Does Algorithmic Trading Attenuate Asset Price Bubbles: An Experiment

Abstract

This study investigates the impact of algorithmic trading strategies on asset price mispricing and relative payoffs of human and algorithmic traders using the Smith- Suchanek-Williams (SSW - (Smith et al. (1988))) framework. A 2x2 treatment design varying algorithmic strategy (market-making or sniper) and speed (instantaneous or 5-second speed bump) is implemented. The benchmark treatment involves only human traders. Results show that the market-making algorithm reduces mispricing by 21% compared to the benchmark. The sniper algorithm reduces mispricing significantly less. The 5-second speed bump has no effect on mispricing. The sniper algorithm out- performs human traders significantly, whereas profits of the market-making algorithm are similar to those achieved by human traders.