Economic Notes from the Australian iGaming Report

For anyone that has yet to see the Gambling Research Australia commissioned study on Interactive Gambling (by Hing et al.), I’d highly recommend flipping through the executive summary or abstract, there is a wealth of information here.

I pulled out some of the more economically related quotes that I found interesting (in italics below), and provide some limited commentary. As always, use some caution when interpreting survey data from gamblers…

There are several figures that identify the importance of price in iGaming selection, both in terms of choosing iGaming over brick and mortar, and in choosing one site over another:

Price differential, including more bonuses, free credits and better odds and payout rates, was the second most commonly cited advantage of interactive gambling, endorsed by over one‐third of online survey interactive gambling participants.

Price, including free credits and bonuses, was the most common factor influencing choice of online gambling sites (by 43% of interactive gamblers in the online survey), followed by site reputation (30%), and number of betting options (26%).

We should be a little cautious in interpreting “Price” since gamblers often interpret price as the denomination of the wager, rather than the theoretical loss. However, the question does at least include “bonuses, free credits, and better odds and payout rates” as part of its description.

“Price” also appears to be becoming increasingly important over time, suggesting that the market for iGaming is becoming increasingly competitive, and the product is becoming more commoditized (e.g. many sites can now provide a decent experience in terms of software stability, security and  features — this was not always the case)…

The second most commonly reported advantage of interactive gambling as compared to land‐based modes by over one‐third of online survey participants was the price differential, including more bonuses, free credits and better odds and payout rates. The value of price was reported more often in the current survey than in a previous survey of Internet gamblers (Gainsbury, Wood et al., 2012), perhaps indicating that this factor has become more important or tangible for interactive gamblers.

Interview responses provide further insight into the ability of online gamblers to search for the best price online by checking current and/or other websites of various operators. This tendency to search online for the best product and price option, including using mobiles at a retail outlet, is similar to trends in other retail sectors that are driving consumer online purchasing behaviours (Grewal, Roggeveen, Compeau, & Levy, 2012). Being able to access a greater number of betting options was reported as an advantage by over one‐fifth of interactive gamblers in the online survey, which is related to the availability of greater information and access through online sites than land‐based venues.

Most gamblers state that they would prefer to play on domestic and regulated sites, but there still is a large portion of players that do not appear to care. Government-run sites have to be prepared to compete with foreign firms, both in their product and their marketing — there are no monopolies on the internet…

Most interactive gamblers preferred to use domestically regulated sites, although this consideration did not influence choice of site for approximately 33% of interactive gamblers and only 6% considered the jurisdiction where the site is regulated when selecting an interactive gambling site.

Ten per cent of interactive gamblers in the online survey indicated that advertisements and promotions influenced their decision to gamble on a particular site with a further 10% reported marketing and promotions to be critical factors in their initial uptake of online gambling. 

Finally, some interesting demographic characteristics to file away:

…interactive gamblers were more likely to work full‐time with household incomes between $90,000 and $119,000 while non-interactive gamblers were more likely to work in casual or part‐time positions.

The majority (81%) of interactive gamblers were also land‐based gamblers.

Measuring casinos’ local value – an alternative to cost/benefit analysis

I read a new article in Growth and Change that uses an approach novel to the casino industry to measure how casinos affect quality of life in a community. First, a little background – normally studies measuring quality of life are done using a cost/benefit analysis, but in casino gambling, it’s very hard to get people to agree what should count as a cost or a benefit, and just as hard as to how we should measure it (everyone seems to have a different sized ruler).

Wenz uses a different approach, which instead of measuring everything from the bottom-up after a casino opens, he looks at two things from the top-down: local real estate prices and local wages (other researchers have looked at these items separately, but never together, afaik). This “Rosen–Roback model” can give some insight into people’s aggregate tastes for having a local casino. The logic is:

Real estate prices up & wages down? The casino is a positive amenity in the community. People are willing to pay a premium to live close to the casino, and are similarly willing to take a pay cut to live nearby. Think of a small resort town with nice weather, real estate prices are high but local wages do not reflect the high cost of living.
Real estate prices down & wages up? The casino is productive/profitable, but makes the area an unpleasant place to live (possibly requiring a premium in wages to encourage people to live in the area). Think of a factory that produces good jobs, but comes with an ugly view of smokestacks.
Real estate prices up & wages up? Casinos are a productive local amenity, driving local real estate prices up and competing to hire talented workers (possibly stimulating other businesses in area). This is like a local hub for an industry, such as Silicon Valley and tech workers – lots of talent, high wages, and high real estate prices.
Real estate prices down & wages down? Not so good…

So what did he find? After modeling every non-NV/NJ casino opening from 1990-2000, he concluded:

for a hypothetical generic casino of average size in an average casino,… casinos are found to generate a $258 annualized increase in house prices and a $695 increase in earnings …[but] Native American casinos have been nearly entirely responsible for the estimated productivity increases shown in the generic specifications…

He later adds,

This would be consistent with the explanation that Native American casinos improve the entertainment opportunities in their neighborhoods, or that Native American casinos are more effective at returning casino profits to local community infrastructure through local tribal ownership or through the terms of their state gaming compacts.

Altogether, the aggregate effects are not that clear in non-Native American casinos, where some of the results on earnings are not statistically significant at normal requirements, and positive vs negative outcomes are dependent on a number of factors about the local market. In a small town, with a handful of casinos, Wenz shows the effects on real estate and wages can both be quite positive, but when casinos are located in large/dense cities, the effects are somewhat muted. This is not surprising, as it is always much tougher to have a meaningful effect on a large market than a small market.

Another wrinkle from the study,

[The] casino size effect seems to suggest that a larger investment in the casino operation creates local productivity increases, either directly through the casino or through some sort of agglomeration effect…non-Native American casinos compete with or cannibalize other local businesses in more developed areas… It is quite easy to see that this approach might lead to commercial casinos that are viewed as positive productive amenities.

Since much research has shown that casinos do not cannibalize local businesses (see here), I would be inclined to think that non-Native American casinos are driving productivity improvements through competition, but as Wenz cautions,

This result should be viewed with some caution, however, as there are no observations in this data set of gaming towns that have located in high-density areas.

My broad takeaway is that Native American casinos seem to have a powerful and positive impact on their communities through what we can infer indirectly in wages & real estate. Commercial casinos need to more carefully consider many aspects of the market and their design before adoption. For example, urban casinos should generally come with a larger capital investment requirement, as the positive effect of casino size on productivity seems to be fairly clear. This  certainly supports the integrated-resort approach to casino expansion. A very nice paper and one worth building upon.

Longitudinal Gambling Studies & Division on Addictions Editorial

Some good points made by members of the Harvard Division on Addiction regarding prevalence and impact study methodologies in this op-ed. They would like to see the Massachusetts Gaming Commission change their impact study approach to focus on comparisons of the same people over time (longitudinal or panel design) instead of across different people in the state (cross-sectional design).

A few points worth making in consideration of their suggestions (disclaimer – I have not gone through all of the MGC documents, and am taking the Division’s claims at face value):

1)  Good longitudinal studies in gambling research require A LOT of participants when using a random sample, due to small numbers of interesting subjects in the general population. For example, if we want to look at problem gamblers, only ~1 out of every 100 people we call will qualify. If we want to talk to problem internet gamblers, maybe only 1 in 10 of those people will qualify. And if we want to see movement in and out of the problem gambler categories to draw any strong relationships with other variables, we need a lot more people!

It can get pretty expensive to follow ~5k-15k people around with surveys for several years.

2) Even drawing causal arguments from longitudinal studies can be challenging when good control groups are not set up properly. For example, with casino expansion, you would like to have another region outside of the casino catchment area that is not developing gambling services at all — sometimes these ‘twins’ are hard to find.

Editorial here:

Op-Ed/Editorials – Can Massachusetts evaluate the impact of its gambling expansion? – The Brief Addiction Science Information Source BASIS.

Notes from the Toronto Executive Committee Debate

I appeared as a debutante at the City of Toronto Executive Committee session on a new casino and convention development in Toronto today. Here are the notes from my (three minute) presentation:

Dear members of the executive committee,

My name is Dr. Kahlil Philander, and it is my pleasure to be here today to speak about a recent series of casino policy study reports written by International Gaming Institute at the University of Nevada Las Vegas. The UNLV International Gaming Institute (IGI) was founded in 1993 to serve as the intellectual hub for the global gaming industry, and now boasts program graduates who lead gaming institutions and governmental agencies in more than 50 different jurisdictions. Faculty from our institute have provided research-based expertise and testimony in front of many governmental entities, including the U.S. Senate, the Nevada State Legislature, and government officials from Canada, Europe, Asia, and Africa.

My personal expertise is as a scholar in the economics of gambling and gambling policy, and I have worked and published extensively in the areas of gambling economics, tourism, and responsible gambling program design. Along with the executive director of the International Gambling Institute, Dr. Bo Bernard, who himself is a world renowned expert in problem gambling and gambling policy, I have written a five policy papers that address specific myths that arise during casino adoption debates, and that I have observed during this debate. These reports were not based on hearsay or testimonial evidence, but instead relied exclusively on academic peer-reviewed research that can be trusted as the most reliable available evidence.

I am not just a Las Vegas scholar, but also have strong ties to Toronto. Several years ago, I actually lived in the Annex community of downtown Toronto, where I studied under some of country’s finest policy scholars in the University of Toronto’s Master of Economics program.

Although I only have time to speak to a couple issues described in the institute reports, I note that they touch on several issues related to industry cannibalization, problem gambling prevalence, casino social costs, and the economic impacts of casino gambling.

To address the oft-cited “cannibalization” debate, we examined all of the relevant peer-reviewed literature on topic, and concluded that there is no strong evidence to suggest that a GTA resort casino will meaningfully cannibalize incumbent businesses. In fact, we expect that many industries will be stimulated by the casino, given that available empirical evidence has tended to show complementary relationships rather than cannibalistic or substitutionary. This finding was repeatedly found in several different studies, which all varied in terms scope of geography and casino design. Industries such as tourism, entertainment, lodging, food and beverage, may actually observe a positive economic bump from the expansion of casino gaming in the GTA. We also expect that these positive effects will be even greater if an integrated-resort property is built, as opposed to a gaming-only facility.

With regards to the likely effect of a casino on crime, I have read several sources involved in the Toronto debates that suggest that research in this area provides unclear conclusions. This is incorrect. The literature on casinos and crime has produced consistent results in the past 15 years of research. The findings of our review support a view that the proposed casino-resort would increase the total volume of crimes in the area, but that there will be no effect on the overall crime rate, when populations are adjusted for the number of people drawn to the area. That is, with respect to the total volume of crime, casinos seem to have an impact similar to other large recreation/tourism draws, such as a hockey game or the Canadian National Exhibition. With respect to the crime rate, casinos are typically found to have no significant effects. The increase in volume has been repeatedly shown in research to be explained by the number of temporary visitors in the area – meaning that there is no evidence of increased risk of crime-related harm to nearby residents. These findings were consistent between studies that focused on jurisdictions within Canada, and in other international locations.

In terms of problem gambling prevalence, in the most recent and comprehensive reviews of the gambling opportunities and problem gambling literature, researchers previously believed that gambling opportunities lead to linear increases in the problem gambling rate. This was called the “exposure” model. However, in the past ten years, experts in the area have made a compelling argument that this perspective is flawed – or at the very least, incomplete. These researchers suggest that evidence for “adaptation” can be observed, where populations adjust after initial exposure. This adaptation curve can be observed with many diseases, whereby more vulnerable groups develop problems first, but then the disease’s spread begins to diminish as the general population learns more about the disease, better understanding risks and preventative measures.

This “adaptation” perspective also has support in the empirical literature. In the United States for example, problem gambling prevalence rates have remained relatively stable over the past 35 years, despite the introduction of numerous new gambling opportunities during this period. A 1979 study found a national lifetime problem gambling rate of 0.7%, and more recent comparable figures have been 0.4% to 0.6%.

So, while this perspective has been more readily accepted than the “exposure” model in recent research circles, we were surprised that there has not been significant discussion of its merit in this debate. Especially since the population of Toronto has already been exposed to many forms of gambling, and the city boasts World-renown experts in problem gambling science and responsible gambling program design.

While I have only been able to highlight a few of the items that are addressed in our series of reports, the full reports are available at igi.unlv.edu/research. I thank you for your attention to this important issue and if you have any questions, I would be happy to respond to them now.

Casinos and Economic Growth

Doug Walker and John Jackson have added to their past studies on casinos and economic growth, providing the strongest evidence to date that casinos have a positive causal effect on economic growth. Their newest study looks at 12 different U.S. states during the period from 1990-2010 and finds that, “there is, in fact, a positive impact of casino revenues on state-level per capita personal income.”

In their past studies, the authors were able to find evidence of short-term positive effects of casinos on growth and positive effects of casinos during the post-Katrina recovery in the Mississippi Gulf Coast, but that the growth effects dissipated over the long-run. This new article suggests that explanation may have been incomplete.

The authors now believe that casinos may not have noticeable effects on long-run growth during periods of relative prosperity, but that the positive growth from casinos may be more visible during periods of economic stress like the Great Recession that began in late 2007. As noted by the authors: “Perhaps the casino industry was more “recession-proof” than many other industries, and in this way the industry was helpful in mitigating some of the negative effects of the recession in host states.”

Online Poker in the States: March 25th, 2013 | Online Poker Report

A nice summary of online poker bills in America via OnlinePokerReport. It will be interesting to see how these bills might change as jurisdictions like Nevada and New Jersey start courting states for inter-jurisdictional play. Nevada is currently looking to identify how the oversight and financial relationships will be defined in inter-state compacts.

Report here:

Online Poker in the States: March 25th, 2013 | Online Poker Report.

How online gambling affects offline gambling

With all of the recent changes (and proposed changes) in the U.S. online gambling landscape, I have received a few inquiries about what kind of effect that this may have on existing – land based – gaming operators. I have done a bit of research in this area (see here and here), and the general response I like to give is that it really depends on the game.

In my work with Ingo Fiedler, we found a pretty clear complementary relationship between online poker and brick and mortar casinos, running counter to some of the brick and mortar industry claims. Intuitively, this makes a lot of sense. Participation in online poker (often at stakes well below any available in a casino) has gone a long way to reinvigorate interest in the game overall. And, online poker sites often partner with casinos through promotions – for example, winning seats to a major poker tournament through an online qualifier. In many cases, poker sites will offer seats to casino tournaments without even formal partnerships in place. With developments like Pokerstars’ move into the brick and mortar poker World, these effects may get even stronger.

In some other forms of gambling, it seems pretty likely that there will not be this same complementary relationship between the offline and online gambling worlds. Lotteries are a good example. They can usually be grouped as much more of an aspiration (wealth increasing) type of gambling. This is in opposition to poker, which is generally considered to be much more recreational and less about a sudden change in wealth. The convenience of being able to either buy lottery tickets on your computer or at the corner store will surely increase overall lottery sales, which may be appealing to lottery groups. But it seems quite unlikely that in-store sales will increase from expanding sales online. Unlike poker, I expect that the net relationship between online and offline lottery sales will be substitutionary rather than complementary.

With all the other types of gambling, the net effects probably lie somewhere between lotteries and poker. Sports betting may be a little closer to lotteries, since the activity is still placing a wager on a future event that does not require the player to be in the casino – although with so few states having any legal form of sports betting, cannibalism is less of an issue. Casino-style games may be a little closer to poker, since familiarity and comfort with the games may create new gamblers that are less intimidated with unknown rules at the casino, and I expect more opportunities for cross-marketing.

One final comments is that most of the markets where I have looked at these effects have been where online operators were different entities from offline operators. I would certainly expect more integration and positive effects to occur in the states where offline and online casinos, poker rooms, etc. are run by the same companies.

Caesars to start charging resort fees, says guests demanding them – Las Vegas Sun News

via Caesars to start charging resort fees, says guests demanding them – Las Vegas Sun News.

My take:

A dubious statement, but this seems like an outcome that has been a long time coming. The challenge with these types of add-on fees is that the firms that avoid charging them have a hard time gaining any advantage in doing so. Gabaix and Laibson put together a formal model of why this happens.

Basically, customers that are diligent in figuring out the full cost of the room will simply sum the price of the resort  fee and the “sticker” room price when making their decision, while non-diligent customers will just look at the sticker room price. The diligent customer’s behavior will therefore not be influenced by the fee, since he is using the full information in either case, while the non-diligent customer’s behavior will be biased to the “lower observed” price at the resort fee hotel.

By not joining the gang with resort fees, Caesar’s was putting itself at a disadvantage in the same way that airlines that held out on fuel surcharges did during late-2007.

Las Vegas Strip Shooting, Tourism, and Crime Policy

The shooting early this morning on the Las Vegas Strip and the obvious negative implications for the City’s tourism image reminded my of a recent Freakonomics podcast on gun crime policy. The gist of the podcast is that there are not a lot of easy solutions to gun violoence but one thing that does have an effect is tough gun laws with mandatory sentences.

Mostly talk on gambling policy and its impacts by Kahlil Philander