Do Tournaments with Superstars Encourage or Discourage Competition?

Abstract

To test and replicate the superstar effect reported by Brown (2011) we empirically study contests where a single entrant has an endogenously higher probability of winning. Unlike the previous literature, we test for the presence of the superstar effect in several different contexts. Ultimately, we collect and explore data from four sources: men's and women's professional golf, and men's and women's professional alpine skiing. Our baseline study of men's professional golf serves as a replication of Brown's (2011) study. Empirically, we find little robust evidence of the superstar effect in any of our datasets. In our replication exercise, we approximate the findings of Brown (2011), however, we cannot reject the null that the presence of a superstar has no impact on high ranked competitors. In our other settings, we cannot reject the null that superstars have no influence on the performances of highly ranked competitors.