Pay and Performance in Major League Baseball

1. Pay and Performance in Major League Baseball Gerald W. Scully The American Economic Review, Vol. 64, Issue 6, Dec. 1974

2. Background Baseball bringing in substantial revenue. Salaries are the main part of contract negotiations and the reserve clause gives considerable power to teams. • First players’ strike over pension fund.1972 • Threat of strike, lockout from issues of minimum salaries, salary cuts, benefits, and the reserve clause. 1973 • Reserve clause retained a player’s rights to his team after contract’s expiration. Abolished in 1975 Reserve clause system ruled the game. It dictated how and where players moved between teams. Owner say clause is “necessary for equalization of playing strength among teams” and insist that player salary demands are unrealistically high.Scully disputes these contentions.

3. “Purpose of this paper is to crudely measure the economic loss to the players due to restrictions of the reserve clause. Relationships between player performance and salary will be estimated and the predicted salary compared to predicted player marginal revenue product.” Scully, p. 915

4. Approach • The individual player is the unit of observation. Players have various performance levels and career lengths • Salary functions are estimated within a monosponisticlabor market characterized by extensive bargaining • Marginal revenue factors are estimated explicitly • Exploitation found by comparing salary and MRP • Outline characteristics of the labor market • Formulate an empirical marginal revenue product and salary determination model • Present results • Calculate monopsonisticexploitation • Discuss policy implications “Aside from providing insights into the operation of a particular labor market, the approach differs from other studies of the wage determination process in the following ways: 1) the individual player is the unit of observation; 2) salary functions are estimated within the framework of a labor market characterized by extensive bargaining; 3) marginal revenue products of the factors of production are estimated explicitly: and 4) rates of monopsonistic exploitation are found by comp-aringsalary and marginal revenue product over various performance levels and career lengths. “First, the institutional characteristics of the base-ball players labor market are briefly outlined. Next, with this institutional framework in mind, a model of marginal revenue product and salary determin-ationis formulated. Then, the results of the empirical investigation into marginal revenue product and salary determination are presented. Finally, the rate of monopsonistic exploitation is calculated and policy implications discussed.” Scully, p. 915

5. Organization of the Baseball Players Labor Market (more about the reserve clause and MRP) • Player is a free agent until a one year contract is signed • Contract is exclusively renewable for the owner (reserve clause) • Owner may renew, sell, or terminate the contract • Player make accept the terms or negotiate • After owner’s “final offer,” player must accept or withdraw from professional baseball • MRP is “the ability or performance he contributes to the team and the effect of that performance on the gate receipts. The effect of a player performance on revenue may be direct or indirect.” • “Ability contributes to team performance and victories raise gate receipts—and broadcast revenues; this is the substantial effect of the individual’s performance.” • “Some players may attract fans over and above their contribution through the team.” Scully, p. 916

6. A Simple Model of MRP and Salary Determination in Major League Baseball The MRP and salary determination model should consider: 1) revenues are related to individual performance through their effect on team standing; 2) the reserve clause reduces player salaries below player MRP • W = Percent wins • Ai = Player skills • Ii = Nonplayer inputs (managers, coaches, capital, etc.) W = W ( A1, A2,…, An; I1, I2…Im)

7. A Simple Model of MRP and Salary Determination (cont.) Teams derive revenue from gate receipts and broadcast rights, which are directly related to the team’s percent wins and population in the area • R = Team revenue • p= Ticket price • T = Number of tickets sold • W = Team performance • Pi= Potential broadcast households • B = Broadcast revenues • W = Percent wins • Ai = Player skills • Ii = Nonplayerinputs (managers, coaches, capital, etc.) R = p · T [ W ( Ai, Ij), Pa ] + B [ W ( Ai, Ij), Pb] i = 1,…, n; j = 1,…, m)

8. A Simple Model of MRP and Salary Determination (cont.) Costs are determined by the level of skill and nonplayer inputs. Because of the monopsonistic labor market, costs are related to the level of skill • C = Team costs • Si (Ai ) = Player supply functions • rj = Nonplayer factor remunerations C = ∑ Ai Si (Ai ) + ∑rjIj i = 1,…, n; j = 1,…, m)

9. A Simple Model of MRP and Salary Determination (cont.) Standard profit model • = Profit • R = Team revenue • C = Team costs = R - C

10. A Simple Model of MRP and Salary Determination (cont.) First-order conditions for a maximum profit are obtained by differentiating with respect to Ai (player skills) and Ii (nonplayerinputs) “Conditions reveal that teams maximize profits by selecting a level of player skills and nonplayer inputs such that players receive a salary equal to their MRP less monopsony rents (Ai S’i).” Scully, p. 917

11. Player MRP Equation1 MRP shall be determined by 1) player performance on team wins and 2) the effect of wins on revenue. Estimated with team data for 1968 and 1968 • = Team percent wins · 1000 • TSA = Team slugging average • TSW = Team strikeout-to-walk ratio • NL = Dummy variable; 1 = the team is in the National League • CONT = Dummy variable; 1 = divisional winners and their closest competitors at the end of the season, if competition was five or fewer games out • OUT = Dummy variable; 1 = teams which at the end of the season were twenty or more games out of placing PCTWINt = 37.24 + .92 TSAt + .90 TSWt - 38.57 NL + 43.48 CONTt - 75.64 OUTt, (.39) (4.37) (5.62) (4.03) (3.77) (6.17) R2 = .88, DF = 38;t = team

12. Player MRP Equation 2 MRP shall be determined by 1) player performance on team wins and 2) the effect of wins on revenue. • = Home attendance times average ticket price plus revenue from broadcasting rights • SMAS = Population size of the Standard Metropolitan Statistical Area • MARGA = Team specific attendance • NL = Dummy variable; 1 = the team is in the National League • STD = Dummy variable; 1 = older stadiums in poor neighborhoods and with limited parking facilities • BBPCT = Percentage of black players on the team REVENUEt = - 1,735,890 + 10,330 PCTWINt + 494,585 SMAS70 - 512 MARGA (1.69) (6.64) (4.61) (4.28) + 580,913 NL CONTt- 762,248 STDt - 58,526 BBPCTt (1.84) (2.42) (3.13) R2= .75, DF = 36; t = team

13. PCTWIN • The coefficient of primary interest in PCTWIN. A one pint increase is estimate to raise REVENUE $10,330. A one point increase in TSA (slugging avg.) or TSW (strike-to-walk) raises PCTWIN by .92 and .90 • However, MRP may be overstated. Omitted factors may include managerial quality, entrepreneurial player drafting and trading abilities, and stadium investment

14. Individual Player’s MRP • “To obtain the individual player MRP, assumptions have to be made about how individual performance affects team averages.” Scully, p. 921 • Assume individual performance carries no externalities (team performance is simply the linear summation of individual performance) • Gross MRPs show that even mediocre players contribute $200,000 to team revenue. Star players easily have MRP double that amount

15. Salary Determination • Is offensive or defense more important? • Hitting or pitching performance • Weight of contributions • Star players

16. Salary Determination (Hitters) • SA= Slugging average • M = Years experience in the majors • DBA= Dummy variable, 1 = Players with below average lifetime Slugging averages but above average lifetime batting averages. Controls for excellent hitter who have low slugging averages. • AB = Total lifetime at bats / number of years in the majors times 5,500 (average season at bats for a major league team) • SMSA, MARGA, and NL to control for monopoly Salary sample is nonrandom and biased toward the upper tail of the salary distribution. There are 148 observations from 1968 and 1969. Salary range is from $10,000 to $125,000

17. Salary Determination (Pitchers) • SW = Lifetime strikeout-to-walk ratio • M= Years experience in the majors • IP= lifetime average percentage of inning pitched out of a total innings • SMSA, MARGA, and NL to control for monopoly Salary sample is nonrandom and biased toward the upper tail of the salary distribution. There are 148 observations from 1968 and 1969. Salary range is from $10,000 to $125,000

18. Degree of Monopsonisitic Exploitation of Professional Ball Players • Players are generally exploited and to a large degree • Over a career length, average players receive salaries equal to about 11% of their gross MRP and 20% of their net MRP • Star players receive about 15% of MRP • Some mediocre players have salaries in excess of their net MRP

19. Policy Implications • Place ball players under a contractual arrangement similar to pro football. A player can play out his option and become a free agent. Player-initiated transfers would still not exist • Long-term contracts. Contracts for 5+ years, followed by free agency • Completely free labor market will all contracts for a full season negotiated off-season. Eliminate economic rents of players.

20. Questions • What is the reserve clause? • What is a monopsony? • In the win model, what are the two independent variables? • Player MRP is determined by 1) the effect of A on B, and 2) the effect of B on C. What are the variables? • What two categories of players were evaluated for salary determination? • Are professional baseball players exploited?