by Eyal Pe’er and Yuval Feldman
One of the main challenges in interactions between regulators and regulatees, as well as between managers and employees or between companies and their customers, is whether mutual trust can be achieved, especially when options for monitoring are limited. Ample evidence on the prevalence of unethical behaviors, be they intentional or not, which exists in most societies in various contexts, has justified the hesitance to trust other people, especially when the consequences of breach of trust might be dire. At the same time, many governments and organizations that (justly) wish to reduce regulatory burden often ask citizens and businesses to state, in advance, that their reports are truthful or that they abide by the relevant rules and regulations (“honesty pledges”). However, while such pledges can indeed reduce unneeded procedures, they might also tempt some agents, aware of the reduced monitoring, to cheat in their reports. We study how ex-ante honesty pledges can be used as a behaviorally-based regulatory tool to enhance the likelihood that people who are entrusted with behaving in a certain way will do so, with as limited as possible regulatory burden. We find that, when done properly, adoption of such an approach is likely to reduce dishonesty, and we believe that could also increase mutual trust and carry other positive effects.
Some previous research has shown that honesty pledges can indeed sometimes reduce unethical behavior. However, other findings claiming that making ethicality salient can curb dishonesty (e.g., by moving the position of a signature on a form) have been recently challenged. Also, a study showed that students who were asked to sign a commitment form before starting their exam actually cheated at an increased rate, and a field experiment on tax-payers in Sweden found that promising to pay taxes on time increased compliance only among those already highly compliant in the past. Furthermore, previous studies neglected to examine how pledges work compared to sanctions (fines), in sequential and repeated (and not just one-shot) decisions, how their effect lasts over time, and whether different people react to pledges differently.
To deal with these challenges, we conducted a series of experiments (N>2,600) to examine a) would pledges reduce dishonesty significantly, b) would their effects persist when fines for cheating are included, c) would pledges affect different levels of cheating, d) across differences in the individuals making the pledges, the periods of time during which the pledges were made, and, lastly, different pledges’ phrasings.
Do pledges work?
In our studies, we asked participants to complete an online calculus task, with 10 problems presented sequentially, each for 20 seconds, under different conditions: Control participants had to report the solution to each problem to earn a bonus, while Self-Report participants received their bonus for any problem they only reported as solved. In the Pledge-only condition (without fines), participants had to first re-type a pledge stating they will report honestly, and were told that 10% of problem reports may be audited. In the combined Pledge & Fine condition, we also told participants that if they failed this audit, their entire bonus would be lost. In our first study (N=480), we found that participants in the Self-Report condition (without a fine or pledge) reported solving about twice as many problems as those actually solved in the Control group (63% vs. 32%). However, participants in the Pledge and the Pledge & Fine conditions reported solving much fewer (49% and 52%), and these differences were statistically significant, showing a reduction of the cheating extent by about 40% on average. We repeated the study with the same sample after a year (N=297), using the same design and conditions, and found similar results for the effect of the pledge. Also, we found no evidence of a decay effect – where repeated exposure during that year to other pledges in the real-world (e.g., health statements people had to sign during COVID-19) could have reduced its efficacy.
How do pledges work with fines and reminders?
Another study extended these findings to a larger sample (N=1,158), this time also including conditions of a fine without a pledge and adding reminders about the pledge and/or the fine in the middle of the task. We again found that participants in the Self-Report condition reported solving twice more than the Control group (60% vs. 28%), and the pledge, without a fine, reduced reports significantly to 49%. When both a fine and a pledge were present, report rate was lowest at 38%. The effects of the pledge and fine were each significant, while their interaction was not, suggesting fines did not crowd-out the effect of the pledge. The reminders did not show any significant effects. We also found that the effect of a pledge was not moderated by individual differences in prior tendency to follow rules and regulations, as measured by several validated scales, and that the pledge also reduced “brazen” lying (those claiming all or almost all problems solved) significantly.
What’s the best text to use?
Lastly, in another large study (N=1,024), we used eight differently worded pledges, and replicated the significant effect of the pledge on reducing dishonesty in all of them. We found a significant interaction between the formality of the pledge’s language and the mention of the sanction: when the pledge used formal language, mentioning the sanction reduced cheating more, but when the pledge used informal language, not mentioning the sanction reduced cheating somewhat more.
Conclusions
Pledges can indeed reduce dishonesty significantly, considerably also when compared to fines, and consistently across different extents of cheating, individual differences, periods of time and pledges’ phrasings. Because previous research on moral reminders, textual primes and signature position has shown mixed results, these findings substantiate the effectiveness of ex-ante honesty pledges, define its scope and boundaries, and can lead to further research that explores the underlying mechanisms of the effects of pledges on dishonesty and compliance. Moreover, we also hope that using behaviorally-based regulation instruments – such as honesty pledges – can help rebuild and foster the important trust between government and the public.
More information can be found at https://eyalpeer.huji.ac.il/honestypledges.
Eyal Pe’er is an Associate Professor at the Federmann School of Public Policy, Hebrew University of Jerusalem. Pe’er’s research focuses on judgment and decision-making processes and their applications for behavioral public policy. Yuval Feldman is the Mori Lazarof Professor of Legal Research, Bar-Ilan University Faculty of Law and Senior Fellow at the Israel Democracy Institute. Feldman’s research focuses on Behavioral and Experimental Analysis of Regulation, Compliance and Ethical Behavior.
Disclaimer
The views, opinions and positions expressed within all posts are those of the authors alone and do not represent those of the Program on Corporate Compliance and Enforcement or of New York University School of Law. The accuracy, completeness and validity of any statements made within this article are not guaranteed. We accept no liability for any errors, omissions or representations. The copyright of this content belongs to the authors and any liability with regards to infringement of intellectual property rights remains with them.