What Determines Movement Across Health Care Plans?
Most non-elderly Americans who have health insurance receive their coverage through an employer. Many workers are offered a choice among several employer-sponsored health insurance plans. These plans can be characterized as more or less generous, where the more generous plans offer greater freedom in selecting providers, and/or more complete coverage of health care expenditures. Such plans, however, have higher premiums.
What factors determine a worker's choice of a particular plan, either initially, or to remain or change plans if previously enrolled? Those decisions and the resulting dynamic features of insurance plan populations are the subject of a new conceptual and empirical study by researchers David Cutler, Bryan Lincoln, and Richard Zeckhauser, Selection Stories: Understanding Movements across Health Plans (NBER Working Paper 15164).
The authors begin with a theoretical discussion of the factors affecting workers' choice of plan. Traditional economic theory suggests that workers consider price, including both premiums and out-of-pocket costs, as well as future expected spending when choosing a health insurance plan. Expected spending is related to both age and current health status, and can differ dramatically across individuals. Average health care costs rise rapidly with age once workers reach their 40s. The distribution of health care costs at any age is highly skewed; thus, a relatively small share of the insured population accounts for a large fraction of total expenditures. The premiums that workers pay for employer-sponsored health insurance typically are not related to their own expected spending.
This choice setting is likely to generate adverse selection, the tendency of those with higher expected spending to select the more generous plan. Such choices render the more generous plan a poor deal financially for healthier workers with lower expected spending, since the plan will have to charge high premiums to cover the costs of the less healthy workers it has attracted. Thus, too few of the healthier workers will enroll in the more generous plan relative to what their risk preferences alone would dictate. In some cases the more generous plan may be forced out of business through a phenomenon known as a "death spiral."
The authors identify other factors that may affect plan choice as well. The costs of switching from one health insurance plan to another are potentially important. Suppose that workers with higher spending are less likely to switch plans, for example because they are concerned about changing doctors in the middle of treatment, or because there is transaction cost or insecurity associated with transferring medical records to a new doctor. In this case, adverse retention will result, meaning that sicker workers would be less likely to move, regardless of the generosity of their current plan.
If switching costs are sufficiently high, workers will rarely or never switch plans. The authors call this phenomenon aging in place. If the initial demographics of the two plans differ substantially, aging in place could result in an increasing premium differential over time, due to the rapid rise in health care costs after age 40.
Having established that these three phenomena-adverse selection, adverse retention, and aging in place - have the potential to affect switching and staying behavior among health plans, the authors turn to an empirical examination of their relevance. For this, they use data on all medical claims during 1994 through 2004 for Massachusetts state employees and their families. In this context, the "more generous plan" was a traditional fee-for-service (FFS) plan and the "less generous plan" was an amalgam of a number of different managed care plans, primarily offered by Health Maintenance Organizations (HMOs).
The authors find clear evidence of adverse selection. For people initially in the HMO, the probability of switching to the FFS plan increases with their health expenditures. The reverse is true for people initially in the FFS plan-their probably of switching to the HMO decreases with health expenditures. This result holds whether the authors use expenditures in the year before switching, which would be consistent with "backwards-looking" adverse selection, or expenditures in the year after switching, which is consistent with "forward-looking" adverse selection.
The authors also find that demographics are important predictors of plan mobility. For example, families with older members are much more likely to switch to the FFS plan than are other families.
The authors show that the probability of switching from one HMO plan to another decreases when health expenditures are higher, which is consistent with adverse retention. But when the authors include all factors in a single analysis, demographics and adverse selection appear to be more important than adverse retention.
Finally, the authors create a simulation model to show how these different factors affect plan enrollments. They find that the equilibrium share in the FFS plan is 30 percent and that adverse selection is responsible for about one-sixth of this enrollment. However, the level of subsidy to the FFS plan is quantitatively more important. If the employee's share of the premium differential between the FFS plan and the HMOs rises from 15 to 50 percent, the share of employees enrolled in the FFS plan would fall to just 13 percent.
In sum, the authors find that adverse selection is more important than adverse retention in explaining insurance plan dynamics for this pool. However, these effects are modest relative to the impact of changing the mix of employer and employee premiums. If the relative price of a FFS plan were increased considerably, the authors suggest that it is "entirely possible that an adverse selection death spiral would set in, and the generous FFS plan would ultimately no longer be available."
Pointing to the importance of demographics such as age and sex in explaining both spending differentials and plan mobility decisions, the authors end on "a note of optimism about the ability to have a competitive choice process for health insurance," since insurers can easily observe these characteristics and price insurance plans accordingly.
The authors gratefully acknowledge financial support from the National Institute on Aging (P01 AG005842).