Self-serving bias across strategic and non-strategic dictator games with production

ABSTRACT

This paper investigates behaviour across strategic and non-strategic dictator games with a preceding production phase. In both treatments subjects remain paired and play a trust game immediately following the dictator game. In the strategic condition subjects are informed about the subsequent trust game, while in the non-strategic condition they are not. Dictators in our non-strategic condition display a self-serving bias. On the other hand, dictators in our strategic condition are more generous and display no self-serving bias. Despite the increased generosity, transfers and earnings in the trust game are lower in the strategic condition. Results suggest that generosity, perceived as having a strategic motivation, can undermine trust.

KEYWORDS:

• Real effort
• dictator
• distributive preferences
• trust

JEL CLASSIFICATION:

• C91
• C72
• D63

Acknowlegments

Funding generously provided by the University of Massachusetts Lowell is gratefully acknowledged.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.

Notes

¹ Empirical evidence regarding self-serving bias is mixed. Cappelen et al. (2007) find no evidence that participants were systematically self-serving. However, contributions towards production in Cappelen et al. (2007) is varied through rates of return on individual investments rather than through real effort tasks, in contrast to Ubeda (2014) and the design presented in this paper.

² Experimental instructions available in supplementary online materials.

³ Throughout the paper, unless otherwise noted, two-sided t-tests and Wilcoxon ranksum tests are used with 66 and 62 observations evenly split across dictators and receivers in Partner Known and Partner Unknown, respectively.

⁴ To provide further support we regress the proportion of the endowment that the dictator kept on a Partner Unknown indicator, the size of the endowment generated, and indicators for whether the dictator was more, less, or equally productive as the receiver. The Partner Unknown indicator is positive and significant $\left(0.091\left(0.045\right)p=0.047\right)$. Neither the size of the endowment nor the productivity indicators effect the proportion kept. Details available in the online supplementary materials.

⁵ To provide further support we regress the transfer on a Partner Unknown indicator and the proportion of the endowment kept by the dictator. The Partner Unknown indicator is positive and significant $(2.68\left(0.849\right)p<0.01)$ and the proportion kept is negative and significant $(-6.76\left(2.34\right)p<0.01)$. Details available in the online supplementary materials.

⁶ However, results in a regression controlling for the transfer amount suggest that the Partner Unknown indicator is not significant. Details available in the online supplementary materials.

Additional information

Funding

This work was supported by the University of Massachusetts Lowell [SHE12994].