Golf Match: The Choice by PGA Tour Golfers of Which Tournaments to Enter

Article excerpt

Abstract

This paper compares two methods of examining the entry choice of professional golfers, focusing on the size of the purse, the strength of the competition, and a newly constructed variable, the match of the player's skills with the skills rewarded at each tournament, while controlling for some dynamic factors such as year end pushes to cross relative earnings thresholds. Logit regressions are one method of examining the entry choice. A second method exploits combinatorial arithmetic. Choosing which n of N tournaments to play is equivalent to choosing n balls without replacement from an urn with N balls. The results show that golfers choose tournaments with higher purses, with a better skills match, and when the competition is fiercest.

Keywords: tournament compensation, intertemporal labor supply, PGA Tour

Golf Match: The Choice by PGA Tour Golfers of which Tournaments to Enter

The issue of intertemporal labor supply has been addressed in an interesting variety of settings including taxi cab drivers (Camerer et.al., 1997; Farber, 2005), stadium vendors (Oettinger, 1999), bicycle messengers (Fehr & Goette, 2007), and professional golfers (Hood, 2006; Rhoads, 2007a). Various aspects of labor supply have been examined, including: how often to show up for work, which days to show up, how long to stay at work once showing up, and how much effort to apply. The varying results of these inquiries with respect to labor supply elasticity has pointed out that careful attention must be paid to the demand and supply shift variables so that the elasticities can be properly identified. The abundant and high quality data available pertaining to professional golfers on the PGA Tour allow several interesting margins to be examined.

The labor supply issue in golf has been addressed from a variety of perspectives. Starting with Shmanske (1992), several economists (Moy & Liaw, 1998; Nero, 2001; Shmanske, 2000, 2007, 2008; Rishe, 2001; Alexander & Kern, 2005; Callan & Thomas, 2007) have estimated production functions wherein tournament scores are a function of the golfer's skills, or earnings functions wherein earnings are a function of the golfer's skills. Thus, the supply in different dimensions of talent leads to the performance or the earnings. Going one step backward in this supply chain, Shmanske (1992) also looked at the production, development, and maintenance of the skills themselves where the input supply is the golfer's practice time. Ehrenberg and Bognanno (1990a, 1990b) consider the golfer's supply of effort, especially in the final rounds of competition. Finally, Gilley and Chopin (2000), Hood (2006), and Rhoads (2007a, 2007b) study what might be called regular old supply, that is, the number of tournaments entered by professional golfers in a given year. The focus in these studies is on supply elasticity in order to determine whether a golfer's individual supply curve could be backward bending.

The research in the current paper draws a little from each of these traditions, but perhaps can be seen most closely as a complement to Rhoads's recent papers.Whereas Rhoads looks at the number of tournaments entered by a golfer, I hold that result constant and attempt to answer the question of which tournaments a given golfer will enter. In a nutshell, the paper does the following. By looking at the statistics from individual tournaments, it can be determined which skills, for example putting or driving, are rewarded more heavily on a tournament by tournament basis. By looking at the statistics for each individual golfer over the course of a year, it can be determined how much of each skill a particular golfer has. By matching the skills that the golfer has to the skills required in each tournament, an expected performance can be calculated for each tournament. These expected performances can be ranked from 1 to N, and if the golfer chooses to play in n of the N tournaments, the ideal set of which n tournaments the golfer should choose can be determined. …