1970-1994

Team performance, attendance and risk for major league baseball stadiums: 1970-1994

William N Kinnard Jr

The market value of a major league baseball (MLB) stadium is, to a large extent, a function of attendance levels1 at MLB team games. Attendance levels determine revenue from ticket sales (and stadium rentals), concession (food and souvenir) and parking.2 What are the identifiable influences on home game attendance for a MLB stadium or team? In an effort to provide supportable answers to this question, we analyzed available published data on average attendance for all MLB stadiums and teams for the period 1970-1994.3 Activities Undertaken By RECGC Sources of Data

With assistance from Research Associates of Virginia, data were assembled on attendance figures, stadium capacity and won-lost records during the regular season (also called “winning percentage” or WINPCT) for all teams in major league baseball. These data were obtained from published sources for 1970 through 1994. We considered this 25-year span adequate to reflect long-term trends as well as cyclical variations over time. Data Gathered

We first organized our information by league (American and National), and then by team, for each year. A data file for each team was developed that included the following information: Team name League

Year franchise began (if after 1970)

Average attendance per home date per year (excludes post-season) Won-Lost record (“Winning Percentage”) Games behind league or division champion at end of season Won league or World Series previous year Capacity of stadium (each year) Date of new stadium (if any) Capacity of new stadium

(if applicable) Dome stadium (Yes-No)

Strike Year (Yes-No)

Games missed (if strike year)

Year (for all annual data) Five teams began franchise operations after 1970: Texas Rangers (1972); Seattle Mariners (1977); Toronto Blue Jays (1977); Colorado Rockies (1993); and Florida Marlins (1993). In addition, there were seven new American League stadiums and two new National League stadiums occupied during the period covered in the analysis (see Exhibit 1). In addition, capacity of the California Angels stadium was increased in 1981 to accommodate National Football League specifications. The New York Yankees played home games at Shea Stadium in 1974 and 1975 while Yankee Stadium was being renovated and its capacity reduced.

We calculated Winning Percentages by dividing games won by total games played each season. The Attendance Percentage was calculated by dividing average attendance per game for each season by the stadium’s official seating capacity occupied during that season. In the case of the Toronto Blue Jays, who occupied the SkyDome in June 1989, a weighted average percentage was calculated, because games were played in two different stadiums during the 1989 season. Tabulations, Graphs And Models Produced After the foregoing information and calculations were assembled, the occurrence and duration of MLB work stoppages (strikes and lockouts) were tabulated, as shown in Exhibit 2.

The Winning Percentage (WINPCT) and Attendance Percentage (ATTPCT) figures for each team were calculated by year. The American League figures are presented in Exhibit 3; the National League figures are in Exhibit 4. From these figures each team’s average Winning Percentage and average Attendance Percentage were calculated for the entire 25-year study period. The Winning Percentage and Attendance Percentage averages also were calculated for 1989-1994 for all 26 teams (a two-year average for the Colorado Rockies and Florida Marlins).

Each of the teams was then ranked by average Won-Lost percentage and by average Attendance Percentage for the two time periods: 1970-1994 and 1989-1994. Those results are presented in Exhibit 5. We plotted the relationships between average Winning Percentage and average Attendance Percentage figures, over the entire 25-year study period, on separate graphs for each team. Eight of those graphs are presented as Exhibit 6. They show figures for the five teams that occupied new stadiums after 1989, plus the Atlanta Braves and Los Angeles Dodgers (high Won-Lost records in recent years), and the Boston Red Sox (an unexplained anomaly). Finally, we developed Multiple Regression models using the entire data set of some 638 separate annual team data files. The most appropriate form and format for the model were identified by testing different combinations of variables, including both Attendance Percentage (ATTPCT) and the Natural Logarithm of Attendance Percentage (LNATTPCT) as the Dependent Variable. SizeAttendance Percentage relationships are typically curvilinear rather than straight-line, since Attendance Percentage has an upper limit: 100%. We therefore chose the model with LNATTPCT as the Dependent Variable.

The best estimator model is presented as Exhibit 7. In that model, most of the independent variables are binary (Yes-No) variables. With the Natural Logarithm of Attendance Percentage as the dependent variable, the coefficients of the Yes-No independent variables can be used as indicators of percentage differences in their impact or influence on Attendance Percentage. Exhibit 7 indicates that such Yes-No variables (Yes = 1; No = 0) included: Dome (Is it a covered dome stadium? Yes-No) American League (as opposed to National League; Yes-No)

Strike (Was it a year in which a strike occurred? Yes-No)

Won League previous year (Yes-No) “Strike” rather than “Games Missed” was selected as the variable to represent work stoppages, because “Strike” proved to be more significant statistically.

True variables with values determined by calculation or observation included: Year (Any year from 1970 through 1994) Winning Percentage and Games Behind (games behind the winner of the league or division at the end of the season) Games Behind rather than Standing were chosen because the former was statistically significant and the latter was not.

For age of stadium, Year of Operation of Stadium (after its opening) was used. The variables were YROPP1 (for Year 1), YROPP2 (for Year 2), YROPP3 (for Year 3), YROPP4 (for Year 4). All years after the fourth year of operations were used as the norm against which the others were compared: 5 years or more. The team variables were also Yes-No variables. To reflect both the team (including the influence of its market area, its reputation and its following) and the capacity of the home stadium for the team, these two factors were combined into a “Team CAP” interactive variable. If a team played in more than one stadium over the 1970-1994 period, that is indicated in the team variables by “CAP 1”, “CAP 2” and (in the case of the New York Yankees) “CAP 3,” as well. Findings

General Findings

First, the Year (“YR”) variable in Exhibit 7 demonstrates quite clearly that, for most teams and for major league baseball generally, there has been an upward trend in Attendance Percentage over time. That increase has been especially evident since about 1980 (see the graphs in Exhibit 6). It is partly explained by some teams moving into new stadiums with smaller capacity: e.g., Baltimore Orioles, Cleveland Indians and New York Yankees. The highly significant positive coefficient value for the “YR” variable indicates an unquestionable underlying upward trend in overall Attendance Percentage.

The graphs in Exhibit 6, particularly when read in conjunction with the new stadium information in Exhibit 1, show unequivocally that (since 1989 at least) a new stadium has resulted in dramatic increases in Attendance Percentage for the affected team. There were smaller increases during the 1970s (in Philadelphia and Montreal), as well as very modest increases in Anaheim (1981), Kansas City (1993) and Minnesota (1982).

It is not at all clear, however, how long the positive effect of a new stadium is likely to last. When is a stadium no longer new and a separate attraction to attend a major league baseball game, irrespective of the field performance of the home team? Exhibit 7 indicates clearly that there is a decline in the positive percentage impact of a new stadium that is statistically significant over the first three years. On average, it is a robust 24 percent in Year 1, still 20 percent in Year 2 and 14 percent in Year 3. By Year 4 the positive impact is no longer statistically significant, but it is still 11 percent.

Conversely, a domed stadium has had a negative (somewhat significant) effect on Attendance Percentage. It is possible that the lackluster longterm field performance of teams with domed home stadiums (except for Toronto) accounts for this. A strike in any year has had a negative and significant effect (3.8 percent, on average). These results, shown in Exhibit 7, indicate further that being in the American League enhances the Attendance Percentage of the home team, but not significantly.

A team’s Winning Percentage is the major (and highly significant) positive influence on Attendance Percentage, after time (YR). In addition, having won the league title and played in the World Series the previous year (PREVLEAG) is a very significant positive influence on Attendance Percentage, while being a greater number of Games Behind at the end of the season is a significant negative influence. Neither of these results is surprising: local fans enjoy seeing the home team win and are not particularly attracted by home teams that are not in contention for the league or division title for much of the season.

The Multiple Regression model in Exhibit 7 is statistically very robust and gratifying. The results are consistent with intuitive expectations. The Coefficient of Multiple Determination (R2) means that over 80 percent of the variability in Attendance Percentage is explained by the model, a notably strong result. Moreover, the F-Ratio is extremely high; its probability indicates there is virtually no chance this model could have emerged randomly. The Standard Error of the Estimate (Std Error [d.f.]), adjusted for degrees of freedom, is lower than for all the other models considered. The model in Exhibit 7 produces the most reliable results. In summary, the results of this model can be used with a high degree of confidence for reliability and statistical significance. Conclusions

1. Not surprisingly, winning is still better than losing; it produces a high level of Attendance Percentage. Indeed, the most important influence on Attendance Percentage, aside from the longterm upward trend for all of major league baseball and the downsizing of new ballparks, is a team’s on-field performance. Higher attendance percentages produce increased revenues from ticket sales and from parking, food and souvenir concessions at MLB stadiums. Dependence of stadium revenues on team performance represents greater investment risk, since third-party owners of MLB stadiums have no control or influence over the on-field performance of the tenant team.

2. Since 1989, a new stadium has been a dramatic stimulus to Attendance Percentage. At first, the ballpark itself is an attraction, almost irrespective of the team’s on-field performance. This initial increase is tempered by general declines over the next five years, unless the team itself remains or becomes a winner. Moreover, some of the increase in Attendance Percentage is artificial when the franchise moves from an oversized stadium to a smaller, more friendly or traditional environment, as in Baltimore and Cleveland. 3. Nevertheless, when combined with a winning team, a new stadium generates notably higher Attendance Percentages for a few years. This is particularly evident for Baltimore, Cleveland and (most especially) Toronto, as shown in Exhibit 6. On the other hand, the new Comisky Park did not help Attendance Percentages for the Chicago White Sox nearly as much, nor as long. Moreover, Attendance Percentages reportedly have fallen off noticeably in Toronto in 1995 and 1996, when the Blue Jays’ Won-Lost percentages and league standing declined.

4. A new franchise helps for a while, but that effect lasts briefly if the team does not win regularly. This occurred with the Florida Marlins and to a lesser degree the Colorado Rockies. Their rankings in Won-Lost records and Attendance Percentages, shown in Exhibit 5, reflect this. 5. It is quite unusual for a team in either league to sustain a high Won-Lost percentage and to win a league/division championship for more than 5-6 years. The resulting cyclical patterns of attendance result in variable stadium revenues. Income variability creates further investment return risk, as well as debt service coverage risk. 6. The investment risk problem is exacerbated when the team franchise regards the stadium’s luxury boxes and club seating arrangements as inadequate. With the exception of the SkyDome, the bulk of MLB luxury box and club seat license fees goes to the team franchise, rather than to stadium ownership. The costs of luxury boxes and the extra amenities of club seating have been borne in recent years by public ownership. These increased costs must be financed and amortized through public debt. Nonparticipation in luxury box or club seat revenues enhances the investment return risk and debt service coverage risk associated with MLB stadium ownership, especially by public bodies. 7. The threat of a move to a new stadium within a team’s franchise area, or the necessity to add or improve luxury boxes and club seating, reduces the reasonably expected economic life of an existing MLB stadium.

8. All of this confirms what has been pointed out regularly in published case studies and anecdotal essays:4 a major league baseball stadium makes no sense as a financial investment. The results of this study reinforce that conclusion and demonstrate that the risks associated with owning event-driven facilities for major league baseball are greater than previously estimated.5

NOTES

1. In this presentation, “attendance” means paid attendance, rather

than turnstile “clicks.”

2. Of course MLB franchises receive local and shared national television and radio revenues, none of which benefit the stadium owner per se. Further, luxury box and club seat license fees usually flow entirely or primarily to the team franchise, rather than to stadium ownership. The SkyDome, at least until 1998, is a notable exception.

3. The major sources for these data were the annual year books for the American and National Leagues, plus the annual Baseball Almanac. Conflicts were resolved primarily through telephone calls to the affected teams.

4. See, for example, Robert Baade’s, “Sports Stadiums and Area Development: A Critical Review,” Economic Development Quarterly, August 1988, pp. 265-275. See also, Kevin Grace, Ballparks: A Research and Reading Guide. Cincinnati, OH: University of Cincinnati, 1994.

5. According to published accounts, SkyDome cost C$475 million to complete in June 1989. It was acquired by a private consortium in the fall of 1994 for approximately C$175 million.

REFERENCES

1. American League Yearbook, 1971-1995. 2. Amusement Business Magazine. 3. Amusement World. 4. Baseball Almanac, 1971-1995.

5. Balliett & Fitzgerald and USA Today Sports Staff, The Complete

Four Sport Stadium Guide, 1995.

6. Financial World, May 10, 1994; February 14, 1995; May 9, 1995; May 20, 19%.

National League Year 1971-1995. USA Today, October 21, 1994.

William N. Kinnard, Jr., Ph.D., CRE, is president of the Real Estate Counseling Group of Connecticut, Inc. (RECGC). He is professor emeritus of real estate and finance at the University of Connecticut and a principal in the Real Estate Counseling Group of America. Kinnard also testifies regularly as an expert witness on methodology for real property and personal property valuation throughout the U.S. and Canada. Mary Beth Geckler, CRE, is vice president of RECGC and a licensed general appraiser in Connecticut. She was a commercial real estate lending officer at two regional banks in Connecticut for nine years after spending nearly a decade in market research, analysis and providing advice to public agencies and educational institutions. Since 1990, she has specialized in market proximity impact studies, evaluation of risks involved in real estate ownership or development and review of special purpose property appraisals.

Jake W DeLottie is a senior research associate at RECGC and has participated in several studies of market rentals in regional and community shopping centers in the United States and Canada. He is a candidate for a Master of Science degree in Computer Science at The Graduate Center, Hartford, Connecticut.

Copyright American Society of Real Estate Counselors Apr 1997

Provided by ProQuest Information and Learning Company. All rights Reserved