HOTEL DEMAND BEFORE, DURING, AND AFTER SPORTS EVENTS: EVIDENCE FROM CHARLOTTE, NORTH CAROLINA CRAIG A. DEPKEN II and E. FRANK STEPHENSON This paper uses daily hotel occupancy data to examine the effects of a wide variety of political and sporting events on the hotel room market of Charlotte, North Carolina from 2005 to 2014. Two political conventions and NASCAR auto races are associated with large increases in hotel occupancy, prices, and revenue, but many other events have no discernable effect on Charlotte’s hotel market. The results also indicate that occupancy effects before or after most events are modest at best. Back-of-the-envelope calculations show incremental hotel-tax receipts fall short of the debt service incurred in constructing and maintaining the city’s sports venues. (JEL H71, Z23, Z28) I. INTRODUCTION days after the event. We test whether there are statistically significant impacts on hotel demand, As the academic literature on the economic hotel prices, and hotel revenue, during the days impact of sports or other events is vast, it is worth leading up to or following the events included in calling immediate attention to this paper’s contri- our analysis. This provides insight into the eco- butions. This paper is the first to use daily hotel nomic effects of sports events and any external occupancy data over a long period to investigate benefits or external costs they may generate. the net impacts of political and sporting events on We distinguish between hotels that are close hotel demand, room prices, and total daily hotel to the Charlotte venues, those that are at an inter- room revenue in a large city in the United States. mediate distance from the Charlotte venues, and High-frequency data allow for granular estimates those that are at some distance from the Charlotte of visitor inflows net of any crowding out or dis- venues although still within the Charlotte metro placement effects (Porter 1999). area. This last point has economic importance While the use of high-frequency hotel occu- because the metro area’s core county has a hotel pancy data improves upon existing research in occupancy tax that is dedicated to debt service several ways, a notable limitation is that hotel associated with a football stadium, a basketball occupancy data alone cannot capture any eco- nomic impact of visitor spending on food and arena, and a sports hall of fame, all in the city drink, souvenirs, and other nonhotel goods and of Charlotte. However, many hotels in neighbor- services. Such activities might be better captured ing counties are within driving distance of the using sales tax data as the measure of economic various venues and are not subject to the hotel activity generated by sports (e.g., Coates and occupancy tax being used to fund the facilities. Depken II 2011). Therefore, the geospatial distribution of hotel The data facilitate a test of the long-touted but untested claim that major political and sporting ABBREVIATIONS events spur tourism spending before and after the event as visitors arrive days before or stay several ACC: Atlantic Coast Conference CIAA: Central Intercollegiate Athletic Association CMS: Charlotte Motor Speedway DNC: Democratic National Convention Depken: Professor of Economics, Department of Economics, University of North Carolina—Charlotte, Charlotte, NC NASCAR: National Association for Stock Car Auto 28223. Phone 704-687-7484, Fax 704-687-6422, E-mail Racing [email protected] NBA: National Basketball Association Stephenson: Gund Professor of Economics, Department of NCAA: National Collegiate Athletic Association Economics, Berry College, Mount Berry, GA 30149. NFL: National Football League Phone 706-238-7878, Fax 706-238-7854, E-mail NHL: National Hockey League [email protected] SOCON: Southern Conference 1764 Economic Inquiry doi:10.1111/ecin.12572 (ISSN 0095-2583) Online Early publication March 9, 2018 Vol. 56, No. 3, July 2018, 1764–1776 © 2018 Western Economic Association International DEPKEN & STEPHENSON: HOTEL DEMAND & SPORTS EVENTS 1765 registrations and revenue for various events can though perhaps small, effects of sporting events influence tax collections in the various counties in cities with large volumes of hotel rentals. For of the metro area and therefore the ability to example, suppose the Toronto Maple Leafs play service the debt incurred to build and renovate four home games in a certain month and that venues in the metro area. each game causes 500 people to have overnight To preview our empirical results, we find the hotel stays. The monthly total of 2,000 fans impacts of various sporting and political events attributable to the hockey franchise might be differ in terms of magnitude, geospatial distribu- difficult to detect since there may well bemore tion, and temporal distribution of hotel registra- than one million hotel room rentals in Toronto in tions, prices, and revenues. The claims that many a given month. events draw tourists for multiple days before and Allmers and Maennig (2009) use monthly after the event are generally not supported; most hotel occupancy data in their analysis of the 1998 events draw, at most, a statistically significant, if FIFA World Cup, hosted by France, and the 2006 not overly large number of people, on the day FIFA World Cup, hosted by in Germany. Their before or after the event. Only large, multiple- results show no net impact on hotel stays arising day events such as a national political convention from either tournament, though their finding that or a multiday National Association for Stock Car Germany had an increase of 708,000 overnight Auto Racing (NASCAR) event appear to gen- stays in June 2006 followed by a drop of 738,000 erate significant net increases in hotel rentals in stays in July 2006 suggests that the World Cup days before and after the event. may have shifted the timing of tourist visits to Germany.1 Similarly, du Plessis and Maennig (2011) II. LITERATURE REVIEW study the effect of the 2010 FIFA World Cup in Studying variations in and factors affecting South Africa using a simple comparison of hotel hotel occupancy is not new. Papers such as occupancy between June 10 and July 11, 2010 Andrew, Cranage, and Lee (1990), Jeffrey and and a comparable period from the preceding year. Hubbard (1994), and Jeffrey et al. (2002) are Their back-of-the-envelope analysis concludes among many examining determinants of hotel that South Africa’s visitors for the World Cup occupancy. However, little attention has been fell well short of ex ante predictions. paid to analyzing the interaction between hotel Coates (2009) uses monthly hotel tax rev- room rentals and sporting events, facilities, and enue from South Carolina counties to examine franchises. Lavoie and Rodríguez (2005) use the impact of Clemson University and University monthly hotel occupancy data from eight Cana- of South Carolina football games and NASCAR dian cities in the 1990s to study the effects of races at the Darlington track. Hotel tax revenue is the 1994–1995 National Hockey League (NHL) a reasonable measure of tourism, but hotel taxes lockout. Also included in their analysis are sev- are typically levied as a percentage of room rental eral events which affected some but not all of rates so hotel tax revenues reflect both any quan- the eight cities in their study: the 1994 baseball tity increases and any price increases caused by players strike, the 1998 National Basketball tourist visits for sporting events. However, to the Association (NBA) lockout, and the departure extent that any price increases accrue to nonlo- or arrival of several NHL and NBA franchises. cal businesses rather than local input suppliers, They find that the NHL lockout is associated the use of hotel tax revenue as a measure of local with significant drops in hotel occupancy in economic impacts is somewhat noisy (Baumann, three of the eight cities affected by the lock- Matheson, and Muroi 2009). Our ability to sepa- out, and in only two of the nine cases are their rately examine the daily number of rooms rented results for the other labor market disruptions and the average rate charged for those rooms, and or franchise arrivals or departures consistent subsequently daily revenue, allows us to separate with increases in hotel room occupancy. Their any price and quantity effects associated with var- results are broadly consistent with the literature ious events. summarized in Coates and Humphreys (2008), Several studies use measures other than hotel which concludes that the evidence of large eco- occupancy to assess the economic impacts of nomic benefits from sporting events is weak. Lavoie and Rodriguez’s use of monthly hotel 1. The 2006 tournament was played between June 9 and data allows for the possibility that their data July 9; that the tournament straddles 2 months is another are not sufficiently granular for detecting real, limitation of using monthly data and month dummy variables. 1766 ECONOMIC INQUIRY sporting events. A common approach is to use had a population of approximately 2.2 million airplane passenger arrivals to isolated destina- of which one million reside in Charlotte (U.S. tions that host sporting events. Nishio (2013) uses Census Bureau 2016). monthly arrival data to New Zealand from 1983 Charlotte hosts two professional sports fran- to 2005 to analyze the effect of that country host- chises, the NBA Charlotte Hornets (formerly ing the 1990 Commonwealth Games, a Cricket Bobcats) and the NFL Carolina Panthers, and World Cup, a Rugby World Cup, two America’s therefore hosts regular season (and occasion- Cups, and tours by the British and Irish Lions ally postseason) NBA and NFL games. The rugby teams. The only event found to increase Charlotte area is also home to the Charlotte inbound traffic overall (i.e., from all countries Motor Speedway (CMS) (formerly Lowes Motor combined) was the 1990 Commonwealth Games, Speedway) which hosts three major NASCAR though increases in inflows from participating races each year.