Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, ECONOMICS AND FINANCE ( (c) Oxford University Press USA, 2019. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 14 December 2019

Health Insurance Plan Choice and Switching

Summary and Keywords

Choice of health insurance plans has become a key element of many healthcare systems around the world along with a general expansion of patient choice under the label of “Consumer-Directed Healthcare.” Allowing consumers to choose their insurance plan was commonly associated with the aim of enhancing competition between insurers and thus to contribute to the efficient delivery of healthcare. However, the evidence is accruing that consumers have difficulties in making health insurance decisions in their best interest. For example, many consumers choose plans with which they spend more in terms of premiums and out-of-pocket costs than in other available options. This has consequences for the individual consumer’s budget as well as for the functioning of the insurance market.

The literature puts forward several possible reasons for consumers’ difficulties in making health insurance choices in their best interest. First, consumers may not have a sufficient level of knowledge of insurance products; for example, they might not understand insurance terminology. Second, the environment or architecture in which consumers make their decision may be too complicated. Health insurance products vary in a large number of features that consumers have to evaluate when comparing options, introducing search or hassle costs. Third, consumers may be prone to psychological biases and employ decision-making heuristics that impede good choices. For example, they might choose the plan with the cheapest premium, ignoring other important plan features that determine total cost, such as copayments. There is also evidence that consumer education programs, simplification of the choice environment, or introducing nudges such as setting smart defaults facilitate consumer decision making.

Despite recent progress in our understanding of consumer choices in health insurance markets, important challenges remain. Evidence-based healthcare policy should be based on an evaluation of whether different interventions aimed at facilitating consumer choices result in welfare improvements. Ultimately, this requires measuring consumer utility, an issue that is vividly debated in the literature. Furthermore, welfare calculations necessitate an understanding of how interventions will affect the supply of health insurance, including supply reactions to changes in demand. This depends on the specific regulatory setting and characteristics of the specific market.

Keywords: health insurance, consumer choice, consumer-directed healthcare, choice architecture, health economics

Choice in Health Insurance Markets

Consumer choice of health insurance plans is a key feature of many healthcare systems around the world. In these systems, health insurers reimburse healthcare providers for their services—in contrast to so-called single-payer systems, such as the English National Health Service or the Canadian healthcare system, in which the state pays healthcare providers. To introduce competition between insurers and thus ultimately contribute to the efficient delivery of healthcare and improve welfare, consumers are given a choice between insurance plans. As both consumers’ needs as well as the supply of health insurance plans may change over time, consumers are granted the option to switch between insurance plans repeatedly, for instance, during an annual open enrollment period. Theoretical studies of the properties of such private, competitive markets typically assume that consumers choose and switch plans rationally. Whether they indeed do is an empirical question of great relevance for healthcare policy.

Interest in consumers’ ability to make health insurance plan choices in their best interest has arisen from an introduction and increase of consumer choice of health plan in many health insurance markets. The increased importance of health plan choice is part of a general expansion of patient choice in healthcare systems around the world under the name “Consumer-Directed Healthcare” (CDHC). CDHC refers to a variety of schemes in which consumers manage and pay for their own care, and have access to the information on prices and quality needed to make informed choices among providers and therapies. Patient choice is introduced to keep health costs in check while satisfying consumers’ healthcare needs (McFadden, Winter, & Heiss, 2008).

Choice in health insurance has consequences for the functioning and design of health insurance markets in a variety of dimensions. McGuire (2011) discusses situations in which restricting consumer choice may be welfare improving, for example, due to adverse selection and transaction costs. He also discusses the role of employer-sponsored health insurance, which is widespread in the United States. In this case, the menu of available plans that consumers choose from is preselected by the employer. In all of these situations, the question arises of how good consumers are in making their choices. Liebman and Zeckhauser (2008) lay out the problem in the wider context of research on behavioral economics, and at the close of their paper they conclude that “health insurance is too complicated a product for most consumers to purchase intelligently” (p. 26). The recent availability of detailed data on individual-level demand for insurance plans in a variety of markets has generated better knowledge of the reasons of imperfect insurance choices. In this article, we describe these reasons in more detail. We further discuss possible interventions targeted at improving consumers’ ability to choose health insurance plans, as well as their consequences for the functioning of insurance markets.

Choice of health insurance plan was introduced and increased in the different health insurance systems in the United States, as well as in European Social Health Insurance Systems, although to different degrees. In the United States, an increase in choice was observed as new forms of health plans; for example, managed-care options with limited provider choice but lower copayments entered the market over time. As U.S. employers started to take up these options in order to contain healthcare costs, the menu of choices that employees faced in employer-sponsored health insurance expanded. While only 18% of employees in the United States had choice of more than one health plan in 1977, this number increased to 66% by 1998 (Gabel, 1999). However, as employers preselect plans for their employees, choice remains limited for many in the U.S. employer-sponsored insurance market (McGuire, 2011). The public insurance for the elderly in the United States, Medicare, started to introduce choice of managed care options in the 1970s. While at first only a few demonstration plans were offered, since 1985 all Medicare beneficiaries have been offered a choice of either the traditional Medicare system, or a range of privately-offered managed-care plans in a program called a variety of names over time from Medicare+Choice, Medicare Part C, or Medicare Advantage (McGuire, Newhouse, & Sinaiko, 2011 provide an overview on this system’s history). Since 2006, Medicare beneficiaries additionally face the choice of privately-offered prescription drug insurance under Medicare Part D, triggering a wave of research on consumers’ ability to make choices in their best interest (e.g., Abaluck & Gruber, 2011; Heiss, Leive, McFadden, & Winter, 2013; Heiss, McFadden, & Winter, 2006; Ketcham, Lucarelli, & Powers, 2015; Zhou & Zhang, 2012).

Additional interest in health insurance choice in the United States arose with the introduction of the “health insurance exchanges” legislated by the Affordable Care Act of 2010. In these online marketplaces, consumers can shop for standardized insurance plans. Importantly, the choices that consumers face in these settings are often complex as consumers have to choose between a large number of options that often vary in multiple dimensions, such as deductibles, copayments, premiums, limits on out-of-pocket payments, and network of healthcare providers or prescription drugs that they cover. (For a recent overview on choice in U.S. health insurance, see also Gruber, 2017.)

Many European Health Insurance Systems introduced choice of health insurance in the 1990s. As further summarized in Laske-Aldershof, Schut, Beck, Shmueli, and Van de Voorde (2004), Belgium, Germany, Israel, the Netherlands, and Switzerland introduced choice during that period. In most of these countries, choice was not only introduced to increase efficiency but also because of equity concerns, as the old systems without choice led to vastly different premiums across consumers. In the case of Germany, for example, most individuals were assigned to insurers depending on their employer. As there was no risk-equalization scheme across insurers, premiums varied greatly depending on each profession’s risk pool while the benefits were the same. This was perceived as unfair (McGuire & Bauhoff, 2007). Compared to the dimensions of choice that consumers confront in the different U.S. insurance systems, however, the degree and complexity of choice is typically much more limited in the European systems. In the German Social Health Insurance that covers 90% of the German population, for example, consumers choose among a large number of so-called sickness funds—not for profit insurers—that offer highly regulated benefits and differ mainly in premiums. Health plan features such as deductibles, copayments, or provider networks do not vary across plans in this system.

Before turning to consumer decisions in health insurance, we briefly point out that behavioral economists have documented imperfect consumer choices in many other markets, ranging from electricity markets (Hortacsu, Madanizadeh, & Puller, 2017) to take-up of social support or tax credit (Bhargava & Manoli, 2015). A particularly close link exists with the streams of literature studying choices with high stakes and under uncertainty, including other patient decisions in the healthcare setting (McFadden et al., 2008), investments in pension plans (Choi, Laibson, & Madrian, 2010; Iyengar & Kamenica, 2010; Madrian & Shea, 2001), the take-up of credit (Agarwal & Mazumder, 2013; Woodward & Hall, 2012), or other insurance products such as automobile liability insurance (Kiss, 2016). An important theme that connects the different strands of literature is the incorporation of insights from psychology and behavioral economics into understanding individual decision making and decision-making errors. As discussed by Chetty (2015), these strands of literature do not typically focus on testing neoclassical models or their assumptions and contrasting them with behavioral economics. Instead, they view the incorporation of insights from psychology as a possibility to improve the understanding of individual decision making, leading to new suggestions on the design of the choice environments to nudge consumers toward making decisions in their best interest.

In the following, we summarize in the section “Evidence on Quality of Health Plan Choice, Reasons, and Interventions” the empirical evidence on the quality of plan choice, possible reasons for imperfect choices, as well as evidence of interventions and how these interventions might impact the functioning of the insurance market, mainly focusing on studies analyzing observational data on actual insurance choices. In the section “Controlled Laboratory Experiments on Complexity of Health Insurance Choice,” we discuss additional evidence on health insurance plan choices stemming from laboratory experiments. In the final section, “Controversies, Debates, and Open Questions,” we discuss key debates and open questions in this field.

Evidence on Quality of Health Plan Choice, Reasons, and Interventions

Consumer Ability to Make Health Plan Choices in Their Best Interest

The evidence is accruing that consumers have difficulties when making health insurance plan choices. A large strand of the literature focuses on consumer choice in Medicare Part D, which offers prescription drug insurance in U.S. Medicare. This program, introduced in 2006, allows beneficiaries a choice between roughly 50 different insurance plans offered by private insurance firms to cover their prescription drugs. Heiss et al. (2006) document that roughly 80% of seniors faced with a choice of a health plan in Part D decided to enroll, which is in line with the authors’ calculation according to which most beneficiaries profit from enrolling (Winter et al., 2006). However, when it comes to evaluating the exact plan that consumers have chosen, the evidence is less favorable: according to Zhou and Zhang (2012) only 5% of consumers chose the cheapest plan given their medication needs; more than a fifth of beneficiaries spent at least USD500 more than they would have spent with the cheapest plan. Similarly, Heiss et al. (2013) calculate that only 25% of seniors enrolled in the plan that minimizes the consumer’s cost in terms of premiums and out-of-pocket payments from an ex ante perspective. On average, consumers could have saved around USD300 had they chosen this cost-minimizing plan. But not only elderly consumers in a choice environment as complex as Medicare Part D, but also younger consumers and consumers in less complex choice environments make choices that are likely not in their best interest: Sinaiko and Hirth (2011) document that one third of employees in the setting of a U.S. employer-based insurance chose a health insurance plan that was dominated by an unambiguously better option. Similarly, Bhargava, Loewenstein, and Sydnor (2017) document that the large majority of employees of a U.S. firm chose dominated insurance plans, in which they spent 24% more than their chosen plans’ premiums, while Abaluck and Gruber (2016a) document that typical school-district employees in Oregon pay over USD600 more for healthcare and insurance because they do not choose the cheapest available option. Furthermore, even in the relatively simple European systems, consumers do not seem to make health insurance choices in their best interest: consumers in the German Social Health Insurance, for example, where options vary mainly in premiums, do not choose the cheapest option available to them, indicating that consumers leave money on the table (Schmitz & Ziebarth, 2017; Wuppermann, Bauhoff, & Grabka, 2014).

Determinants of Consumer Choice

Consumers may not make plan choices in their best interest for several reasons. First, in particular settings like the United States where many consumers face choices between plans that vary in multiple features such as premiums, deductibles, copayments, and covered benefits, consumers have to be able to understand these features in order to accurately evaluate and compare plan value. Many consumers, however, lack an understanding of even basic cost-sharing features like deductibles (see, e.g., Barcellos et al., 2014; Handel & Kolstad, 2015a; Loewenstein et al., 2013). Hoerl et al. (2017) show that this lack of understanding predicts failure to take up insurance among the uninsured in the United States after the introduction of new insurance options through the Affordable Care Act, also known as Obamacare, in 2014. Furthermore, Bhargava et al. (2017) demonstrate that it can explain consumers’ choices of dominated insurance plans.

Second, even if consumers did understand plan features, consumers may not choose the best available plan either initially or when faced with a decision to switch plans due to search or hassle costs that arise when gathering information, investigating alternatives, or filling out paper work (e.g., Handel, 2013; Heiss, McFadden, Winter, Wuppermann, & Bo, 2016; Maestas, Schroeder, & Goldman, 2009). However, even if information on plans is readily available at relatively low cost, as it is the case through internet comparison websites in the German Social Health Insurance, in the U.S. Medicare Part D program, and the U.S. Health Insurance Exchanges for example, consumers may fail to take this information into account. In the context of Medicare Part D, Kling, Mullainathan, Shafir, Vermeulen, and Wrobel (2012) sent readily available personalized cost information to a treatment group of beneficiaries who subsequently were more likely to switch their prescription drug plan and save money from switching compared to a control group who did not receive this information. The authors present further suggestive evidence indicating that these “comparison frictions” arise as consumers underestimate their potential savings from switching plans. Furthermore, Heiss et al. (2016) investigate the role of switching or hassle costs in Medicare Part D. More specifically, they separate switching costs from inattention as reasons for inertia, defined as the failure of beneficiaries to switch plans although a better option is available. Using a two-stage dynamic discrete choice model, they find that although switching costs do play a role, a large part of inertia likely stems from inattention to plan choice.

The lack of attention to choice as well as consumers’ tendency to ignore information on options that is readily available at low cost may be a result of cognitive biases. These biases more generally constitute a third factor impeding consumers’ ability to make health plan choices in their best interest. Ericson and Starc (2012), for example, present evidence that consumers in the Massachusetts Health Insurance Exchange make choices in line with a simple “choose the cheapest plan” heuristic. Similarly, Abaluck and Gruber (2011) find that seniors in Medicare Part D put too much weight on plan premiums compared to other cost-relevant plan features, indicating that consumers are not fully rational in making their choices. Additional evidence on the importance of heuristics in health plan choice comes from a laboratory experiment in which Besedeš, Deck, Sarangi, and Shor (2012) find that their subjects make decisions using the tallying heuristic, according to which individuals weigh all attributes equally and thus choose the option with the largest number of attributes, irrespective of each attribute’s potential future usefulness. Besedeš et al. (2012) additionally find evidence for choice overload, a behavioral bias according to which individuals make worse choices or even abstain from making a choice when the number of options they face increases. Frank and Lamiraud (2009) and Wuppermann et al. (2014) present further evidence of choice overload in the context of the Swiss and German health insurance systems. Ketcham et al. (2015) and Bhargava et al. (2017), however, do not find evidence that consumer choice worsens with the number of available options in the U.S. context of Medicare Part D, the prescription drug insurance for the elderly, and an employer-based insurance in the United States, respectively. In addition to choice overload, health insurance plan choices may be subject to status quo or confirmation bias (a tendency to stick to one’s choices or opinions), which may at least partially explain the observed stickiness of choices over time (e.g., Afendulis, Sinaiko, & Frank, 2015).

Interventions to Facilitate Consumer Choice and Market Implication

Given that consumers have difficulties in making health plan choices in their best interest, the questions arise how to facilitate consumer choice and how interventions targeted to simplify choices will affect the functioning of the market. Interventions that are discussed include consumer education or information interventions and decision aids; simplification of the choice architecture, that is, the environment in which consumers make their health plan choices, including a reduction of the number of choice options; and the use of specific (smart) defaults.

Given that many consumers lack an understanding of health insurance terms and thus the competence needed to make diligent decisions concerning health insurance (Barcellos et al., 2014; Loewenstein et al., 2013), educating consumers could improve decision making. Furthermore, consumer choice could benefit from decision support, for example, through decision-making aids or consumer assistance programs. Similar programs exist, for example, in the U.S. Health Insurance Exchanges legislated under the Affordable Care Act (Obamacare). First evidence suggest that consumer education programs and the Affordable Care Act’s certified assister programs can make a difference to insurance enrollment (Bartholomae, Russell, Braun, & McCoy, 2016; Grob & Schlesinger, 2015). Abaluck and Gruber (2016a), however, analyzed data from a randomized trial of decision support tools among school-district employees in Oregon and found that they did not improve consumer choice, largely because consumers ignored the given recommendations. In addition, there is evidence on the effectiveness of information interventions that remind consumers of the possibility to choose a health insurance plan and provide them with personalized cost information.

Field experiments on information interventions delivered mixed results. While Kling et al. (2012) present evidence that a consumer-information campaign in Medicare Part D increased switching rates and led to significant savings among beneficiaries, Ericson, Kingsdale, Layton, and Sacarny (2017) find no evidence for increased plan switching among consumers who received personalized cost information in a field experiment conducted in the setting of the Affordable Care Act’s health insurance exchange in the U.S. state of Colorado. These mixed results may reflect differences in the information that consumers received in the two studies, as the first provided a suggested lowest cost alternative while the second did not. They may also be related to the higher number of choices that consumers face in the second study, indicating a likely interaction between the design of the intervention and the specific choice environment.

Simplifying how information on the available plans is presented to consumers might also improve choices. In the case of Medicare Part D in the United States, the Centers for Medicare and Medicaid Services (CMS) already provides an internet-based decision support tool (“Plan Finder”) that directs consumers to the lowest cost plan upon entering their current prescription drug needs, but poor choice outcomes in this market, reviewed in section “Consumer Ability to Make Health Plan Choices in Their Best Interest,” suggest that this tool has limited effect. McGarry, Maestas, and Grabowski (2018) argue that the tool is too complex and difficult to interpret. They conducted a randomized experiment with hypothetical Part D plan choices to test the effect of simplifying the financial information provided by Plan Finder. They found that reducing the amount of financial information indeed helped consumers pick low-cost plans; importantly, this effect was not associated with a decrease in average plan quality. McGarry et al. argue that modifications to the existing Part D Plan Finder design have the potential to improve beneficiaries’ plan choices in this specific market.

Proposed changes in the choice architecture to help consumers make health insurance plan choices in their best interest include restricting the number of available plans, standardization of plan features across options, and changes in framing of these features. The evidence on the impact of restricting the number of available option is mixed. Results in the European contexts of Switzerland (Frank & Lamiraud, 2009) and Germany (Wuppermann et al., 2014) as well as from laboratory experiments (Besedeš et al., 2012; Schram & Sonnemans, 2011) suggest that the number of options does influence consumers’ ability to make choices. Similarly, Abaluck and Gruber (2016a) show that restrictions in the number of health plan choices to school-district employees in Oregon have led to improved consumer choice. This effect, however, does not result from improvements in consumer choice per se but largely results from changes in the quality of the available plans. As larger choice sets in their setting include worse plans, the outcomes of consumer choice look worse with the larger choice sets. This result is largely in line with Ketcham et al. (2015) and Bhargava et al. (2017) who indicate that choice overload resulting from large number of options cannot explain consumers’ difficulty in making health insurance plan choices. Concerning the standardization of plan features, Ericson and Starc (2016) present evidence according to which limiting the number of combinations of features has helped consumers make better choices and counteracted effects of a simultaneous increase of the number of available plans. The authors conclude that standardization across plans may be a way to improve consumer choice without limiting the number of available options. Wuppermann et al. (2014) and Schmitz and Ziebarth (2017) provide additional evidence underlining the importance of the way choices are presented to consumers. The authors find that the way premiums are framed has a large effect on consumers’ decision to switch health plans in the setting of the German Social Health Insurance.

Setting specific defaults is a third way to influence consumer choice. Instead of designing health insurance as an opt-in, it could be designed as opt-out. This is, for example, the case for the roughly 40% of U.S. Medicare Part D enrollees who are identified as being eligible for the program’s Low Income Subsidy (LIS). As policymakers were concerned that LIS-eligible individuals could fail to enroll in Medicare Part D and thus forgo prescription drug coverage, individuals identified as being LIS eligible are assigned to a randomly chosen plan if they do not make an active choice themselves. While this opt-out default ensures that even inactive LIS-eligible beneficiaries are covered by a prescription drug plan, it does not ensure that beneficiaries are covered by the plan that best meets their needs. Zhang, Zhou, and Baik (2014) estimate that designing a smart default which assigns beneficiaries to plans depending on their previous prescription drug use would be superior to random assignment. In addition to the question of initial take-up of insurance and choice, a large concern in the literature is that consumers do not switch plans over time although better options become available (e.g., Heiss et al., 2016). As Heiss et al. (2016) indicate, a large part of this observed consumer inertia can be attributed to inattention. One way to force consumer attention to plan choice could be to remove the default of staying in the previously chosen plan and thus force consumers to make a choice. Research that compares choices of consumers who are forced to choose, for example, because their previously chosen plan leaves the market, and those of consumers for whom the default is to stay with their current plan, however, indicates that just removing the default is unlikely to improve consumer choices to a large extent (Abaluck & Gruber, 2016a). Again, designing smart defaults instead of automatically re-enrolling consumers to the plan chosen in the previous year has likely a larger potential for improving consumer choices. Instead of reassignment to the previously chosen plan, individuals could, for example, be reassigned to the cheapest plan in their choice set given their healthcare needs (Handel & Kolstad, 2015a). Similar smart defaults are already employed in a public health insurance program in California (Cal MediConnect), which is offered for Medicare beneficiaries who are also eligible for Medicaid, the U.S. insurance program for the poor (Sinaiko & Zeckhauser, 2016). Although smart defaults theoretically have the potential to overcome the worsening of healthcare choices over time due to inertia, their impact on choice or health outcomes has not been scientifically evaluated up to date.

An important question that arises concerning the different strategies to improving consumer choice is how they will affect welfare. First, depending on the strategy used and the specific setting, not all consumers may benefit equally from it. Results by Heiss et al. (2016), for example, indicate that forcing all consumers to make a choice each year in order to overcome consumer inattention may lead to an increase in overall costs for some consumers, most likely those who have difficulties in making choices in their best interest and were—for whatever reason—lucky in their first pick.

Second, in order to calculate overall effects of these interventions, not only heterogeneity across consumers, but also potential supply-side reactions to changes in consumer behavior should be taken into account. Again in the context of Medicare Part D, Ericson (2012) demonstrates that plans respond to consumer inertia with a simple “invest-then-harvest” strategy: firms set relatively low prices for newly introduced plans, but then raise prices as plans age and introduce new, low-cost plans each year. Furthermore, Ho, Hogan, and Scott Morton (2017) analyze plan pricing in Medicare Part D and demonstrate that insurers set prices above the level they would choose if all consumers were attentive and thus extract high rents due to consumer inertia. The authors indicate that reducing consumer inattention in Medicare Part D would lead to a reduction in premiums and thus further increase consumer benefits. Similarly, Polyakova (2016) indicates that reducing switching costs in Medicare Part D would lead to overall benefits to consumers. In her analysis she explicitly takes into account that simplifying consumer choice may lead to an increase in adverse selection across plans, as consumers are enabled to choose plans that better meet their needs and thus individuals with higher risks choose more generous coverage. Although an increase in adverse selection could potentially lead to increases in premiums among more generous plans and thus hurt consumers, Polyakova shows that in the specific setting of Medicare Part D insurers cannot pass through enrollees’ risk profiles to premiums. Therefore, she finds that losses due to adverse selection are moderate and dwarfed by gains due to better matches of enrollees to plans in the setting of Medicare Part D. This, however, need not always be the case. On the contrary, Handel (2013) presents evidence according to which enabling consumers to make better choices would aggravate adverse selection in an employer-based insurance setting in the United States with overall negative welfare consequences, as insurance firms would raise premiums in response.

Not only helping consumers to actively make better choices by reducing switching costs or increasing attention, but also other means to improve consumer choice, such as smart defaults, can have unintended consequences. For the case of smart defaults, Handel and Kolstad (2015a) discuss that some inertial consumers may be harmed by smart defaults if they are defaulted to a plan that is worse than their previous option. Furthermore, if the default algorithm favors certain plans (independent of whether this happens inadvertently or because certain firms are successful in lobbying the design of the default algorithm), smart defaults may reduce rather than increase competition across health plans. Additionally, consumers could be harmed if firms react to the smart default by improving their plans in dimensions valued by the default algorithm while at the same time delivering low quality on dimensions not sufficiently valued by the algorithm.

Overall, the effects of strategies to improve consumer choice are thus not only different for different consumers but also depend on the scope of reactions on the supply side, which in turn depend on the specific regulatory setting.

Controlled Laboratory Experiments on Complexity of Health Insurance Choice

Individual decision making has been studied in controlled laboratory experiments for a long time. Historically, most of these studies used neutral framing, that is, the decision tasks were described abstractly in terms of the payoffs and other constraints. More recently, experimental economists have turned to framing experimental tasks with specific applications in mind, including applications to decisions in healthcare markets. Also, health economists have started to use controlled laboratory experiments to study health-related decisions. Cox, Green, and Hennig-Schmidt (2016) point out that the fact that this relatively recent trend is somewhat surprising given that “Fuchs (2000) and Frank (2007) proposed using behavioral economics and experimental methods to complement traditional approaches in healthcare research more than a decade ago” (p. A1). Controlled experiments should be particularly useful to study decisions on health insurance markets. There are many papers, some of which we reviewed above, which show that consumers make suboptimal health insurance choices using field data. However, real-world markets are often so complex that the specific behavioral mechanisms that generate suboptimal choice cannot be determined. Two recent papers illustrate how laboratory experiments can be used to investigate behavioral aspects of health insurance choices.1

Schram and Sonnemans (2011) implement several complexities that individuals face when choosing health insurance plans in a laboratory experiment. These features include a large number of alternatives; the fact that these alternatives differ on a variety of dimensions; uncertainty; and the need to make repeated choices. The laboratory setting allows them to monitor individuals’ decision strategy (e.g., which information they look up). Experimental treatments introduce exogenous variation in the decision environment in a variety of dimensions: the number of alternatives; switching costs; and the speed at which health deteriorates. Schram and Sonnemans find that an increase in the number of alternatives increases decision-making time; makes subjects consider a lower fraction of the available information; makes it more likely that subjects will switch; and decreases the quality of their decisions. The introduction of positive costs of switching makes people switch less often but improves the quality of their decisions. These findings are in line with those of a variety of studies that use field data, including Frank and Lamiraud (2009) and Wuppermann et al. (2014) on the impact of the available number of options—but see the discussion above on the role of number of options—and Heiss et al. (2013) and Heiss et al. (2016) on switching costs and inattention.

Kairies-Schwarz, Kokot, Vomhof, and Weßling (2017) study insurance choice in conjunction with risk preferences. In that experiment, subjects first face stylized insurance choices. They are then presented with a separate task that reveals their risk preferences. In this second task, the choices of the majority of subjects are consistent with Cumulative Prospect Theory rather than Expected Utility Theory. Decisions of most subjects are in line with these risk preferences, only about 14% of all participants show inconsistent behavior and low decision quality. These results suggest that studies that analyze field data on health insurance choice under the maintained assumption of expected utility might find spurious choice inconsistencies—while those choices are inconsistent with expected utility, they might not be inconsistent with cumulative prospect or other theories of decision making under uncertainty. This raises the issue of evaluating the welfare changes associated with the effect of interventions in choices (which is of course directly observed).

Controversies, Debates, and Open Questions

As described in the previous sections, a large part of the literature on health insurance plan choice and switching concludes that individuals have difficulties in making these choices and often fail to make choices in their best interest. These conclusions rely on measures of the quality of plan choice. As health plans are often complex products and vary in many dimensions that may not all be observable to the researcher and as consumer choice naturally depends on consumers’ preferences, which are equally hard to observe, a key question and debate in the literature on health plan choices evolves around the ability to measure the quality of plan choice (see, e.g., the debate between Ketcham, Kuminoff, & Powers, 2016, and Abaluck & Gruber, 2011, 2016b). This question becomes even more important in the evaluation of impacts of interventions to improve consumers’ ability to choose as welfare calculations necessarily require assumptions on consumer preferences. Broadly speaking, the recent literature uses three approaches to measure the quality of health insurance choices.

A first approach mainly focuses on financial aspects of health insurance plans and thus evaluates whether consumers choose plans that minimize their costs. In the European context, for example, in Germany where health plan benefits are highly regulated and plans vary mainly in premiums, low price sensitivity of health plan choice is classified as a “mistake” (e.g., Wuppermann et al., 2014). In the context of Medicare Part D, Zhou and Zhang (2012) focus mainly on financial aspects of insurance and document that consumers do not choose the cheapest plan in their choice set, measured as the plan with the lowest premiums and out-of-pocket spending, given each consumer’s medication needs. Similarly, Ketcham et al. (2012, 2015) focus on overspending, that is, the amount of money in terms of premiums and out-of-pocket spending that consumers pay more than they would have in the cheapest available option. This approach ignores non-financial plan features such as reputation or service quality, and also abstracts from the fact that some plans, in particular in Medicare, may provide higher risk protection valued by risk-averse consumers.

Risk-aversion and non-financial plan features such as service quality can be taken into account in more complex models of consumer choice that are based on random utility maximization (e.g., McFadden, 2001). Such models are sometimes called “structural.” Many recent studies follow this second approach, to various degrees. Abaluck and Gruber (2011), Heiss et al. (2013), Heiss et al. (2016), Polyakova (2016), and Abaluck and Gruber (2016b) among others estimate logit models of health plan choice, in which consumers are assumed to choose the plan with the highest value to them. Plan value, in turn, can depend on financial plan features, such as premiums, deductibles, and copayments, the plan’s out-of-pocket costs that depend on each consumer’s healthcare needs, service quality, or the plan’s brand, and risk protection offered by the plan. The latter can, for example, be measured empirically by simulating the variance of out-of-pocket costs across similar consumers had they chosen the specific plan (see Abaluck & Gruber, 2011 and Heiss et al., 2013). This approach delivers evidence for choice inconsistencies if consumers put more weight on some plan features, such as a plan’s premium, than, for example, on out-of-pocket costs that are measured in the same metric (e.g., Abaluck & Gruber, 2011). Another possibility is that consumers might sometimes be too sensitive to short-run fluctuations in premiums, so that they would effectively switch plans too often. If that were the case, insurance firms might cut preventive programs that are associated with short-run costs and long-run health benefits in order to reduce prices and attract consumers. There is little systematic evidence on the empirical relevance of such consumers and firm behaviors.

The third approach characterizes choices of dominated plans as mistakes, that is, choices of plans that are the same or worse than another plan in the choice set on all dimensions and worse than that plan on at least one dimension for all possible health states. Although this approach does not make explicit assumptions on consumers’ utility, it requires observing and taking into account all dimensions of health plans that are relevant to consumers. This approach is thus particularly suitable for laboratory choice experiments in which the researcher can control all relevant plan dimensions (Besedeš et al., 2012). There are, however, also specific situations in observational settings that allow observing dominated options. For example, Sinaiko and Hirth (2011) study a case in which one health plan available to employees in a U.S. firm dominates another as it offers additional benefits at the same premium as the other plan. Importantly, both plans are offered by the same insurance company, suggesting that service quality or the insurer’s reputation—which is typically not directly observable to researchers but may matter to consumers—can be assumed to be the same across plans. Afendulis et al. (2015) argue that in the early years of Medicare Advantage, the traditional Medicare Program was dominated since certain plans in Medicare Advantage offered the same or more benefits as the traditional program at lower premiums. Similarly, Bhargava, Loewenstein, and Sydnor (2017) focus on plans offered by the same insurance firm that also cover the same benefits but vary in premiums and cost-sharing features. In their setting, many plans with low deductibles were financially dominated because the additional premium charged for the reduction in the deductible was extremely high and exceeded the maximum savings that consumers would get from the reduction in the deductible.

While approaches using choices of dominated options to measure decision mistakes do not need an explicit model of consumer preferences or utility, calculations of impacts of possible interventions on consumer or overall welfare require explicit assumptions on consumers’ utility. As discussed above, studies in laboratory settings (e.g., Schram & Sonnemans, 2011 or Kairies-Schwarz et al., 2017) explicitly elicit consumer preferences and then evaluate health plan choices according to these preferences. This, of course, is hardly feasible with observational data. But even based on the results from the laboratory settings, the question arises whether the overall situation of consumers with “non-standard” preferences that are, for example, inconsistent over time would be improved by interventions, such as nudging consumers toward making different health insurance choices. A related issue is which determinants of health plan choice should be considered in welfare calculations. For example, consumers may have preferences for specific brands and thus choose plan A over plan B not because plan A offers higher benefits, lower premiums, or better service but just because it is offered by one specific firm—in the extreme with a nicer sounding or more familiar name. Should these brand preferences that consumers may exhibit enter welfare calculations; in other words, should consumer utility following plan choice be measured including utility from choosing plans of specific brands, or not?

Economists need to make decisions—whether explicitly or implicitly—about features of the models they use to calculate consumer welfare and the effects of interventions aimed at improving choices. A key question and debate that remains open in the literature thus concerns normative assumptions with respect to the measures of choice quality and consumer welfare. The issue arises not only in the analysis of health insurance markets, but also in behavioral economics more generally. Reviewing this vast and growing literature is beyond the scope of this article, but we point the reader to the debate between Ketcham et al. (2016) and Abaluck and Gruber (2011, 2016b) and to the paper by Handel and Kolstad (2015b) for insightful discussions of these challenging conceptual questions.

An important implication of poor initial plan choices and reluctance to switch plans is that competition in health insurance markets might not work as well as the proponents of consumer-directed healthcare argue. Further, insurance firms might be able to exploit poor plan choices and inertia, thus extracting rents from consumers. Quantifying such effects is another important area of research in which researchers from the fields of health economics and industrial organization could cooperate fruitfully.


Abaluck, J., & Gruber, J. (2011). Choice inconsistencies among the elderly: Evidence from plan choice in the Medicare part D program. American Economic Review, 101(4), 1180–1210.Find this resource:

Abaluck, J., & Gruber, J. (2016a). Improving the quality of choices in health insurance markets. NBER Working Paper, 22917.Find this resource:

Abaluck, J., & Gruber, J. (2016b). Choice inconsistencies among the elderly: Evidence from plan choice in the Medicare part D program: Reply. American Economic Review, 106(12), 3962–3987.Find this resource:

Afendulis, C. C., Sinaiko, A. D., & Frank, R. G. (2015). Dominated choices and Medicare advantage enrollment. Journal of Economic Behavior & Organization, 119, 72–83.Find this resource:

Agarwal, S., & Mazumder, B. (2013). Cognitive abilities and household financial decision making. American Economic Journal: Applied Economics, 5(1), 193–207.Find this resource:

Barcellos, S. H., Wuppermann, A. C., Carman, K. G., Bauhoff, S., McFadden, D. L., Kapteyn, A., . . . Goldman, D. (2014). Preparedness of Americans for the affordable care act. Proceedings of the National Academy of Sciences, 111(15), 5497–5502.Find this resource:

Bartholomae, S., Russell, M. B., Braun, B., & McCoy, T. (2016). Building health insurance literacy: Evidence from the smart choice health insurance program. Journal of Family and Economic Issues, 37(2), 140–155.Find this resource:

Besedeš, T., Deck, C., Sarangi, S., & Shor, M. (2012). Age effects and heuristics in decision making. Review of Economics and Statistics, 94(2), 580–595.Find this resource:

Bhargava, S., Loewenstein, G., & Sydnor, J. (2017). Choose to lose: Health plan choices from a menu with dominated options. Quarterly Journal of Economics, 132(3), 1319–1372.Find this resource:

Bhargava, S., & Manoli, D. (2015). Psychological frictions and the incomplete take-up of social benefits: Evidence from an IRS field experiment. The American Economic Review, 105(11), 3489–3529.Find this resource:

Chetty, R. (2015). Behavioral economics and public policy: A pragmatic perspective. American Economic Review, 105(5), 1–33.Find this resource:

Choi, J. J., Laibson, D., & Madrian, B. C. (2010). $100 bills on the sidewalk: Suboptimal investment in 401(k) plans. The Review of Economics and Statistics, 93(3), 748–763.Find this resource:

Cox, J. C., Green, E. P., & Hennig-Schmidt, H. (2016). Experimental and behavioral economics of healthcare. Journal of Economic Behavior & Organization, 131(November), A1–A4.Find this resource:

Cox, J. C., Sadiraj, V., Schnier, K. E., & Sweeney, J. F. (2016). Higher quality and lower cost from improving hospital discharge decision making. Journal of Economic Behavior & Organization, Experimental and Behavioral Economics of Healthcare, 131(November), 1–16.Find this resource:

Ericson, K. M. (2012). Consumer inertia and firm pricing in the Medicare part D prescription drug insurance exchange. Working Paper 18359. National Bureau of Economic Research.Find this resource:

Ericson, K. M., Kingsdale, J., Layton, T., & Sacarny, A. (2017). Nudging leads consumers in Colorado to shop but not switch ACA marketplace plans. Health Affairs, 36(2), 311–319.Find this resource:

Ericson, K. M., & Starc, A. (2016). How product standardization affects choice: Evidence from the Massachusetts health insurance exchange. Journal of Health Economics, 50, 71–85.Find this resource:

Ericson, K. M., & Starc, A. (2012). Heuristics and heterogeneity in health insurance exchanges: Evidence from the Massachusetts connector. American Economic Review: Papers & Proceedings, 102(3), 493–497.Find this resource:

Frank, R. G. (2007). Behavioral economics and health economics. In P. Diamond, & H. Vartianinen (Eds.), Behavioral economics and its applications. Princeton: Princeton University Press.Find this resource:

Frank, R. G., & Lamiraud, K. (2009). Choice, price competition and complexity in markets for health insurance. Journal of Economic Behavior & Organization, 71(2), 550–562.Find this resource:

Fuchs, V. R. (2000). The future of health economics. Journal of Health Economics, 19(2), 141–157.Find this resource:

Gabel, J. R. (1999). Job-based health insurance, 1977–1998: The accidental system under scrutiny. Health Affairs, 18(6), 62–74.Find this resource:

Galizzi, M. M., & Wiesen, D. (2018). Behavioral experiments in health economics. Oxford Research Encyclopedia of Economics and Finance.Find this resource:

Grob, R., & Schlesinger, M. (2015). Educating, enrolling, and engaging: The state of marketplace consumer assistance under the affordable care act. Health Affairs, 34(12), 2052–2060.Find this resource:

Gruber, J. (2017). Delivering public health insurance through private plan choice in the United States. Journal of Economic Perspectives, 31(4), 3–22.Find this resource:

Handel, B. R. (2013). Adverse selection and inertia in health insurance markets: When nudging hurts. American Economic Review, 103(7), 2643–2682.Find this resource:

Handel, B. R., & Kolstad, J. (2015a). Getting the Most form marketplaces: Smart policies on health insurance choice. The Hamilton Project.Find this resource:

Handel, B. R., & Kolstad, J. T. (2015b). Health insurance for ‘humans’: Information frictions, plan choice, and consumer welfare. American Economic Review, 105(8), 2449–2500.Find this resource:

Heiss, F., Leive, A., McFadden, D., & Winter, J. (2013). Plan selection in Medicare part D: Evidence from administrative data. Journal of Health Economics, 32(6), 1325–1344.Find this resource:

Heiss, F., McFadden, D., & Winter, J. (2006). Who failed to enroll in Medicare part D, and why? Early results. Health Affairs, 25(5), w344–354.Find this resource:

Heiss, F., McFadden, D., Winter, J., Wuppermann, A., & Zhou, B. (2016). Inattention and switching costs as sources of inertia in Medicare part D. Working Paper 22765. National Bureau of Economic Research.Find this resource:

Hennig-Schmidt, H., Selten, R., & Wiesen, D. (2011). How payment systems affect physicians’ provision behaviour—An experimental investigation. Journal of Health Economics, 30(4), 637–646.Find this resource:

Ho, K., Hogan, J., & Morton, F. S. (2017). The impact of consumer inattention on insurer pricing in the Medicare part D program. RAND Journal of Economics, 48(4), 877–905.Find this resource:

Hoerl, M., Wuppermann, A. C., Barcellos, S. H., Bauhoff, S., Winter, J., & Carman, K. G. (2017). Knowledge as a predictor of insurance coverage under the affordable care act. Medical Care, 55(4), 428–435.Find this resource:

Hortacsu, A., Madanizadeh, S. A., & Puller, S. L. (2017). Power to choose? An analysis of consumer inertia in the residential electricity market. American Economic Journal: Economic Policy, 9(4), 192–226.Find this resource:

Iyengar, S. S., & Kamenica, E. (2010). Choice proliferation, simplicity seeking, and asset allocation. Journal of Public Economics, 94(7–8), 530–539.Find this resource:

Kairies-Schwarz, N., Kokot, J., Vomhof, M., & Weßling, J. (2017). Health insurance choice and risk preferences under cumulative Prospect theory—An experiment. Journal of Economic Behavior & Organization, 137(May), 374–397.Find this resource:

Kesternich, I., Schumacher, H., & Winter, J. (2015). Professional norms and physician behavior: Homo Oeconomicus or homo Hippocraticus? Journal of Public Economics, 131(November), 1–11.Find this resource:

Ketcham, J. D., Kuminoff, N. V., & Powers, C. A. (2016). Choice inconsistencies among the elderly: Evidence from plan choice in the Medicare part D program: Comment. American Economic Review, 106(12), 3932–3961.Find this resource:

Ketcham, J. D., Lucarelli, C., Miravete, E. J., & Christopher Roebuck, M. (2012). Sinking, swimming, or learning to swim in Medicare part D. The American Economic Review, 102(6), 2639–2673.Find this resource:

Ketcham, J. D., Lucarelli, C., & Powers, C. A. (2015). Paying attention or paying too much in Medicare part D. American Economic Review, 105(1), 204–233.Find this resource:

Kiss, A. (2016). Salience and switching. Working Paper.Find this resource:

Kling, J. R., Mullainathan, S., Shafir, E., Vermeulen, L. C., & Wrobel, M. V. (2012). Comparison friction: Experimental evidence from Medicare drug plans. The Quarterly Journal of Economics, 127(1), 199–235.Find this resource:

Laske-Aldershof, T., Schut, E., Beck, K., Greß, S., Shmueli, A., & Van de Voorde, C. (2004). Consumer mobility in social health insurance markets: A five-country comparison. Applied Health Economics and Health Policy, 3(4), 229–241.Find this resource:

Liebman, J., & Zeckhauser, R. J. (2008). Simple humans, complex insurance, subtle subsidies. National Bureau of Economic Research Working Paper Series, 14330.Find this resource:

Loewenstein, G., Friedman, J. Y., McGill, B., Ahmad, S., Linck, S., Sinkula, S., . . . Volpp, K. G. (2013). Consumers’ misunderstanding of health insurance. Journal of Health Economics, 32(5), 850–862.Find this resource:

Madrian, B. C., & Shea, D. F. (2001). The power of suggestion: Inertia in 401(k) participation and savings behavior. Quarterly Journal of Economics, 116(4), 1149–1187.Find this resource:

Maestas, N., Schroeder, M., & Goldman, D. (2009). Price variation in markets with homogeneous goods: The case of Medigap. National Bureau of Economic Research Working Paper Series No. 14679.Find this resource:

McFadden, D. (2001). Economic choices. American Economic Review, 91(3), 351–378.Find this resource:

McFadden, D., Winter, J., & Heiss, F. (2008). Consumer-directed health care: Can consumers look after themselves? Swiss Journal of Economics and Statistics, 144(3), 285–307.Find this resource:

McGarry, B. E., Maestas, N., & Grabowski, D. C. (2018). Simplifying the Medicare plan finder tool could help older adults choose lower-cost part D plans. Health Affairs, 37(8), 1290–1297.Find this resource:

McGuire, T., & Bauhoff, S. (2007). A decade of choice: Tracking the German national experience with consumer choice of sickness fund. In P. Oberender & C. Straub (Eds.), Auf Der Suche Nach Der Besseren Loesung (pp. 145–159). Baden-Baden: Nomos.Find this resource:

McGuire, T. G. (2011). Demand for health insurance. In M. Pauly, T. McGuire, & P. Barros (Eds.), Handbook of health economics (Vol. 2, pp. 317–396). Amsterdam: Elsevier.Find this resource:

McGuire, T., Newhouse, J. P., & Sinaiko, A. D. (2011). An economic history of Medicare part C. Milbank Quarterly, 89(2), 289–332.Find this resource:

Polyakova, M. (2016). Regulation of insurance with adverse selection and switching costs: Evidence from Medicare part D. American Economic Journal: Applied Economics, 8(3), 165–195.Find this resource:

Schmitz, H., & Ziebarth, N. R. (2017). Does Price framing affect the consumer Price sensitivity of health plan choice? Journal of Human Resources, 52(1), 88–127.Find this resource:

Schram, A., & Sonnemans, J. (2011). How individuals choose health insurance: An experimental analysis. European Economic Review, 55(6), 799–819.Find this resource:

Sinaiko, A., & Hirth, R. (2011). Consumer health insurance and dominated choices. Journal of Health Economics, 30(2), 450–457.Find this resource:

Sinaiko, A. D., & Zeckhauser, R. J. (2016). Forced to choose, again. In I. Glenn Cohen, H. F. Lynch, & C. T. Robertson (Eds.), Nudging health: Health law and behavioral economics (1st ed., pp. 326–338). Baltimore, MD: Johns Hopkins University Press.Find this resource:

Winter, J., Balza, R., Caro, F., Heiss, F., Jun, B.-H., Matzkin, R., & McFadden, D. (2006). Medicare prescription drug coverage: Consumer information and preferences. Proceedings of the National Academy of Sciences, 103(20), 7929–7934.Find this resource:

Woodward, S. E., & Hall, R. E. (2012). Diagnosing consumer confusion and sub-optimal shopping effort: Theory and mortgage-market evidence. American Economic Review, 102(7), 3249–3276.Find this resource:

Wuppermann, A., Bauhoff, S., & Grabka, M. (2014). The price sensitivity of health plan choice among retirees: Evidence from the German social health insurance. Munich Economic Discussion Paper 34.Find this resource:

Zhang, Y., Zhou, C., & Baik, S. H. (2014). A simple change to the Medicare part D low-income subsidy program could save $5 billion. Health Affairs, 33(6), 940–945.Find this resource:

Zhou, C., & Zhang, Y. (2012). The vast majority of Medicare part D beneficiaries still Don’t choose the cheapest plans that meet their medication needs. Health Affairs, 31(10), 2259–2265.Find this resource:


(1.) Other topics in health economics studied in laboratory experiments include treatment choice (Cox, Sadiraj, Schnier, & Sweeney, 2016), the effects of reimbursement schemes on physician behavior (Hennig-Schmidt, Selten, & Wiesen, 2011), and physician preferences more generally (Kesternich, Schumacher, & Winter, 2015). See Galizzi and Wiesen (2018) for a review of this active area of research.