We conduct a natural field experiment in a retail chain to test predictions of tournament theory regarding prize spread and noise. A random subset of the 208 stores participates in two-stage elimination tournaments. Tournaments differ in the distribution of prize money across winners of the first and second rounds of the tournament. As predicted, we find that a more convex prize spread increases second-round performance at the expense of first-round performance, although the magnitude of these effects is small. Moreover, the treatment effect is larger for stores with more stable past performance.

We ran a field experiment in a Dutch retail chain consisting of 128 stores. In a random sample of these stores, we introduced short-term sales competitions among subsets of stores. We find that sales competitions have a large effect on sales growth, but only in stores where the store’s manager and a sufficiently large fraction of the employees have the same gender. Remarkably, results are alike for sales competitions with and without monetary rewards, suggesting a high symbolic value of winning a tournament.

Preferences for monetary and nonmonetary job attributes are important for understanding workers' motivation and the organization of work. Little is known, however, about how those job preferences are formed. We study how macroeconomic conditions when young shape workers' job preferences for life. Using variation in income-per-capita across U.S. regions and over time since the 1920s, we find that job preferences vary in systematic ways with experienced macroeconomic conditions during young adulthood. Recessions create cohorts of workers who give higher priority to income, whereas booms make cohorts care more about job meaning for the rest of their lives.