Gains from Choice in Health Insurance

Featured in print Bulletin on Aging & Health

A majority of non-elderly Americans receive health insurance through an employer. Once individuals decide to take up employer-provided health insurance, their choice of plan is restricted to whatever menu is offered by the employer. Often, those choices are quite limited. A recent Kaiser Family Foundation survey found that 80 percent of firms offered a single plan, while only 6 percent of firms offered a choice of three or more plans.

How much would consumers benefit if they had greater choice among insurance plans? This question has special relevance now, in light of the recent debate over health care reform. The proposed changes to the health care system could lead to a sizeable reduction in employer-sponsored health insurance and a rise in insurance purchased on the individual market. Such a move could potentially increase the cost of insurance, since prices on the individual market are typically higher, but also provide individuals with a greater choice among plans.

The value of choice is explored in a recent study by researchers Leemore Dafny, Katherine Ho, and Mauricio Varela , Let Them Have Choice: Gains from Shifting Away from Employer-Sponsored Health Insurance (NBER Working Paper 15687).

In the paper, the authors estimate how much employees would be willing to pay for the right to apply their employer subsidy to the insurance plan of their choice. As an analogy, the authors give the example of an employer who offers employees a heavily-subsidized new vehicle, but only two specific models from which to choose. The question is what share of the subsidy the employee would forego to be able to pick a different car.

To address this question, the authors use a unique proprietary data set of employer plan offerings and employee plan choices for a sample of over 800 large employers. The data cover the period 1998 through 2006 and represent over 10 million employed workers per year.

The authors' analysis proceeds in several steps. First, they estimate a model of employee health plan choice to learn how much employees value different characteristics of the plan, such as the level of co-payments, the type of insurance plan (PPO, HMO, POS), and the specific insurance carrier (e.g. Aetna). Second, they estimate a model that describes the relationship between plan characteristics and premiums - for example, how much more a plan with low co-payments costs relative to a plan offered by the same employer in the same market that has high co-payments. Finally, they use the results of both analyses to predict what plans employees would choose and how much their welfare would increase under scenarios with greater choice.

The authors simulate several different scenarios. First, they keep the total number of plans the same, but in place of the menu actually offered by the firm they use the plans the firm's employees would have liked best out of all those offered in the same market and year ("plan swapping"). Second, they give employees access to their preferred plan within each plan type, that is, access to the best PPO, HMO, and POS plans for them given their characteristics and preferences ("all plan types"). Finally, they give employees access to all plans available in that market and year ("all plans"). The "plan swapping" scenario will generate the smallest welfare gain from switching, while the "all plans" scenario will generate the largest, since it provides employees with the widest range of choices.

Turning to the results, the authors find that the estimated annual gains for the median covered person would be $504, $970, and $2,045 for the plan swapping, all plan types, and all plans scenarios, respectively. The authors conservatively select the plan swapping scenario as their preferred estimate, and find that it generates a welfare gain equivalent to 21 percent of average premiums.

As the authors note, to interpret these results they need to project how premiums might increase when employees have more choices. Such increases might occur, for example, because administrative costs are higher when consumers shop on an individual basis rather than in groups. While estimates of the potential premium increase vary, the authors conclude "the value of choice is rather likely to exceed its cost, as least for most employees and their dependents."

The authors note some caveats to their analysis. They do not model some costs of increased choice such as consumer shopping costs and switching costs. Conversely, their estimates may represent a lower bound on the value of increased choice because they observe only a subset of the plans available in the market. Furthermore, they do not model differences in preferences within employer groups, and doing so would tend to increase the gains from better matching of plans to specific employees. They conclude that "the value of choice is a nontrivial benefit" of a transition away from employer-sponsored insurance, and "may more than offset the higher costs associated with the individual marketplace."