Health insurance competition and consumer welfare

Katherine Ho, Robin Lee

16 September 2016



Four major US health insurers are trying to merge into two. Multiple insurers have announced their withdrawal from the Affordable Care Act health insurance exchanges in several states. Should we be worried? Would fewer insurers necessarily harm consumers?

The answer is not obvious. Because the US health care system is dominated by a small number of large insurance companies operating alongside powerful hospital systems and physician groups, standard economic insights gleaned from well-functioning competitive markets need not apply. While a more concentrated insurance market may allow existing insurers to charge higher premiums or administrative fees, it may also enable them to negotiate better reimbursement rates with medical providers, and pass along savings to consumers. Further complicating matters, lower reimbursements may lead providers to cut back on the quality or availability of care.

These tradeoffs are at the heart of the US Department of Justice’s recent challenges to the proposed Anthem-Cigna and Aetna-Humana mergers. New economic research can help us understand them better. Two recent studies suggest that insurer mergers and exits can harm consumers.  The first, by Dafny et al. (2016), exploits variation in market-level competition induced by United Healthcare’s decision not to participate in any of the federally facilitated marketplaces during the first year of open enrolment. The authors conclude that premiums would have fallen significantly – by approximately 5% on average – if United had instead decided to participate. Consistent with this, Dickstein et al. (2015) find that an increase in the number of active insurers, generated when states combined small counties with neighbouring urban areas into a single market, led to $200-300 savings in annual premiums.

However, Scheffler et al. (2016) paint a more complicated picture. They argue that reduced insurance competition, when paired with appropriate and thoughtful legislative oversight (e.g. by having exchanges negotiate premiums with plans, and standardising plan characteristics to make it easier for consumers to compare them), can limit premium increases. The authors investigate insurer premium growth on state exchanges in two US states. In New York, which did little to limit the number of plans or standardise coverage, areas with more insurers had lower premium growth than those with fewer insurers. In California, the state exchange did more to constrain the characteristics and number of offered plans, contracted selectively with insurers, and actively negotiated plan premiums. Here, the authors find that areas with fewer insurers had lower premium growth.

Recent research

We investigate these offsetting effects in more detail in a recent paper (Ho and Lee 2016), and also study the impact of insurer competition on negotiated hospital reimbursement rates. Rather than basing our analysis on a single policy change or difference that is observed in the data, we develop a model of how insurers set premiums and negotiate with health providers to help us understand the relationship between insurance competition and health care spending across a variety of settings. We focus on employer-sponsored insurance, a market that shares many similarities with state insurance exchanges, and use detailed claims, enrolment, and plan data from a large benefits manager in California to ensure that our model matches a well-defined empirical setting.

Using our model, we examine the effect of removing one out of three insurance plans from a benefit manager’s choice set, thereby reducing insurance competition for consumers. We find that if insurers can freely set premiums, then reduced insurer competition leads to a substantial premium increase for the remaining insurers. This is consistent with arguments that a more concentrated insurer market tends to harm consumers. However, if the manager can constrain premium increases by effectively bargaining with insurers, then the removal of an insurer can lead to a premium decrease.

For example, one of our simulations indicates that the removal of Anthem Blue Cross of California (a relatively small insurer by market size in our population) leads to a significant premium reduction for both remaining insurers when the benefits manager negotiates premiums. Yet even in this case, we find that consumers are harmed on average due to the removal of an insurance option.

Partly driving these predictions are the effects of changes in insurer competition on negotiated hospital reimbursement rates. These effects can differ across hospitals. When an insurance plan is removed, the remaining insurers have more leverage to negotiate better reimbursement rates with providers. However, because premiums can rise, it’s also possible that overall reimbursement rates can increase as some providers can extract part of these greater profits.  Overall, we predict that even if premiums increase, some markets are likely to experience average hospital rate reductions.

However, we caution that our analysis holds fixed other characteristics of insurance plans, such as the breadth of their hospital networks, which may also respond to these interventions. We also do not explicitly consider how hospitals respond to changes in the insurance market, and we acknowledge that adjustments in their quality, accessibility, and investment need to be considered. We are investigating these issues in ongoing related work.

Overall, our findings suggest that the relationship between insurer competition and health care spending depends on institutional and market structure.  If premiums can be constrained, either through effective regulation (such as medical loss ratio requirements) or by strong employers and exchanges ‘playing off’ insurers against one another, then reduced insurer competition might lead to lower negotiated reimbursement rates, lower premiums, and lower spending. Absent such constraints, these savings may not be realised and consumers will likely be harmed. This second case may be the situation in the local commercial markets identified by the US Department of Justice as areas of concern in its Anthem-Cigna complaint.

We conclude by stressing that although effective premium setting constraints – imposed by large employers in the commercial market or by state exchanges – may mitigate the extent to which a reduction in insurer competition harms consumers, they also incur risks. Profit-maximising firms may adapt to such constraints by adjusting more than just their prices (e.g. by reducing investment or quality), and this may cause additional harm. Ultimately, regulators and policymakers would be wise to devise policies that preserve their flexibility to adapt to potential issues as they arise.


Dafny, L, J Gruber, and C Ody (2015), “More Insurers Lower Premiums”, American Journal of Health Economics 1(1), 53-81.

Dickstein, M, M Duggan, J Orsini, and P Tebaldi (2016), “The Impact of Market Size and Composition on Health Insurance Premiums: Evidence from the First Year of the ACA”, American Economic Review: Papers and Proceedings 105(5), 120-125.

Ho, K, and R S Lee (2016), “Insurer Competition in Health Care Markets”, NBER Working Paper 19401.

Scheffler, R, D Arnold, B Fulton, and S Glied (2016), “Differing Impacts of Market Concentration on Affordable Care Act Marketplace Premiums”, Health Affairs 35(5), 880-888.



Topics:  Competition policy Health economics

Tags:  health, insurance, health insurance, competition, US, welfare, efficiency

Associate Professor of Economics, Columbia University

Assistant Professor of Economics, Harvard University