Thursday, July 21, 2011

Testing the Efficiency of Sports Gambling Markets

George Diemer and I recently completed a chapter in a forthcoming book edited by Leighton Vaughn Williams and Donald Siegel entitled The Oxford Handbook on the Economics of Gambling. The focus of our chapter was on the economics (and legal aspects) of online sports betting. As part of our research, we delved into the efficiency of online/offshore sports books vis-à-vis terrestrial sports books in Las Vegas and London. It was a fascinating inquiry, as we came across a number of studies that analyzed specific aspects of the sports betting industry. One such study was recently published in the Journal of Prediction Markets (published version here and SSRN version here). Author Jeremy Arkes investigated whether gamblers correctly price momentum in the NBA gambling market. The abstract is below:
There is little research on whether new information is correctly synthesized in prediction markets. Previous studies have found evidence consistent with, but have not proved, gambler misperceptions on the existence of momentum effects in the NBA. I use novel momentum measures that, unlike prior studies, incorporate the strengths of the opponent and the wins (or losses). With these measures, I test whether gamblers correctly synthesize information on momentum in the NBA. Contrary to previous studies, I find strong evidence for the existence of a momentum effect. Furthermore, gamblers incorporate momentum into their beliefs on the game outcomes. Gamblers, however, significantly overstate the importance of momentum. But, there is little evidence that the extent of this gambler misperception is large enough to generate market inefficiencies, or profit opportunity. Still, the gambler mis-pricing of the information has implications for how well new information is synthesized in other types of prediction markets.