TABLE 3

Average and Median Percentages of Total Expenditures Paid
Out-of-Pocket by Nonelderly Nongroup and ESI Enrollees
with Health Expenditures

For people with health expenditures, Table 3 shows that nongroup enrollees on average paid a much higher share of their health expenditures out-of-pocket than ESI enrollees. The median nongroup enrollee with expenditures also has a much higher out-of-pocket share than the median ESI enrollee with expenditures.  The pattern is similar whether expenditures for vision and dental services are included or excluded.

In Table 4 we extend the analysis to see if the difference in out-of-pocket shares persists for nongroup and ESI enrollees with high health care expenditures.  We looked at the out-of-pocket shares of enrollees who were among the top 20% and the top 5% in total health expenditures.  These groupings are not exclusive:  enrollees in the top 5% of spending are by definition also in the top 20% of spending.  We find that average out-of-pocket shares and median out-of-pocket shares became smaller among both nongroup and ESI enrollees as total health spending rises (compare Tables 3 and 4), but that significant differences between nongroup and ESI enrollees persist. 

TABLE 4

Average and Median Percentages of Total Expenditures Paid
Out-of-Pocket by Nonelderly Nongroup and ESI Enrollees
with High Health Care Expenditures

Tables 2 through 4 suggest that nongroup insurance is less likely than employer-sponsored health insurance to pay for health care expenditures that people have, leaving nongroup enrollees to pay for a relatively high share of their health care expenditures out-of-pocket.  This is true whether looking at total spending over all nongroup and ESI enrollees or looking at the average or median out-of-pocket shares of enrollees with health spending.  Even among enrollees with very high total health care expenditures, nongroup enrollees have higher average and median out-of-pocket shares than ESI enrollees.  Again, these results are consistent with the information from insurance industry sources indicating that nongroup insurance policies, on average, have higher cost sharing than employer-sponsored coverage.  These results also show that the size of the difference in protection is meaningful.

Health Status of Nongroup and ESI Enrollees

This section briefly looks at the self-reported health status of nongroup and ESI enrollees.  Another potential reason why nongroup coverage might cost less than employer-sponsored coverage would be that nongroup enrollees are healthier on average than ESI enrollees. 

The MEPS HC asks respondents to classify their health status and their mental health status on 5-point scales:  excellent, very good, good, fair, or poor.  Table 5 shows that nongroup enrollees are more likely than ESI enrollees to classify their health status as excellent (43% vs. 35%).  Table 6 shows a similar result for mental health status.  The differences in the distributions of responses between nongroup and ESI enrollees are statistically significant in both tables; the differences in the percentage of respondents reporting that their health or mental health status is excellent also are statistically significant. 

Table 5

Perceived Health Status of Nonelderly Nongroup and ESI Enrollees

Table 6

Perceived Mental Health Status of Nonelderly Nongroup and ESI Enrollees

These differences in perceived health may reflect the populations who are able to choose employer-sponsored and nongroup coverage, or they may reflect market rules in many states which restrict people in poorer health from enrolling in nongroup coverage.  In any event, if nongroup enrollees are healthier than ESI enrollees, they may use fewer health care services, which could help explain the relatively low nongroup premiums that we see in the convenience surveys from industry sources.

Discussion

This analysis of national survey data from MEPS shows that average payments by private insurance for people with nongroup insurance are much lower than average private insurance payments for people with employer-sponsored coverage.  This is consistent with information from insurance industry sources suggesting that premiums for nongroup insurance are lower on average than premiums for employer-sponsored insurance.

The analysis also suggests several reasons why private insurance payments for nongroup enrollees would be relatively low.  First, nongroup enrollees pay a much higher proportion of their health expenditures out of pocket, which indicates that on average nongroup insurance provides less coverage than employer-sponsored insurance.  Nongroup coverage may have higher cost sharing, may cover fewer services, or both.  The supposition that nongroup insurance has relatively high cost sharing is consistent with insurance industry surveys which indicate deductible levels that are higher than we see for employer-sponsored insurance.⁹ Second, nongroup enrollees are more likely than ESI enrollees to report that their health status and mental health status are excellent.  If nongroup enrollees are healthier on average than ESI enrollees, it would help explain why private nongroup insurance is relatively less expensive than employer-sponsored coverage.

These findings raise some important considerations for policy makers evaluating different options for covering the uninsured.  One is the relatively high share of total health spending that is required in the nongroup market.  Public policies that would provide current nongroup policies to low-income people may still leave them with quite burdensome out-of-pocket costs, and policy makers may want to consider approaches to limit the cost sharing and other out-of-pocket expenses under coverage expansions targeted to those with low incomes. 

Another consideration for policymakers is that the premium levels observed in the nongroup market today may be lower than the premiums that would be necessary to provide coverage to a greater share of the population, including people who are chronically uninsured.  Premiums in the nongroup market may reflect the relatively good health of the nongroup population, and the target population for expanded coverage may not share the same good health. Appendix A shows the perceived health status and mental health status of nonelderly people who do not report having any insurance during the year: people without insurance are much less likely to report their health and mental health as excellent and are more likely to report their health and mental health as fair or poor than either nongroup or ESI enrollees.¹⁰  Policymakers should be careful not to assume that they could insure many of the currently uninsured at the premium levels now observed in the nongroup market.

APPENDIX A

Perceived Health Status of Nonelderly Uninsured

Perceived Mental Health Status of Nonelderly Uninsured

1. America’s Health Insurance Plans (AHIP) (January 2006). Census Shows 3.2 Million People Covered by HSA Plans. Retrieved on October 17, 2006 from www.ahip.org/content/default.aspx?docid=15302.

2. A recent report from America’s Health Insurance Plans (AHIP), based on information from member companies, reported average premiums for 2004 of $2,268 for single coverage and $4,424 for family coverage; by comparison, data from the 2004 KFF/HRET Employer Health Benefits Survey shows average premiums for employer-sponsored coverage of $3,695 for single coverage and $9,950 for family coverage.

3. For example, AHIP reports that over 30% of nongroup enrollees with single coverage had deductibles of $2,000 or more in 2004.  For more information on this report, see http://www.ahip.org/content/default.aspx?bc=31|130|136|259|261.  Data from the 2004 KFF/HRET Employer Health Benefits Survey show that, for PPO plans, only 2% of enrollees with single coverage in employer-sponsored plans had a deductible of $2,000 or more for preferred providers. 

4. For more information on the Medical Expenditure Panel Survey, see http://www.meps.ahrq.gov/mepsweb/.

5. To normalize expenditure amounts over the four years, we divide expenditure amounts in each year by the weighted per capita mean expenditure amount for that year.  The resulting ratios are then multiplied by the weighted average mean per capita expenditure amount for the four years combined.

6. The ESI group includes those with TriCare coverage.

7. Without expenditures for vision and dental services, average expenditures paid by private insurance for ESI enrollees are 257% higher for people ages 0 to 17, 53% higher for people ages 18 to 34, 62% higher for people ages 35 to 49, and 83% higher for people ages 50 to 64.

8. America’s Health Insurance Plans (AHIP) (January/February 2004). Individual Health Insurance: New Studies Shed Light on Affordability, Access, and Plan Design. Retrieved from http://www.ahip.org/content/default.aspx?bc=31|130|136|259|261

9. See note three for a comparison of deductibles for nongroup coverage compared to employer-sponsored coverage.

10. The differences in the distributions of health status between the uninsured and either people with nongroup insurance or ESI are statistically significant at p<.05. 

Errata: The November 2006 publication of this Snapshot did not include TriCare expenditures with private expenditures, despite those with TriCare coverage being included in the ESI group. This error has been corrected, and the revision did not change the results significantly.