Taking Stock: Gains in Health Insurance Coverage under the ACA as of March 2015
Sharon K. Long, Michael Karpman, Genevieve M. Kenney, Stephen Zuckerman, Douglas Wissoker, Adele Shartzer, Nathaniel Anderson, and Katherine Hempstead
April 16, 2015
Since the Affordable Care Act’s (ACA) first open enrollment period began in late 2013, there has been rapid enrollment growth in Medicaid and in private health insurance plans purchased through the new Marketplaces.1 However, administrative data on Medicaid enrollment and enrollment in Marketplace health plans do not show how health insurance coverage is changing under the ACA, because not all of those enrolling were previously uninsured. Household survey data are needed to track changes over time in the share of the population that is uninsured.
The Urban Institute has used its Health Reform Monitoring Survey (HRMS) to examine trends in health insurance coverage since the first quarter of 2013. The HRMS provides early feedback on changes under the ACA to complement the more robust impact assessments that will be possible as federal survey data become available.2 HRMS data through December 2014, midway through the second Marketplace open enrollment period, show a decline of 4.9 percentage points in the uninsurance rate among nonelderly adults since September 2013, just before the first open enrollment period. In this brief, we use HRMS data to provide estimates of coverage changes between September 20133 and early March 2015, just after the close of the second open enrollment period.4
We find a further drop in the uninsurance rate since December 2014: according to HRMS data, the uninsurance rate among nonelderly adults has declined 7.5 percentage points between September 2013 and March 2015, representing 15 million fewer adults without health insurance.
What We Did
We examine changes in insurance coverage for nonelderly adults (ages 18 to 64) overall and by state Medicaid expansion status going back to the first quarter of 2013.5 We also report coverage changes by family income, age, race and ethnicity, and gender. Each round of the HRMS is weighted to be nationally representative. We use these weights and regression adjustment to control for differences in the demographic and socioeconomic characteristics of the respondents across the different rounds of the survey.6 This allows us to remove changes in insurance coverage caused by changes in the types of people responding to the survey over time rather than by changes in the health insurance landscape. The basic patterns shown for the regression-adjusted measures are similar to those based solely on simple weighted estimates. In presenting the regression-adjusted estimates, we use the predicted rate of coverage in each quarter for the same nationally representative population. For this analysis, we base the nationally representative sample on survey respondents from the most recent 12-month period of the HRMS (i.e., quarter 1 of 2015 and quarters 2–4 of 2014). We focus on statistically significant changes in insurance coverage over time (defined as changes that are significantly different from zero at the 5 percent level or lower) and highlight changes relative to September 2013, just before the first Marketplace open enrollment period began. We provide a 95 percent confidence interval (CI) for key estimates.
There are several limitations to this analysis. Though HRMS estimates capture the changes in insurance coverage from the first open enrollment period under the ACA, the estimates understate the full effects of the ACA because the estimates do not reflect the effects of some important ACA provisions (such as the ability to keep dependents on health plans until age 26 and early state Medicaid expansions) that were implemented before 2013. In addition, these change estimates might not reflect only the effects of the ACA, because they do not control for long-term trends in health insurance coverage that predate the ACA nor do they control for changes in the business cycle. Further, the difference in coverage gains between states that expanded Medicaid and states that did not should not be entirely attributed to the ACA; there were other policy choices that likely affected enrollment. For example, many of the nonexpansion states did not set up their own Marketplaces and therefore did not get the same access to outreach and enrollment assistance funding.
What We Found
Fifteen million nonelderly adults gained coverage between September 2013 and March 2015 as the uninsurance rate fell from 17.6 percent to 10.1 percent.
In September 2013, which was just before the initial Marketplace open enrollment period, we estimate that 82.4 percent of nonelderly adults had health insurance coverage. By early March 2015, which was just after the second open enrollment period, 89.9 percent of the adults were insured. This implies that the uninsurance rate for nonelderly adults fell from 17.6 percent to 10.1 percent over that period, a drop of 7.5 percentage points (95% CI [5.9, 9.1]) or 42.5 percent (figure 1). Applying the 7.5 percentage-point decrease in the uninsurance rate to the estimated 2015 national population of nonelderly adults yields a net coverage gain of 15.0 million (95% CI [11.9 million, 18.1 million]) adults during this period.7
The uninsurance rate in states that expanded Medicaid by March 2015 fell by more than half (52.5 percent), from 15.8 percent to 7.5 percent. In states that did not expand Medicaid, the uninsurance rate fell from 20.7 percent to 14.4 percent, a decline of 30.6 percent.
There have been large coverage gains for low- and middle-income adults targeted by key ACA provisions.
Low- and middle-income adults targeted by the Medicaid expansion and by the tax credits available for Marketplace coverage (those with family income at or below 138 percent of the federal poverty level [FPL] and those with family income between 139 and 399 percent of FPL, respectively) experienced the largest coverage gains. Coverage for low-income adults increased 16.2 percentage points (95% CI [12.6, 19.8]; figure 2), from 61.0 percent with insurance in September 2013 to 77.3 percent in March 2015 (data not shown). The share of middle-income adults with insurance also increased 7.6 percentage points (95% CI [5.0, 10.2]). The estimated change in coverage for adults with family income at or above 400 percent of FPL, who historically have had very high coverage rates, was less than 1 percentage point and was not statistically significant.
There were gains in coverage for adults in each age, gender, and racial and ethnic group examined, but adults who are young, nonwhite, or Hispanic saw especially large percentage-point gains.
The coverage gains under the ACA have been particularly strong for population subgroups that had high levels of uninsurance before the ACA, including young adults and those from racial and ethnic minority groups. For example, the share of insured adults ages 18 to 30 increased 9.8 percentage points (95% CI [6.5, 13.1]), compared with 6.4 percentage points (95% CI [4.3, 8.6]) for adults ages 31 to 49 and 6.7 percentage points (95% CI [5.0, 8.5]) for adults ages 50 to 64 (figure 2).
Hispanic adults experienced a 14.8 percentage-point increase in coverage (95% CI [10.9, 18.6]), compared with a 9.1 percentage-point increase among nonwhite, non-Hispanic adults (95% CI [6.4, 11.8]) and a 5.0 percentage-point increase among white, non-Hispanic adults (95% CI [3.4, 6.5]). The gap in coverage between white, non-Hispanic adults and nonwhite, non-Hispanic adults narrowed from 7.2 percentage points in September 2013 to 3.1 percentage points in March 2015 (data not shown).
Adults from subgroups with high rates of uninsurance before the ACA also had the largest relative gains in both Medicaid expansion and nonexpansion states.
Within Medicaid expansion states, the share of low-income adults with coverage increased 19.1 percentage points (95% CI [14.0, 24.2]) between September 2013 and March 2015, representing a 53.4 percent decline in the uninsurance rate for adults in the income group targeted by the Medicaid expansion. Over the same period, coverage increased 11.5 percentage points (95% CI [7.7, 15.2]) for adults ages 18 to 30, 9.5 percentage points (95% CI[7.0, 12.0]) for nonwhite, non-Hispanic adults, and 17.1 percentage points (95% CI [13.4, 20.8]) for Hispanic adults in expansion states (figure 3).
Though coverage increases were smaller in nonexpansion states than in expansion states, low-income adults, young adults, and adults from minority backgrounds also experienced large coverage increases in nonexpansion states. For example, the share of low-income adults with coverage rose 12.5 percentage points (95% CI [8.6, 16.3]) in nonexpansion states. About one-third (31.0 percent) of low-income adults remained uninsured in March 2015 in nonexpansion states, compared with about one-sixth (16.7 percent) of low-income adults in expansion states (data not shown).
Finally, men had larger coverage increases in expansion states than in nonexpansion states; coverage increases were similar for women across the two state groups. The larger gains for men found in expansion states likely reflect the greater expansion of Medicaid eligibility for men under the ACA: men were less likely than women to be eligible for Medicaid before the ACA expansion.
What It Means
The share of nonelderly adults who are uninsured continued to decline following the ACA’s second open enrollment period. As of March 2015, the number of uninsured adults was estimated to be 15.0 million lower than the number of uninsured adults in September 2013, a decrease in the uninsurance rate of 42.5 percent. Coverage gains have been especially large among adults in states that expanded Medicaid and among adults targeted by the ACA’s Medicaid expansion and Marketplace subsidies. However, not all of the coverage gains are necessarily the result of these ACA provisions because this analysis does not control for other external factors that affect insurance coverage, such as changes in the business cycle.
In the coming months, there may be additional Marketplace enrollment during special enrollment periods (SEPs) for individuals who have to pay a tax penalty for not having coverage in 2014 but were unaware of the requirement to have coverage and for people with a qualifying life event, such as a move to a new state, marriage, loss of coverage, or birth of a child. A recent study found limited capacity of state-based Marketplaces to conduct outreach for SEPs in 2014 (Wishner et al. 2015). To the extent that Marketplaces work to improve SEP enrollment systems and outreach efforts, SEPs may become a more important source of coverage gains between open enrollment periods (Wishner et al. 2015). At the same time that some may gain coverage under the SEPs, others who signed up for coverage under the most recent open enrollment period may not maintain their coverage over the course of the year.
Our results are broadly consistent with recent Gallup-Healthways Well-Being Index data that tracks changes from the third quarter of 2013 through the first quarter of 2015.8 Though the study time frames differ across the two data sets, the estimated changes in the uninsurance rate between September 2013 and March 2015 based on the HRMS are similar to the estimated changes for roughly the same period using the Gallup data for all nonelderly adults.9 Our results have also tracked closely with early estimates from the National Health Interview Survey through September 2014 (Cohen and Martinez 2015). All three of these surveys suggest there has been a substantial decline in the nation’s uninsurance rate for nonelderly adults since full implementation of the ACA coverage provisions began in 2014.
Though providing evidence of the ACA’s significant success at expanding coverage, these new HRMS findings also highlight that roughly 1 in 10 adults remained uninsured nationwide in March 2015, indicating considerable scope for additional reductions in coverage. This is particularly true in nonexpansion states, where coverage is lagging behind coverage in expansion states. According to HRMS data from March 2015, adults in states that had not expanded Medicaid were 1.9 times as likely to be uninsured as adults in states that had expanded Medicaid (14.4 percent vs. 7.5 percent). Of the 22 states that have yet to expand Medicaid, only a handful appear to be actively debating whether to expand.10 Therefore, it is likely that the coverage gap between expansion states and nonexpansion states found here will persist, at least in the near future.
Office of the Assistant Secretary for Planning and Evaluation. 2015. Health Insurance Marketplace 2015 Open Enrollment Period: March Enrollment Report. Washington, DC: US Department of Health and Human Services.
Centers for Medicare and Medicaid Services. 2015. Medicaid & CHIP: January 2015 Monthly Applications, Eligibility Determinations, and Enrollment Report. Baltimore, MD: US Department of Health and Human Services.
Cohen, Robin A., and Michael E. Martinez. 2015. Health Insurance Coverage: Early Release of Quarterly Estimates from the National Health Interview Survey, January 2010–September 2014. Hyattsville, MD: National Center for Health Statistics.
Karpman, Michael, Sharon K. Long, and Michael Huntress. 2015. Nonfederal Surveys Fill a Gap in Data on ACA. Washington, DC: Urban Institute.
Long, Sharon K., Michael Karpman, Genevieve M. Kenney, Douglas Wissoker, Nathaniel Anderson, and Stephen Zuckerman. 2015. QuickTake: Taking Stock: Health Insurance Coverage under the ACA as of December 2014. Washington, DC: Urban Institute.
Long, Sharon K., Michael Karpman, Adele Shartzer, Douglas Wissoker, Genevieve M. Kenney, Stephen Zuckerman, Nathaniel Anderson, and Katherine Hempstead. 2014. Taking Stock: Health Insurance Coverage under the ACA as of September 2014. Washington, DC: Urban Institute.
Long, Sharon K., Genevieve M. Kenney, Stephen Zuckerman, Dana E. Goin, Douglas Wissoker, Frederic Blavin, Linda J. Blumberg, Lisa Clemans-Cope, John Holahan, and Katherine Hempstead. 2014. The Health Reform Monitoring Survey: Addressing Data Gaps to Provide Timely Insights into the Affordable Care Act. Health Affairs 33 (1): 161–67.
State Health Access Data Assistance Center. 2013. Comparing Federal Government Surveys that Count the Uninsured. Princeton, NJ: Robert Wood Johnson Foundation.
Wishner, Jane B., Sandy Ahn, Kevin Lucia, and Sarah Gadsden. 2015. Special Enrollment Periods in 2014: A Study of Selected States. Washington, DC: Urban Institute.
About the Series
This brief is part of a series drawing on the HRMS, a quarterly survey of the nonelderly population that is exploring the value of cutting-edge Internet-based survey methods to monitor the ACA before data from federal government surveys are available. The briefs provide information on health insurance coverage, access to and use of health care, health care affordability, and self-reported health status, as well as timely data on important implementation issues under the ACA. Funding for the core HRMS is provided by the Robert Wood Johnson Foundation and the Urban Institute.
For more information on the HRMS and for other briefs in this series, visit www.urban.org/hrms.
About the Authors
Sharon K. Long is a senior fellow, Genevieve M. Kenney and Stephen Zuckerman are senior fellows and codirectors, and Michael Karpman, Adele Shartzer, and Nathaniel Anderson are research associates in the Urban Institute’s Health Policy Center. Douglas Wissoker is a senior fellow at the Urban Institute. Katherine Hempstead is a senior program officer at the Robert Wood Johnson Foundation.
1 Administrative data show that nearly 11.7 million individuals were enrolled in Marketplace plans as of February 22, 2015 (Office of the Assistant Secretary for Planning and Evaluation 2015). In addition, as of January 2015, approximately 11.2 million additional individuals were enrolled in Medicaid or CHIP relative to average monthly enrollment in July through September 2013 (Centers for Medicare and Medicaid Services 2015).
3 We establish September 2013 as our baseline period for measuring changes in coverage because survey respondents may have begun reporting coverage gains resulting from ACA implementation after this date even if their coverage did not take effect for several more months. Marketplace coverage for people enrolling between October 2013 and December 2013 did not start until January 2014. Some who signed up in the fall, however, may have reported having coverage during the December 2013 HRMS survey. Further, some of those seeking coverage through the Marketplace between October 2013 and December 2013 were enrolled in Medicaid before January 2014.
4 The second Marketplace open enrollment period began November 15, 2014, and ended February 15, 2015. A special open enrollment period was offered between February 15, 2015, and February 22, 2015, for people who started applications in the federal Marketplace but experienced technical issues or waited in line at the call center. Most state-based Marketplaces offered similar special enrollment periods, with most ending in February (Office of the Assistant Secretary for Planning and Evaluation 2015). In addition, a special enrollment period was extended from March 15, 2015, to April 30, 2015, to individuals in states with a federally facilitated Marketplace if they paid a tax penalty for not having coverage in 2014 and only became aware of the requirement to have coverage after open enrollment ended and in connection with preparing their 2014 tax return (see Centers for Medicare and Medicaid Services, “CMS Announces Special Enrollment Period for Tax Season,” February 20, 2015). Eight state-based Marketplaces also extended special enrollment periods to these individuals (see Manatt Health Solutions, “Tax Season Special Enrollment Periods” February 25, 2015).
5 We focus on changes in insurance coverage because estimates of the level of coverage often vary across survey programs because of differences in the surveys that are unrelated to the ACA (State Health Access Data Assistance Center 2013). For this analysis, we focus on state decisions to expand Medicaid by March 1, 2015. The states that had expanded Medicaid by this date are AZ, AR, CA, CO, CT, DE, DC, HI, IL, IN, IA, KY, MD, MA, MI, MN, NH, NV, NJ, NM, NY, ND, OH, OR, PA, RI, VT, WA, and WV. Several of those states, including CA, CT, DC, and MN, expanded Medicaid under the ACA before 2013.
6 Specifically, we control for the variables used in the poststratification weighting of the KnowledgePanel (the Internet-based survey panel that underlies the HRMS) and the poststratification weighting of the HRMS. These variables are sex, age, race and ethnicity, language, education, marital status, whether any children are present in the household, household income, family income as a percentage of FPL, homeownership status, Internet access, urban or rural status, and census region. In this analysis, we also control for citizenship status and participation in the previous quarter’s survey (i.e., whether the respondent completed the survey in the previous quarter, was sampled in the previous quarter but did not complete the survey, or was not sampled in the previous quarter).
7 To extrapolate our estimates of changes in uninsurance rates to the number of adults who have gained coverage over the same period, we use projections for the size of the 2015 population from the US Census Bureau. These files give population projections by race, ethnicity, and sex of all ages from 2014 to 2060 based on estimated birth rates, death rates, and net migration rates. Using the “Table 1” file (which has a 2015 projected population of 321,368,864), we summed the 2015 population projections for all 18-to-64-year-olds to arrive at 199,903,264 nonelderly adults in 2015. See US Census Bureau, “2014 National Population Projections: Downloadable Files,” US Department of Commerce, last modified December 10, 2014. Because we used population projections for 2014 in earlier briefs in this series, a small portion of the change in the estimated number of adults gaining coverage between those analyses and the current analysis reflects population growth. The estimated coverage gain between September 2013 and March 2015 would be about 108,000 fewer nonelderly adults if we applied the population projection used in Long, Karpman et al. (2014) and about 75,000 fewer if we applied the projection used in Long et al. (2015).
8 Jenna Levy, “In U.S., Uninsured Rate Dips to 11.9% in First Quarter,” Gallup, April 13, 2015; Jenna Levy, “In U.S., Uninsured Rate Sinks to 12.9%,” Gallup, January 7, 2015.
9 In addition to the different survey designs (Karpman, Long, and Huntress 2015), there are several differences between the HRMS and Gallup surveys in their measurement of changes in health insurance coverage. For instance, the HRMS estimate for the third quarter of 2013 is based on data collected in September 2013 and its estimate for the first quarter of 2015 is based on data collected in March 2015. In contrast, the Gallup estimate for the third quarter of 2013 is based on daily surveys conducted in August and September 2013, and the estimate for the first quarter of 2015 is based on surveys conducted between January 2, 2015, and March 31, 2015. This study time frame for the Gallup data may lead to an underestimate of the total decline in uninsurance by the end of the first quarter of 2015 because many surveys were completed before the end of the second Marketplace open enrollment period on February 15, 2015.
10 Kaiser Family Foundation, “Status of State Action on the Medicaid Expansion Decision,” last updated March 6, 2015.