An empirical test of Tobit model robustness in estimating online auction prices over various distributions (Q2204296)

Summary: Data censoring is a common issue in estimating demand and pricing data. The issue is often handled by Tobit models with normal distribution being assumed for its maximum likelihood function. Realistically, datasets can deviate from normal distributions. In this research, we specifically tested Tobit model robustness under distribution variations in online auction markets. We collected data from online auction markets and tested Tobit model robustness against various distributions. Our conclusion showed that Tobit model turned out to be fairly robust. This research provided empirical evidences for the robustness of Tobit estimations in online auction markets.