Understanding the Plott-Wit-Yang Paradox

Plott, Wit & Yang (2003) conduct a betting market experiment and nd: First, information was aggregated. This suggests that traders updated their private information based on observed market odds. Second, a model based only on the use of private information seems to fit their data best. The authors call this paradoxical. Because the original data are lost, we replicate their experiment. Our results suggest that the paradox seems due to aggregate rather than individual level data analysis. We analyze the individual level data and explain the paradoxical results reported in Plott et al. (2003).