Abstract
- Issue: The Affordable Care Act (ACA) made it easier for older adults and those with medical conditions to enroll in individual-market coverage by eliminating risk rating and limiting age rating. While the ACA also encourages young and healthy people to enroll through subsidies and the individual mandate, it’s not clear whether these incentives have been sufficient to prevent the risk pool from becoming disproportionately old and sick.
- Goal: To assess whether patterns in individual-market participation changed following ACA implementation.
- Methods: Comparison of Medical Expenditure Panel Survey (MEPS) data for the periods 2003–09 and 2014–15.
- Findings and Conclusion: The analysis found few differences in individual-insurance market participation before and after the ACA. Adverse selection occurred during both: people switching into individual insurance coverage after being uninsured were higher utilizers prior to the switch than were those who remained uninsured. Those who disenrolled from individual plans tended to be lower utilizers of care before switching compared with those who kept their coverage. The main difference was that more people — especially young adults — switched from Medicaid to individual insurance, and vice versa, after the ACA. Adverse enrollment or disenrollment in the individual market did not increase following ACA implementation. The combination of easing rating rules and encouraging participation appears to have maintained market stability.
Background
Much attention in recent years has focused on the impact of the Affordable Care Act (ACA) on the stability of the individual health insurance market. The law’s elimination of medical underwriting and health-rating restrictions, and its limits on age rating, have made the market more attractive to older, sicker people. In an effort to promote a balanced risk pool, the law also included an individual mandate to purchase coverage, intended to incentivize young and healthy adults to buy coverage, as well as subsidies to make coverage less costly for lower-income Americans.1 (The mandate penalty was set to zero under the 2017 Tax Cuts and Jobs Act.)
Health insurers, while generally supportive of the ACA, have expressed concerns that those who entered the individual market following its implementation, and those who retained coverage throughout the year, were less healthy than those who chose not to enter or who exited the market.2 Some insurers have argued that federal regulation has constrained their ability to serve the young and healthy population,3 and advocated for deregulation to allow them to offer less comprehensive and cheaper plans that would appeal to those with fewer medical needs. Some also argue that ACA’s special enrollment periods should be modified to discourage the enrollment of people with greater health needs.4 They have advocated for broader age bands, more limited special enrollment periods, and higher mandate penalties.
This issue brief explores whether the ACA’s provisions, on balance, led to a deterioration in the risk pool in the individual market. Using the Medical Expenditure Panel Survey (MEPS), we examined the patterns of those who enrolled in individual coverage from other types of coverage and from being uninsured and their utilization of health care services before (2003–09) and after (2014–15) the implementation of the ACA’s individual-insurance market provisions. We focus on adults under age 63.
Findings
We used the MEPS to track how many people changed insurance types — across Medicaid, Medicare, employer-sponsored insurance, individual insurance, and no coverage — over two-year periods (Exhibits 1 and 2).
In the pre-ACA period, most people who joined the individual market had previously been uninsured. On average, about 1,305,000 people switched from being uninsured to holding individual-market coverage during each two-year MEPS panel. Such coverage, however, attracted only a small minority of those who had been uninsured — some 3.4 percent of the over 38 million uninsured adults. Nearly as many people, about 908,000, switched from employer-sponsored to individual-market coverage in the average two-year period. However, because far more people began with employer coverage, this group constituted less than 1 percent of those who held employer coverage. In the average two-year period before the ACA, just 41,000 people (0.5% of those who began with Medicaid) switched from Medicaid to individual-market coverage.
In the post-ACA period, as expected, there was a substantial increase in the number of previously uninsured adults entering the individual market. About 5,030,000 people switched from being uninsured to holding individual coverage during the two-year period beginning January 2014, or 12 percent of those who began the period uninsured. The numbers switching from employer to individual coverage doubled in the post-ACA period, as people leaving employer coverage enrolled in individual-market coverage, rather than becoming uninsured (Appendix Exhibit A1).
Switching from Medicaid increased by a factor of 10. Nearly 469,000 people moved from Medicaid into individual-market coverage in the post-ACA period, amounting to 3.2 percent of those who initially held Medicaid coverage. This increase corresponded with a decrease in the proportion of people leaving Medicaid and becoming uninsured (Appendix Exhibit A2). Because of the availability of subsidized individual-market coverage and the mandate to have coverage, people who lost eligibility for Medicaid because their incomes rose bought individual-market coverage instead of becoming uninsured.
There was much less change in terms of people leaving the individual market. In the pre-ACA period, a roughly similar number (and percentage) of people left the individual market for employer insurance or became uninsured: 727,000 (18.7%) and 737,000 (18.9%), respectively. A much smaller group left individual coverage for Medicaid: 45,000 people, or 1.2 percent of the individual market.
In the post-ACA period, the pattern among those who switched from individual to employer coverage (1,308,000 people; 17.1%) or who became uninsured (1,463,000 people; 19.1%) was similar to that seen in the earlier period (though the baseline number of people in the individual market was much higher). By contrast, there was a large increase in the share of people with individual coverage who left that market and enrolled in Medicaid during the post-ACA period: 525,000 people, or 6.8 percent of those who had been covered through the individual market.
Switching by Age Group
We also examined these patterns by age group (Exhibit 3). In the pre-ACA period, those entering the individual market who had previously been uninsured were disproportionately older adults. Among those initially uninsured ages 50 to 63, 4.2 percent switched to individual insurance over a two-year period. By contrast, just 3.4 percent of those between ages 35 and 49 and 2.7 percent between ages 25 and 34 switched from no coverage to individual coverage over a two-year period. Across age groups, the share of people switching to individual from employer insurance was consistently very low. In addition, very few people switched from Medicaid to individual coverage in the pre-ACA period.
In the post-ACA period, uninsured people in all age groups were more likely to switch into individual-market coverage, but the relative gains were greatest among younger people. About 11.4 percent of those initially uninsured between the ages of 25 and 34 switched to individual coverage, as did about 11.5 percent and 13.6 percent of the initially insured ages 35 to 49 and 50 to 64, respectively. There was little overall change in the number of people switching from employer to individual coverage after the ACA, but the composition of this population shifted markedly toward older adults. This may reflect switching from COBRA policies (which allow people to retain their employer coverage for 18 months after leaving a job) to marketplace coverage among older adults eligible for subsidies. By contrast, the large increase in switching from Medicaid to individual-market coverage was concentrated among young adults. About 4 percent of young adults initially covered by Medicaid switched to individual-market coverage, as did 2.8 percent of middle-aged adults and 2.7 percent of older adults.
The stability of the individual market also depends on patterns of disenrollment, because the market can become unstable if a disproportionate number of young and healthy people leave (Exhibit 4). Many people switch out of individual-market coverage each year, most often into employer coverage. This behavior is particularly common among young adults. About one-third (32.8%) of people ages 25 to 34 who initially held individual coverage switched to employer coverage over each two-year panel, as they entered the labor market or got jobs with benefits. This pattern was somewhat less prevalent among older people: 20.2 percent of those ages 35 to 49 and 17.1 percent of those ages 50 to 63 with individual coverage switched to employer coverage. After the ACA, rates of switching from individual to employer coverage remained relatively stable among young adults. The absolute number of middle-aged and older adults switching from individual to employer coverage rose, as the total number with individual coverage increased. But among people who began with individual insurance, the share of those who switched to employer coverage declined.
After the ACA, there was a substantial increase in the number of people who left individual-market coverage for Medicaid, because of the expansion of eligibility for that program. About 18 percent of those ages 25 to 34 who had been initially enrolled in individual coverage switched to Medicaid in 2014–15, as did 4.7 percent of those ages 35 to 49 and 2.9 percent of those ages 50 to 63.
Overall, the main change in terms of age related to Medicaid. Pre-ACA, very few people of any age switched from individual coverage to Medicaid or from Medicaid into individual coverage. Post-ACA, a substantial number of people ages 25 to 34 switched from individual coverage to Medicaid and a slightly smaller number of people in this age group switched from Medicaid to individual coverage.
The main consequence of this was that the proportion of young people with individual coverage who maintained that coverage over time fell (Appendix Exhibit A3). Prior to the ACA, about 40 percent of those ages 25 to 34 who held individual coverage retained it through the end of the two-year period. In the post-ACA period, only 31 percent of those in this age group with individual coverage held it until the end of the period. There was almost no difference in the share of people ages 35 to 49 who held coverage continually in both periods (it increased from 56% to 59%). The share of those ages 50 to 63 who held individual coverage continually declined modestly, from 66 percent to 63 percent.
Switching Patterns by Prior Health Service Utilization
To understand how switching patterns may have varied with health characteristics beyond age, we looked at the expenditures of people who switched coverage in the three months prior to a switch, comparing patterns before and after the ACA took effect.
To see whether those who were high users of health care services were more likely to join the individual market, we recorded total health expenditures incurred by people in the three months prior to a switch and compared them with the average three-month expenditures for people who did not switch (Exhibit 5). In the pre-ACA period, those who switched from employer to individual coverage had somewhat lower service use prior to their switch than those who did not switch ($395 vs. $710), while those entering the individual market after having been uninsured had higher average service use prior to the switch than did the typical continuously uninsured person ($295 vs. $190).
Health care costs grew rapidly between 2003–09 and 2014–15, so those who continuously held employer coverage had average spending about 2.5 times higher in the post-ACA period than in the pre-ACA period ($1,795 vs. $710) and those who continuously held individual coverage had average expenditures about 1.7 times higher in the later period ($1,150 vs. $670). Despite these increased expenditures among the continuously insured, the patterns of expenditures among switchers were very similar in the two periods. As in the pre-ACA period, those who switched from employer to individual coverage were somewhat lower utilizers than were those who did not switch, while those who switched from being uninsured to holding individual coverage had higher average service use before the switch than the typical uninsured person — suggesting that less healthy uninsured people were more likely to enroll in individual coverage. Likewise, those who switched out of individual coverage to either employer or no coverage had relatively lower spending than those who continuously held individual coverage in both periods — suggesting that healthier people and lower utilizers were more likely to disenroll from coverage.5 The patterns of enrollment and disenrollment were nearly the same before and after the ACA.6
Discussion
Implementation of the ACA coverage expansions in 2014 led to a substantial increase in coverage in the individual market and in Medicaid. Overall, however, we see little evidence that implementation of these coverage expansions led to meaningful changes in patterns of enrollment in or disenrollment from individual-market coverage by age or health status.
After ACA implementation, many people switched from being uninsured to holding individual coverage, a substantial number switched from Medicaid to individual-market coverage, and a nearly equal number switched from individual coverage to Medicaid. Young adults were disproportionately likely to switch between Medicaid and individual coverage. This is consistent with the findings of Commonwealth Fund Affordable Care Act Tracking Surveys, which show increases in Medicaid coverage among young adults who, prior to the passage of the ACA, could not afford insurance.7 Our results are also consistent with the finding of a 2017 analysis, which found that the ACA “increased transitions to Medicaid and nongroup coverage among the uninsured, while strengthening the existing employer-sponsored insurance system and improving retention of public coverage.”8 We did find some evidence that, after the ACA, older adults who might otherwise have participated in COBRA plans chose individual coverage instead.
We examined whether there was an increase in adverse enrollment or disenrollment in individual-market coverage by comparing expenditures among switchers in the pre- and post-ACA periods. In both periods, we found that people switching among types of coverage are lower utilizers than those who remain in the same type. In both periods, people who switched from being uninsured to purchasing individual coverage were slightly higher utilizers prior to the switch than those who remained uninsured and that, in both periods, those who switched from individual to no coverage were somewhat lower utilizers prior to the switch than those who remained covered by individual insurance. In other words, there has always been adverse selection and retention in the individual market relative to having no coverage, and that remained the case in the post-ACA period.
The ACA led to large numbers of people entering the individual market. Patterns of exit from the individual market remained relatively stable, except that more young people departed individual coverage for Medicaid. While health care service use patterns among those entering and exiting didn’t change, changes in the volume of entrances and departures may have led insurers to perceive that more adverse selection was occurring.
Policy Implications
The package of changes made by the Affordable Care Act provided greater incentives to purchase individual-market plans through subsidies, the individual mandate, and changes in risk-rating rules. This approach appears to have accomplished what was intended: expanding enrollment in the individual market without substantially increasing adverse selection.
Adverse enrollment and disenrollment have always been challenges in the individual market. Maintaining affordable access to coverage for people with preexisting health conditions requires offsetting measures to maintain market stability. As Congress and the Trump administration consider and implement policy reforms that weaken some of these measures (such as eliminating the penalty for not having health insurance or permitting the sale of plans that do not meet all ACA requirements), it will be important to consider their effects to avoid destabilizing the individual market.
How We Conducted This Study
We used data from the Medical Expenditure Panel Survey (MEPS) to conduct these analyses. Our sample included adults ages 25 to 63. We counted the number of switches in a panel or year by creating dummy variables that would count an individual if they had one type of health insurance in one month and a different type the following. The counts are recorded in Exhibit 1. The pre-ACA period, from 2003–09, represents three panels. Accordingly, counts are divided by three to give a single panel average.
We next analyze patterns of utilization and health care expenditures among those who changed their insurance coverage before and after implementation of the ACA.
We recorded the average spending in the three months prior to making a certain kind of coverage switch and compared them. To get data on individual’s expenditures, we merged the event files, which contain expenditure data by month, with the full-year consolidated files by Dwelling Unit/Person ID. Using the same dummy variables for switches as above, we then recorded the sums of expenditures reported in the three months prior to a switch in insurance type. We then recorded the mean expenditure for each switch type. For those respondents that did not report switching, we recorded the average three-month expenditure over their two-year panel. These results are recorded in Exhibit 5. Expenditures were rounded to the nearest $5.
We appended the populations of the two periods together and ran linear regressions to ascertain the statistical significance of switching effects on utilization, and the effects of change in period.
To ascertain the statistical significance of the effects of different switching categories and the change in period on utilization, we ran linear regressions with dummy variables representing each switching category, and then ran regressions recording their interaction with a post-ACA variable.