Elsevier

Economics Letters

Volume 120, Issue 1, July 2013, Pages 53-56
Economics Letters

Third party assessments in trust problems with conflict of interest: An experiment on the effects of promises

https://doi.org/10.1016/j.econlet.2013.03.028Get rights and content

Highlights

  • We study the effect of promises and incentives on truthful assessments in a 3-player trust game.

  • An assessor assesses the trustworthiness of his friend (i.e. the trustee).

  • Trustors can condition their choice on this assessment.

  • Assessor’s promises to give a truthful assessment sign reduce favorable assessments.

  • Promises reduce misreporting to the same extent as incentivized assessments.

Abstract

Is it possible to elicit reliable assessment from an assessor having a conflict of interest (e.g. a professor that writes a recommendation letter for a formal PhD student)? We propose an experimental test and show that compared to a not-incentivized assessment, a promise to give a truthful assessment reduces misreporting to the same extent as an incentivized assessment (i.e. when the assessor gains higher payoff if the assessment is correct).

Introduction

Recruiting new employees is an unavoidable task for firms, governments as well as universities. One difficulty is that information about the applicants’ abilities is usually limited. In many universities, taking an example, it is common to ask for a recommendation letter by a professor that has worked together with the applicant. For such information to be useful, honest reporting is required. Some professors, however, develop a friendship with their Ph.D. students, which creates a conflict of interest and the risk that the information is biased in favor of the applicant (Leising et al., 2010). Assessments play a crucial role in many job markets to inform the employs about the applicants’ abilities, and biased recommendations may have serious implications: Exaggerating the applicant’s abilities increases the risk that the employer’s expectations remain unfulfilled, often to the disadvantage of other, better candidates.

In this study we test a simple mechanism to overcome such biased assessments. We observe the behavior of Assessors in three treatments: when they are not (monetary) incentivized to tell the truth, when they have monetary incentives to tell the truth, and when they are not incentivized but sign a statement of honesty (oath henceforth).

There is evidence that promises work. Charness and Dufwenberg (2006) experimentally examine the impact of communication on trust and cooperation. They suggest that a promise works because of guilt aversion: A guilt-averse person does not want to let down others’ expectations and is therefore motivated by beliefs about others’ beliefs. An alternative explanation is that people may have a taste for keeping their word (e.g. Ellingsen and Johannesson, 2004). Using a novel design, Vanberg (2008) found support for the latter explanation, i.e. people have preferences for promise-keeping per se.1 Note that free form communication was used in all three cited studies. Interestingly, in Charness and Dufwenberg (2010) the free-form communication is replaced with an opportunity to send an exogenously given message of promise, and limited support for lying- and guilt aversion is found.

Jacquemet et al. (2013) explore the impact of oaths in an incentive-compatible second price auction. The oath treatments presented subjects with the opportunity to sign an oath prior to participating in the auction. By signing the oath, subjects “swear on their honor” to tell the truth and provide honest answers. Subjects who took the oath were on average less likely to either overstate or understate their bids. Carlsson et al. (forthcoming) tested an oath-like procedure in the field using non-market valuation surveys and found that the share of zero WTP responses and extremely high WTP responses decreases, which could be interpreted as reduced dishonesty. Shu et al. (2012) find that signing a statement of honesty at the beginning instead of at the end of a self-report serves as a commitment and leads to significant reductions in misreporting.2

We extend the existing literature on promises and oaths by asking for statements not about own intentions, but about information concerning the trustworthiness of friends.

Section snippets

The trust game and treatments

In this study we extended a binary trust game to include an additional third player, called the Assessor. The game starts with the Assessor, who has private information about the trustee, because the Assessor and the Trustee are friends and know each other prior to the experiment, which is common knowledge. The Assessor has to assess whether the Trustee will later return the trust or not (i.e. give a positive or negative assessment of the Trustee’s trustworthiness).3

Results

The results are presented starting with the choices of the Assessors, followed by the Trustors’, and the Trustees’ choices. Fig. 2 gives a visual impression of the assessments made by the Assessors.

The share of Assessors giving a positive assessment of their friend’s trustworthiness is greatest in treatment NotIncent (89% positive assessments). Incentivizing honest reporting in the Incent treatment significantly reduces positive assessments to 56% (z=3.13,p<0.006).5

Concluding discussion

The truthfulness of an assessment by a professor considering the abilities of a former Ph.D. student, or an employer about a former employee, are two applications highlighting the importance of reliable communication in economic interactions. Relying on such assessments involves the risk that assessments may be biased in favor of the applicants, and hired applicants could turn out to be unreliable themselves.

In this study, we augmented a trust game to test if elicitation under oath can mitigate

Acknowledgments

Helpful comments by an anonymous reviewer and the help of Adrian Liebtrau, Andreas Matzke, Lisa Schöttl and Ria Stangneth in conducting the experiments is gratefully acknowledged.

References (11)

  • G. Charness et al.

    Bare promises: an experiment

    Economics Letters

    (2010)
  • N. Jacquemet et al.

    Preference elicitation under oath

    Journal of Environmental Economics and Management

    (2013)
  • Carlsson, F., Kataria, M., Krupnick, A., Lampi, E., Löfgren, A., Qin, P., Sterner, T., 2013. The truth, the whole...
  • G. Charness et al.

    Promises and partnership

    Econometrica

    (2006)
  • T. Ellingsen et al.

    Promises, threats and fairness

    The Economic Journal

    (2004)
There are more references available in the full text version of this article.

Cited by (0)

View full text