«Forthcoming, American Economic Review Abstract We conduct an experiment assessing the extent to which people trade off the economic costs of ...»
In the main text, we found that PV serves as a useful measure to organize the data. Here, we provide a quantitative interpretation of the findings for PV. (The other candidate measures for intrinsic costs of lying were insignificant.) For PV, a useful way to interpret the results quantitatively is offered by slightly rewriting the preference specification in equation (1). Let Vi T PVi bm ECOST bECOST 1 k i 1 T i 1 T . The difference between the utilities of truthtelling and of lying is here given by Yi * i bECOST 1 k i . (A1) In this formulation, the role of intrinsic costs of lying is split up even more explicitly than in the formulation in the main paper, although they are mathematically isomorphic. Here, first, protected values generate moral costs of lying PVi i. A second role of protected values is that the actual marginal utility an agent ascribes to monies depends on the process by which the funds were obtained and on the agent’s moral evaluation of that process. In the present context, a dollar obtained by an act of lying may be regarded as less valuable than a dollar obtained by telling the truth. This is reflected in a preference parameter k PVi k i. An agent discounts funds obtained by lying if ki is positive.
We posit for parsimony the simple linear parametric specifications i 0 1 PVi and
is unknown. One way to make progress is to determine (speculatively) ki at some PV level. As a perhaps reasonable benchmark, assume that k is zero for an individual with an average PV level (3.82). From this, we can infer the implied b = 2.91. (Recall that we are assuming that b is independent of protected values.) This in turn then allows us to plug in a range of PV levels to obtain the implied lying discounts (or premia). (Another approach is to simply posit values for b and calculate ki accordingly.) Table S-I shows the results of these calculations. 95% confidence intervals are shown in brackets below the point estimates. For the moral cost of lying, i, the estimates suggest an average value of around unity. If the average participant is indifferent to the process by which funds are obtained, our estimate of b is 2.91. For the discount parameter, ki, our experiment then implies that an individual with a protected value in the 75th percentile will discount funds acquired dishonestly by about 21%, with even the lower bound of the 95% confidence interval being about 8%. Thus, the trade-off resistance implied by protected values is not only a statistically, but also an economically, significant factor in decision making.7 Of course, assuming that the average ki is zero implies that an economic model using the proposed preference specification becomes interesting mostly when there is variation in protected values across agents.
Table S-I: Quantitative implications for preference parameters
Note that factors reflected neither in the model nor in the experiment may shift the attractiveness of truthfulness. For example, an anticipated loss of reputation in case of cheating may make lying less attractive at a given cost of truthfulness, thus adding to the personal moral cost of lying. These external factors would, therefore, complement and enhance the power of the discount factor ki.
Appendix references Ajzen, Icek and Martin Fishbein. 1980. Understanding Attitudes and Predicting Social Behavior. Englewood Cliffs: Prentice-Hall.
Ariely, Dan; Anat Bracha and Stephan Meier. 2009. "Doing Good or Doing Well? Image Motivation and Monetary Incentives in Behaving Prosocially " American Economic Review, 99(1): 544–55.
Musch, Jochen; Robbi Brockhaus and Arndt Bröder. 2002. "Ein Inventar zur Erfassung von zwei Faktoren sozialer Erwünschtheit." Diagnostica, 48(3): 121-29.
Paulhus, Delroy. 1984. "Two-component models of socially desirable responding." Journal of Personality and Social Psychology, 46: 598-609.
Tanner, Carmen; Bettina Ryf and Martin Hanselmann. 2009. "Geschützte Werte Skala:
Konstruktion und erste Validierung eines Messinstrumentes (Protected Values Measure:
Construction and first validation of an instrument to assess protected values)." Diagnostica,