Risk Preference, Correlation Choice, Sabotage, and the Design of Promotion Tournaments

I examine properties of worker behavior under promotion tournaments, and discuss their implications for the design of promotion tournaments. Given their tasks, workers will decide how much effort they exert. In addition to this decision-making, they have the opportunity to decide their approaches to perform their tasks within their delegated authority. I define the term approach as a method, way, procedure, plan or project to perform his task, into which a worker infuses his effort. Through their approach choices, they can control the riskiness and correlation of their performances. It is shown that under the loser-selecting tournament, which means the promotion ratio is more than one-half, workers prefer a low risk approach or a common approach which peers also know well; on the other hand, under the winner-selecting tournament, which means the promotion ratio is less than one-half, they prefer a high risk approach or their own original approaches. These results suggest that the loser-selecting tournament is more efficient than the winner-selecting tournament in terms of the cost for implementing high efforts of risk-averse workers. I rigorously show this in the case of three workers. I argue that the winner-selecting tournament is better suited for the upper job levels or for firms in innovative and immature markets or industries; on the other hand, the loser-selecting tournament is better suited for the lower job levels or for firms in stable, mature, or strictly regulated markets or industries. Furthermore, I investigate properties of sabotage under tournaments. Under the one-winner tournament, each worker attacks his peers so as to minimize the maximal value of their expected performances. The sabotage operates to create more homogeneity. Under the one-loser tournament, each worker intensively attacks one peer so as to minimize the minimal value of his peers' expected performance. One worker is intensively attacked by all peers. I argue that the one-loser tournament is more subject to damage by sabotage than the one-winner tournament.