A++ Work (quality, no plaigarism, on time)

Applied Economics, 2009, 41, 3209–3214

Attendance and promotions in

minor league baseball: the

Carolina League

Richard J. Cebula, Michael Toma* and Jay Carmichael

Economics Department, Armstrong Atlantic State University, Savannah,

GA 31419, USA

This empirical study investigates determinants of attendance at minor

league baseball games in the Carolina League in 2006. The focus of the

analysis is on the effect of a wide variety of game-day promotions on

attendance on a game-by-game basis, rather than aggregate attendance

during the season. The Ordinary Least Square (OLS) results imply that

attendance is positively a function of per capita income in the city or

county hosting the team, runs scored by the home team, Friday and

Saturday games, and promotions that provide cost-reduced food or

beverages, low- and high-value merchandise and post-game fireworks.

Attendance is negatively a function of home team errors, Monday games

and possibly rainy conditions during the game. An unusual finding with

respect to minor league baseball is that team performance variables affect

attendance. However, home team runs scored and home team errors

contribute to the overall entertainment experience for the home team fans,

and thus yield plausible effects on attendance.

I. Introduction

The operation of Major League Baseball (MLB)

teams is a remarkably complex enterprise involving

the marketing of a diverse multi-dimensional enter-

tainment commodity (Demment, 1973; Scully, 1974,

1989; Baade and Tiehen, 1990; Quirk and Fort, 1992;

Zimbalist, 1992; Burger and Walters, 2003). Indeed,

as a consequence, there has developed rather sophis-

ticated theoretical as well as empirical literature

dealing not only with baseball but also with other

professional sports as well as amateur sports, partic-

ularly in the US (El-Hodiri and Quirk, 1971; Koch

and Leonard, 1978; Grimes and Chressanthis, 1994;

Vrooman, 1995; Solow and Krautmann, 2007). 1

At the uppermost level is the MLB franchise team’s

playing games in either the National League or

American League. This level of marketing involves

myriad forms of de facto ‘services’/‘commodities’,

especially the playing of MLB games (predominantly

in the form of regular season games), which generates

revenues not only through ticket sales, television

revenue and radio revenue but also through conces-

sion sales (soft drinks, beer, hot dogs, popcorn,

candy) and merchandising, for example, the sale of

team baseball caps, shirts of star players, baseballs,

bats, pennant flags and the like. At another level of MLB is the multi-tiered system

of minor league teams, a mechanism through which

screening of players with greater potential for MLB

playing occurs and through which development of

players with talent occurs such that at least some

*Corresponding author. E-mail: [email protected] 1 The reader is also referred to the innovative survey by Fort and Quirk (1995).

Applied Economics ISSN 0003–6846 print/ISSN 1466–4283 online � 2009 Taylor & Francis 3209 http://www.informaworld.com

DOI: 10.1080/00036840903286323

portion of minor league players eventually, and sometimes quickly, are ‘called up’ to the MLB team for a chance to make the MLB team roster.

Managers of minor league teams want to maximize team success, as well as to help develop players to reach their potential. Arguably, the most successful minor league teams develop players through a com- bination of coaching/direction, conditioning and other means. Arguably, more successful minor league teams help to avert their own extinction over the long run by attracting larger crowds. Presumably, these larger crowds serve to generate favourable attendance data and revenues that make them less of a financial burden to the MLB franchise. Teams with poor attendance records are more likely to be a financial burden and ultimately become candi- dates for phasing out. Moreover, in theory, when ‘successful’ minor league teams attract larger crowds, they can in effect use the ‘roar of the crowds’ to encourage (‘psych’) young would-be MLB candidates to respond to the crowd and play to their capacity so as to attract the attention of their host MLB team while becoming more accustomed to playing in front of larger and perhaps more vocal audiences. Indeed, learning to adjust to heckling may be yet another side benefit of performing in front of larger (and arguably more vocal) crowds.

Attendance at minor league games is the focus of this study. In particular, the objective of this study is to identify key factors that determine the attendance record of minor league baseball teams. To ensure greater comparability of data between teams and hence relevance of the results, this study focuses upon a single grouping of teams, the Carolina League, and a single minor league baseball season, 2006.

2

II. The Framework

The framework of analysis is one in which attendance at minor league baseball games is largely a reflection of factors influencing the demand for home team tickets for game j. To begin this analysis, it is argued that the higher the per capita income in the host

county (or host city) for a minor league team, the greater the demand for tickets in that county, ceteris paribus, as implied directly or indirectly in a number of prior studies (Baade and Tiehen, 1990; Fort and Quirk, 1995; Cebula and Belton, 1996; Solow and Krautman, 2007). The term PCIj represents the 2005 per capita income in the host county or host city where game j was played. Naturally, the demand for minor league tickets is expected to be a decreasing function of ticket price, ceteris paribus. The term TPj represents the price of a general admission ticket on game day for the home team’s j-th game. Team performance has been argued/found to profoundly affect the economic well-being of professional base- ball teams (Baade and Tiehen, 1990; Fort and Quirk, 1995; Cebula and Belton, 1996; Solow and Krautman, 2007). This study measures team perfor- mance for the j-th minor league team in two ways: the cumulative mean number of home team fielding errors per game over the course of the season (ERRj); and the mean number of runs scored per game by the home team over the course of the season (RUNj). The demand for tickets is expected to be a decreasing function of ERRj, ceteris paribus, and an increasing function of RUNj, ceteris paribus. Arguably, home team fans prefer their team to make fewer errors (manifest good fielding/defense) and score more runs (manifest good offense). Next, minor league baseball fans presumably prefer to attend games when the weather is not rainy, ceteris paribus. The variable RAINj is a binary variable indicating whether there was precipitation present during the course of game j.

Arguably, the demand for minor league game tickets might reflect various marketing efforts direc- ted at attracting fans by making attendance a more pleasurable family experience. General data reflecting such marketing efforts for each of the teams in the Carolina League assume the following four forms: LOWVALj (a binary variable reflecting whether low value merchandise was ‘given away’ upon entrance to the stadium at game j, e.g. key chains or magnetized team schedules),

3 HIGHVALj (a binary variable

indicating whether higher value items were given away upon entry into the stadium at game j, e.g. hats, jerseys or helmets),

4 FOOD/DRj (a binary variable

indicating whether discounts or specials on

2 The teams in the Carolina League (and their respective MLB affiliations and county or city plus state where located) are as

follows: Frederick Keys (Baltimore Orioles, Frederick County, MD); Kinston Indian (Cleveland Indians, Lenoir County, NC); Lynchburg Hillcats (Pittsburgh Pirates, Lynchburg City, VA); Myrtle Beach Pelicans (Atlanta Braves, Horry County, SC); Potomac National (Washington Nationals, Prince William County, VA); Salem Avalanche (Houston Astros, Salem City, VA); Wilmington Blue Rocks (Kansas City Royals, New Castle County, DE) and the Winston-Salem Warthogs (Chicago White Sox, Forsyth County, NC). 3 Also included in this category of promotions are mugs, bobble heads, calendars, water bottles, mouse pads, posters, team

photos, baseball cards and stadium replicas. Such items can, in theory, tend to generate a degree of spectator loyalty. 4 Also include in this category are shirts, blankets, backpacks, gym bags, baseball caps and more.

3210 R. J. Cebula et al.

concession items such as two-for-one hotdogs at or before game time j were offered) and FIREWKSj (a binary variable indicating whether a fireworks show/display occurred following the conclusion of game j). In each of these four cases, the expected impact of the marketing policy/tool is expected to be positive, ceteris paribus.

Finally, there are the temporal control variables, that is, variables that reflect the day during the week when game j was played. Arguably, such a variable is needed to control for the fact that families are more likely to attend games on certain days of the week, especially Friday and Saturday, when the working adults in the family are relatively more available, than other days. Accordingly, dummy variables to reflect whether game j was played on Monday (MONj), Tuesday (TUj), Thursday (THj), Friday (FRj), Saturday (SATj) or Sunday (SUNj) are included in the model.

III. Empirical Model

Based upon the arguments provided above, the following reduced-form equation is to be estimated

PERCAPACITYj

¼ a0 þa1PCIj þa2TPj þa3ERRj þa4RUNj

þa5RAINj þa6LOWVALj þa7HIGHVALj

þa8FOOD=DRj þa9FIREWKSj

þa10MONj þa11TUj þa12THj þa13FRj

þa14SATj þa15SUNj þu ð1Þ

where

PERCAPACITYj the total attendance at game j, expressed as a percentage of the seating capacity of the stadium where game j was played during the 2006 season for all of the games in the Carolina League, j¼1, . . . , 975;

a0 constant term; u stochastic error term.

The Carolina League consists of eight teams that played 975 games during the 2006 season. The effective demand for tickets is described as a per cent of the stadium capacity in each of the venues where Carolina League games were played. Expressing the dependent variable thus permits comparison across stadiums of different capacities. All variables are for the year 2006. Table 1 provides the data sources, and Table 2 provides basic descriptive statistics

for the variables in the model. Observe that the

mean percentage attendance at the 975 Carolina

League games in 2006 was 52.29%, with a SD

of 27.6%. Based on the arguments in the previous section

of this study, the following are the expected signs on

the coefficients for the economic, team-performance,

weather and marketing variables

a1 4 0, a2 5 0, a3 5 0, a4 4 0, a5 5 0,

a6 4 0, a7 4 0, a8 4 0, a9 4 0 ð2Þ

As for the days-of-the-week control variables, it is

expected that

a13 4 0, a14 4 0 ð3Þ

based on the argument that families can most easily

‘get together’ on Fridays (especially during the

evenings) and on Saturdays, when working parents

are more available. Interestingly, the 2 days of the

Table 1. Data sources

Variable Source

PERCAPACITYj Ballparks of Baseball (2007), http://www.ballparksofbaseball. com/aballparks.htm

PCIj US Department of Commerce, Bureau of Economic Analysis (2005)

TPj Team contacts* ERRj http://www.minorleaguebaseball.

com/milb/stats/ RUNj http://www.minorleaguebaseball.

com/milb/stats/ RAINj http://www.minorleaguebaseball.

com/milb/stats/ LOWVALj Team contacts* HIGHVALj Team contacts* FOOD/DRj Team contacts* FIREWKSj Team contacts* MONj Team contacts* TUj Team contacts* THj Team contacts* FRj Team contacts* SATj Team contacts* SUNj Team contacts*

Notes: *Team contacts – Frederick Keys, Deanna Davis (2006), Assistant General Manager of Ticket Operations; Kinston Indians, Katrina Carter (2006), Director of Sales and Promotions; Lynchburg Hillcats, Erica Marcum (2006), Ticket Manager; Myrtle Beach Pelicans, Dan Kurland (2006), Director of Ticket Sales and Services; Potomac Nationals, Doug McConnell (2006), Box Office Manager; Salem Avalanche, Jeanne Boester (2006), Director of Ticket Operations; Wilmington Blue Rocks, Jared Forma (2006), Director of Ticket Operations; Winston-Salem Warthogs, Brian Shollenberger (2006), Director of Ticket Operations.

Attendance and promotions in the Carolina League 3211

week having the highest percentages of Carolina League games are Friday and Saturday. By contrast, the signs on the control variables for MONj, TUj and THj should not be significantly positive because Mondays, Tuesdays and Thursdays are generally working days for most working parents. The argu- ment regarding SUNj is unclear because although most employed parents are not working on Sunday, the family often has other family obligations, reli- gious attendance and related activities, and possibly preparation for the coming workweek during the evening and/or afternoon on Sundays.

IV. Empirical Results

Estimating Equation 1 by OLS, adopting the White (1980) heteroscedasticity correction yields Equation 4

PERCAPACITYj

¼ 1:35 ðþ0:22Þ

þ0:0012PCIj ðþ8:00Þ

�0:109TPj ð�0:12Þ

�7:831ERRj ð�3:76Þ

þ2:071RUNj ðþ3:65Þ

�15:89RAINj ð�1:87Þ

þ9:579LOWVALj ðþ6:23Þ

þ13:693HIGHVALj ðþ4:87Þ

þ5:662FOOD=DRj ðþ2:31Þ

þ31:696FIREWKSj ðþ11:20Þ

�4:67MONj ð�2:00Þ

�4:23TUj ð�1:62Þ

þ4:25THj ðþ1:65Þ

þ6:40FRj ðþ2:29Þ

þ19:62SATj ðþ7:25Þ

þ0:81SUNj þ0:33ð Þ

;

R2 ¼ 0:39; adj: R2 ¼ 0:38; F ¼ 41:48 ð4Þ

where terms in parentheses beneath coefficients are signed t-values. Seven of the nine noncontrol variables exhibit the expected signs and are significant at the 5% level or beyond; only the ticket-price variable fails to be significant at an acceptable (i.e., 5%) level. Among the control variables, the estimated coefficients on FRj and SATj are both positive and statistically significant at the three and one percent levels, respectively. None of the other control variables are positive and statistically signif- icant at the 5% level. The coefficient on MONj is actually negative and significant at the 5% level. Overall, ceteris paribus, the results for the control variables suggest that Friday and Saturday games are the most likely to attract large turnouts whereas Monday is perhaps a day to avoid scheduling games (if feasible). The coefficient of determination is 0.39, so that the model explains nearly two-fifths of the variation in the attendance variable as defined. The F-statistic is significant at the 1% level, attesting to the overall strength of the model.

As shown in Equation 4, the coefficient on the per capita income variable is positive and significant at beyond the 1% level, suggesting strongly that locating a team in a venue with a higher per capita income acts to raise attendance. By contrast, the coefficient on the ticket-price variable is not statisti- cally significant, suggesting that in the proximal price range of general admission tickets, the ticket price is not a significant factor in determining ticket pur- chases. As a comparison, the average price of a general admission ticket is less than that of an adult ticket to a movie theatre. As for the home team performance variables, ERRj and RUNj, the esti- mated coefficients are both statistically significant at beyond the 1% level and, respectively, negative (as hypothesized) and positive (as hypothesized). Thus, attendance at/ticket purchases to Carolina League games are inversely impacted by fielding errors committed by the home team and positively impacted by runs scored by the home team. Team performance counts! However, the loyalty of Carolina League fans may be tested by inclement weather. Namely, the coefficient on the RAINj variable is negative and marginally statistically significant, implying that rainy weather conditions may dampen attendance, indeed, by as much as 16%.

Lastly, there are the impacts of the marketing mechanisms. The estimated coefficients on each of the four marketing variables, LOWVALj, HIGHVALj, FOOD/DRj and FIREWKSj are posi- tive and statistically significant at beyond the 1% level. Thus, in 2006, when Carolina League home teams offered fans enticements that fell under the umbrella of LOWVAL, attendance rose on the

Table 2. Descriptive statistics

Variable Mean SD

PERCAPACITYj 52.29 27.6 PCIj 21.880 3.379 TPj 6.46 0.86 ERRj 1.153 1.1479 RUNj 4.997 4.234 RAINj 0.008 0.09 LOWVALj 0.217 0.412 HIGHVALj 0.0636 0.244 FOOD/DRj 0.071 0.257 FIREWKSj 0.138 0.346 MONj 0.127 0.333 TUj 0.1456 0.353 THj 0.131 0.338 FRj 0.161 0.368 SATj 0.1589 0.366 SUNj 0.135 0.342

3212 R. J. Cebula et al.

average by roughly 10 percentage points for the game in question, ceteris paribus. Alternatively, when home teams in the Carolina League offered the enticements that fell under the more costly umbrella of HIGHVAL, attendance rose for the game in question on average by roughly 14 percentage points, ceteris paribus. Enticements falling under the umbrella of FOOD/DRj on average for the game in question acted to raise attendance by 6 percentage points, ceteris paribus. Finally, home teams in the Carolina League that chose to offer the costly displays of fireworks (FIREWKSj) on average experienced a 32 percentage point boost in attendance for that game, ceteris paribus. Thus, on average, taken one at a time, each of these marketing options offered by itself had a significant effect on attendance, ceteris paribus.

In pursuing higher attendance levels, however, management must be cognizant of increased operat- ing costs associated with each option and must be very circumspect as to how these (or other) marketing tools might be optimally combined. Clearly, sound marketing strategy would seem to require that management simultaneously consider all of the factors that influence ticket purchases, including day of the week. It is likely, for example, that Monday night fireworks displays or some other costly marketing strategy for Monday night games might well yield negative net benefits. Furthermore, the repeated, commonplace use of any of these marketing tools could easily lead to diminishing returns and disappointing outcomes. For example, offering fireworks displays, which are quite expen- sive, every night over any extended period, would clearly yield diminished benefits over time as fans came to take the displays for granted. The same could prove true for any of the marketing tools. Finally, it would be folly to simply assume that the same marketing strategy that worked in one year would necessarily work as well in subsequent years. Anything from demographic changes to volatile economic conditions to increased competition for the entertainment dollar could act so as to require a dynamic process of marketing formulation and re-formulation.

V. Conclusion

This study has investigated determinants of atten- dance at minor league baseball games. Using data from the Carolina League for 2006, it was found that attendance, expressed in relative terms as the number of persons in attendance as a percent of stadium

capacity, was directly related to the per capita income of the county or city of the host team, the home team’s runs scored record, the enticement of low value marketing, the enticement of high value marketing, fireworks displays, special-offer food and drink deals, and scheduling on Fridays or Saturdays. It was also found that attendance for the home team was negatively affected by a poor fielding record (in terms of fielding errors) and possibly by inclement (rainy) weather. In addition, Mondays are a poor choice for scheduling a home game, whereas general admission ticket prices are not a factor. Finally, it is emphasized that the strength of the marketing tools at management’s disposal offers not only opportu- nities but complex challenges.

References

Baade, R. and Tiehen, L. (1990) An analysis of major league baseball attendance, 1969–1987, Journal of Sport and Social Issues, 14, 14–32.

Ballparks of Baseball (2007) The fields of major league baseball: single a ballparks. Available at http://www.ballparksofbaseball.com/aballparks.htm (accessed March 2007).

Boester, J. (2006) Director of ticket operations, Salem Avalanche. Available at http://www.salemavalanche. com (accessed August 2006).

Burger, J. and Walters, S. (2003) Market size, pay, and performance: a generalized model and application to major league baseball, Journal of Sports Economics, 4, 108–25.

Carter, K. (2006) Director of sales and promotions, Kinston Indians. Available at http://www.kinstonin dians.com http://www.kinstonindians.com/schedule/ show_calendar.asp?caledate¼4/1/2006, http://www. kinstonindians.com/schedule/show_calendar.asp? caledate¼5/1/2006, http://www.kinstonindians.com/ schedule/show_calendar.asp?caledate¼6/1/2006, http://www.kinstonindians.com/schedule/show_ calendar.asp?caledate¼7/1/2006, http://www.kinsto- nindians.com/schedule/show_calendar.asp? caledate¼8/1/2006, http://www.kinstonindians.com/ schedule/show_calender.asp?caledate¼9/1/2006 (accessed August 2006).

Cebula, R. J. and Belton, W. (1996) Economics of the Sports Industry, McGraw-Hill Book Co, New York, NY.

Davis, D. (2006) Assistant general manager of ticket operations, Frederick Keys. Available at http:// www.frederickkeys.com (accessed August 2006).

Demment, H. (1973) The Economics of Professional Sports Teams, Lexington Books, Lexington, MA.

El-Hodiri, M. and Quirk, J. P. (1971) An economic model of a professional sports league, Journal of Political Economy, 79, 1302–19.

Forma, J. (2006) Director of ticket operations, Wilmington Blue Rocks. Available at http://www.bluerocks.com (accessed August 2006).

Attendance and promotions in the Carolina League 3213

Fort, R. and Quirk, J. P. (1995) Cross subsidization, incentives, and outcomes in professional team sports, Journal of Economic Literature, 33, 1265–99.

Grimes, P. W. and Chressanthis, G. A. (1994) The role of intercollegiate sports and NCAA sanctions on alumni contributions, The American Journal of Economics and Sociology, 53, 27–40.

Koch, J. V. and Leonard, W. M. (1978) The NCAA: a socioeconomic analysis, The American Journal of Economics and Sociology, 37, 225–39.

Kurland, D. (2006) Director of ticket sales and services, Myrtle Beach Pelicans. Available at http://www.myrtle beachpelicans.com, http://www.myrtlebeachpelicans. com/schedule/promos/?month¼04andyear¼2006, http://www.myrtlebeachpelicans.com/schedule/promos/ ?month¼05andyear¼2006, http://www.myrtlebeachpe- licans.com/schedule/promos/?month¼06andyear¼2006, http://www.myrtlebeachpelicans.com/schedule/promos/ ?month¼07andyear¼2006, http://www.myrtlebeachpe- licans.com/schedule/promos/?month¼08andyear¼2006, http://www.myrtlebeachpelicans.com/schedule/promos/ ?month¼09andyear¼2006 (accessed August 2006).

Marcum, E. (2006) Ticket manager, Lynchburg Hillcats. Available at http://www.lynchburg-hillcats.com (accessed August 2006).

McConnell, D. (2006) Box office manager, Potomac Nationals. Available at http://www.potomacnationals. com/promotions.html (accessed August 2006).

Quirk, J. P. and Fort, R. D. (1992) Pay Dirt: The Business of Professional Team Sports, Princeton University Press, Princeton, NJ.

Scully, G. (1974) Pay and performance in major league baseball, American Economic Review, 64, 915–30.

Scully, G. (1989) The Business of Major League Baseball, University of Chicago Press, Chicago.

Shollenberger, B. (2006) Director of ticket operations, Winston–Salem Warthogs. Available at http:// www.warthogs.com (accessed August 2006).

Solow, J. L. and Krautmann, A. C. (2007) Leveling the playing field or just lowering salaries? The effects of redistribution in baseball, Southern Economic Journal, 73, 947–58.

US Department of Commerce, Bureau of Economic Analysis (2005) Regional economic accounts, local area personal income, Table CA04. Available at http:// www.bea.gov/regional/reis/default.cfm?selTable=CA04 (accessed August 2006).

Vrooman, J. (1995) A general theory of professional sports leagues, Southern Economic Journal, 61, 971–90.

White, H. (1980) A heteroskedasticity-consistent covar- iance matrix estimator and a direct test for hetero- skedasticity, Econometrica, 48, 817–38.

Zimbalist, A. (1992) Baseball and Billions, Basic Books, New York, NY.

3214 R. J. Cebula et al.