In the debate surrounding the economic benefits of hosting mega-sporting events such as the Super Bowl or the Olympics, economists have been toiling away at the data to see if the economic impacts are as boosters and promoters advertise. Unfortunately (or fortunately?) we haven't found much support for the grandiose claims that occur before the event occurs.

I have spent countless hours thinking about this discrepancy (and I am not the only one) and in the past month or so have spent several hours talking with reporters about the expected impacts of the Super Bowl in South Florida, the 2010 NBA Allstar game and 2011 Super Bowl in Arlington and Dallas-Fort Worth, and the Olympics in Vancouver. I point out that I am not the only one thinking about these issues, co-author Dennis Coates, Victor Matheson, Robert Baade, and Phil Porter (to name a few) have also made great(er) contributions in this area of research.

In a soon-to-be resubmitted paper looking at the net impact of sporting events on taxable activity in host cities in Texas, Dennis and I outline four reasons for the disparity between ex post estimates and ex ante predictions (all abstracting away from principal-agent problems that present themselves in ex ante predictions):

1. Tourist substitution: some tourists choose not to visit the host city during the event and are replaced by other tourists. Whether replacement tourists spend more per-capita per-day or more in aggregate is not immediately clear.

2. Participation of locals: locals might participate in the festivities associated with the big event. Some locals might spend their leisure money on attending the event rather than on other local leisure activities, which would contribute nothing new to the local economy. Some locals might transfer future leisure spending to attend the event, thereby perhaps reducing spending in the host city in the future (for example, an individual might spend $600 to attend a World Series game but not spend on greens fees for the next six months). In aggregate the spending might not change over the course of the year, but the impact on the golf sector might be non-trivial.

3. Locals sit out the event - This effect I term the "hunker down" effect. Locals might choose not to go out and about during the event, perhaps not choosing to visit their favorite restaurant, because of the perception or reality of facing huge crowds. If, for example, an individual doesn't visit their restaurant during the event and they are not replaced by a visiting tourist, this might harm the local restaurant because it is unlikely that the local will visit the restaurant twice immediately after the event. The "hunker down" effect therefore introduces a negative influence on local spending during and perhaps immediately following the event.

4. Locals leave the city - This effect I term the "skedaddle" effect. In this case, some locals leave the city entirely and spend money in a different city during the event. This introduces yet another drain on local spending during the event.

The upshot is that even if ex ante predictions were not biased they are likely incorrect because the traditional input-output approach to economic impact studies abstracts away from these potential drains on local spending during the event. On the other hand, ex post studies look at the actual data in "reduced form," that is, ex post studies look at NET effects rather than GROSS effects.

The recent streams of research suggest that the ex ante predictions might be anywhere from 8 to 10 times larger than the ex post estimated net effects. The ex post studies likely miss some spending associated with the event, most likely because it is not possible to know what the relevant geographic market is: is it the host city, the metropolitan area, the general area of the state? However, one reason to focus on the host city, rather than the region, is because generally the host city fronts the expenses to host the event (such as police, fire, bomb squads, stadium construction, beautification efforts). Thus, even if the ex post studies are somewhat off, they are not likely to be 8 to 10 times off.

I like the tags "hunker down" and "skedaddle" because most reporters and students can recognize them and they are somewhat easy to remember. However, there is always a lurking question: just how big are these effects? Just how many people hunker down? Just how many people skedaddle? It is impossible for us to know at this point, there haven't been any studies that I am aware of that have tried to measure any temporary diaspora associated with big-time sporting events.

Anecdotal evidence appeared during the 1996 Olympics in Atlanta hundreds (if not thousands) of people rented out their houses during the games (to the chagrin and consternation of the USOC) and planned to leave town during the games. Unless only those who rented their house left town, the skedaddle effect would seem to be non-trivial as well.

During home football games in Athens, Georgia, there are a large number of student apartments close to the stadium/campus for rent during the weekend. During the same weekends there are several houses for rent that are a bit bigger and a bit further from campus. Student apartments are generally not for rent during the basketball season, suggesting that students are willing to forego the weekend of football related activities. I am not sure if the only people leaving town during a football weekend are those who rented their apartment.

I admit that the possible magnitudes of these effects are currently the focus of thought experiments, butI am confident that future research by myself and others will continue to flesh the effects out in more detail and help us contribute to the public debate in a meaningful (and helpful?) way.

Given my thoughts in this area and the work of others who suggest similar substitution effects associated with big-time sporting events, it was reassuring to read this post by George Washington University professor Lisa Neirotti at Sports Business Daily:

Based upon informal conversations, it appears half of the population considers the Games an expensive inconvenience and are leaving town — newspapers report approximately 250,000 residents flying the coop — with the other 50 percent excited to welcome the world.

50% are leaving town? Now that might introduce a huge drain on local spending relative to a normal February, thereby reducing the net impact of the games on the local economy. Moreover, because tourists generally stick to "touristy" areas and do not necessarily frequent those establishments that the locals know and love, the impact of the Olympics can be very disparate throughout the city. Some restaurants and bars are jammed to the gills while other restaurants across town are severely worse off than they normally would be during a February.

I love economics - more right than wrong, notwithstanding recent navel-gazing by some economists.