Sometimes You Know They Are There, Sometimes You Don't. When Does Algorithmic Trading Matter?

Abstract

This work studies the behavioral and economic effects of the (un)certainty about algorithmic trading (AT) presence in asset markets. We consider two opposite AT trading strategies commonly employed by high-frequency trading, spoofing (SP), identified as a deceptive activity, and market making (MM), seen as a market liquidity provider. We run laboratory asset market experiments to study the influence of (1) the information (public knowledge vs. uncertainty) about AT presence; (2) the type of AT strategy employed (SP vs. MM); on human behavior and price dynamics. From these experiments, besides confirming that the type of AT strategy has a significant influence on human behavior and price dynamics, we find that the (un)certainty about the presence of AT, regardless of the type of AT strategy, matters more than the actual presence of AT. This work contributes to the more general discussion of the effects of the interactions between human traders and AT by suggesting that these effects crucially depend on the available information about AT presence and strategy. Lastly, our results have major implications for market regulation by hinting at a possible trade-off between market transparency and market quality when designing effective HFT regulation.