Skidmore College Skidmore College

Creative Matter Creative Matter

Economics Student Theses and Capstone

Projects

Economics

2016

Analyzing Managerial E5ciency in Major League Baseball: A Analyzing Managerial E5ciency in Major League Baseball: A

Sabermetric Approach Sabermetric Approach

Jebediah C. Clarke

https://creativematter.skidmore.edu/econ_studt_schol/18

This Thesis is brought to you for free and open access by the Economics at Creative Matter. It has been accepted

for inclusion in Economics Student Theses and Capstone Projects by an authorized administrator of Creative

Matter. For more information, please contact [email protected].

Analyzing Managerial Efficiency in Major League Baseball:

A Sabermetric Approach

Jebediah C. Clarke

Thesis Advisor: Joerg Bibow

Modern statistical analysis has allowed for teams to more accurately measure Major

League Baseball player performance. However, other than tracing wins there are few ways to

track the performance of on-field managers whose strategies, decisions, and expertise

fundamentally influence the outcome of each game. I begin this paper by investigating and

critiquing prior empirical analyses that have attempted to quantify the effect of managerial skill

on team performance. Using Stochastic Frontier Analysis and data from the 2008-2015 MLB

seasons, I expand on previous research by calculating managerial efficiency estimates while

including control variables that better objectively measure player performance. I find that the

least efficient managers achieve winning percentages that are around 80% of what is possible,

The goal of any manager, in any industry, is to efficiently combine inputs to generate

output. When firms in two industries use the same amount of inputs, yet have different output

in professional sports provides scholars with an advantage in studying managerial impact on

team performance.

The rise of advanced statistical analysis in Major League Baseball has fundamentally

changed the game. Michael Lewis’ (2003) Moneyball: The Art of Winning an Unfair Game,

brought the use of advanced statistics in baseball to the attention of the public masses. Lewis

(2003) wrote the story of the 2002 Oakland Athletics and General Manager Billy Beane— and

how a small-market team used unconventional player-evaluation methods to field a successful

group of players. Beane favored lesser-used statistics like on-base percentage over the more-

traditional counts of a player’s homeruns and runs-batted-in. His wisdom was simple— the

Athletics’ payroll could not afford to acquire high-profile players, so they needed to find an

In Major League Baseball, the on-field leader is called the “manager,” while the NFL, NHL, and

NBA use the term “head coach.”

via hits, walks, or hit-by-pitches. While other teams valued a player’s physical makeup— his

arm strength, bat speed, and 60-yard dash time, Beane could not care less. As Lewis (2003)

writes of Beane’s right-hand-man, Paul DePodesta,

Since then, a new field of statistical analysis has enveloped baseball. Coined “sabermetrics,”

after the Society for American Baseball Research (SABR)— of which many of baseballs’ current

writers and statisticians are members. The abundance of freely-available data has allowed player

(WAR), fielding independent pitching (FIP), and weighted on-base average (wOBA) allows

teams to better objectively evaluate player performance.

While the sabermetric movement has revolutionized how MLB teams evaluate players,

there is still one component of the team that front-offices know little about: the manager. Other

than tracing wins, there are few ways to evaluate the performance of the on-field manager. The

job of a baseball manager is to make decisions that will benefit the team. More specifically, to

pitch calling, baserunning tactics, and hitter and pitcher substitution, among others. Successful

managers are also required to make sure the right player is in the right situation. For example,

certain pitchers are more effective against left-handed hitters, and vice-versa. It is up to the

manager to put his players in situations where they are more likely to succeed. Furthermore,

skillful managers also know the tendencies of their opponent, and thus create a game plan that

implications of managerial performance. Section V offers further areas of possible research.

Section VI summarizes and concludes the paper.

The history of measuring managerial skill in Major League Baseball can be divided into a

few distinct methodologies. The earliest practice, introduced by Porter & Scully (1982), uses

Stochastic Frontier Analysis (SFA) to calculate managerial efficiency in MLB via a production

function consisting of an output (winning percentage) and inputs (player and managerial

performance statistics). This method has been further developed and applied to MLB by

Ruggiero, Hadley, & Gustafson (1996) and Smart, Winfree, & Wolf (2008). In a similar fashion,

Envelopment Analysis (DEA) has also been used to quantify managerial efficiency in

professional sports. Fizel & D’itri (1996) use DEA to measure managerial efficiency in college

basketball. Volz (2009) and Lewis, Lock, and Sexton (2009) apply DEA to managers in MLB,

while Young Han Lee (2009) uses the same method to calculate managerial efficiency in Korean

professional sports. Lastly, Horowitz (1994) and Ruggiero and Hadley (1997) use the

1. Stochastic Frontier Analysis (SFA) Literature

Despite the early knowledge that the role of management was crucial to the overall

production function of a business, very little was known about the actual impact managers had

on output. This was due to the fact that little applicable data existed and the difficulty of

separating the outputs and inputs of a traditional business (Porter & Scully, 1982). However, the

availability of baseball data and knowledge of outputs (wins) and inputs (player and managerial

performance statistics) made measuring managerial efficiency both appealing and feasible.

Porter and Scully (1982) created a Cobb-Douglass production function that compared

wins (output) to inputs of player skills, which measure team hitting performance and team

pitching performance. Team Slugging percentage (SLG) was used as the hitting input and team

strikeout to walk ratio (K/BB%) was used as the pitching input. These statistics were used, over

beyond the control of the starting pitcher.

From the aforementioned production function, Porter and Scully (1982) calculated

management resulted in improved efficiency, at a decreasing rate. Efficiency increased by 0.8%

scores compared to American League teams with similar winning percentages.

While Porter and Scully’s (1982) research was a breakthrough in the evolution of

measuring managerial skill, their methodologies had flaws. Ruggiero et al. (1996) used a

production function with player and managerial inputs to calculate managerial efficiency during

the years 1982-1993. In essence, Porter and Scully (1982) defined player skill as a factor of team

winning percentage but did not offer any insight into how managers may affect team

performance. Results suggest that major league teams can achieve a winning season in two ways.

A team can win with superior player talent while having an inefficient manager, such as the 1990

Cincinnati Reds. Or, a team can win with inferior player talent so long as inputs are managed

efficiently, as proven by the 1987 Minnesota Twins (Ruggiero et al., 1996).

Joe Torre, manager the St. Louis Cardinals and New York Yankees, had highest average

efficiency score (.8809 over 14 seasons). Jim Riggleman of the Chicago Cubs and San Diego

greatest impact on efficiency, and what qualities teams are paying their managers for. Total

offensive resources (a weighted version of on-base percentage, plus net stolen bases) was used as

the offensive/hitting input and team earned run average was used as the defensive/pitching input.

winning percentage, years of MLB managerial experience, and number of MLB teams managed.

The managerial change group included dummy variables for first year managing current team,

in-year managerial change, and manager’s first year in MLB. Interestingly, only the managerial

change inputs were significant at a 10% level. The others were insignificant.

characteristics (measured by the aforementioned qualities). All but the manager’s MLB playing

experience variables were significant at the 5% level. In other words, more efficient managers

often have more managerial experience but do not necessarily need to have MLB playing

experience. The only problem with this conclusion is that it is hard to know if managerial

experience actually contributes to efficiency, or whether managers who are efficient early in their

Padres had the lowest (.7273 over 7 seasons). This can be interpreted as—Joe Torre achieved a

winning percentage that was 88% of what was possible, given his players’ talent level.

Number of seasons managed (at the beginning of the current season) – a measure of managerial

experience. Experience is expected to reflect accumulated expertise in player motivation and

development and in adapting to the movement of players (i.e. injuries, trades, releases) (Smart &

Again, it is expected to increase efficiency and thus have a positive sign.

Manager-of-the-year award – this variable is equal to 1 if the manager has won a MLB manager

of the year award (yearly award given to best manager in both the American and National

Leagues-- as voted on by the Baseball Writers Association of America), 0 otherwise. The award

indicates previous success and management skills. This variable is expected to have a positive

sign.

otherwise. The hypothesis is that a manager who was formerly a player may have a better

In-season managerial change – this variable is equal to 1 if there was a mid-season managerial

change, 0 otherwise. While managers may change during the season for a variety of reasons (i.e.

illness or change in ownership), most managerial change is a result of poor performance.

However, it is unlikely that a interim manager will be a able to improve the team drastically

(Smart & Wolfe, 2003). From 2008-2015, there were 26 mid-season managerial changes. This

NL, so managers often pinch-hit for pitchers late in games. NL managers make more decisions,

and thus, have more opportunities to appear less efficient. This variable is expected to have a

negative sign.

I use STATA to estimate managerial efficiency via the following production function,

where logwin% represents the winning percentage expressed in log form, loghWAR represents

FanGraph’s position-player WAR, logpWAR represents FanGraph’s pitching WAR, logyearexp

represents years of managerial experience prior to the current season, logcareerwinpct represents

managerial career winning percentage prior to the current season, award is a dummy variable for

whether the manager has won a manager-of-the-year award or not, player is a dummy variable

for whether the manager has MLB playing experience or not, change is a dummy variable for

whether the team had a in-season managerial change or not, NL is a National League dummy

variable, and ε is an error term.

(1) (2)

(0.0107) (0.0101)

(0.0106) (0.0101)

(0.00863) (0.00823)

(0.0142) (0.0140)

(0.0602) (0.0582)

Observations 220 220

R-squared 0.806

Prob >= chibar2 0.010

Standard errors in parentheses

logpWAR and loghWAR are significant at the 1% level. Logyearsexp, change, and NL are

significant at the 5% level. Finally, logmanwpct, award, and MLBplayer are insignificant. The

positive and the most efficient manager will have an error term of zero. (p.313)

I can reject the null hypothesis that OLS is the best method to measure this production function

statistically significant at the 1% level, indicating that teams with greater pitching talent tend to

win more, all else equal. In other words, a 1% increase in pitching WAR results in a 0.18%

increase in winning percentage, on average. Similarly, the coefficient for loghWAR is positive

and statistically significant at the 1% level, indicating that teams with greater position-player

talent tend to win more, all else equal. A 1% increase in position-player WAR increases winning

percentage by 0.15%, on average.

Turning to the coaching measures, logyearsexp is positive and statistically significant at

multicollinearity with the logyearsexp variable. Smart et al. (2008) found similar results.

Managers who win more tend to stick around for longer periods of time. Intuitively, this makes

sense— a manager’s retention is decided primarily by his ability to win. This insignificance of

award implies that managers who have won at least one manager-of-the-year award do not win

more than trophy-less managers. The results of MLBplayer suggest that playing experience has

statistically significant at the 1% level. As expected, a mid-season managerial change has a

negative impact on winning percentage. Teams that change managers mid-season likely do so

because of poor performance, and therefore, a lower winning percentage than a team that did not

make a change is expected, on average.

Equation 1 was estimated as follows:

In other words-- Major League Baseball

+-------------------------------------------------------------------------------------------------------------+

---------------------------------------------------------------------------------------------------------------

Mattingly Dodgers 5 .9461148 .0125879

Guillen White Sox, Marlins 4 .9424755 .0183252

Gardenhire Twins 7 .9420212 .0431298

Farrell Blue Jays, Red Sox 5 .9363688 .044357

Gibbons Blue Jays 4 .9329651 .0110666

Yost Royals, Brewers 6 .9278911 .0443375

Table 3 ranks MLB managers by their average efficiency during years 2008-2015. Mean

efficiency scores, standard deviations, years managed in sample, and the teams for which they

managed are presented for managers with at least four years of experience. 26 managers were

identified. Mike Matheny of the St. Louis Cardinals ranks as the most efficient manager during

the time period (0.977) and also is the most consistent (sd = 0.005). Clint Hurdle of the Pittsburg

Pirates is the second most efficient manager during that span (0.962, sd= 0.026). On the other

extreme, Terry Collins of the New York Mets ranks as the least efficient manager (0.905), while

Bud Black of the Sad Diego Padres is the most inconsistent (sd= 0.059).

0.977), there is a statistical difference between the best and worst performers. On the other hand,

these results imply that even the worst managers do a credible job.

71 managerial salaries from 2008-2015 were obtained from Cot’s Baseball Contracts.

Due to the fact that MLB is not required to release managerial salary figures, not all salary data

could be found. To observe the relationship between managerial salary, efficiency, and

experience, and to get a sense of what MLB teams are basing managerial compensation from, I

estimate the following equation:

 = 



 + 



 + 



 +

 + 

 +  (3)

contracts, managerial salaries are guaranteed in spite of whether the manager is still employed by

the team.

with poor on-field performance. Similarly, teams are required to pay player contracts in full

whether the player gets injured, sees his performance levels decrease, or is released. Thus, only

the salary figure in the contract’s first year is included because it best-represents the player’s

value to the team. A salary figure from any subsequent year may not represent a player’s true

value because of the risk of injury, decrease in performance, etc. However, it may be appropriate

Jennings is currently employed with the Washington Nationals as a special advisor to the

General Manager (Axisa, 2016).

References

Aigner, D., Lovell, C. A. K., & Schmidt, P. (1977). Formulation and Estimation of Stochastic

Castrovince, A. (n.d.). Wedge named AL’s top manager. Retrieved April 5, 2016, from

http://m.mlb.com/news/article/2300277/

Gains, C. (2015). The Rays gave the youngest manager in MLB an unusual contract and it was a

brilliant move. Retrieved May 4, 2016, from http://www.businessinsider.com/kevin-cash-

contract-tampa-bay-rays-2015-3

Cot’s Baseball Contracts. (n.d.). Retrieved from

http://www.baseballprospectus.com/compensation/cots/

Crasnick, J. (2015). Who is baseball’s best manager? Retrieved May 2, 2016, from

http://espn.go.com/mlb/story/_/id/13186480

Fizel, J. L., & D’itri, M. (1996). Estimating Managerial Efficiency: The Case of College

Huzzard, D. (2015). The Industry Standard on Managers Contracts. Retrieved May 4, 2016, from

http://www.scout.com/mlb/nationals/story/1607331-the-industry-standard-on-managers-

contracts

Kahane, L. H. (2005). Production Efficiency and Discriminatory Hiring Practices in the National

Hockey League: A Stochastic Frontier Approach. Review of Industrial Organization,

175.

Lewis, H. F., Lock, K. A., & Sexton, T. R. (2009). Organizational capability, efficiency, and

effectiveness in Major League Baseball: 1901–2002. European Journal of Operational

Research, 197(2), 731–740. http://doi.org/10.1016/j.ejor.2008.07.002

Lewis, M. (2003). Moneyball: The Art of Winning an Unfair Game. New York: W.W. Norton &

Company.

Meeusen, W., & Van Den Broeck, J. (1977). Efficiency Estimation from Cobb-Douglas

Production Functions with Composed Error. International Economic Review, 18(2), 435.

MLB Manager Contracts. (n.d.). Retrieved from http://www.otherleague.com/contracts/mlb-

manager-contracts-salaries/

Pollis, L. (2014, April). If You Build It: Rethinking the Market for Major League Baseball Front

Office Personel. Retrieved from http://sabr.org/sites/default/files/Pollis-Lewis-

If_You_Build_It-2014-04.pdf

Porter, P. K., & Scully, G. W. (1982). Measuring Managerial Efficiency: The Case of Baseball.

Southern Economic Journal, 48(3), 642.

Baseball. In J. Fizel, E. Gustafson, & L. Hadley (Eds.), Baseball economics: Current

research (pp. 191–200). Westport, Conn. and London:

Salary Arbitration | FanGraphs Sabermetrics Library. (n.d.). Retrieved from

http://www.fangraphs.com/library/business/mlb-salary-arbitration-rules/

Scully, G. W. (1994). Managerial Efficiency and Survivability in Professional Team Sports.

Managerial & Decision Economics, 15(5), 403–411.

Singell, L. D., Jr. (1993). Managers, Specific Human Capital, and Firm Productivity in Major

League Baseball. Atlantic Economic Journal, 21(3), 47–59.

http://doi.org/http://link.springer.com/journal/volumesAndIssues/11293

Smart, D. l., & Wolfe, R. a. (2003). The contribution of leadership and human resources to

Journal of Sport Management, 22(3), 303–321.

Vargovick, D. (2015, June 13). How is Brad Ausmus managing the Tigers’ bullpen? Retrieved

April 7, 2016, from http://www.blessyouboys.com/2015/6/13/8771935/detroit-tigers-

ausmus-bullpen-evaluation

Volz, B. (2009). Minority Status and Managerial Survival in Major League Baseball. Journal of

Sciences, 21(2), 93–112.