Compensation schemes when workers have different characteristicsthree experimental essays /

Resum

Summary This dissertation is composed by three chapters that may be summarized as follows: In the first chapter of the thesis we present results from three-player experiments aimed at studying distributional concerns in how owner-managers compensate themselves and workers of different productivities and effort costs. We use a game in which workers first exert effort and owner-managers then decide on bonuses for themselves and workers. Our motivation is to contribute to understanding both the vertical and the horizontal dimensions of the distributional concerns of owner-managers. We are also interested in how owner-managers decisions’ are affected by pay secrecy. Our design includes four treatments: 1) different productivities of workers with complete information; 2) different productivities of workers with pay secrecy among workers; 3) different effort cost of workers with complete information; and 4) different effort cost of workers with pay secrecy among workers. We find that, on average, managers do not pay relative wages in accordance to relative production levels. In our data about 1/3 of additional production translates into additional compensation. The equalizing tendency of managers’ compensation policies together with the fact that high-productivity workers exert more effort in all cases leads to ex post similar income levels among workers. We also find that pay secrecy does not affect wage differences among workers. Across all treatments about 50% of all manager choices are compatible both with ‘production equity’ and with effort- cost equity. In the second chapter, we present an experimental analysis that investigates compensation policies and its effects when workers differ in its distribution impact on profits. Specifically, we set three types of workers according to their impact in profits. We introduce one standard worker (routinely used in gift-exchange experiments); a star worker who outperforms the standard worker by doubling its productivity for every level of effort; and a guardian worker who generates losses unless he/she exerts high effort. Managers will face combinations of these workers and set compensations. Our objective, hence, it is to analyze the determinants of different types of workers on wage setting decisions. We observe that guardians get more compensation in relation to the soldier than stars do since managers seek to avoid negative production. Moreover these differences change the effort decisions by the negatively discriminated worker when a guardian and a star are paired together, but it does not change the effort decision of the positively discriminated worker in any situation. We interpret this as evidence of unacceptable unfair wage distributions and give a reason for its cause and prevalence. In the third chapter, we study the incentive effect of firing threats when bosses have limited information about workers. We show that, regardless of the amount of information possessed by the boss, firing threats substantially boost workers’ production and reduce on-the-job leisure. Even when the boss has no individual information about workers’ effort and production levels, firing threats have strong incentive effects. Any minimal amount of individual information about workers individual effort such as the time they spend at their work station is sufficient to ensure strong incentive effects.