Nearly all analysts predicted that enrollment in insurance plans offered through the healthcare marketplace would be higher in 2015 than in 2014, though projections varied. Using the latest enrollment data, we evaluate these projections—including our own—and look at what this might mean for the next open enrollment period beginning November 1.

When open enrollment started last year, policymakers and researchers attempted to estimate 2015 marketplace enrollment. The Congressional Budget Office (CBO) projected 13 million enrollees in 2015, the Department of Health and Human Services revised its projection downward to 9 million, and Citi Investment Research projected around 11 million. Using the Urban Institute's Health Insurance Policy Simulation Model, we projected that 10.9 million people would be enrolled by the end of 2015 open enrollment. 

Plan selections compared with actual enrollment

By March 2015, 11.7 million people had selected marketplace plans.  Some people who selected a plan, however, never paid the first month’s premium or failed to submit required documentation.  As a result, the number of plan selections is higher than the number of people who were actually covered by health plans.  So, while 11.7 million people had selected marketplace plans, the number of people actually covered by marketplace plans fell to 9.9 million by June 2015, a 15 percent decrease.  It is this actual enrollment that should be compared with the projections.

About 87 percent of people who selected plans through HealthCare.gov qualified for advanced premium tax credits (APTCs), which reduce premiums and are available to those who earn up to four times the poverty limit but are not eligible for Medicaid or other health coverage. There was much greater attrition among those eligible for APTCs than among those who weren’t eligible.  About 7.7 million people selected marketplace plans using APTCs, but only 6.2 million were covered by those plans in June, a 20 percent decrease.  By comparison, about 1.1 million people selected plans without tax credits through healthcare.gov, and 1.0 million were covered by June, a 10 percent decrease.

There was considerable variation among states in the share of people who selected plans and were still enrolled in June.  Mississippi and Hawaii lost 30 percent of people who selected plans.  Twelve states lost between 20 and 30 percent of plan selections: Arizona, Arkansas, DC, Georgia, Indiana, Louisiana, Minnesota, New Jersey, Ohio, South Carolina, Tennessee, and Texas.

How do the projections compare?

Before open enrollment began, we projected that 2015 marketplace enrollment would be 10.9 million.  Although 2015 plan selections were above that estimate, enrollment fell to 9.9 million by June, about 8 percent lower than our projections and 10 percent higher than HHS’ projections.   

The states that exceeded our projections the most are Florida (43 percent higher), Vermont (37 percent higher), DC (30 percent higher), and Maine (29 percent higher).  The states that fall below our projections the most are Hawaii (69 percent lower), Minnesota (64 percent lower), Iowa (57 percent lower), South Dakota (51 percent lower), and Arkansas (47 percent lower).  These are both diverse groups of states, with no obvious patterns in terms of Medicaid expansion decisions, size, region, and use of healthcare.gov.

There is more data available for states that use healthcare.gov for their marketplace enrollment, so we analyzed those states for a more in-depth review of our projections. Some patterns were notably different.  Among states that expanded Medicaid, enrollment with tax credits was only about 3 percent lower than our projections, while among states that did not, enrollment with tax credits was 34 percent higher than projected. 

We recently released a detailed analysis of marketplace participation rates by state and income level.  The high enrollment in non-expansion states was driven by particularly high enrollment among those with incomes between 100 and 150 percent of the poverty level.  We projected lower take-up because people below 138 percent of the poverty level in states that have not expanded Medicaid qualify for an exemption to the individual mandate penalties. It is possible that many were not aware of that exemption.

There was lower marketplace enrollment among those not eligible for tax credits than we projected.  This likely means that more people chose to enroll in a health plan outside the marketplace than a plan inside it. There are few economic or health care incentives for those not eligible for APTCs to choose a marketplace plan over an outside plan, or vice versa, so that decision was difficult to predict.  Also, these decisions may have been heavily influenced by broker participation, IT system performance, and other factors difficult to model.

What lies ahead for 2016? 

We have just released three reports that examine 2015 marketplace enrollment in more detail and discuss consequences for the 2016 open enrollment period:

• An analysis of the rates at which people eligible for APTCs actually enrolled, by state and income level.
• Factors that contributed to high marketplace enrollment in five states.
• Factors that contributed to low marketplace enrollment in five states.

A recent HHS report predicts marketplace enrollment would likely be between 9.4 million and 11.4 million by the end of 2016.  We have just revised our predictions based on analysis of 2015 enrollment rates.  Assuming a modest improvement in enrollment rates in 2016, we now project 11.2 million people could enroll in the marketplace in 2016.

A spreadsheet that compares 2015 plan selections, effectuated enrollment and projected enrollment for each state is available.