Is default risk acceptable when purchasing insurance? Experimental evidence for different probability representations, reasons for default, and framings
Introduction
People buy insurance to protect themselves against different kinds of potential losses. However, when deciding to purchase insurance, consumers must keep in mind (or should) that the insurance contract itself might be exposed to the risk of default, i.e., the insurance policy might involve a small probability that the policyholder will not be reimbursed partially or totally by the insurer in case of a loss. Experimental research by Wakker et al. (1997) and Albrecht and Maurer (2000) shows that the awareness of default risk has an influence on consumers’ insurance purchase behavior. People dislike insurance contracts that might default when indemnity payments are needed. In their study, Wakker et al. (1997) demonstrate that people demand a greater than 20% premium reduction when facing a 1% default probability. Similar results are reported by Albrecht and Maurer (2000), who, in addition, find that the greater the insurer’s default risk, the more people will refuse to pay anything at all for such an insurance policy.
This paper tests the robustness of these findings. We analyze consumers’ reactions to default risk dependent on the way an insurer’s default situation is communicated. Specifically, we examine whether the way the default risk is represented influences reactions to it. Furthermore, we vary the level of default probability and the reasons for default. Finally, we check for framing effects and reference dependence. Most of these factors have not been analyzed previously with respect to their impact on reactions to default risk. We also investigate what actually drives these behaviors.
The factors we analyze are of relevance for managerial decisions and might have policy implications as well. First, we incorporate the realistic feature that policyholders do not have numeric information on the probability of default. Rating agencies usually provide only a verbal explanation of an insurer’s financial situation. In principle, the insured can obtain numeric information, but doing so involves the considerable effort of working through empirical studies, e.g., annual default reports provided by financial institutions and rating agencies (see, e.g., Standard and Poor’s, A.M. Best, and Moody’s). A few studies suggest that being given verbal probabilities will have a very different impact on behavior than being provided with numeric probabilities. For example, Wallsten, Budescu, and Zwick (1993) show that individuals exhibit greater overconfidence when provided with verbally expressed probabilities. In Budescu, Weinberg, and Wallsten (1988) study of bidding behavior, bids made under the influence of verbal probabilities turned out to be less optimal than those based on numeric ones. Subjects would have earned about 24% less under the verbal than under the numeric condition.1 We therefore analyze the effects of verbal and numeric probabilities on the willingness to pay for insurance with default risk.
Second, we examine the effect of positive and negative framing of default risk, i.e., whether default described in terms of nonpayment of claims has a different effect on purchase decisions than default described in terms of payment of claims. The way default risk is framed might direct consumers’ attention toward the potential occurrence or nonoccurrence of default and thus might affect their decision making (see, e.g., Teigen and Brun (1999) for the effects of positive and negative phrases on decision making). Furthermore, there is wide evidence that individuals put different weights on small and high probabilities (see, e.g., Kahneman & Tversky, 1979). Thus, the choice of high positive and low negative probabilities for the description of default risk might affect policyholders’ reaction to the risk.
Third, Tversky and Kahneman (1981) show that a change in reference point can lead to a reversal of individual preferences. Similarly, the choice of a reference alternative (or the status quo) can have a major impact on individuals’ decisions (see Burmeister and Schade, 2007, McKenzie et al., 2006, Samuelson and Zeckhauser, 1988). We analyze whether the willingness to pay for a specific insurance contract is influenced by information provided about the default risk and price of another, concurrently available contract.
Fourth, there are reasons other than the insurer’s insolvency that may result in a total or partial default in paying claims (see Doherty and Schlesinger (1990, pp. 243–244), Kahneman and Tversky, 1979, Wakker et al., 1997). One prominent example is the recent “wind-water controversy” in respect to hurricane losses. Victims of Hurricane Katrina have been very disappointed about the unexpectedly low indemnity payments they received for repairing or rebuilding their damaged houses. They were not aware of the fact that a standard homeowner policy in the US covers losses caused by fire, hail, winter storms, tornadoes and wind damage but not from rising water due to floods and hurricanes (see Kunreuther (2006, p. 3)). Another highly relevant source of default risk is nonpayment due to the insurer’s claim settlement practices, something that has not been investigated in the literature to date. Specifically, claim settlement practices2 have to do with the insurer’s willingness to meet its financial obligation rather than its ability to do so. It is unclear whether consumers will react in the same way to different reasons for default because each type of default may evoke different emotions (see Kunreuther et al. (2002)).3 For example, individuals are likely to be more angry about a default based on claim settlement practices, than they are about one due to insolvency as they may believe that insolvency is something beyond the insurer’s control. Affect regarding the insured object have been shown to have an impact on insurance decisions – affect regarding the reason for default might have an effect, too (Hsee and Kunreuther, 2000, Slovic et al., 2001).
To analyze consumer reactions to insurance default risk we conduct three experiments. The main aim of Experiment 1 is to examine the impact of probability representation and reasons for default on the willingness to pay for household insurance involving different levels of default risk. In this first experiment, the reference insurance is default free. Experiment 2 is designed to test whether the results of Experiment 1 hold when the reference insurance contains a small probability of default. Finally, Experiment 3 examines the effect of positive and negative framing of default risk on willingness to pay.
We find that in the presence of default risk, individuals either refuse to purchase insurance or they demand a considerable reduction in insurance premiums and are willing to pay a substantial loading on the expected claims to avoid such risk. These behaviors are observed over a variety of situations and, except for the level of default risk, the factors analyzed in our experiments have no systematic impact. In all experiments, decisions appear to be largely driven by the individuals’ degree of security concerns.
The paper is organized as follows. In the next section we present a detailed explanation of our experimental designs and the statistical analyses of the results. We continue with a discussion of our findings in relation to existing evidence. Finally, in Section 4, we draw conclusions and discuss the practical impact of our findings.
Section snippets
Design characteristics
The goal of all three experiments is to elicit individuals’ willingness to pay for different household insurance contracts,
Discussion
The results of this study confirm previous results that individuals either will not accept insurance with default risk or they will ask for a higher than expected reduction in insurance premiums. But why do people refuse to accept insurance with default risk – even as a gift? This cannot be explained by any rational model of choice. Instead, this behavior looks like an extreme certainty effect (Allais, 1953): Willingness to pay is not only heavily reduced but the entire risky prospect is
Conclusions, implications, and limitations
We carried out three experiments on insurance with default risk that varied the characteristics of the decision situation. We find that irrespective of the choice of default probability representation, the reason for default, the framing of the default risk, and the default-proneness of the reference insurance, default risk is a very unattractive feature of insurance contracts for most individuals. When purchasing insurance, individuals seem to be in search of peace of mind and this desire is
Acknowledgements
This research was supported by the Berlin Association for the Advancement of Insurance Sciences. We are very grateful for this support and the excellent cooperation we were provided by the Association. Furthermore, we are very grateful to Howard Kunreuther for his insightful comments and suggestions.
References (49)
- et al.
Verbal probabilities: Ambiguous, context-dependent, or both?
Organizational Behavior and Human Decision Processes
(1988) - et al.
Consistency in interpretation of probabilistic phrases
Organizational Behavior and Human Decision Processes
(1985) - et al.
Are entrepreneurs’ decisions more biased? An experimental investigation of the susceptibility to status quo bias
Journal of Business Venturing
(2007) - et al.
Are humans good intuitive statisticians after all? Rethinking some conclusions of the literature on judgment under uncertainty
Cognition
(1996) - et al.
Risk perception and worries after the Chernobyl accident
Journal Environmental Psychology
(1990) - et al.
Active information search and complete information presentation in naturalistic risky decision tasks
Acta Psychologica
(1997) - et al.
“I think I can, I think I can”: Overconfidence and entrepreneurial behavior
Journal of Economic Psychology
(2007) When equal chances = good chances: Verbal probabilities and the equiprobability effect
Organizational Behavior and Human Decision Processes
(2001)- et al.
The directionality of verbal probability expressions: Effects on decisions, predictions, and probabilistic reasoning
Organizational Behavior and Human Decision Process
(1999) - et al.
Base rate effects on the interpretations of probability and frequency expressions
Journal of Memory and Language
(1986)
Zur Bedeutung der Ausfallbedrohtheit von Versicherungskontrakten – ein Beitrag zur Behavioral Insurance
Zeitschrift für die gesamte Versicherungswissenschaft
Le Comportement de l’Homme Rationnel devant le Risque: Critique des Postulats et Axiomes de l’Ecole Americaine
Econometrica
Determinants of priority for risk reduction: The role of worry
Risk Analysis
Decisions based on numerically and verbally expressed uncertainties
Journal of Experimental Psychology: Human Perceptions and Performance
Rational insurance purchasing: Consideration of contract nonperformance
Quarterly Journal of Economics
Media reputation of the insurance industry: An urgent call for strategic communication management
Geneva Papers on Risk and Insurance Issues and Practice
Ambiguity and uncertainty in probabilistic inference
Psychological Review
Risk management, capital budgeting, and capital structure policy for Insurers and reinsurers
Journal of Risk and Insurance
The psychology of good judgments: Frequency formats and simple algorithms
Medical Decision Making
How to improve bayesian reasoning without instruction: Frequency formats
Psychological Review
Do decision quality and preference order depend on whether probabilities are verbal or numerical?
American Journal of Psychology
To hedge or not to hedge: Managing demographic risk in life insurance companies
Journal of Risk and Insurance
Preference and belief: Ambiguity and competence in choice under uncertainty
Journal of Risk and Uncertainty
Risk taking and problem context in the domain of losses: An expected utility analysis
Journal of Risk and Insurance
Cited by (51)
Managing reputational risk in the decumulation phase of a pension fund
2023, Insurance: Mathematics and EconomicsCitation Excerpt :For instance, in non-life insurance a considerable fraction of individuals refuses any default risk. In general, consumers demand more than 20% reduction in premium to offset a 1% default risk (see, for example, Wakker et al. (1997) and Zimmer et al. (2009)). Although it is clear that a decrease in pensions could occur if there is no contractual guarantee for the height of future pension payments, the reputation of a pension fund provider will seriously suffer if pensions have to be decreased without an objective reason (such as deflation).
Do you trust your insurer? Ambiguity about contract nonperformance and optimal insurance demand
2020, Journal of Economic Behavior and OrganizationCitation Excerpt :Kahneman and Tversky (1979) introduce ‘probabilistic insurance’, which incorporates an element of nonperformance risk but also exhibits probabilistic repayment of the premium. Subjects appear to dislike both probabilistic insurance as well as pure nonperformance risk and adjust insurance demand accordingly, as suggested by various surveys (Kahneman and Tversky, 1979; Wakker et al., 1997; Zimmer et al., 2009) and evidenced in incentive-compatible laboratory experiments (Herrero et al., 2006; Zimmer et al., 2018). What’s more, nonperformance risk reduces welfare of risk-averse individuals directly, because contracts are less efficient, and indirectly, because the compound nature of nonperformance risk induces violations of the reduction of compound lotteries axiom (Harrison and Ng, 2018).
Contract nonperformance risk and uncertainty in insurance markets
2019, Journal of Public EconomicsCitation Excerpt :At least, the issue does not seems to disappear easily. Insurance clients have been shown to be sensitive to the insurer's influence on the decision not to pay claims (Kunreuther et al., 2002; Zimmer et al., 2009), and we are interested in whether these framing effects are relevant in our uncertain contract nonperformance setting. We consider two main channels by which framing could impact insurance demand.
Risk- and value-based management for non-life insurers under solvency constraints
2018, European Journal of Operational ResearchCitation Excerpt :In addition, assessing consumer preferences and willingness to pay is vital for strategic decisions and is not sufficiently considered in the insurance sector yet, as Braun, Schmeiser, and Schreiber (2017) criticize. Particularly, experimental and empirical research (Lorson, Schmeiser, & Wagner, 2012; Wakker, Thaler, & Tversky, 1997; Zimmer et al., 2014; Zimmer, Schade, & Gründl, 2009) shows that an insurer's default risk can have a strong influence on customer demand, where lower safety levels can lead to a considerable reduction of achievable premiums. In this context, Zimmer et al. (2014) develop a risk management model assuming that the insurer's default risk is fully known to consumers, and based on this derive the solvency level that maximizes shareholder value, which is the case for a ruin probability of zero.
Systematic longevity risk: To bear or to insure
2020, Journal of Pension Economics and FinanceAnnuity selection in the presence of insurer default risk and government guarantees
2024, Journal of Risk and Insurance