Using Ex-ante Pledges to Reduce Dishonesty and Improve Regulation

Reducing the regulatory burden is a key objective for many government ministries -but how can this be achieved while maintaining honest and ethical behaviour

The following is a summary of the full Hebrew analysis. Link to the Hebrew report.

Abstract

Reducing the regulatory burden is a key objective for many government ministries, and in recent years, steps in this direction have been taken in Israel and abroad. However, in some cases, an easing of regulatory processes is liable to expose the public to risks, given the possibility that those who are regulated (citizens, entrepreneurs, and business persons) might behave in undesirable and unethical ways, in the absence of any oversight or supervision. This problem can be addressed by the use of behavioral tools as a substitute or supplement to the more traditional methods of enforcement. This approach adopts principles of “responsive regulation,” which adapt the demands of regulation to the entity being regulated. . One of these behavioral tools is the use of declarations (certified by an attorney) or honesty pledges (signed by the entity being supervised, such as the individual applying for a business license) who makes a commitment in advance to comply with regulatory directives and rules, and in return- is granted a significant easing of the bureaucratic process. For example, under the Ministry of Interior's business license reform, certain types of low-risk businesses can receive a temporary permit to open, on the basis of a declaration that includes advance commitments. . Behavioral economics research has shown that when individuals are asked to make an ex-ante declaration, they will behave ethically, and their tendency to file false reports drops sharply. The use of declarations or pledges can thus achieve two goals simultaneously: Reducing the regulatory burden on those being supervised, and deterring them from behaving unethically.

But whereas the potential for reducing the regulatory burden—for example, a switch from requiring the submission of certifications, to relying on advance pledges —seems clear and is not controversial, the actual influence of such pledges on the extent of potential unethical behavior remains uncertain. In addition, there is a need to investigate and determine under which conditions, if any, pledges can reduce unethical behavior and justify curtailing the use of supervisory tools, enforcement, and punishment, while still protecting the public interest. Moreover, to date, no research in Israel or anywhere in the world has studied the long-term effect of pledges and the possibility of using this tool on a regular basis, without harm to public safety or the public interest. It is also important to identify s population groups among whom pledges might be especially effective and also those where it may be ineffective, or especially e, all, mainly in the context of a prior tendency to compliance with the rules.

The present document describes two large-scale experiments conducted under laboratory conditions to study these questions. Both revealed that pledges indeed reduce unethical behavior significantly in comparison with traditional sanctions (fines). They also found that the influence of pledges is consistent and does not diminish over time. The document provides details of the implications for public policy of the use of pledges in the real world. We recommend that an extensive field study be conducted in order to examine the feasibility of using pledges, and their influence on other aspects of regulation, such as in restoring trust between inspectors and those they monitor, as a way to bolster government agencies’ ability to reduce the regulatory burden, encourage growth, and enhance the public welfare in the future.

Summary and Findings

This document presents innovative research that was- the first to conduct an integrated and systematic study on the effect of ex-ante pledges on actual ethical behavior that leads to real and long-term benefits. The research was conducted in such a way as to make it possible to generalize its findings to apply to some parallel real-world situations. For the purposes of the research we conducted two experiments. Here we present the conclusions of the research and both its implications and limitations.

1) Pledges Significantly Reduce Dishonesty

The drop in dishonesty subsequent to a pledge (in contrast to a situation in which a pledge is not required) was statistically significant and meaningful from a practical standpoint. as well. In the first experiment, the pledge reduced cheating by about one-third and had the same effect (in the absence of a pledge) as the threat of a large fine that would deprive the participant of all profits if caught cheating. In the second study, the effect of the pledge was even greater (reducing cheating by more than half); in this case too, the effect did not disappear when there was a threat of fines. Similar findings emerged from an examination of the rate of failure on the question that was used as a control in both experiments: The ex-ante pledge markedly reduced the percentage of subjects who were caught cheating on this question. This reduction was stable over time. These findings, which are evidence of the stability and strength of the effect of pledges, are especially important. Had we detected only a weak or relatively unstable effect of pledges, it would have been difficult to justify their use as a means to reduce the regulatory burden of inspection and enforcement. Note that the pledges examined in these experiments were of the “weaker” variety, as compared to those with legal pledges that support the filing of criminal charges, when necessary. Consequently we might expect that actual pledges and declarations would have a greater deterrent effect, which implies that we should relate to the study’s findings as conservative estimates of the real effect of declarations and pledges, when backed by legal tools. Nevertheless, in order to obtain more precise estimates of the expected influence of pledges and declarations, it is important to plan and carry out formal and systematic field studies of the actual behavior of those subject to regulation. Such studies would allow policymakers to weigh the use of this tool responsively and effectively in diverse cases.

2) Pledges and Fines Applied in Tandem Can Reduce Cheating

The first experiment found that the combination of a pledge and a fine resulted in the largest drop in dishonesty (about 70%), compared to the control group. That is, the combination of a pledge and a fine has an aggregate effect, and the existence of one does not eliminate or significantly diminish the other’s impact. In the second experiment (conducted in Hebrew with Israeli subjects), the addition of a fine reduced the pledge’s impact only slightly, and the reduction was not statistically significant. This is an important finding, since one might have hypothesized that the possibility of a (relatively large) fine might totally suppress the effect of the pledge through a "crowding-out" mechanism, in which the existence of an external motivation reduces the impact of the internal motivation (Frey and Oberholzer-Gee, 1997). But we found just the opposite: The pledge and the fine functioned independently and in the aggregate to reduce cheating; the combination of the two was the most effective with regard to the final result. Only in the second experiment did we find weak evidence of a negative impact of the fine on the pledge’s effectiveness, but as stated- it was not statistically significant.

With regard to policy, the combination of pledges and fines can appease contradictory voices on current regulatory policy around the world. Because regulators are by their nature risk-averse, and because public trust in regulators is linked to the degree of their willingness to impose heavy punishments, it is certainly possible that this combination might also have institutional advantages that are compatible with responsive regulation as well as with the principle attributed to Theodore Roosevelt, “speak softly and carry a big stick.” In their ongoing daily practice, those regulated would behave on the basis of their pledges, thereby increasing the regulator’s ability to trust their self-reports. At the same time, those regulated would know that a monitoring mechanism is being applied, that may examine their behavior at random and punish them severely should they violate their pledge. Another possible interpretation of the findings might lead to the opposite conclusion: Because the use of fines requires an investment of resources for enforcement, their relative efficiency is less than that of pledges, which require only modest resources if any; and so, it is possible that even though the combination of a pledge and a fine would produce a greater reduction in cheating in absolute terms, given cost-benefit considerations the use of pledges alone might produce a better outcome.

3) The Effect of Pledges Does Not Wane Over Time

In neither experiment did we find evidence of an increase or “compensation” in the level of cheating as time elapsed after the pledge was made. In fact, both experiments found that the cheating patterns of participants in all groups remained consistent throughout the series of problems, and their inclination to cheat did not change significantly over time in either direction. That is, it seems that the main effect of the pledge is on the initial level of cheating; once this is set, individuals continue to behave in a similar fashion over time. This conclusion is especially important in cases in which pledges cannot have a significant deterrent value in and of themselves. In general, this relates to an innate drawback of ethical nudges, such as pledges, whose effectiveness is expected to decline more rapidly over time. In contrast to this idea, in the two experiments we conducted we found no evidence of a decline in the effectiveness of the pledge. However, it must be acknowledged that the experiments’ time frame was relatively short. Furthermore, the time elapsed between the signing of the pledge and the opportunity to behave unethically was not varied. We recommend that these aspects be studied in future research.

4) There Were Differences in the Influence of Pledges

The first experiment studied the variance in the influence of pledges among individuals with different tendencies towards compliance with the rules; to put it schematically, between “good people” and “bad people;” that pledges mainly reduced cheating by those who are relatively “good” (near and above the median on the compliance scale), whereas a pledge had no statistically significant effect among the “bad” people who were at the lower end of the compliance scale. On the surface, the finding that pledges were ineffective precisely among individuals with the greatest potential to cheat, corroborates the prevalent fear among those opposed to soft regulation. By contrast, the use of disincentives (fines) had a similar effect on all types of individuals -another advantage of the use of traditional forms of regulation. The finding on the heterogeneity of public preferences increases the need, described in the introduction, for responsive regulation. Such regulation can distinguish between situations in which most people have an inherent tendency to comply, among whom we can be content with only a pledge (because of the high ratio between effectiveness and costs), and situations in which most people are not inclined to comply, among whom it would make sense to invest resources in enforcement and heavier punishment.

5) The Use of Pledges is Perceived as Attractive and Desirable

In both experiments, participants who were asked up front to pledge ethical behavior displayed a stronger preference to make such a pledge in a future experiment. These findings also coincide with those of a survey conducted in the past by the Israel Democracy Institute, which found that 92% of business owners in Israel support a shift to automatic permits ("fast track"), and that 75% support the issuing of a permanent license, even if this is accompanied by more stringent enforcement and punishment (Pe’er, Tikotsky, and Feldman 2018). Even though these findings are based only on correlations, they suggest that the use of pledges can have a long-term effect on the relations between regulators and the regulated, and that pledges can serve as initial confidence-building measures. Nevertheless, one must look into whether the declared preference, found in both experiments, is indeed translated into actual choices, and how, if at all, it influences behavior over time.