No one likes a losing streak, whether you are the owner of a hockey team or a member of a company’s board of directors. Usually, a ten-game losing streak or several consecutive quarters of poor financial performance gives one permission to use the adjective “former” before the word coach or CEO. But new research shows that there are certain exceptions to this rule, leading to a new trivia question: What do a CEO and an NHL coach have in common? Read on.
Not meeting Wall Street’s expectations? Going through a rough patch of poor performance? When in doubt – fire the CEO!!!
Conventional wisdom in today’s bottom-line “What have you done for me lately?” environment says that no combination of factors is so sure to lead to the dismissal of a CEO than the toxic mix of poor performance and failing to meet expectations. However, a just-released1 study throws a wrench into this conventional wisdom by showing that this combination is lethal only some of the time.
Before we start clearing out corner offices because of “not meeting expectations” and “poor performance”, we can take heart. It turns out that, as we find in so many other situations, things are a little more complicated than they appear. By “more complicated” we mean that the circumstances leading up to a CEO’s departure may affect, in different ways, the market value of the firm.
In order to determine, in as impartial and clinical a way as possible, the factors and their relationships surrounding CEO firings, the new research looked at the National Hockey League, specifically to the firing of coaches.. While it may not seem intuitive that NHL coaches’ firings could serve as a good surrogate for CEO firings, using them is, in some ways, superior to conducting research on firings in the business world. Consider the following:
- it is easier to observe all of the coaching changes and the factors that go into these changes (not so easy in the business world);
- there is no ambiguity and greater similarity in the goals of all the NHL teams (in business this is less certain across markets and even within the same market);
- turnover is higher among NHL coaches than it is in managerial positions in the business world, thus giving us a lot of data to study; and,
- coaches are very visible leaders, recognized as important to the teams they represent, and are the subject of much public discussion and media coverage (again this discussion is clearer than it is in the business world).
The coach and the CEO: Different yet alike
Although being the coach of an NHL team is not exactly like being the CEO of a Fortune 500 company, both a coach and a CEO are ultimately responsible for the performance of their organizations. Of the two, the circumstances surrounding coaches’ departures are a little less fuzzy. Therefore, for the purposes of trying to shed light on how, when, and why CEOs get fired, the study of coaches’ departures can be instructive. Joe Moglia, the CEO of Ameritrade, argued that coaches and CEOs have much in common, in that they have to “set strategies, maximize results using people with diverse skills, and recruit and motivate subordinates and that they can learn from each other.”2
When the legendary Sam Pollock, who built 9 Stanley Cup winners in 14 seasons (1964–1978) with the Montreal Canadiens, began his front-office career he believed that it was necessary that general managers have 50 percent hockey acumen and 50 percent business skills. When he left the game in 1978, he had come to believe that GMs needed to have 15 percent hockey acumen and 85 percent business skills. He believed that to be a successful hockey general manager (GM), one needed to know and follow sound business principles.3 The arguments by Moglia and Pollock imply that some of the findings from the research on coaching changes in the NHL are applicable to non-sports business organizations. This means that it is reasonable to suggest that maybe a few CEOs could learn from Sam Pollock or his modern-day counterparts, Lou Lamoriello, the GM of the New Jersey Devils, and Ken Holland, the GM of the Detroit Red Wings, both of whom have won three Stanley Cups, the most of any GM since 1987.
From a researchers’ perspective, and without resorting to administering truth serum to everyone who may have contributed to a decision to fire a CEO, comparing the NHL coaches’ departures to the dismissals of CEOs is a reasonable analogy.
There are a few theories that are more prominent than others in understanding why coaches, CEOs and, anyone else for that matter, get fired, and which have been the subject of much research. It would be good to review these before getting into the meat of the study.
The Common Sense Theory of Firing: Companies fire their leaders during periods of poor performance in the hope that a replacement who will improve performance will be found.
The Ritual Scapegoating Theory: If one assumes that performance actually depends on things other than what the leader does, then in times of poor performance the board fires the leader in an effort to deflect criticism and appease the market.
The Vicious Circle Theory: A change in leadership in a corporation is disruptive and such disruption tends to make matters worse, a circumstance that precipitates another change in leadership.
If you are nodding your head ruefully at these aptly named theories, you are probably either a cynic or have spent more than two years working in a large company.
The study we referred to in the second paragraph examines the reasons why coaches are fired in relation to a number of variables. It is left to the reader to translate these results into business realities that may be applicable to his or her particular company:
- the expectations of how the team would do in preseason predictions versus its actual performance at some point during the season;
- the team’s actual performance in the ten to fifteen games preceding the firing versus its expected performance in these games (considering who they were playing, whether they were at home or away, or whether they were fighting for a playoff spot);
- how well the team did in the previous year’s playoff run;
- the coach’s age; and,
- whether the coach previously played on the team he was coaching.
Importantly, the study then looked at the performance of the team after the coach was fired.
One might have expected some of the results, such as the higher a coaches’ winning percentage the longer a coach’s tenure, or the coaching turnover rate among the lower-ranked teams (obviously if the team performs well then coaches are kept on for longer tenures). These results support each of the theories described above.
Coaches were given approximately 200 games (roughly two and one-half seasons) to prove themselves. The study indicates they were more likely to be fired after 200 games than before that number of games. For example, in the 2007-2008 season, both Bob Hartley of the Atlanta Thrashers and Glen Hanlon of the Washington Capitals were fired after they had coached 200 games. In the business world, Brent Willis was fired from Cott Corp. after 3 years of non-performance (even when Cott’s performance relative to the S&P 500 was dismal after 1 year (-45 percent) and two years (-86 percent)).
Similarly, Robert Nardelli was fired from Home Depot after 3 years of non-performance versus the S&P 500, even though his most recent quarterly financials mirrored the index.
Short-term performance prior to the coach’s firing was important. With every possible definition of “short term” tested, the one that most accurately predicted coaches’ firings was the team’s performance during the previous 15 games. While many coaches stay on past their 200th game, having the misfortune of being at the helm for an unimpressive 15-game stretch precipitates fallout that is almost immediate. Bob Hartley, 0 and 6 before being fired at the beginning of the season, and Glen Hanlon, 3-11-1 before being fired, fall into this category. Jim Donald of Starbucks was shown the door after Starbucks performed terribly versus the Nasdaq index for the most recent quarterly financials preceding his ouster.
The researchers then assessed whether the general managers and owners considered the quality of the opposition in these particular 15 games. Apparently they did not; only the actual results in these games were deemed relevant. In fact, it was found that the results for these 15 games mattered more than the coaches’ career won-loss record. A good case of “What have you done for me lately?”
Not surprisingly, it was found that the team’s current standing relative to its preseason, predicted finish was important in predicting when a coach would be fired: the lower in the standings versus the preseason prediction, the greater the chance of a coach being fired. The study indicates that not meeting expectations is a very good predictor of a coach’s tenure behind the bench. For example, John Paddock was in his first year, but went through a 4-8-3 streak before being fired by the Ottawa Senators, who eventually finished in 7th place, even though they were expected to finish 2nd in their conference. Both John Tortorella, of the Tampa Bay Lightning, and Joel Quenneville, of the Colorado Avalanche, were fired after the season. Their teams were expected to finish 6th and 4th, but instead finished 15th and 6th, respectively.
A coach could reduce the pressure or likelihood of being fired by doing well in the previous season’s playoffs, or better yet, by winning the Stanley Cup. Leading a team to do better in the previous year’s playoffs lessens the likelihood of being fired. However, getting one’s team into the playoffs but not progressing deeper into the playoffs leads to coaches departing after the season is over. One needs to look no further than the San Jose Sharks who, despite finishing 1st in their division, disappointed their fans by going out in the second round of the playoffs. Ron Wilson was subsequently fired by San Jose and hired during the off-season by the Toronto Maple Leafs. When Jonathon Miller was fired from AOL Time Warner, he was replaced by Randy Falco, who, arguably, was a star at NBC Universal.
Interestingly, a coach who had previously been a player on the same team had a higher probability of being fired – quite possibly because other factors such as sentimentality had entered into the hiring decision instead of pure coaching ability. Consider the case of Chicago Blackhawk coach Denis Savard, who was fired after just four games into the 2008-2009 season.
Common Sense Theory (you have permission to quote this theory around the water cooler) says that if a coach is going to be fired for the team’s poor performance in the previous 15 games, the owners of the team would expect the performance of the team to pick up as a result of the coaching change. However, the research actually found that team performance was even worse for the four games immediately after the firing than it was for teams that did not experience a coaching change. Additionally, performance was the same after these four games, both for teams that had made the coaching change and the teams that did not. In fact, there were few instances where teams that made a coaching change in the middle of a season outperformed teams that did not make a change in that season. One very notable exception was when Lou Lamoriello dismissed Robbie Ftorek and replaced him with Larry Robinson, who led the New Jersey Devils to a Stanley Cup win that season.
In short, the research demonstrated that the chance of a coach being fired increased:
- with poor short-term performance (15 games prior to the firing – notwithstanding the fact that it did not matter whether or not the team should have won these games based upon difficulty, home/away etc),
- with how poorly the team was performing as compared to the preseason estimates of where the team would finish in the standings, and,
- if the coach had previously played on that team.
The chance of a coach being fired decreased;
- if the team did well in the previous year’s playoffs; and,
- if the coach was older.
Significantly, teams that fired their coach in mid-season tended to perform worse in the short term (first four games after the firing) than teams that did not replace their coach.
So, aside from providing some possible advantages in betting on NHL games, what does this have to do with the firing of a CEO?
The research on coaches’ dismissals in the National Hockey League suggests that the combination of “not meeting expectations” and “poor performance” alluded to earlier is probably toxic for a CEO. This is especially true when you examine the firm’s performance relative to the medium- to long-term expectations of the board and the firm’s short-term performance (whether it meets the board’s short-term expectations or not), regardless of the firm’s overall performance during a CEO’s tenure. Putting a positive spin on this, the firm needs to perform (regardless of expectations) in the short term and meet the board’s expectation in the medium to long term if the CEO wants to have a longer-than-average tenure.
For those who believe that the value of their firm’s stock may be worth less because the CEO was fired, and may, therefore be inclined to sell their stocks before they become even less valuable, the study suggests that this may not be advisable. While performance may be worse in the very short-term, the stock is likely to return to a level similar to what it would have been if the board had left things alone.
The authors’ hope is that this research assists in predicting and explaining CEO departures, and, incidentally, gambling on NHL games (where permitted by law, of course).
- Audas, R., Goddard J., & Rowe WG. 2006. Modelling Employment Durations of NHL Head Coaches: Turnover and Post-succession Performance, Managerial and Decision Economics 27: 293-306.
- Kalawsky, K., & Maich, S. (2004, March 20). Double standard: Shareholders in publicly traded sports empires are much more patient with corporate losses than team losses. Financial Post, National Post, FP1.
- Goyens, C., & Turowetz, A. (1986). Lions in winter. Scarborough, ON7 Prentice-Hall, Inc.