Is there a Consensus?: An Experimental Trial to Test the Sufficiency of Methodologies Used to Measure Economic Impact

Abstract

This research utilizes local GDP in a panel data set of 383 MSAs in the U.S. during 2012-2017 to determine whether historical methods in the academic literature to measure the economic impact of sporting events, facilities, and teams are sensitive enough to generate conclusive statistically significant results (i.e., are the methods used able to avoid making a Type I error, rejecting a true null hypothesis). An experiment is created where a random set of MSAs receives an injection of various levels of economic impact to their GDP (treatments range from $25 million to $1 billion). Standard panel regression techniques utilized in the sports economics literature to measure economic impact are then tested on both the baseline model (all MSAs with their actual GDPs) and the experimental model (containing all MSAs with the experimental group having received the treatment). The findings show that the historical methods used in the literature fail to be able to detect the built-in-by-design injections of economic activity for the experimental group until very high levels of treatment of at least $300 million to $1 billion annually are present, thus providing evidence that Type I errors are likely to have occurred in some of the literature.