ABSTRACT
While amenity effects generated by sports stadiums or facilities have been studied extensively for housing markets, there has been significantly less attention focused on team effects generated by sports franchises alone. The objective of this analysis is to estimate the impact of Major League Soccer (MLS) expansion on property values, using nearby condominium sales from 2003–2016 in Seattle, Washington. Econometric results from hedonic pricing method and repeat sales regression indicate that property values depreciated after the Seattle Sounders Football Club was promoted to the MLS in 2009. The distance-decaying depreciation in condominium values occurs within a mile of the facility.
Acknowledgments
We thank Xiaoxue Li, Alok Bohara, and Xuanhao He for valuable comments, and Adam Nowak, Brad Humphreys, Johanna Catherine Maclean, and Bern Dealy for valuable data assistance. We also thank an anonymous referee for valuable suggestions and comments.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 See Ahlfeldt and Kavetsos (Citation2014), Ahlfeldt and Maennig (Citation2009, Citation2010), Dehring, Depken, and Ward (Citation2007), Feng and Humphreys (Citation2018), Feng and Humphreys (Citation2012), and Tu (Citation2005).
2 Statistics from http://inrix.com/scorecard-city/?city=Seattle%2C%20WA&index=58 (accessed 5/15/2019).
3 Information from https://news.gallup.com/poll/4735/Sports.aspx (accessed 5/1/2019).
4 Information from https://www.statista.com/statistics/207458/per-game-attendance-of-major-us-sports-leagues/(accessed 5/1/2019).
5 Information from https://www.statista.com/statistics/477837/team-value-of-mls-soccer-teams/(accessed 5/1/2019).
6 Information also available on A-League archives http://a-leaguearchive.tripod.com/(accessed 5/1/2019).
7 Information available from https://www.statista.com/statistics/249667/average-home-attendance-of-the-seattle-seahawks/(accessed 5/1/2019).
8 Given Safeco Field (opened in 1999, and now named T-Mobile Park), home of Seattle Mariners of the Major League Baseball (MLB), is also located next to CenturyLink Field, we also investigated for any drastic alteration in average attendance. Our search found no evidence of drastic alteration in average attendance, and the Mariners also failed to make the playoffs throughout our period of study.
9 To ensure changes in property values are not a direct result of changes in housing characteristics, we disregard difference in sales price between two consecutive transactions that occurred directly before and after renovation.
10 Differing assumptions can be made. Case and Shiller (Citation1988) assume a heteroskedastic error distribution based on the time interval, t – r, between two observed sales, and they propose a Weighted least square (WLS) method to correct for such heterogeneity.
11 One of the limitations of RSR is the exclusion of single sale observations. To address this limitation, Case and Quigley (Citation1991) propose a pooled generalized least squares (GLS) that utilizes and stacks different natures of sales, i.e., single sales or repeat sales, into a single reduced formed equation. Hill, Knight, and Sirmans (Citation1997) propose a pooled maximum likelihood estimation (MLE) approach to correct for an autoregressive component in the error term.
12 To address possible spatial correlation between transactions, we also conduct a spatial lag model (Anselin and Bera Citation1998).
13 Data available: https://info.kingcounty.gov/assessor/DataDownload/default.aspx (accessed 3/19/2019).
14 Geocoder available at: https://geocoding.geo.census.gov/.
15 Specification of treatment and comparison region discussed in the following section.
16 Nominal sale prices adjusted for inflation using Consumer Price Index (CPI). Data available: https://fred.stlouisfed.org/series/CUURA101SAH (accessed 8/10/2018).
17 Analysis of the log price residuals gradient, Appendix ., also indicates depreciation of property values, approximately, within a mile of the stadium.
18 Percentage estimated using .
19 The estimates obtained from WLS, Pooled MLE and Pooled GLS methods are presented in Appendix .
20 Appendix . provides estimated coefficients of all housing characteristics from column 1 of .
21 See Appendix .
22 The HPM analysis conducted with yearly and monthly time fixed effects, Appendix ., resulted in similar estimates.
23 The estimates obtained for alternate HPM analyses are presented in Appendix .
24 Appendix . presents the results from this analysis. Comparatively, our analysis identifies more sale transactions within 1-mile of the KeyArena than Humphreys and Nowak (Citation2017). However, this discrepancy could potentially be attributed to differences in study period specification and data restrictions.