The Perils of Guaranteed Long Term Contracts. Evidence From

Boom & Bust: The Perils of Guaranteed Long Term Contracts
Evidence from OPS100 Performance Over the Contract Cycle
Heather O’Neill – Ursinus College
Abstract
Model & Hypotheses
This study focuses on panel data of 256 MLB free agent hitters under the 2006-2011
Collective Bargaining Agreement (CBA) to demonstrate that hitters , on average, increase
their offensive production, measured by OPS100, during the last year of their contract and
subsequently underperform the first year of the newly signed long term contract. The
contract year phenomenon arises from the incentive to land a lucrative guaranteed contract
for players not intending to retire. Signing a long term guaranteed contract creates an
incentive to shirk (underperform) the first year of the new contract because performance and
pay become unlinked and the need to boost performance for another contract is years away.
A player’s intention to retire negates the incentive associated with a new contract, which
mitigates the contract year boost and leads to greater shirking.
The population regression model for OPS (or OPS100)
for player i in season t:
(+) (+)
(+) (+)
(+)
(+)
OPS100i,t = β0 + β1*GAMESi,t + β2*PLAYOFFi,t β3*CONTYRi,t +
(+) (+)
(-)
(-)
(-)
(β4*FIRST
) i,t + β5*SHIRKLENGTHi,t + β6*PROBRETi,t + Ai + µi,t
Modeling OPS100 over the contract cycle requires focusing on within player behavior using
seasonal statistics, intention to retire, and incorporating unobservable player traits, such as
innate ability, work ethic, family background, etc. that may differ across players. Fixed
effects, as opposed to OLS or pooled OLS, estimation addresses potentially biased results
from omitting unobserved traits. Controlling for games played and team play-off contention,
players at the end of a multiyear contact show a 6% boost in OPS100, but it declines by 1%
for each percentage point increase in the probability of retiring. Players on one year
contracts, however, boost their OPS100 by 13% without retirement effect. Signing a five year
contract leads to a 4% drop in OPS100 during the first year of the newly signed contract and
each additional year on a new contract creates another 2.5% decline in OPS100. While
there is no evidence of shirking for two to four year contracts given no change in “retiring”,
statistically significant shirking occurs for two to four year contracts if retiring intention
increases by 8 and 3 percentage points, respectively.
GAMES: Playing more correlates with better performance
PLAYOFF: Being on playoff contending team correlates with
better performance
CONTYR: Player boosts performance to garner future contract
FIRST:
In first year of a one year contract player boosts
performance
SLENGTH: Defined as FIRST*LENGTH
Poorer performance the longer the contract
PROBRET: Greater intention to retire reduces performance
These findings suggest that general managers negotiating free agent contracts may wish to
sign less than five year contracts to mitigate shirking behavior and estimate the likelihood of
a player retiring since it impacts future performance.
Ai:
Unobserved time-invariant player effects, such as
innate ability, work ethic and family background
µit:
Stochastic error term
Previous Regression Model Findings
”Just give me 25 guys on the last year of their contract;
I’ll win a pennant every year.” Sparky Anderson
Contract Year Only Papers
•
•
•
•
Ordinary Least Squares (OLS) and Pooled OLS generally find no boost
Ignoring retirement finds no boost
Narrowly measured performance stats find no boost
Fixed Effects + Retirement Intention + OPS find boost
"The experience of individual clubs, and the industry as a
whole, is that for whatever reason, the player's
performance is not the same following the signing of a
new multi-year contract.“ Dan O’Brian, VP (Indians)
Shirking Only Papers
•
•
•
OLS studies with Marginal Revenue Product find shirking
OLS studies with non-revenue defined performances do not find
shirking
Some find shirking for contracts of 4 or more years
Models with Contract Year & Shirking
•
•
Fixed Effects with broad performance measure find both for 2 or more
year contracts
Greater retirement intention reduces boost and increases shirking
Tales of Two Phillies
NAME
IBANEZ
IBANEZ
IBANEZ
IBANEZ
IBANEZ
IBANEZ
IBANEZ
WERTH
WERTH
WERTH
WERTH
WERTH
CODE YEAR TEAM PLAYOFF OPS100 GAMES DL AGE YRSEXP CONTYR LENGTH FIRST SLENGTH PROBRET
155 2006 SEA
0
125
159
0
32
11
1
1
0
0
-26.4723
155 2007 SEA
0
121
149
0
35
12
0
2
1
2
-17.1998
155 2008 SEA
0
124
162
0
36
13
1
2
0
0
-9.3194
155 2009 PHL
1
132
134
23 37
14
0
3
1
3
-1.3026
155 2010 PHL
1
111
155
0
38
15
0
3
0
0
17.1231
155 2011 PHL
1
91
144
0
39
16
1
3
0
0
37.171
155 2012 NYY
1
104
130
0
40
17
1
1
1
1
37.171
325 2008 PHL
1
122
134
15 29
7
0
1
1
1
-7.9896
325 2009 PHL
1
129
159
0
30
8
0
2
1
2
-8.1475
325 2010 PHL
1
144
156
0
31
9
1
2
0
0
-9.5566
325 2011 WSN
0
97
150
0
32
10
0
7
1
7
13.272
325 2012 WSN
1
125
81
87 33
11
0
7
0
0
12.4216
Jason Werth’s OPS100
Over His Contract Cycle
No Shirk in 2009 +7
Boost in 2010
+15
Shirk in 2011
-47
Anecdotal Evidence: 4 of 7 Cases Conform to Expectations
Let’s Play Ball with Robust Data & Modeling
Multiyear Contract: ∆OPS100 = β3+β6*PROBRET^
1 Year Contract: ∆OPS100=β3+β4+β5*LENGTH+β6*PROBRET^
Shirking Impact
Multi-yr Contract: ∆OPS100=β4+β5*LENGTH+β6*PROBRET^
where LENGTH = 2, 3, …10
Empirical Results
(One tail p-values)
(1) PROBRET^= .215 - .061*YEARSEXP + .006* YEARSEXP
(.0539)
(.0003)
+ .00034*DL - .0037*OPS100
(.1726)
(.0001)
Percent Correctly Predicted= .94
PROBRET = α0 + α1*YRSEXP + α2*YRSEXP2 +α3*DL +α3*OPS100
Likelihood of retiring increases at an increasing rate with MLB
experience, increases with more days on the DL and falls with
better performance
(2) OPS100^ = 104.707 + .0499*GAMES + 1.882*PLAYOFF
(.0307)
(.0556)
Role of age depreciating performance enters through retirement
Data Sources & Simple Statistics
256 free agent hitters 6 or more years in majors yields 1,106 player/year
observations - ESPN.com Major League Baseball Free Agent Tracker
All playing under 2006-2011 CBA so that players face same incentive
structure and employment rules
OPS100 controls for home field and league, Games, Age - Baseball
Reference.com
Days on DL - Josh Hermsmeyer 2006-09; BackseatFan 2010; Fansgraph
2011
FIRST = 0 (Not in First Year of Multiyear Contract)
Variable
N Mean Std Minimum Maximum
Dev
OPS100
402 99.34 28.36
-8.00
192.00
PROBRET 402 8.47 20.76 -53.91
66.06
AGE
402 32.93 3.00
26.00
48.00
PLAYOFF 402 0.37 0.48
0.00
1.00
GAMES
402 115.19 37.45 10.00
162.00
LENGTH
402 3.85 1.96
1.00
10.00
SALARY($M) 402 8.88 5.86
0.40
33.00
FIRST = 1 ( In First Year of Multiyear Contract)
Variable
N Mean Std Minimum Maximum
Dev
OPS100
706 86.15 31.49 -39.00
188.00
PROBRET 706 16.25 22.20 -28.94
83.43
AGE
706 33.14 3.33
24.00
47.00
PLAYOFF 706 0.29 0.46
0.00
1.00
GAMES
706 97.34 40.69
7.00
162.00
LENGTH
706 1.44 1.08
1.00
10.00
SALARY($M) 706 3.19 3.57
0.33
28.00
OPS100 lower in first year of multiyear contract 86.15 < 99.34
At face value suggests shirking
Raul Ibanez’s OPS100
Over His Contract Cycle
Shirk in 2007
-4
Boost in 2008
+3
No Shirk in 2009 +8
No Boost in 2011 -10
Contract Year Impacts
BUT
The average likelihood of retiring is twice as great in first year, 16.25 > 8.47
Retiring players have lower OPS100 stats
Role players earn less $3.19 < $8.88 & have shorter contracts, 1.44 < 3.85
Role players have lower OPS100 stats
Shirking is getting too much credit!
Model correctly to yield unbiased results
+ 5.396*CONTYR + 9.166*FIRST – 2.530*SHIRKLENGTH +
(.0001)
(.0001)
(.0001)
- .907*PROBRET^
(.0001)
Buse R-Square = .77
Conclusions
 Incorporate likelihood of retiring & contract length for unbiased results
 Use fixed effects modeling to account for unobserved traits
 Contract year boost, independent of retiring :
5.4 point increase OPS100 at end of 2 or more year contract
12 point increase OPS100 at end of 1 year contract
 Contract year boost decreases OPS100 by 1 point per retirement
increase
 Contracts of five years lead to shirking of 3.8 OPS100 drop, regardless
of retirement, and 2.5 additional declines for each additional year on
contract
 Two year contracts show shirking if 8+ point increase in retire
probability
 Three year contracts indicate shirking with 5+ point increases in retiring
 Four year contracts show shirking starting with 3+ point increases
Selected Works Cited
Berri, D., & Krautmann, A. (2006).Shirking on the Court: Testing for the Incentive Effects of
Guaranteed Pay.. Economic Inquiry, 44, 536-46 .
Krautmann, A., & Solow, J. (2009). The Dynamics of Performance Over the Duration of Major
League Baseball Long-term Contracts. Journal of Sports Economics, 10(1), 6-22.
Maxcy, J., Fort, R., & Krautmann, A.( 2002). The Effectiveness of Incentive
Mechanisms in Major League Baseball. Journal of Sports Economics, 3(3),
246-55.
O’Neill, H. (2013). Do Major League Baseball Hitters Engage in Opportunistic Behavior?
International Advances in Economic Research, 19(3), pp. 215-232..
O’Neill, H. (2014) . Do Hitters Boost Their Performance During Their Contract Years?”
The Baseball Research Journal, 43(2) , 78-85.
Stiroh, K. (2007). Playing for Keeps: Pay and Performance in the NBA. Economic
Inquiry, 45 (1), 145-161.