Discussion Paper
No. 2016-21 | May 20, 2016
Jing Hua Zhang, Kwo Ping Tam and Nan Zhou
Do Smoking Bans Always Hurt the Gaming Industry? Differentiated Impacts on the Market Value of Casino Firms in Macao (China)

Abstract

The gaming economy has expanded rapidly in East Asia during the past decade. Despite the public health hazards of Second Hand Smoking (SHS), smoking bans in casinos remain controversial due to concerns over potential economic losses for casino firms. Applying an event study method, the authors examine the abnormal returns of casino stocks in response to three unexpected announcements of smoking bans in casinos from 2011 to 2015 in Macao. The analysis reveals that the announcements were associated with differentiated abnormal returns of casino stocks. The stocks of the traditional casinos in Macao suffered significant cumulative abnormal losses of 1% to 6%, while the Las Vegas themed casinos in Macao enjoyed significant cumulative abnormal excess returns of 1.4% to 4.8%. Furthermore, the authors find that the low air quality in gaming venues and high dependence on gaming revenues are associated with abnormal losses, while positive management initiatives are significant correlated with positive abnormal returns. This study provides a full picture of the impacts of smoking bans on casinos and will thus be a useful policy reference for the Macao government, as well as for the rapidly growing gaming industry in Asia and other developing economies.

Data Set

JEL Classification:

L83, K32, Z33

Links

Cite As

[Please cite the corresponding journal article] Jing Hua Zhang, Kwo Ping Tam, and Nan Zhou (2016). Do Smoking Bans Always Hurt the Gaming Industry? Differentiated Impacts on the Market Value of Casino Firms in Macao (China). Economics Discussion Papers, No 2016-21, Kiel Institute for the World Economy. http://www.economics-ejournal.org/economics/discussionpapers/2016-21


Comments and Questions



Anonymous - Invited Reader Comments
June 06, 2016 - 08:59
This interesting paper asks an unusual but important question: what is the smoking ban on casino’s stock price? The results suggest that the policy impact can have opposite on different types of casino (negative on traditional casinos while positive on resort-style casinos). This study can help policy makers design more responsive and effective smoking ban policies in Macau, especially targeting traditional casinos. The paper can make more contributions to the field with a theoretical model. The casinos have at least four types of customers (smoking and non-smoking gamblers and non-gamblers). Given the structure of these four types of customers in each casino, management team of the casino will optimize their strategies towards the smoking ban. For example, if traditional casino has the revenues mostly from smoking gamblers, then it will spend resources to lobby against smoking ban. The smoking and non-smoking customers will also take different responses towards the smoking ban. If non-smoking gamblers overweigh smoking gamblers in the traditional casino, perhaps the department of health in Macau has more leverage to convince the casinos to be more cooperative in smoking. The authors can think about the game theory approach to develop a solid theoretical foundation for the paper. It is also important to clarify what is known or unknown about the share of smoking/non-smoking gamblers/non-gamblers in the literature. This evidence will help the policy maker and the casino itself optimize the smoking ban policies. Some field experiments will be interesting to demonstrate the scale of the smoking ban on smoking customers’ behavior change and the spill-over effects on non-smoking customers’ behavior change.

Jing Hua Zhang - theoretical model+smoking share
June 30, 2016 - 10:31 | Author's CV, Homepage
We are grateful for the comments of the anonymous reader. Please refer to the attached file for our detailed response.

Anonymous - Referee Report
June 23, 2016 - 06:44
As an invited referee, I am very happy to read this interesting study. I hereby attach my referee report below. Thank you.

Jing Hua Zhang - event study models
June 30, 2016 - 10:34 | Author's CV, Homepage
We are grateful for the comments of the Referee. Here we have attached our detailed response.

Anonymous - Referee Report
July 04, 2016 - 10:52
This is an interesting paper that examines the impact of smoking bans on the market value of gambling firms in Macao using an event study approach. The topic is motivated well and relevant literature is cited. However, the paper is methodologically weak and its analysis may be incorrect. Cumulative abnormal return (CAR) is typically calculated over the length of the event window/forecast interval. In the dummy variable approach (Karafiath 1988) there is a dummy variable for each day of the event window giving rise to as many dummies as the number of days in the event window. These “coefficients may then be aggregated to provide the traditional cumulative prediction error (abnormal return) over a desired interval” (op. cit. p. 354). But the authors for some inexplicable reason and contrary to established practice add the abnormal return from different calendar years and thus different event window for the same firm in years 2011, 2014 and 2015 and report this as CAR (Table 4). This doesn’t make sense. Further no significance test is done for CAR. The authors address non-normality and heteroscedasticity of residuals by using bootstrapped standard errors. But it is not clear how other issues such as correlation between residuals and Rmt are addressed. More seriously the authors have not controlled for event clustering which will lead to contemporaneous correlation and thus over rejection of null when it is in fact true (Kolari and Pynnonen Rev. Financ. Stud. 2010). In fact, Karafiath (op. cit.) proposes combining the dummy variable technique with Zellner’s seemingly unrelated regressions (SUR) estimation procedure as a solution to the problem of event clustering. The authors have decided to restrict the event window to the actual event day (1 day). This relies heavily on the assumption of efficient markets. At the same time the market model they use (equation 4) has lagged market return which would suggest markets have memory and contradicts the assumption of efficient markets. The explanatory regressions on abnormal returns (AR) in Table 6 are based on very small sample size (10/16) and their relevance may be limited.

Jing Hua Zhang - response letter
July 17, 2016 - 08:21
We are very grateful for the Referee's valuable feedback and comments, which have brought up good questions to our attention for further discussion and clarification. Please refer to the attached file for our detailed response. Hope our answers have sufficiently addressed your concerns and your feedback is sincerely appreciated.

Anonymous - Referee Report
July 18, 2016 - 08:15
see attached file

Jing Hua Zhang - response letter2
July 24, 2016 - 09:44
We are deeply grateful for the Referee's detailed advice and in-depth feed-back to our manuscript. We highly value your constructive suggestions. Attached here please find our response, which reflects our best efforts. Again, we sincerely appreciate your time and your great advice.