BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Overwhelming Evidence That Obamacare Caused Premiums To Increase Substantially

Following
This article is more than 7 years old.

Two scholars at the renowned Brookings Institution, Loren Adler and Paul Ginsburg, have published an analysis finding that “average premiums in the individual market actually dropped significantly upon implementation of the ACA [Affordable Care Act].” This contrasts with a plethora of evidence, including a rigorous 2014 Brookings study, showing that the ACA significantly increased premiums. In this post, I discuss methodological concerns with the Adler and Ginsburg approach as well as evidence that leads most scholars to reach a very different conclusion.

While I will discuss the relevant evidence of the ACA’s effect on premiums in depth, there are three data points worth emphasizing. First, unlike Adler and Ginsburg's approach, Brookings 2014 study used actual data and found that “enrollment-weighted premiums in the individual health insurance market increased by 24.4 percent beyond what they would have had they simply followed…trends.” Second, S&P Global Institute found that average individual market medical costs increased substantially between 2013 and 2015, up an estimated 69%. Third, 2014 insurer data shows that premiums for individual market Qualified Health Plans (QHPs), ACA-compliant plans certified to be sold on exchanges, were much higher than premiums for individual market non-QHPs, mostly plans in existence before 2014 that did not comply with the ACA. Relative to non-QHPs, insurers collected more than $1,000 per enrollee in higher premiums and more than $2,300 in higher premium revenue per enrollee in 2014 after accounting for large premium subsidy programs for their QHPs.

Adler and Ginsburg do not discuss this previous research, and their analysis also fails to account for other crucial factors. They do not account appropriately for substantial subsidies insurers received to discount individual market ACA plan premiums. They also do not consider the trend in medical claims costs, which is presumably a better measure of the ACA’s effect on the individual market thus far given both the large subsidies and large losses insurers have incurred selling ACA plans. Instead, most of their analysis relies upon crude back-of-the-envelope estimates for the average individual market premium in 2009 as well as well as for an annual growth rate to inflate their 2009 premium estimate.

Finally, it is worth noting that if Adler and Ginsburg were correct in their analysis, we would expect to see very different results than have occurred. If, as they claim, the ACA was delivering better coverage at lower cost, the exchanges would have attracted a wider cross-section of enrollees and more insurers would be looking to enter these markets. Instead we see adverse selection in the individual market, with spiraling premiums, sizeable insurer exits, and enrollees generally attracted to ACA plans only if they are either highly subsidized or relatively old or unhealthy.

Ample Evidence Shows Average Premiums and Spending Exploded After ACA Implementation in 2014

While it is important to look at data for several years after 2013 to assess the impact of the ACA, comparing individual market premiums in 2013 with those in 2014—the year its key changes took effect—provides an approximation of the initial change. The Manhattan Institute compared the average of the five least expensive pre-ACA plans in 2013 with the least expensive plans available on exchanges in 2014. Manhattan’s researchers adjusted the pre-ACA plan premiums upward to account for the population facing surcharges or denied coverage because of a pre-existing condition. Manhattan estimated that the average state individual market premium increased 41% between 2013 and 2014. A county-level analysis suggested that premiums increased by 49%.

The 2014 Brookings study on this same subject by Amanda Kowalski—and unaddressed by Adler and Ginsburg—used actual pre-ACA individual market premium data, finding that “[a]cross all states, from before the reform to the first half of 2014, enrollment-weighted premiums in the individual health insurance market increased by 24.4 percent beyond what they would have had they simply followed state-level seasonally adjusted trends.” According to Kowalski, the Manhattan Institute estimates are higher likely because they were not enrollment-weighted, and individuals in areas with high premiums likely selected cheaper plans.

Economists at the University of Pennsylvania, also using actual pre-ACA individual market data, estimated that the total expected price of individual market coverage (premiums plus out-of-pocket payments) increased by 14% to 28% as a result of the ACA. According to their findings, “the pre-ACA average premium was lower than the lowest silver plan premium.” Penn’s economists also estimated that plans in the individual market before the ACA had similar actuarial value to silver plans, not an average actuarial value that was 10 percentage points less (and 17% less) than assumed by Adler and Ginsburg. Importantly, all of these studies compare gross premiums before and after the ACA without accounting for subsidies that lowered ACA plan premiums.

Perhaps the most striking visual data that suggest Adler and Ginsburg’s conclusions are wrong is an S&P Global Institute analysis of individual market per member per month (PMPM) costs from May. The figure below from the S&P report shows trends in PMPM costs for the individual and employer-provided markets. The ACA is responsible for the huge spike, clearly shown in the figure, in the individual market PMPM costs, which by early 2015 exceeded PMPM costs in the employer markets.

The data shows a huge increase in PMPM costs in the individual market between 2013 and 2015. According to S&P, PMPM costs increased 38% between 2013 and 2014, and another 23% between 2014 and 2015. The two-year increase (69%) is the product of the two single-year increases.

The comparable PMPM cost increase in the employer market, which the ACA affected much less, amounted to about 11%. Assuming an 11% increase would have happened in the individual market absent the ACA, a very rough initial guess would be that the ACA increased individual market PMPM costs by about 58% between 2013 and 2015. As previously noted, premiums did not increase by this full 58% because of insurers’ large losses and the large subsidies they received.

It is worth noting that the individual market includes both ACA-compliant plans as well as non-ACA-compliant plans. If only ACA-compliant plans were included in the post-2013 data, the spike would likely be much larger.

Brookings Comparison Fails to Account for Important Subsidy Program that Reduced ACA Premiums

In their analysis, Adler and Ginsburg fail to account for the ACA’s reinsurance program—a program that allowed insurers to reduce premiums since it compensated them for a large share of the cost of their most expensive enrollees. In an April Mercatus working paper I coauthored, we computed that net reinsurance payments paid by government to insurers selling individual market QHPs totaled about 20.4% of gross premiums. Therefore, regardless of the other problems with Adler and Ginsburg’s methodology, just accounting for reinsurance payments negates their central finding that “2014 premiums in the ACA market were 10-21 percent lower than 2013 individual market premiums.”

In addition to the need to account for government subsidy programs, it is also important to consider that insurers incurred substantial losses in both 2014 and 2015, which I have estimated, after accounting for reinsurance, averaged about $400 per enrollee in 2014 and about $1,000 per enrollee in 2015. The fact that ACA plan premiums have been much too low to cover insurer expenses is another reason that it is problematic to compare face value premiums of ACA plans with pre-ACA individual market plans. 

2014 Insurer Data Shows QHPs Much More Expensive than Other Individual Market Plans

Individual market QHP premiums were much higher than individual market non-QHP premiums in 2014, a data point that is further evidence that the ACA increased premiums. As noted earlier, QHPs are plans that satisfy all the ACA requirements and are certified to be sold on exchanges. Non-QHPs include ACA-compliant plans but mostly consist of plans in effect prior to 2014 that did not conform to the ACA’s rules and regulations. The following table contrasts premium income for individual market QHPs and non-QHPs.

The table shows that even when excluding reinsurance payments and cost sharing reduction (CSR) payments,  average QHP premiums exceeded average non-QHP premiums by more than $1,000 in 2014. CSR payments essentially represent additional government premium payments for insurers to reduce deductibles and other copayment amounts for certain exchange enrollees with lower incomes. When net reinsurance and CSR payments are included, average QHP premium revenue exceeded average non-QHP premium revenue by more than $2,300, or about 76%.

It is necessary to include CSR payments in comparing total effective premiums between pre-ACA individual market plans and ACA plans. Adler and Ginsburg’s decision to not include CSR payments in their analysis is understandable given that they are mainly looking at the price of the second lowest cost silver (SLS) plan.

Brookings Study Compares Lower-Cost ACA Plans to Inflated Estimates of Average Pre-ACA Costs 

In their analysis, Adler and Ginsburg compare fundamentally different measures of premiums by contrasting the cost of the SLS plan available in ACA exchanges to their estimated measure of the average individual market premium that would have existed without the ACA. Importantly, the SLS plan is a lower than average price exchange plan. As shown in the table, the SLS plan premium they use ($3,800) is about 10% less than the average 2014 individual market QHP gross premium and 43% lower than the premium income received by insurers when accounting for reinsurance and CSR payments.

Brookings Failed to Properly Account for CBO’s Premium Projections

Adler and Ginsburg support their claim of the ACA leading to lower premiums by pointing to the fact that the 2014 SLS plan was 15% below CBO’s 2009 estimate of what it would be in 2014. This comparison is problematic and misleading because CBO’s 2009 estimate involved significant and generally unforeseeable errors in key underlying assumptions.

First, starting in the mid-2000s, health care inflation slowed and the extent of this slowdown, largely the result of the recession and weak economic recovery, was not anticipated by CBO in 2009. Adler and Ginsburg cite that the 2014 per enrollee cost of workplace coverage, which was much less affected by the ACA, was 12% lower than CBO expected in 2009. Second, CBO also erred in projecting that the reinsurance program would only reduce premiums by about 10% in 2014; in fact, the per enrollee effect of the reinsurance program was double what CBO expected. Third, as late as February 2014, CBO projected that insurers would be profitable in the aggregate selling ACA plans—meaning they thought premiums would adequately cover expenses. Finally, CBO did not anticipate how narrow ACA plan provider networks would be, and provider network size is positively related to average premiums, all else equal. It’s uncertain what CBO assumed about ACA plan deductibles, but they are higher than many ACA advocates expected, and deductible size is inversely related to premiums, all else equal.

Without adjusting for all of CBO’s inaccurate assumptions—and I have listed at least four major ones—it is misleading to compare CBO’s 2009 projections for SLS premiums with actual SLS premiums. Properly adjusting for all the above mentioned factors would plausibly result in CBO’s 2009 estimate of future SLS effective premiums being too low rather than too high.

Conclusion

Adler and Ginsburg argue that the ACA both improved coverage and lowered costs. If true, the exchanges would undoubtedly be more successful than they are. Thus far, the result is significantly worse than expected, with far fewer enrollees—particularly younger and healthier enrollees—than projected, and substantial market instability. Another Brookings Institution scholar, Stuart Butler, recognizes the problem, writing just two weeks ago that “the ACA might be more appropriately labeled the ‘Medicaid Expansion Act’” because “enrollment in the ACA exchanges has been disappointing, with an estimated 10 million fewer people enrolled compared with earlier expectations.”

The new Brookings study does not mention the numerous studies, including the rigorous 2014 Brookings study, that come to opposite conclusions, does not use actual pre-ACA individual market data, does not consider the huge increase in the cost of medical claims in the individual market after 2013, and makes a number of questionable methodological choices. The authors arrive at a different conclusion than most scholars who have examined the effect of the ACA on health insurance premiums. Most scholars and analysts conclude that, particularly when fully accounting for the various government subsidies for individual insurance coverage, the ACA significantly increased individual market premiums.