This story was originally published May 31, 2013. Once again, two U.S. teams, the Chicago Blackhawks and Tampa Bay Lightning, are headed to the Stanley Cup finals.
The N.H.L.’s conference finals will begin on Saturday, and fans in Boston, Chicago, Los Angeles and Pittsburgh can still dream of a Stanley Cup championship. But this year is an unhappy anniversary for America’s northerly neighbor. It is 20 years since the Montreal Canadiens defeated the Los Angeles Kings to win the 1993 Stanley Cup; no Canadian team has won the championship since.
Just how unlikely has Canada’s Stanley Cup drought been? And is there anything in the league’s economics that might help to explain it?
The answer isn’t simple, so let me give you the executive summary before we parse the details.
First, bad luck is a major component. Even after accounting for the fact that Canadian teams have rarely been among the league’s best in recent years, you would still have expected Canada to pick up at least a couple of Stanley Cups at some point.
Second, the N.H.L.’s economic structure changed at an unfavorable time for Canada. During the first half of the 20-year drought, the league allowed teams to spend freely, but Canadian teams were hampered by the weak Canadian dollar. Since 2005, the Canadian dollar has recovered substantially, and Canadian teams are now turning large profits. But they are limited in their capacity to invest those profits in superior players because the league has instituted a hard salary cap.
Third, there is almost certainly a shortage of N.H.L. teams in Canada relative to the demand for hockey there and the revenues that Canada contributes to the league. Teams in nontraditional hockey markets like Raleigh, N.C., Tampa, Fla., and Anaheim, Calif., have won Stanley Cups since 1993, but without doing especially well financially. So have the Colorado Avalanche, who relocated from Quebec City in 1995-96. Had the distribution of N.H.L. teams more closely matched fan interest in the sport across the United States and Canada, Canada would have more teams in the league and – very probably – at least one Stanley Cup championship.
Finally, and related to the excess demand for hockey in Canada, Canadian teams routinely sell out their arenas at high ticket prices — whether or not they are any good. This may reduce their incentive to compete.
The simplest way to conceive of the situation is to assume that each N.H.L. team begins with an equal chance of winning the championship and to estimate the odds of a Canadian team winning the Stanley Cup as a function of the number of Canadian teams in the league.
Eight of the 26 N.H.L. teams, or 31 percent, were from Canada in the 1993-94 season. The percentage fell to 20 (6 of 30 teams) by 2000-1, before increasing to 23 percent after the Atlanta Thrashers moved to Winnipeg in 2011.
If a champion were randomly chosen from all N.H.L. teams active each season, the odds that a Canadian team would have won at least one Stanley Cup since 1993-94 are 99.2 percent. (This period comprises 19 seasons, excluding 2004-5 when the N.H.L. canceled its season because of a labor dispute.) That means the chance of Canada failing to have won a Stanley Cup is just 0.8 percent, yielding odds of about 125-to-1 against.
But it is naïve, of course, to assume that the Stanley Cup champion is chosen randomly. Major League Baseball’s Kansas City Royals, for example, aren’t merely unlucky to have failed to reach the playoffs in 28 years: they have usually stunk. Perhaps the Canadian N.H.L. teams just haven’t been very good either.
At first glance, the difference in competitiveness seems minor. Canadian teams have averaged 35.3 wins and 81.5 points per regular season over this period, compared with 36.7 wins and 84.4 points for the American teams. Canadian teams made the playoffs 54 percent of the time, compared with 56 percent for the United States-based clubs.
However, the Stanley Cup is not normally won by an average team that just sneaks into the playoffs — instead, it usually goes to one of the best teams in the league. Eleven of the 18 Stanley Cup champions since 1993-94 have been seeded No. 1 or No. 2 in their conference. And there has been a lack of elite Canadian teams; only 14 percent of the No. 1 and No. 2 seeds since 1993-94 have been from Canada.
A more exacting way to evaluate the odds is by estimating each team’s chance of winning the Stanley Cup as a function of its playoff seed and its regular-season power rating. (Power ratings, in this case, are taken from hockey-reference.com’s Simple Ratings System, which accounts for each team’s goal differential and strength of schedule during the regular season.) We can accomplish this by means of a logistic regression analysis. The probabilities for each season are recalibrated such that the 16 teams that make the playoffs have exactly a 100 percent chance of winning the Stanley Cup, in total.
In the 1996-97 season, for instance, I estimate that the Canadian teams had just a 2.1 percent chance of winning the Stanley Cup on the basis of their regular-season performance. Three of the six Canadian teams made the playoffs that year — but all as No. 7 or No. 8 seeds, and none of them were much good.
In 15 of the past 19 N.H.L. seasons, in fact, the probability of a Canadian team winning the Stanley Cup, as calculated by this method, was lower than you would get if you drew from among all N.H.L. teams at random. Only twice — the Ottawa Senators in 2005-6, and the Vancouver Canucks in 2010-11 — did a Canadian team enter the playoffs as the best bet to win the Stanley Cup on the basis of regular-season performance.
You would still have expected a Canadian team to win the Stanley Cup sooner or later. The chance that at least one would do so was 97.5 percent, according to this method. Nonetheless, the lack of dominant Canadian teams has helped to make the 0-for-19 drought merely unlikely rather than almost impossible — raising the chance of the Canadian shutout to 2.5 percent from 0.8 percent. So it is worth asking why Canadian teams have rarely been among the N.H.L.’s elite.
Perhaps Canadian teams are competing on an unlevel playing field? Among the seven current Canadian N.H.L. teams, the median television market is 1.7 million people. By contrast, the median American team competes in a television market with 5.4 million people.
Professional hockey, however, has an extremely regional following: it is extraordinarily popular in almost all of Canada, reasonably popular in parts of the northeastern and midwestern United States, and quite obscure elsewhere in the United States. These differences more than make up fo
r the smaller size of Canadian markets. In fact, the typical Canadian team has considerably more N.H.L. fans in its market than the typical United States team does.
In 2011, I estimated the number of fans for each N.C.A.A. football team by comparing Google search traffic for the term “college football” in different media markets. We can do the same thing for hockey by evaluating the number of people who searched for the term “N.H.L.”
On a per-capita basis, searches for the term “N.H.L.” were more than seven times as common in Canada as in the United States. Although the United States population is about nine times larger than Canada’s, this makes up for a lot of the difference.
How many N.H.L. fans are there in each country? A 2010 YouGov poll estimated that 11 percent of American adults were serious N.H.L. fans; a 2007 Scarborough Research study, using a stricter definition that was limited to the most avid fans, put the total at just 4 percent. Using the midpoint of those two surveys yields an estimate that 7.5 percent of Americans are N.H.L. fanatics, or about 23.5 million Americans among a population of 314 million.
The frequency of Google searches in Canada for “N.H.L.” implies that about 54 percent of Canadians are serious N.H.L. fans, or about 18.5 million people in a country of 34.5 million.
In other words, the two countries are reasonably close in their number of N.H.L. fans, despite the United States’ vastly larger population. I estimate that Canada accounts for something like 45 percent of the N.H.L. interest between the two countries. (The N.H.L., incidentally, makes about 40 percent of its national television licensing fees from Canadian networks — and the percentage is likely to increase after the league negotiates new Canadian TV contracts.)
If N.H.L. teams were distributed between the countries in proportion to fan interest, Canada would have something like 12 to 14 teams out of a 30-team N.H.L. Instead, Canada has seven teams. This undersupply means that the Canadian N.H.L. teams have larger numbers of fans, on average.
In the chart below, I have used the Google search data to estimate the proportion of people in various United States and Canadian media markets who are serious N.H.L. fans. (The method uses Google search data at the metro-area level for the United States and at the provincial level for Canada as metro-level data is unavailable for Canada.)
The results demonstrate how much difference hockey avidity can make. Compare the cities of Los Angeles, Boston and Calgary. There are about 17 million people in the Los Angeles media market, 6.2 million in Boston’s and 1.6 million in Calgary’s. Based on their Google search traffic, however, I estimate that only 6 percent of the population in Los Angeles are serious N.H.L. fans, compared with 17 percent in Boston (high by United States standards) and 67 percent in Calgary (high even by Canadian standards). As a result, the three markets are roughly equivalent on hockey terms: each has about 1.1 million serious N.H.L. fans.
Another astonishing result can be obtained by comparing Houston and Saskatoon, Saskatchewan — two markets that are occasionally mentioned as possible sites for N.H.L. expansion teams. Houston’s media market has about six million people, compared with not quite 340,000 for Saskatoon. However, I estimate that only 2 percent of Houston residents are serious N.H.L. fans, versus 46 percent of the population in Saskatoon. Thus, they are roughly equal at about 150,000 N.H.L. fans each.
Accounting for the far greater propensity of Canadians to follow hockey, the median Canadian team plays in a market with about 1.1 million N.H.L. fans, while the median United States team plays in a market with about 530,000 fans.
The American cities that have had their hockey teams for a long time generally place well above the United States median, however. Of the United States-based markets that had an N.H.L. team continually since 1990, all but St. Louis have at least 500,000 fans.
So a couple of decades ago, there was a nice balance in the N.H.L. between American cities with large populations but moderate hockey interest, and Canadian cities with moderate populations but intense hockey interest. There were a few outlying cases — Los Angeles has very low hockey interest but a very large population, while Buffalo has a modest population but levels of hockey affinity that approach Canadian levels. These teams functioned well enough, and no team was dragging down the average that much. The American markets that the N.H.L. has ventured into since 1990 have only about 250,000 hockey fans on average, however.
These differences in hockey interest have a profound effect on each team’s bottom line. In the chart below, I have compared the number of hockey fans in each market against their estimated operating profit or loss in 2011-12, according to Forbes magazine. (For the markets with multiple hockey teams, I have divided the fans as follows: in New York, 55 percent of the market to the Rangers, 25 percent to the New Jersey Devils and 20 percent to the Islanders; and in Los Angeles, two-thirds to the Kings and one-third to the Anaheim Ducks. These estimates are derived by evaluating the number of Google searches for the team names among the residents in each metropolitan area.)
This measure of hockey fans does a very good job of predicting each team’s profitability. The 13 teams with 700,000 or more N.H.L. fans in their markets all made money and totaled $357 million in operating profits. The six teams in markets with fewer than 300,000 fans all lost money, totaling $77 million in operating losses. The teams between 300,000 and 700,000 fans had varied results but roughly broke even, on average. (This is the range in which a quality of a team’s management matters, along with other factors like per-capita income in the region and a team’s appeal outside its immediate metropolitan area.)
Six of the seven Canadian teams are above the 700,000-fan threshold (and therefore made money). So did Winnipeg, with roughly 560,000 fans. In total, Canadian teams brought in $219 million in operating profits in 2011-12 — whereas the American teams made a net of just $31 million. (Outside of the highly profitable Rangers, in fact, the United States-based teams lost money that year.)
So why have the Canadian teams struggled to win Stanley Cups?
Part of it stems from the same reason that operating profits are so closely tied to the number of N.H.L. fans in each team’s media market. N.H.L. teams vary greatly in how man
y fans they have available to them. But under the league’s current collective bargaining agreement, there is relatively little revenue sharing. There is a hard salary cap, however, along with a high salary floor, so each team’s player expenditures are about the same.
The good news for the Canadian teams is that, with nothing else to spend them on, those extra revenues flow through straight through to the bottom line. The bad news is that they could not spend them by investing in better player talent, even if they wanted to.
Before the lost season of 2004-5, the N.H.L.’s economics were different: there was no salary cap. But something else was different as well: the Canadian dollar had been historically weak against the American dollar. Thus, the Canadian teams could not take full advantage of the excess demand for hockey in the country.
In the next chart, I have estimated per-game ticket revenues for American and Canadian N.H.L. teams since 1995. Ticket revenues are calculated as per-game attendance during the regular season, multiplied by average ticket prices as reported by Team Marketing Report. (Since 2003, Team Marketing Report has designated a portion of tickets in each arena as premium seats; I assume that 20 percent of ticket sales are priced at premium levels.) All figures are denominated in current (inflation-adjusted) United States dollars. The chart also shows how much revenue the Canadian teams would have brought in had the American and Canadian dollars been trading roughly at parity throughout the period, as they have been recently.
Per-game ticket revenues for the Canadian teams lagged somewhat behind the United States teams from 1995 through 1999. This was mostly because of the weak Canadian dollar; the Canadian teams would have run even with or slightly ahead of the American teams otherwise. From 2000 through 2003, the American and Canadian teams made about the same amount of money from a typical game despite a still fairly weak Canadian dollar. Since then, the Canadian dollar has strengthened, and Canadian teams have moved considerably ahead of the American teams: the average Canadian team now makes about 50 percent more than the average American team on a per-game basis. However, it is during this period that the N.H.L.’s salary cap has been in place.
But even if the Canadian teams cannot spend their extra revenues on better player talent, this does not fully explain why they have been underachieving in the Stanley Cup and not at least winning their fair share of championships.
Much of the reason, I must emphasize again, boils down to bad luck. Canadian teams have reached the Stanley Cup finals five times in the 19 seasons since 1992-93 but have come up short on each occasion, including in four cases where the series went to the seven-game maximum.
One other factor, however, may be that there is so much excess demand for hockey in Canada that the Canadian franchises do not have to field especially strong teams to sell out their stadiums or to make a considerable profit. In the next chart, I have compared the per-game ticket revenues for each team in the 2012-13 season against the number of points they tallied between this season and the last one.
For the United States teams, there is an imperfect but reasonably clear and statistically significant relationship between on-ice success and ticket revenues. There are lots of fair-weather American hockey fans, and they may not turn out unless their team is pretty good. In Canada, there is less competition from other sports, and there are many die-hard hockey fans who attend games almost no matter what. Consider, for instance, that the Toronto Maple Leafs increased their average ticket prices from about $75 (in inflation-adjusted United States dollars) in 2005-6 to $140 this season (or from about $85 to $140 in inflation-adjusted Canadian dollars) — despite never qualifying for the playoffs in the interim. They have almost always sold out their stadium anyway.
These results can be generalized across other seasons. I ran a regression analysis that estimates per-game ticket revenues as a function of the number of hockey fans in a given media market, the number of points that a team averaged between the current season and the previous one, and a time trend. (Point totals are adjusted relative to the league-average number of points to correct for seasons shortened by labor disputes and changes to N.H.L. rules; the number of fans is expressed as a natural logarithm.)
For American teams, an additional regular-season win (worth 2 points in the standings) yields about $10,600 in extra ticket revenues per game, or roughly $430,000 over a full regular season. For Canadian teams, the marginal value of an additional regular-season win is about $1,400 per game (or $60,000 per season), which is not statistically or practically significant.
So the Canadian N.H.L. teams may suffer from a version of the problem that the Chicago Cubs faced during the “bleacher bum” years. Their fans are so loyal — happy enough to turn out for the spectacle and the beer even if the team stinks — that the franchises don’t have all that much incentive to put out a competitive product.
Whether this has actually translated into complacency or mismanagement is harder to demonstrate. Even if N.H.L. teams are constrained in how much they can spend directly on player salaries, they could make other expenditures that should eventually translate into on-ice success — for example, by investing more in scouting and development. The Toronto Maple Leafs, the Montreal Canadiens and the Vancouver Canucks each employ 20 or more scouts, according to their Web sites — well above the N.H.L. average of about 15. So there is some evidence that they are making a good-faith effort to produce a quality team. (I wouldn’t want to slander any professional sports franchise by comparing it with the Chicago Cubs.)
Nevertheless, a strong case can be made that these teams could stand some additional competition from N.H.L. franchises in their media markets and elsewhere in Canada.
I estimate, for instance, that there are five million N.H.L. fans in the greater Toronto region (which I define fairly liberally to include the outskirts of the Golden Horseshoe): about twice as many as in the New York metropolitan area, which has three N.H.L. teams. And Montreal and Vancouver are not far from New York in their number of N.H.L. fans.
Even acknowledging that additional N.H.L. teams in Toronto would not expect to draw as well as the Maple Leafs — just as the Islanders and the Devils do not have nearly as many fans as the Rangers —
; adding a second team in the Toronto area seems all but certain to produce another profitable N.H.L. franchise in a league where many teams are unable to break even. The new team would need to capture only about 14 percent of the Toronto market to reach the 700,000-fan threshold that seems to guarantee profitability in the N.H.L. A good case could also be made for a third Toronto-area team, along with second teams in Montreal and Vancouver.
At the other end of the fan-interest scale, it seems very difficult for N.H.L. teams with fewer than about 300,000 hockey fans in their media markets to turn a long-run profit under the league’s current economic system. One might hold out hope that the newer hockey markets in the United States will grow to provide more revenue for their teams, but the evidence has not supported this conjecture so far. Per-game ticket revenues for United States teams in nontraditional media markets have grown at a rate of just 1.4 percent per year (inflation-adjusted) over the past 18 seasons, compared with 2.6 percent annually for traditional United States hockey markets and 4.2 percent annually for Canadian teams. These struggling United States teams hurt the Canadian teams both directly by diluting the share of Canadian teams in the league, and indirectly by compelling a salary cap structure that is meant to protect the struggling American teams (but which has yet to make most of them profitable).
My best guess is that the economically optimal distribution of N.H.L. franchises would look something like the schema in the chart below. This would include two new teams in the greater Toronto area, one new team in Montreal and one new team in Quebec City. In lieu of a second team in Vancouver, Seattle — a marginal hockey market but probably better than several United States cities that already have N.H.L. teams — would get a franchise in the hope that support might spill over into British Columbia and other parts of the Pacific Northwest. New York would retain its three N.H.L. teams as the Islanders sought to find success in Brooklyn, while Los Angeles (which has no more N.H.L. fans than Philadelphia or Boston and fewer than Vancouver or Montreal) would be shaved to one. The six United States markets with fewer than 300,000 N.H.L. fans would lose their teams.
This would yield a league with 11 Canadian franchises out of 28 — just shy of 40 percent — a level that comes much closer to the share of N.H.L. interest and revenues between among the two countries. And it would all but ensure that Canada’s Stanley Cup drought ends sooner rather than later.