The Census Bureau released a report last week that showed a drop in uninsurance between 2013 and 2014 of 2.8 percentage points — which translates into 8.5 million fewer uninsured people based on information from the American Community Survey (ACS). The same release also showed a comparable decline in the full-year uninsured as measured by the Current Population Survey (CPS).
These new estimates confirm findings that have been reported in other private and public surveys. These other surveys focused on non-elderly adults and showed larger decreases than the Census, because the Census release focused on the entire population. Since the coverage provisions of the Affordable Care Act (ACA) are targeted at lowering uninsurance among non-elderly adults, the Census approach results in a smaller percentage point change. The key point, however, is that a substantial reduction in uninsurance has been confirmed by these two large, highly respected federal surveys.
Underlying Changes In Private And Public Coverage
Much of the decline is likely due to the implementation of key provisions of the ACA, though the improving economy could also have contributed to some degree. Data from the ACS showed only a slight decline between 2010 and 2013 when the economy was recovering, followed by a sharp decline between 2013 and 2014 when the main coverage provisions of the Affordable Care Act took effect. Moreover, the coverage increases that contributed to the reduction in the uninsured were highly consistent with the coverage provisions of the ACA.
According the CPS, the percent of the population with Medicaid coverage at some point over the course of the year increased from 17.5 percent in 2013 to 19.5 percent in 2014. This is consistent with the ACA's Medicaid expansion, although this increase would have been greater had all states expanded Medicaid coverage. There was also an increase in nongroup coverage as a source of insurance, growing from 11.4 percent of the population in 2013 to 14.6 percent in 2014. This is consistent with the provision of income-based tax credits to purchase nongroup coverage through the new health insurance marketplaces. For those under 400 percent of the federal poverty level (FPL), these tax credits are available regardless of whether or not states adopted their own marketplaces.
Insurance Changes Across Demographic And Socioeconomic Groups
Further evidence from the CPS in the Census report indicates that the ACA is indeed contributing to the observed coverage gains, because the specific population subgroups targeted by the ACA provisions experienced particularly large reductions in uninsurance rates. For example, low-income individuals had substantial percentage point drops in the rate of uninsurance. Those below 100 percent FPL saw a drop from 23.5 percent to 19.3 percent, while those between 100 and 199 percent FPL saw a decline from 20.4 percent to 15.1 percent. These are the income ranges targeted by either the Medicaid expansion or by the most generous marketplace tax credits.
Age groups with the highest uninsurance rates historically, those between 19-24 and 25-34, and minorities, particularly blacks and Hispanics, had larger reductions in uninsurance than other subgroups. Similarly, individuals with less than a high school education or who just graduated from high school had the highest uninsured rates in 2013 and 2014, but had larger reductions between years than those who were more highly educated. Finally, part-time workers and non-workers had the highest uninsurance rates in 2013 and saw larger declines in 2014 than full-time workers. This suggests that the coverage provisions of the ACA are serving to narrow longstanding gaps in coverage.
At the same time, the rate of employer-sponsored insurance (ESI) was relatively constant. Some predicted that ESI rates would fall as Medicaid coverage and income-related subsidies became available. In previous work with the Health Reform Monitoring Survey (HRMS), we showed no change in ESI rates. The recent Census report confirms that at least through 2014, ESI rates were holding steady.
State Level Trends
Finally, state-level estimates for non-elderly adults from the ACS also reflect the likely effects of ACA policy changes related to state decisions regarding Medicaid expansion. ACS data confirm the general pattern previously reported by the Urban Institute using data from the HRMS.
As shown in Table 1, the states that expanded Medicaid by January 1, 2014 had lower uninsured rates on average among adults aged 18-64 in 2013 and also experienced larger drops in the uninsured by 2014. These states saw their average uninsured rate fall from 18.5 percent in 2013 to 13.7 percent in 2014, a decline of 4.8 percentage points (the estimates in Table 1 were not included in the Census release but were derived from the American FactFinder). Several of these states had reductions of more than 6.0 percentage points in the uninsured rate among non-elderly adults and Oregon, Kentucky, Washington, and West Virginia had reductions of 7.0 percentage points or more.
In contrast, states that did not expand Medicaid had an average uninsured rate for those ages 18-64 of 22.2 percent in 2013, which fell to 19.0 percent in 2014, a decline of 3.2 percentage points. It is important to note that the non-expansion states still had significant reductions in their uninsured rates. This may be due to the fact that uninsured rates were higher to begin with and that there was heavy enrollment in marketplaces or in Medicaid among those eligible under pre-ACA rules (the woodwork effect) in many of the non-expansion states; of note, the uninsured rate among non-elderly adults in Florida declined by 5 percentage points.
Survey Issues
All in all, the new Census data confirm results reported earlier by other surveys, which is extremely important. The Census surveys have important strengths relative to these other surveys, including a larger sample than the National Health Interview Survey and higher response rates than the private surveys by Gallup and the Urban Institute. They also provide much more reliable state level information than is available from other sources.
There is one methodological issue that is concerning. The Census reports an uninsured rate at the time of the survey from the ACS of 11.7 percent for 2014 and an uninsured rate for the entire calendar year from the CPS of 10.4 percent.
These rates are closer together than one would expect given evidence from prior research. Despite changes in the CPS questionnaire, the CPS may still be overstating the full-year uninsurance rate. In addition, it is not clear why the Census release provides so few point-in-time coverage estimates based on the CPS — with the revamped instrument, one would expect those estimates to be very robust and informative about the types of coverage people have at the time of the survey. Moreover, they provide the most timely look at coverage available from the Census — namely, conveying coverage status just six months ago in early 2015, which would have reflected coverage gains through the second open enrollment period.
What's Next?
Even with this limitation, the results reported by the Census in this release provide compelling evidence that uninsurance is declining for the groups targeted by the ACA. At the same time, an uninsured rate of 10.4 percent whether it is point-in-time or for an entire calendar year is still relatively high and above what was originally projected for the first year of the ACA by the Congressional Budget Office.
While these results are encouraging, they show how much more there is to be done. The uninsured rate should fall further as marketplaces mature, the tax penalty for not having coverage increases, and as more states expand Medicaid. Policies that increase premium and cost sharing subsidies, provide stronger incentives for states to expand Medicaid, and provide additional support for the administration of the program at both the federal and state levels could help assure continued progress.
Table 1: Non-Elderly Adult (18-64) Population Without Health Insurance Coverage By State: 2013 And 2014
Source: U.S. Census Bureau, 2013 and 2014 American Community Survey 1-year estimates from Table S2701 in American Fact Finder. September 2015.
Notes: ¹Medicaid expansion status as of January 2014
*Difference between 2013 and 2014 is significant at the .10 level
^The change in uninsurance rate between 2013 and 2014 in expansion states is significantly different than the change in non-expansion states at the .10 level .
Authors' Note
The authors appreciate the assistance of Patricia Solleveld, Timothy Waidmann, Michael Karpman, Laura Skopec, and Erik Wengle.
from Health Affairs Blog http://ift.tt/1KOChYy
No comments:
Post a Comment