Do honesty‐nudges really work? A large‐scale field experiment in an insurance context

Abstract

Businesses want their customers to self-report information honestly. One increasingly popular way to stimulate desired behavior is by using nudge interventions. But can customers be nudged to self-report information more honestly? This is currently a debate in the literature, where empirical results are inconclusive. Building on related literature on nudges, we add to this debate with a controlled field experiment (N = 5704). We used data from actual customers making real decisions when they file claims online to a large Nordic insurance provider. To the best of our knowledge, this is the first study to investigate the effects of honesty-nudges on self-reported information when filing insurance claims using a controlled field experiment. We designed and tested three honesty-nudges on insurance customers: (1) signing-atthe-beginning, (2) a descriptive social norm message, and (3) a solidarity message.Across five outcome measures, we found that the honesty-nudges, standalone or in any possible combination, do not have significant effects in reducing indicators of insurance claims fraud. But interestingly, customers in all treatment groups used significantly more characters to describe losses than customers in the control group. Also, in post hoc analyses, we found signs that the direction of nudge effects varies across customers' age and customer loyalty. | INTRODUCTIONInsurance companies want their customers to not falsely report or exaggerate damages when they file claims. Fraudulent claims cost the insurance provider because those increase settlement costs. These can also cost honest customers as providers increase premiums to cover extra costs. In 2017, fraudulent claims cost European insurers and customers an estimated 13 billion Euros (Insurance Europe, 2019a). Finance Norway (Finanse Norge, 2019) estimates the total value of detected fraudulent non-life insurance claims in Norway at NOK 495.1 million (approximately 49 million Euros) in 2019. The total value of undetected yearly insurance fraud in Denmark is estimated to be 10 times that of detected-300 million Euros versus 32.75 million Euros 1 (Insurance Europe, 2019b). The Federal Bureau of Investigation (2020) estimates the cost of non-health insurance fraud in the United States at over 40 billion USD. Dealing with insurance claims fraud is difficult because providers rely on self-reported information and customers have economic incentives to report false or exaggerated damages. 1 Estimates were in Euros than in Danish Kroners for uniformity in comparing figures across countries.

Meta-nudging honesty: Past, present, and future of the research frontier

“…This is supported by empirical research, both in the lab and in the field, indicating that mitigating dishonest behavior through behavioral interventions alone, e.g. norm-nudges or defaults, is challenging [6,9,35,[50][51][52][53]. This is further augmented by theoretical arguments stressing that the nudgeability of individuals matters, which is often overlooked in existing research and deserves more scholarly attention moving forward [54].…”

Investigating the Impact of Social Commitment on Dishonest Behavior

“…It is possible that different forms of expressing social norm commitment (e.g., signing instead of copying text) might produce stronger effects (see Skowronek, 2022 for a framework to improve honesty nudges). As previous studies did not find evidence for the effectiveness of commitment to oaths (Kristal et al, 2020;Koning et al, 2020;Martuza et al, 2022), it seems important to study the conditions under which such an intervention could reduce dishonesty.…”

Prior behavior and wording of norm nudge requests shape compliance and reciprocity

“…Furthermore, recent work has questioned the effectiveness of moral nudges. In a large field study, Martuza et al (2022) found that asking people to sign before submitting insurance claims, providing social norm messages, and even evoking solidarity had no effect on honesty. Taken together, these findings make the prediction that norm nudges should increase compliance less straightforward and worth examining.…”