In October of 2013 President Obama faced a political firestorm over the cancellation of millions of individual insurance plans — plans not compliant with the requirements of the Affordable Care Act (ACA). The Associated Press (AP) estimated that 4.8 million persons with non-group coverage had their policies cancelled, and this estimate was widely quoted in the media and the Congress. In headline stories, the media also reported that policyholders of the canceled plans were now offered alternative plans, often at premiums more than double of their current plans.

When the controversy over cancelled policies broke, no surveys were available to estimate the number and the cost of cancelled policies. In October, and many state-based marketplace websites were virtually non-functional, so assessing comparative cost and benefits of cancelled and Marketplace plans largely was precluded. In this post I highlight information from subsequent surveys and analyses conducted in late 2013 and 2014 that measure the number of cancelled plans and the comparative cost of coverage in the pre-ACA and post ACA-Markets. The next two paragraphs summarize findings.

Recent survey data indicates the number of persons affected by cancelled policies was about 1.9 million persons, less than the often cited 4.8 million estimate. When persons with group health insurance are included in the denominator, these cancellations affected less than one percent of persons holding comprehensive private insurance. The number of people with non-group policies who became uninsured following last October’s cancellation of policies is similar to what occurs in the normal churn of the non-group market.

Persons eligible for subsidies, 56 percent of the non-group population, could choose a bronze or second-lowest-cost silver plan that cost considerably less than pre-ACA plans. People in the 19-34 age group not receiving subsidies faced higher premiums for the second-lowest silver plan than they would have faced in the pre-ACA market. Premiums for bronze plans were lower on the Marketplace for this group. Persons in  the 35-54 and 55-64 age groups faced lower premiums for bronze and second-lowest silver plans, compared to premiums in the pre-ACA market.

Without a doubt, many people who received cancellation policies had to pay more, and in some cases, far more, for their ACA policy than their pre-ACA plan. But this analysis suggests higher premiums were not an example of central tendency, but a deviation from the mean.


The individual insurance market covers an estimated 10.1 million Americans. Fifty-six percent of these Americans earn less than 400 percent of federal poverty-level income and are thus eligible for subsidies, according to the 2012 Health Interview Survey.

In response to the public outcry, on November 14, 2013, President Obama announced an administrative fix to the cancellation of individual insurance policies. The fix allowed insurers to renew for one year individual and small group policies in effect on October 1. The insurer and broker community designated renewed non-compliant policies “grandmothered policies.”  Grandmothered policies are not permitted to change plan benefit options and cannot be sold to new enrollees with effective dates after December 31, 2013.

State insurance commissioners have the authority to allow or disallow insurers to renew grandmothered plans in the state. Kevin Lucia of Georgetown University reported on the Commonwealth Fund website that commissioners in 22 states declined to authorize such renewals of non-compliant plans. Grandmothering also depends on insurers’ willingness to reissue, re-underwrite, and resubmit to the state insurance department the cancelled policies.

Churning in the Individual Market

Benjamin Sommers published a recent article in Health Affairs that provides context for the cancellation controversy. Using data from the Census Bureau’s Survey of Income and Program Participation (SIPP), Sommers followed 4,199 respondents with individual coverage in the pre-ACA period from May 2008 to April 2010. After six months, 65 percent still held non-group policies, nine percent were uninsured, and 22 percent held employer-based coverage. After one year, only 42 percent of the sample still held individual coverage, and after two years only 27 percent held non-group coverage. Sommers estimates that about 6.2 million persons annually transition out of the non-group market, and thus the disruption of cancelled non-ACA policies is little different from the normal workings of this “bridge” market.

How Many Cancelled Policies?

The Urban Institute and the Commonwealth Fund conducted household surveys in late 2013, and RAND did so in March 2014. Each survey offers a more direct measurement of the number of persons who had their non-group policy cancelled than what was previously available.

Urban Institute. In December 2013, the Urban Institute asked on its Health Reform Monitoring Survey, “Did you receive a notice in the past few months from a health insurance company saying that your policy is cancelled or will no longer be offered at the end of 2013?” The question was asked only of households reporting non-group coverage (n=522). In a Health Affairs Blog post, Lisa Clemans-Cope and Nathaniel Anderson reported that 18 percent of households with non-group coverage indicated their policy was canceled because it did not meet coverage requirements, and another six percent indicated it was canceled for other reasons. Based on an estimate of 14 million Americans with non-group coverage, the authors estimate that 2.6 million Americans had their policy canceled because the plan was not compliant with the ACA.

The Medical Expenditure Panel Survey, Household Component estimates 9.6 million in the individual insurance market and the Health Interview Survey (HIS) figure is 10.1 million in 2012. Hence, I estimate 1.8 million persons lost coverage due to ACA requirements based on the Urban household survey.

Commonwealth Fund. Sara Collins and her colleagues reported similar results from a survey of 622 households published by the Commonwealth Fund in December 2013. Among those with individual coverage (n=156), 22 percent had received a notice of cancellation. Another 60 percent said the plan had given the option of renewing their current coverage. Because of the relatively small sample size, Collins declined to provide a point estimate of the number of persons with cancelled policies.

Pooling the Urban and Commonwealth data and weighting by the sample sizes of each survey, I estimate that 1.9 million persons had their non-group coverage cancelled. Expressed in more technical terms, we can be 95 percent certain that between 1.7 and 2.1 million persons had their non-group coverage cancelled.

RAND. Katherine Carman and Christine Eibner analyzed data from the RAND American Life Panel in September 2013 and Mid-March 2014. A total of 2,425 adults age 18-64 participated in both rounds of the survey. The authors found that less than one million adults who had non-group health insurance in September were uninsured in March 2014. This estimate is little different than the normal churn that Sommers observed in the non-group market in the pre-ACA years, where nine percent of the 10.1 million (909,000) would be uninsured after six months.

Carman and Eibner note that although the surveys do not permit them to estimate the number of persons who are uninsured due to ACA-related cancelations as opposed to inability to pay premiums, the one million figure represents less than one percent of the persons between the ages of 18 and 64. Carman and Eibner also report that of the 3.9 million persons enrolled in the individual Marketplaces at the time of the March survey, only one-third were previously uninsured. The implication is that some of the persons whose policies were terminated purchased coverage in the Marketplaces. Enrollment in Marketplaces doubled to over 8.0 million persons in mid-April, and the late surge of enrollees included a larger share of younger Americans who presumably were uninsured.

Sticker Shock or Bargain Basement?

Two factors should increase premiums for post-ACA plans over pre-ACA ones. First, the ACA requires plans to have an actuarial value of 0.6 or more, and to cover 10 essential benefits. (Actuarial value is the percentage of expenses that the insurer will pay for a large standardized population.)  Based on a study I and colleagues published in Health Affairs in 2012, more than half of pre-ACA non-group plans did not pass the 0.6 threshold.

Secondly, insurers are prohibited under the ACA from excluding and charging higher premiums to applicants based on health status. Health spending is heavily concentrated among a few sick individuals. For example, in a typical employer-based plan, the highest five percent of spenders incur the majority of claims expenses. Including the sick in health plans should increase medical care expenses and thus premiums for non-group coverage.

But there are countervailing forces that should reduce premiums for non-group plans under the ACA. First, the Marketplace structure should reduce administrative expenses. Marketing expenses should be lower and underwriting expenses should fall to zero. Increased enrollment in non-group plans on and off the exchange will give insurers greater purchasing power when negotiating with hospitals and doctors and spread fixed costs over more enrollees.

Secondly, the ACA builds a transparent and price-competitive market structure. Consumers can compare premiums for plans with near-identical actuarial value. Third, insurers realized from their marketing studies that buyers were price-sensitive, so rather than offering provider networks that covered virtually all physicians and hospitals in the area, they offered narrow networks. With narrow networks, insurers have been able to obtain 20 percent discounts from hospitals, and they have sometimes excluded the large and more expensive teaching hospitals.

A few studies have examined premiums for individual policies on Marketplaces. The major challenge in comparing pre- and post-ACA plans is the absence of a large national database with data on both premiums and benefits. In subsequent analysis, I use data from a survey by America’s Health Insurance Plans (AHIP) of member companies in the individual insurance market and trend it forward to estimate 2014 premiums for pre-ACA plans. I trend 2009 premiums at five percent a year, a conservative estimate as will be shown later. Pre and post-ACA comparisons are not “apples-to-apples,” because over one-half of pre-ACA plans did not meet the actuarial value requirement of 0.6, and pre-ACA enrollees had to pass underwriting screens, thereby excluding the sickest persons from purchasing health insurance.

The assumption of five percent trend is based on a set of studies my colleagues and I conducted for the Assistant Secretary of Planning and Evaluation of the Department of Health and Human Services about premium increases in the non-group market. We examined insurers’ rate increases in the non-group market for the years 2008-2012 in 28 states by extracting insurers’ rate filling data from state insurance websites. We found average premium increases of 9.9 percent in 2008, 10.8 percent in 2009, 11.7 percent in 2010, 7.7 percent in 2011, and 7.4 percent in 2012. The NORC/ASPE studies suggest that trending premiums at five percent a year is a plausible, if not conservative assumption.

For premiums sold on the Marketplace for 2014, we examine the work of John Holahan, Linda Blumberg, and Matthew Buettgens of the Urban Institute. Holahan and his colleagues calculated premium costs to enrollees for the lowest-cost bronze plan and the second-lowest-cost silver plan by age and income group for each state. Table 1 shows national averages by income level and age group.

Comparing Premiums for Marketplace Plans with Pre-ACA Plans, by Age Group and Income Level


Sources: J. Holahan, L. Blumberg, M. Buettgens, Will Those With Cancelled Insurance Policies Be Better Off in ACA Marketplaces?” January 24, 2013; America’s Health Insurance Plans, Center for Policy Research, “Individual Insurance 2009: A Comprehensive Survey;” NORC, “Effects of Implementing State Insurance Market Reform,” Final Report to ASPE, 2013. 

On average, individuals earning less than 200 percent of poverty income, regardless of age, can purchase Bronze coverage for $19 a month or less. Earlier we noted that about 56 percent of persons with non-group coverage in 2012 earned less than 400 percent of the poverty level, and therefore would receive subsidies. Individuals age 19-34 whose income exceeds the 400 percent threshold, in contrast, receiving no subsidy, would pay on average $162 per month; the corresponding average for persons age 55-64 is $404 per month.

Columns four and five show premiums from the AHIP survey and for individual plans in 2009, and estimated 2014 figures calculated with a five percent trend each year. For 2014, ages 19-34, estimated 2014 premiums for AHIP plans ($185) are higher than premiums for bronze plans, not only for persons receiving subsidies, but persons not receiving subsidies ($162).

In all three age groups – 19-34, 35-54, and 55-64 – those earning less than 200 percent of the federal poverty income would pay on average less than $100 a month for the second-lowest-cost silver plan. Individuals age 19-34 whose income exceeded 400 percent of poverty and thus received no subsidy would pay on average $219 per month, which exceeds the $185 figure for pre-ACA plans.

Persons ages 35-54 and 55-64 age groups would pay on average $351 and $569 per month for the pre-ACA plans. This figure exceeds premiums with subsidies and without subsidies. It greatly exceeds the amount anyone receiving subsidies would pay for premiums, even for silver plans.

I note two other data sources on the cost of individual coverage — MEPS-HC and a report by the Government Accountability Office (GAO) in January 2014. Unfortunately, neither data set allows a comparison with the Urban Institute’s analysis of the cost of individual coverage on the Marketplaces. MEPS-HC provides data on what policyholders under age 65 paid out-of-pocket for non-group insurance in 2011. Regrettably, the figure also includes some individuals who purchased Medigap insurance for persons under age 65, and some persons whose non-group coverage may have been subsidized by their employer. The average cost of a policy was $258 in 2011 according to MEPS-HC. Trended at five percent a year, I estimate the cost at $300 a month in 2014.

Downloading data from HealthCare.Gov Path Finder, the GAO calculated state medians for base premiums in 2013 for 19 year-olds and 64 year-olds. For non-smoking 19 year-old males, the median premium base rate (before medical underwriting) was $157. The cost for a 64 year-old male who was a non-smoker was $529. Trending these figures by five percent a year yields premiums of $165 and $555 respectively.

Monthly Cost of Individual Insurance Coverage, 2005-2012


Sources: Medical Expenditure Panel Survey – Household Component; NORC, “Effects of Implementing State Insurance Market Reform,” Final Report to ASPE, 2013.

Concluding Remarks

The data referred to in this post does not always meet the standards of large government surveys, such as the Health Interview Survey, but it is better information than was available last October. The shortcoming of large national surveys is their timeliness. It is the role of research to reduce uncertainty. In so doing, it also serves to educate policymakers, and hopefully lead to better decisions.