There is no ‘I’ in team: Career concerns, risk‐taking incentives, and team outcomes

• Date: 01 February 2021
• Author: Steven Roberts,Phong T. H. Ngo
• Published date: 01 February 2021
• DOI: http://doi.org/10.1111/jems.12396

J Econ Manage Strat. 2021;30:122–138.wileyonlinelibrary.com/journal/jems122

|

© 2020 Wiley Periodicals LLC

Received: 20 September 2018

|

Revised: 23 March 2020

|

Accepted: 21 July 2020

DOI: 10.1111/jems.12396

ORIGINAL ARTICLE

There is no ‘I’in team: Career concerns, risk‐taking

incentives, and team outcomes

Phong T. H. Ngo |Steven Roberts

Research School of Finance, Actuarial

Studies and Statistics, ANU College of

Business and Economics, Acton, Australia

Correspondence

Phong T. H. Ngo, Research School of

Finance, Actuarial Studies and Statistics,

ANU College of Business and Economics,

Acton ACT, 2601 Australia.

Email: phong.ngo@anu.edu.au

Abstract

The National Basketball Association contracting rules provide plausibly exo-

genous variation in career concerns near contract end. We use this setting to

study how individual career concerns affect risk‐taking behavior and can

sabotage team performance. Using the frequency and duration of player injuries

from 1991 to 2013 we measure individual risk‐taking behavior. We find that

the average player's likelihood of missing a game due to injury falls by

0.06 percentage points (or over 100% relativeto the mean injury rate) in the final

3 months of his contract, and when missing games due to injury is unavoidable,

his recovery time drops by 22 days. However, “elite”players with virtually no

career concerns actually miss more games due to injury. Finally, we find that

elite players missing too many games and “average”players playing before

healthy, combine to hurt team performance. For each additional player in the

last 3 months on contract, the win probability for that team falls by over 2.6%.

1|INTRODUCTION

In a modern workplace, the majority of production is carried out in teams and teamwork is emphasized as a key tenet

for success. For example, according to a survey by Clear Company, about three quarters of employers who were

surveyed rate teamwork and collaboration as “very important”and 97% of employees and executives believe that the

lack of alignment within a team directly impacts the outcome of a task or project.

1

While prior work has shown various factors influence the effectiveness of teams (e.g., Cohen, Ledford & Spreit-

zer, 1996; Dunphy & Bryant, 1996; Jones & George, 1998), little is known about what incentivizes individuals to work

effectively in teams. Accordingly, in this paper, we use unique player data from the National Basketball Association

(NBA) for the 1991–2013 period to study how incentives created by individual career concerns (i.e., implicit incentives)

impact individual risk taking, which in turn, influences team outcomes.

The following key feature of the NBA contracting process make it an excellent setting for investigating career

concerns: once a player has signed a contract, his salary is guaranteed for the duration of the contract. There is no

incentive pay

2

and players receive full payment regardless of performance or injury. However, given the very high rate

of player turnover, a player's future salary becomes risky as he nears the end of his contract, implying that his future

expected salary is typically lower toward the end of his contract.

3

Since the contract expiry date is known well in

advance and the date itself is random—usually coinciding with the end of a season—the timing of contract expiration

provides us with plausibly exogenous variation in player career concerns.

Since our focus is on identifying perverse player behavior, the challenge for us is to find a suitable empirical

measure. While looking at on court performance is possible (and we do), Green and Zwiebel (2017) point out that in a

setting where opposition teams can endogenously respond, on court performance measures can be biased. We therefore

construct an off‐court variable based on player injury data—we observe the date an injury occurs and the number of

days it takes to recover from the injury and can therefore estimate the propensity for a player to play injured. Playing

injured or returning to play with insufficient recovery time is inherently risky behavior because such behavior can

increase the risk of exacerbating the injury or turning a minor injury into a chronic condition (Greene et al., 2001;

Hootman, Dick, & Agel, 2007). A player is willing to bear this risk if he believes the rewards (e.g., signing a new

contract) outweigh the costs (e.g., not signing a new contract and/or exacerbating the injury). The literature has

documented the propensity for athletes to “play through”injuries or to return early from injuries for possible financial

gains (Bauman, 2005; Lehn, 1982; O'Connell & Manschreck, 2012).

Our primary conjecture is that as players approach the end of their contract and are exposed to career concerns,

their likelihood of missing a game due to injury decreases (i.e., an increase in their willingness to play injured).

Moreover, playing injured may not only be costly for the individual, but also for the team. Concealing (the extent of) an

injury to play may be associated with poorer performance, on average, thus ought to have negative implications for

team performance overall.

Anecdotal evidence of this type of perverse behavior is provided by a (formerly) little‐known Australian NBA player

Patty Mills. In July 2012, Mills, then a journeyman NBA player, signed a 2‐year contract with the San Antonio Spurs as

the team's second‐string point guard. Mill's last year of this contract turned out to be his breakout season, and he

resigned with the Spurs after the season for 3 years and $12 million.

4

Notably, Mills admitted to playing with a torn

rotator cuff for most of the season, which required immediate postseason surgery.

5

Mill's decision to play injured

provides a striking example of a player with an uncertain NBA playing future being prepared to take significant risk to

secure a new contract.

6

His behavior is consistent with the “gambling for resurrection hypothesis”popularized in the

economics literature: when you have nothing to lose and everything to gain, then bet big.

7

Using data on 7,220 contract signings and 8,454 various types of player injuries for 1,978 unique players we find

strong evidence of behavior among NBA players that is consistent with predictions of the career concerns literature.

First, compared with other times in his contract cycle, the average NBA player is approximately 0.06 percentage points

(or over 100% relative the sample injury rate) less likely to miss a game due to injury in the final 3 months of his

contract—we interpret the reduction as an increase in risk taking (i.e., players are likely to be playing injured).

Similarly, in comparison with other times in his contract cycle, the average NBA player takes approximately 22 fewer

days to recover from an injury in the last 3 months of his contract.

Second, this effect differs in the cross‐section of player value or quality (i.e., degree of exposure to the career

concerns shock). For high‐value or elite players—superstars like Kevin Durant, Michael Jordan, and LeBron James who

have virtually no career concerns—the propensity to miss games as the result of injury in the final 3 months of a

contract actually increases. This difference between how elite players and other players behave—elite players minimize

risk, whereas other players increase risk or “gamble”—is consistent with a “bang‐bang”risk‐taking strategy: depending

on future earnings opportunities players optimally choose minimal or maximal risk (see e.g., Gan, 2004).

Finally, we investigate whether players' risk‐taking behavior has a broader impact on team outcomes. A clear

implication of playing injured is that, other things equal, the individual's and therefore the team's performance falls.

Consequently, we show that a team's likelihood of victory falls by over 2.6% for each additional player on the team who

is in the last 3 months of his contract.

We highlight the possible need for team management to stagger player contract dates or to provide explicit team

level incentives to avoid negative repercussions for the team. Our findings also have implications for team production

outside professional sports. Professional settings like marketing/sales teams or venture capital/private equity teams are

good examples where teamwork is crucial but individual career concerns may have perverse consequences for the team

outcome. For example, individual promotion based on sales targets may result in perverse effects like “sales stealing”

from team members. We argue that detecting that such behavior exists and finding that it has significant implications

for team outcomes in the NBA—a high stakes environment where players are heavily scrutinized—places a lower

bound on the perverse effects of career concerns one might expect to find in other professional settings.

Our paper contributes to two literatures. First, it is related to the career concerns literature examining how implicit

incentives might reduce agency conflicts (Dewatripont et al., 1999a, 1999b; Fama, 1980; Holmstrom, 1999) or how

career concerns might create perverse incentives to agents to take suboptimal actions that make themselves “look good”

(Siemsen, 2008) especially in a team setting (Auriol, Friebel, & Pechlivanos, 2002). While the theory is well established,

testing the predictions from these models is problematic given career concerns are unobservable. As a result, the

empirical literature—to which we speak most closely—has focused on examining cross‐sectional differences in the

behavior of CEOs, fund managers and security analysts according to their age (Chevalier & Ellison, 1999; Gibbons &

NGO AND ROBERTS

|

123