Casinos: Looking beyond the hyperbole – A special report...Casino impact is anyone's guess

Editor's Note: In an attempt to evaluate the impact of a proposed casino in Oxford, The Advertiser Democrat presents a multi-part report that examines claims from opponents and supporters, as well as independent research. In today’s edition, we explore the claims about the amount of money that the casino would generate, and the impact on municipal, county, and state government.

OXFORD – Supporters of a local casino say that building a casino will bring an age of prosperity to Oxford County, while opponents claim that a casino would suck money from area residents and sink existing small businesses.
Independent studies suggest that the truth is somewhere in the middle, with a small net gain for the local economy that will redistribute wealth among area residents.
The most reliable study on the impact of gambling, the National Gambling Impact Study Report (NGISR), was conducted by the federal government in 1999, which recommended that “jurisdictions considering [introducing or expanding gambling] should sponsor comprehensive gambling impact statements. Such analyses should be conducted by qualified independent research organizations and should encompass, in so far as possible, the economic, social, and regional effects of the proposed action.”
The National Opinion Research Center found “no change in overall per capita income after the introduction of casinos, as the increases in certain industries are offset by reductions in welfare and transfer payments as well as a drop-off in income from restaurants and bars.”
Overall, the NGISR recognized that, in many communities, casinos do deliver some economic benefits to the areas in which they reside. However, the study says that a cost-benefit analysis of casinos cannot be adequately predicted, because not enough research has been done to measure the costs a casino imposes on those same areas.
A recent study conducted by Harvard University concluded that, for all the hype, casinos have both positive and negative impacts on the counties in which they reside, but that both of these types of impacts tend to be small.
In other words, the casino may have no significant impact on the residents of Oxford County. It may be much ado about nothing.

The report against

In a report submitted to the Oxford County Commissioners, casino opponents Savvy Incorporated argue that the casino will have huge negative impacts on the area.
The report predicts that the casino will hurt the economy of both the state and the region.
Citing a “lengthy track record and dismal relationship to economic development,” the study gives dozens of examples of economic indicators that, in total, seem to strongly correlate the presence of a casino with economic downturns.
However, the study relies largely on anecdotal evidence, even though those anecdotes are statistical in nature. For example, the study notes that in Bangor, the unemployment rate “has risen since the opening of Hollywood Slots, from 4.6% in 2005 to 7.2% in September 2009.”
While this statement it true, the study makes no attempt to document this unemployment rate as it relates to the casino. The unemployment rate in the country is higher in September 2009 than it was in 2005, and there is no evidence that suggests what Bangor’s unemployment rate would have been like without the casino.
Similarly, the Savvy study says that of more than 26 states that announced major budget shortfalls in 2009, 20 of them have casino gambling or slot machines. This statistic holds no weight, because it does not directly track the relationship between a statewide budget shortfall and the presence of a casino. It is possible, and even likely, that a different factor, such as widespread poverty, caused both the budget shortfall and a political climate that made the creation of a casino more likely.
The report draws on a wide range of sources to support its conclusion, some of which are reliable, such as a the NGISR, and some of which are less definitive, such as newspaper editorials from Indianapolis and out-of-context statistics.

The report in favor

A “Yes on 1” white paper entitled “Economic Contribution of Proposed Oxford Resort Casino” touts big revenues and double the number of jobs that casino investors were citing just one month ago.
The study, paid for in part by Maine Taxpayers Taking Charge – the “Yes on 1” group -- and the University of Maine, has been cited as evidence that the casino will bring millions to Maine and its state coffers and 2,784 new jobs.
However, the 18-page report carries numerous caveats – some in tiny footnote print and some clearly in the text.
The study was conducted by Todd Gabe, associate professor, School of Economics, University of Maine. Gabe states clearly in his executive summary that the results of his study “do not imply – one way or the other – whether this (the $126.7 million in gaming revenue and $51.3 million in restaurant and lodging) can be considered 'new' economic activity to the state or region.”
The casino-funded report has been used to make bold predictions about jobs, state revenues, and the local economy, but the report itself is much more modest, noting that it was not constructed in such a way as to measure “new economic activity” in the state.

Gaming revenue estimates

In July of 2010, Black Bear Entertainment, the group of investors that is trying to bring a casino to Oxford County, was estimating that total gambling revenues would be $70.8 M. In September, they revised their figures to estimate a whopping $126.7 M in gambling revenues.
This number, $126.7 million, is used as the basis for every other economic claim made in the report. This is the source of the stream of money that will potentially fill the coffers of the state, county, local, and school governments, create jobs, and offset any potential negative social ramifications created by the presence of a gambling site in the area.
Where does this number come from?
In his research, Gabe found a relationship between gaming revenues in Bangor with hospitality sales, which is the sum of restaurant and hotel purchases, in the area surrounding Bangor.
There is an assumption that this same relationship will hold true in Oxford. Every dollar that is currently spent on food or lodgings within a two and a half hour drive of Oxford is regarded as a potential indicator of gambling money.
Gabe documents that the Oxford market has roughly twice the hospitality sales. The expectation is that slot machine revenue will be roughly double what is spent at Bangor. Where Bangor experiences $59.2 M in slot revenues, Gabe sees $111.2 M in Oxford.
This prediction closely matches a prediction made by a different method, which Gabe employed in his 2007 study. He found an average amount of money that is generated per slot machine. He did two averages, one of which eliminated outliers that had unusually high or low revenues per machine. The remaining states have an average of $71,641 per machine. This leads to an estimate of $107.4 M, which is very close to his current estimate of $111.2 M.
Despite this, there are reasons to doubt that the number will be accurate. And if the number is not accurate, all of the promises that are based on this number start to look a little less promising.
Gabe’ study documents a relationship between hospitality sales and slot machine revenues in the Bangor area. Every time a dollar is spent in Bangor’s Hollywood Slots, 35 cents is spent on a restaurant, and 12 cents is spent on a hotel bed. There is an assumption that the casino drives this relationship, which may not be entirely correct. It assumes that a visitor to the region always comes to the casino and then spends money on area food and lodging, while they never come to a region for another reason (such as a museum, a golf tournament, or a festival), and then spend money at the casino.
In an earlier 2007 study on Bangor’s casino, Gabe makes this very clear.
“Our results do not tell us for certain whether the racino patrons ‘cause’ growth in local restaurants and hotels, or whether activity in these other hospitality sectors contribute to an expansion of gaming,” Gabe wrote in 2007. “In some cases, the racino may attract new visitors to the area who spend additional money at local dining and lodging facilities. In other cases, the racino may serve locals and tourists whose presence in Bangor (and spending in other hospitality businesses) is not solely attributable to Hollywood Slots.” Gabe does not make this as explicit in his current report, although it is almost certainly as true today, three years later, as it was when Gabe wrote it.
The failure to assign any value to sales that come to the casino from local restaurants, rather than the other way around, is one flaw in the study, but it may be a relatively small inflation. Overall, Gabe’s method seems to be a fairly accurate way of determining the raw revenues.
However, there is another problem that may falsely inflate the figures. Right now, Hollywood Slots is the only casino in Maine. Oxford will certainly pull away some revenues from Bangor, which will negatively impact the numbers that Oxford is basing its numbers on.
The market for each casino is defined as the area within a two and a half hour drive of that casino. Given that Bangor and Oxford are only two hours away, clearly there is some major overlap in their market.
It would be like looking at a McDonalds in Norway, figuring out its revenues, and then assuming that you would have the same revenue potential when you opened your McDonalds in neighboring Oxford. You won’t get as big a slice of the pie as the Norway McDonalds did, because the slice of pie just got a little smaller.
There is another problem with Gabe’s research. He arrived at a number of $111.2M based on slot machine revenues. Oxford, if approved, will have not just slot machines, but table games, which have no equivalent in Bangor. He decides to look at the proportion of slot dollars to table gaming dollars in other casinos across the country, and then apply that average ratio to Oxford.
So, he assumes that, in addition to the relatively well-documented relationship between slot machine sales and hospitality sales, there will be additional dollars spent on table games that have no direct relationship to hospitality sales. He is assuming that the presence of a certain number of slot machines will drive new traffic to the casino’s 50 table games, which is not substantiated in his research.
His sampling of other casinos is based on states that generally have 10 to 25 times as much gambling activity as Maine does, and so may not apply. The only other state that has a similar range of gambling revenue (South Dakota) sees a higher percentage of slot machine revenue, 90.5 percent, rather than 87.7 percent.
Whether the number is 87.7 or 90.5 is a small matter. The larger issue is that there is an unsupported expectation that this gaming market exists in addition to the market established by hospitality sales. In other words, there is an expectation that, in addition to the market that will come to play the slot machines, there is an entirely different market, not linked to hospitality sales, that will materialize and play table games.
It might have been a safer bet to consider the total gaming revenue at Hollywood Slots, and equate it to total gaming revenue at the planned Oxford casino. There is much less evidence to support the notion that an additional $15.5 million in sales will happen (although this number does fall within the range of average takes for table games in other casinos).
To use the McDonalds example again, if your local franchise began offering root beer floats and sold a thousand dollars worth of them in a month, that’s not to say that they took in an additional thousand dollars. Most, if not all, of that thousand dollars would have been spent on other drinks at the McDonalds.
This is significant not only because it would adjust the overall number of sales downward, from $126.7M to $111.2M. It is also significant because it would change the tax rate on those dollars, from the 46 percent that comes from slot revenues to a relatively paltry 16 percent that comes from table games. In other words, the casino has an interest in convincing customers to spend their money on table games instead of slots, because the casino gets to keep an extra 30 cents of every dollar spent that way.
Using IMPLAN (Impact Analysis for Planning), which is a widely accepted economic model, a multiplier effect is determined. All of these numbers are valid if, and only if, the original numbers estimating gaming revenue are correct.
State revenues are a simple-calculated percentage of the revenues, and thus are solid if the revenue projections are solid. They, also, are not distinguished from new or existing monies.
It is difficult to precisely measure the impact of these considerations, and come up with a lower number that would be any more accurate than Gabe’s study. There are also other factors that might have a huge impact, such as the founding of other, competing casinos in other states, which would be another big negative, or a rebounding economy, which would be a big positive.
Gabe’s number of $126.7 M, which is critical to all other calculations involving the economic impact of the casino, might be inflated, or it might be too low.
There is simply not enough data to make an accurate prediction.

Cash for state, county, town?

Gabe’s report estimates that that state will reap $52.5 million in additional tax revenues, much of which is targeted for education at colleges and K-12 schools throughout Maine.
An academic study that appeared in the academic journal Economic Bulletin in 2003 looked at data in 48 states from 1988 to 2000, using tax receipts from all forms of commercial casinos. They found that an increase of $1 in state casino tax revenues reduces net lottery revenues by 56 cents. In other words, if the state gets $52.5M in tax revenues (excluding what goes to the county and the town of Oxford), as suggested, it will also lose $30M in lottery revenues, for a net gain of only $22.5 M.
One of the key incentives for support from the county and Oxford’s municipal government has been a cut of the taxes that will be assessed on the casino’s revenues. For every dollar gambled at the casino, three cents will go into county and municipal coffers.
If Gabe’s revenue projections are accurate, that would mean a whopping $3.8 M for local governments. If his projections are low, that would mean even more money. If they are inflated, that would mean less money.
Not enough data exists to fully measure the economic cost that would balance out this benefit. However, there are some indicators.
An independent 2004 study by Baylor University Professor Earl Grinols, whose work has been cited by both sides of the debate, looked at the relationship between crime and casinos, using data from every county in the US (excluding Nevada) both before and after casinos entered those counties.
“We find that crime increases over time in casino counties, and that casinos do not just shift crime from neighboring regions, but create crime,” concluded Grinols. “We estimate the crime-related social costs in casino counties at approximately $75 dollars per adult [resident] per year.”
In her paper, “The Economic Winners and Losers of Legalized Gambling,” Melissa Kearney noted that there was a sharp increase in most crimes after the introduction of casinos…the effect on crime is low shortly after a casino opens, and grows over time… roughly eight percent of crime in casino counties in 1996 was attributable to casinos
If this holds true in Oxford County, with its approximately 43,000 adult residents, the casino will cost $3.75 M in increased crime costs. Compare this to the three percent that is promised to county and municipal governments, and you get a net gain of about $30,000 for the town of Oxford, and $15,000 for the county.
This finding is supported by a 2006 report from the University of Winona, which measured the bottom line impact of casino gambling on the welfare of local residents, using key economic variables. It concluded that “casinos are found to have no statistically significant net impact on the quality of life in their host counties.”
There are also other costs to the county and the town that have not been calculated.
There is a drain on public services such as sewers and road maintenance, according to Rose.
Rose points out that the casinos themselves could more than pay for the increased costs if the municipalities are wise and set this up beforehand via tax rates, etc.
However, “no one has documented longer-term negative economic impacts of sustained casino operation,” the Rose report continues, “Given the long-run payoffs of relatively high investment in education, infrastructures and redevelopment of casino operations versus a comparably sized factory (due to government spending of relatively higher taxes), there is every indication that the long-run impacts will be positive. Even if casino gambling fails, there is no indication that an already depressed area would be any worse off.”