Abstract
The Winter Olympics occur once every four years during the first few weeks of February, which is in the middle of the National Hockey League (NHL) regular season. Prior to 1998, the International Olympic Committee prohibited participation by professional athletes. Beginning in that year, the governing body changed the policy to explicitly permit professionals to participate. For the next 20 years, the NHL allowed players to represent their country by participating, in part based on the hope that it would increase demand for the sport. In 2017, the league reversed this decision citing fatigue and significant disruption to the regular season. Previous policy analysis has focused on post-Olympics team performance effects of player participation. To our knowledge, no previous studies have considered the effects on demand for NHL games. Controlling for a variety of other possible factors, we study the effects of the Olympics on game-level attendance. We find that having players on the team roster who participated in the Games does increase attendance in the short run. Thus, it is possible that the original policy shift by the IOC was correct as the benefits to Olympic participation may outweigh the costs.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
The basketball competition is held during the Summer Games.
Helene Elliott. 1995. NHL Stars will compete in 1998 Olympics. Los Angeles Times. September 30, 1995. At https://www.latimes.com/archives/la-xpm-1995-09-30-sp-51793-story.html. Accessed March 2, 2021.
2019 State of the league address. NHL.com. At https://www.nhl.com/video/2019-state-of-the-league-address/t-277350912/c-68311803. Accessed January 25, 2021.
Shortly before the start of the 2022 Games, the NHL changed their policy and did not permit players to participate due to COVID concerns.
Dan Rosen, July 10, 2020. NHL, players agree to return to Olympics in 2022, 2026. NHL.com. At https://www.nhl.com/news/nhl-players-return-to-olympics-in-2022-2026/c-317375420. Accessed January 25, 2021. Greg Wyshynski. February 13, 2020. The Olympics Debate is a Sham. ESPN.com. https://www.espn.com/nhl/story/_/id/28692763/the-nhl-olympics-debate-sham. Accessed January 25, 2021.
Nicholas J. Cotsonikas. 2017. “Plenty of reasons for the NHL to pass on the Olympics.” NHL.com. April 3, 2017. At https://www.nhl.com/news/history-shows-olympics-dont-help-nhl/c-288389004. Accessed March 2, 2021.
Dan Rosen. 2017. NHL will not participate in 2018 Olympics. NHL.com. April 2, 2017. At https://www.nhl.com/news/nhl-will-not-participate-in-2018-winter-olympics/c-288385598. Accessed March 2, 2021.
Another possible strategy would be to test the impact of Olympic participation on international media rights and licensing. Unfortunately, we do not have access to the data required do so, as it would require detailed country data on such arrangements.
The 2012–13 season is excluded here as it was shortened significantly by a labor stoppage. Outdoor games, which typically have attendance several times that of an indoor game and are not evenly distributed across teams, are excluded from these averages and from our analysis below.
Mike Chairi. 2014. Olympic hockey 2014. Tracking stats for every NHL player at Sochi. February 11, 2014. At https://bleacherreport.com/articles/1955884-olympic-hockey-2014-tracking-stats-for-every-nhl-player-at-sochi. Accessed March 2, 2021. 2014 Men’s Olympic Hockey. At https://www.hockey-reference.com/olympics/years/2014. Accessed March 2, 2021. Attendance data can be found at https://www.hockey-reference.com/leagues/NHL_2014_games.html. While we regard Hockey-reference.com as a highly reputable source for NHL data, we acknowledge their disclaimer that “Attendance figures and arena capacities may not be 100 percent accurate and are subject to minor errors.” We use only attendance figures in our analysis and do not make use of their reported capacities.
We also tested whether our model would change if we added a dummy variable indicating when the teams that did not make it to the playoffs became officially eliminated. The coefficient for this variable was statistically insignificant and did not alter our main results. In an effort to keep our final model parsimonious, we did not include this variable in the final specification. We are grateful to an anonymous referee for this suggestion.
To mitigate concern that Olympians might have an impact on winning in the post Olympics period, we estimated a probabilistic model to analyze the impact of having more Olympians on the probability of winning a game. After controlling for accumulated points of home and away team prior to each game, we found no statistically significant impact of the number of Olympians on the probability of winning.
Notice that our model proposed in Eq. (1) estimates a non-linear relationship between attendance and the number of Olympians. Therefore, if we want to compare the marginal impact of Olympians on attendance starting at a different point than zero, then we need to adapt the change and consider the decreasing marginal returns of Olympians. For instance, if we want to measure the marginal attendance of having a third Olympian in a team, i.e., from 2 to 3 Olympians, we should calculate it in the following way: two Olympians = (531.23 × 2) + (−51.61 × 22) = 856.02. And three Olympians = (531.23 × 3) + (− 51.61 × 32) = 1,129.2. Thus, the marginal effect of a third Olympian on attendance is equal to: three Olympians—two Olympians = 273.18.
We create seven dummies to identify each of the eight possibilities to quantify the number of Olympians that each NHL team had, i.e., a number between 2 and 10, except for 8 as no team had 8 Olympians. Formally, we regress the following model:
$${Attendance}_{it}=\lambda {Post}_{t}+ {Post}_{t}.\sum_{g=3, g\ne 8}^{10}{\beta }_{g}.{Treatment}_{ig}+{X}_{it}+{\rho }_{i}+{\theta }_{t}+{\mu }_{d}+{\delta }_{y}+{\epsilon }_{it}.$$To test for the existence of an attendance break in other years, we conducted a falsification test in which we ran the same regression as the one represented in Fig. 3 for the seasons 2011–12 and 2014–15. We included a dummy variable for the Olympic break and a treatment variable measuring the number of Olympians in 2014. Except for an increase from two to three Olympians in 2011–12, we found no increase in attendance for the teams with more Olympians.
It is possible that some of the increase in attendance could simply be pent up demand to attend NHL games. However, if this were the primary driver of the attendance change, we would not observe differences based on the number of Olympians per team. While such an effect is possible, we are not able to test for a baseline case of increases in attendance simply due to the break because no team had fewer than two Olympians.
In 2013–14, the NHL All Star game was not played because of the Olympic Games. First and second team All Stars were named by the Professional Hockey Writers Association. We use this list of names as our All Stars. NHL Announces 2013–14 All Star Teams. NHL.com at https://www.nhl.com/sharks/news/nhl-announces-2013-14-all-star-teams/c-723852. Accessed May 4, 2021.
We found similar results when we considered 30 days after the actual end; 45 days before the actual date; and 45 days after the actual date. Using 15 days (before or after), we found statistically significant results similar to those found in Table 3. This is not unexpected, as 15 days is close to the actual end of the Olympics and, therefore, can still be affected by the outcome we are trying to identify and measure.
References
Acemoglu D, Autor DH, Lyle D (2004) Women, war, and wages: The effect of female labor supply on the wage structure at midcentury. J Polit Econ 112(3):497–551
Berri DJ, Schmidt M, Brook SL (2004) Stars at the gate: The impact of star power on NBA gate revenues. J Sports Econ 5(1):33–50
Buraimo B (2008) Stadium attendance and television audience demand in English league football. Manag Decis Econ 29(6):513–523
Buraimo B, Simmons R (2015) Uncertainty of outcome or star quality? Television audience demand for English Premier League football. Int J Econ Bus 22(3):449–469
Coates D, Humphreys BR (2012) Game attendance and outcome uncertainty in the National Hockey League. J Sports Econ 13(4):364–377
Coates D, Humphreys BR, Zhou L (2014) Reference-dependent preferences, loss aversion, and live game attendance. Econ Inq 52(3):959–973
Humphreys BR, Johnson C (2020) The effect of superstars on game attendance: evidence from the NBA. J Sports Econ 21(2):152–175
Jane W-J (2016) The effect of star quality on attendance demand: the case of the National Basketball Association. J Sports Econ 17(4):396–417
Krautmann AC, Berri DJ (2007) Can we find it at the concessions? Understanding price elasticity in professional sports. J Sports Econ 8(2):183–191
Leeds MA, Sakata S (2012) Take me out to the Yakyushiai: determinants of attendance at Nippon Professional Baseball games. J Sports Econ 13(1):34–52
Longley N (2012) The Impact of international competitions on competitive balance in domestic leagues: the case of the National Hockey League’s participation in the Winter Olympics. Int J Sport Financ 7(3):249–261
Paul RJ (2003) Variations in NHL attendance: the impact of violence, scoring, and regional rivalries. Am J Econ Sociol 62(2):345–364
Rockerbie DW (2016) Fighting as a profit maximizing strategy in the National Hockey League: More evidence. Appl Econ 48(4–6): 292–299. https://doi-org.lib-proxy01.skidmore.edu/http://www.tandfonline.com/loi/raec20
Rosen, D. (2017, April 4). NHL will not participate in 2018 Olympics. At https://www.nhl.com/news/nhl-will-not-participate-in-2018-winter-olympics/c-288385598.
Scelles N (2017) Star quality and competitive balance? Television audience demand for English Premier League football reconsidered. Appl Econ Lett 24(19):1399–1402
Treber J, Mulcahy L, Sharma MB (2018) Empty seats or empty threats? Examining the effects of the 1994–1995 and 2004–2005 lockouts on attendance in the National Hockey League. J Sports Econ 19(5):677–695
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix A
Appendix A
Variable | (1) | (2) | (3) |
|---|---|---|---|
Attendance | Attendance | Attendance | |
After Dummy | − 1,311.83* | − 1,075.49 | − 1,092.60 |
(691.94) | (682.53) | (681.04) | |
\(\widehat{{\beta }_{1}}\) | 583.57*** | 490.95** | 531.23** |
(220.06) | (217.76) | (227.06) | |
\(\widehat{{\beta }_{2}}\) | − 50.06*** | − 48.72*** | − 51.61*** |
(19.03) | (18.51) | (19.42) | |
Year = 2014 | − 632.35*** | − 670.61*** | − 860.37*** |
(230.39) | (225.27) | (231.02) | |
2.february | 93.02 | 63.77 | 122.53 |
(263.36) | (255.39) | (249.85) | |
3.march | − 465.30 | − 525.58 | − 467.85 |
(502.38) | (488.62) | (479.50) | |
4.april | − 1,037.41* | − 1,229.56** | − 1,201.83** |
(574.52) | (561.91) | (580.71) | |
10.may | 1,156.94*** | 1,078.84*** | |
(371.57) | (358.92) | ||
11.november | 462.96** | 644.06*** | |
(231.61) | (225.42) | ||
12.december | − | − | − |
1.monday | − 522.06** | − 529.74*** | − 783.53*** |
(213.54) | (204.77) | (219.90) | |
2.tuesday | − 597.93*** | − 572.53*** | − 840.55*** |
(180.50) | (173.39) | (192.29) | |
3.wednesday | − 255.57 | − 435.74** | − 484.25* |
(222.56) | (216.39) | (260.59) | |
4.thursday | − 384.69** | − 423.18** | − 634.96*** |
(183.23) | (176.81) | (196.05) | |
5.friday | 559.46*** | 506.45*** | 452.04** |
(192.54) | (185.21) | (213.26) | |
6.saturday | 614.36*** | 571.45*** | 330.06* |
(179.80) | (173.60) | (193.56) | |
2.team | 9,157.52 | 9,338.47 | 8,165.72 |
(395,756.27) | (400,788.71) | (453,541.45) | |
3.team | 2,671.71*** | 3,682.65*** | 4,413.42*** |
(310.72) | (484.46) | (892.77) | |
4.team | 2,960.90*** | 3,766.10*** | 4,033.08*** |
(306.41) | (390.18) | (676.85) | |
5.team | − 988.95*** | − 377.57 | 8.11 |
(296.54) | (361.05) | (576.33) | |
6.team | 5,419.12*** | 5,543.08*** | 5,288.57*** |
(311.02) | (312.62) | (374.14) | |
7.team | − 200.22 | − 59.95 | − 465.75 |
(294.57) | (296.03) | (350.40) | |
8.team | − 1,826.64*** | − 1,383.15*** | − 475.36 |
(297.09) | (357.74) | (526.17) | |
9.team | − 1,040.68*** | − 575.48* | − 119.83 |
(301.62) | (340.68) | (496.04) | |
10.team | 11,612.12 | 11,890.26 | 11,084.98 |
(399,844.24) | (406,428.43) | (446,508.18) | |
11.team | 8,470.37 | 9,085.37 | 8,534.26 |
(394,252.82) | (396,208.16) | (429,204.95) | |
12.team | − 1,803.73*** | − 931.12** | − 215.77 |
(312.07) | (407.38) | (710.17) | |
13.team | 1,614.69*** | 1,870.97*** | 1,675.41*** |
(299.83) | (313.08) | (400.09) | |
14.team | 2,182.55*** | 2,463.73*** | 2,648.23*** |
(297.30) | (320.73) | (444.66) | |
15.team | 12,933.03 | 13,148.86 | 11,944.66 |
(396,926.80) | (399,752.74) | (467,340.03) | |
16.team | 196.79 | 701.46* | 845.90 |
(305.17) | (362.43) | (563.50) | |
17.team | − 1,629.36*** | − 1,145.09*** | − 536.36 |
(297.76) | (365.16) | (551.11) | |
18.team | − 2,585.44*** | − 1,837.01*** | − 1,254.59* |
(305.55) | (395.82) | (680.19) | |
19.team | 9,544.79 | 9,864.89 | 8,897.25 |
(391,579.87) | (390,619.88) | (414,380.11) | |
20.team | 1,818.80*** | 2,398.43*** | 2,808.60*** |
(310.80) | (361.34) | (559.20) | |
21.team | 3,291.72*** | 3,746.79*** | 3,684.22*** |
(295.59) | (354.65) | (495.00) | |
22.team | − 2,806.14*** | − 2,389.93*** | − 1,838.18*** |
(299.34) | (331.19) | (480.78) | |
23.team | 2,161.93*** | 2,338.67*** | 1,863.02*** |
(292.50) | (297.89) | (337.57) | |
24.team | 9,132.20 | 9,311.36 | 7,993.83 |
(399,363.42) | (415,500.27) | (417,425.82) | |
25.team | 1,197.80*** | 1,275.94*** | 1,635.97*** |
(303.37) | (304.99) | (339.04) | |
26.team | 2,502.03*** | 2,923.43*** | 3,170.02*** |
(315.03) | (332.00) | (443.22) | |
27.team | 3,057.10*** | 3,502.77*** | 3,451.22*** |
(301.95) | (332.07) | (483.81) | |
28.team | 10,390.20 | 10,825.97 | 9,941.02 |
(401,027.24) | (397,185.82) | (442,866.51) | |
29.team | 10,121.20 | 10,481.91 | 9,325.57 |
(394,445.05) | (406,856.02) | (456,187.27) | |
30.team | 6,503.87 | 6,906.25 | 6,225.37 |
(397,216.29) | (406,633.69) | (449,866.33) | |
Time sequence | 23.29*** | − 1.54 | − 13.02 |
(5.55) | (8.26) | (12.32) | |
Division | 242.04** | 281.77** | |
(115.63) | (127.93) | ||
Conference | 21.96 | 63.37 | |
(119.11) | (133.14) | ||
Goals favor | − 33.22 | − 26.04 | |
(28.42) | (32.41) | ||
Goals against | − 31.50 | − 39.95 | |
(30.80) | (33.81) | ||
Cumulative points (home) | 31.61*** | 47.90*** | |
(10.90) | (18.43) | ||
Cumulative points (away) | 23.01*** | 21.72*** | |
(4.79) | (4.64) | ||
Distance home-away | − 0.11* | − 0.09 | |
(0.06) | (0.07) | ||
Constant | 19,234.65*** | 14,085.04*** | 12,198.51*** |
(734.26) | (1,291.67) | (1,603.22) | |
Observations | 1,223 | 1,184 | 751 |
Number of team | 30 | 30 | 30 |
Rights and permissions
About this article
Cite this article
Schneider, R., von Allmen, P. & Munck, G. The impact of winter Olympic games participation on NHL attendance. Econ Gov 23, 253–270 (2022). https://doi.org/10.1007/s10101-022-00264-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10101-022-00264-3