Humans in charge of trading robots: the first experiment

Elena Asparouhova,Peter Bossaerts,Xiaoqin Cai,Kristian Rotaru,Nitin Yadav,Wenhao Yang
DOI: https://doi.org/10.1093/rof/rfae007
IF: 4.4
2024-03-15
Review of Finance
Abstract:Abstract We present results from an experiment where participants have access to automated trading algorithms, which they may deploy at will while still trading manually. Treatments differ in whether robots must not be halted, deployment is compulsory, or robots can be halted and replaced at will. We hypothesize that robot trading would reduce mispricing, and that the effect would be more pronounced as commitment degree increases. Yet, compared to manual trading only, we observe equally large and frequent mispricing and, in early trading, significantly higher bid–ask spreads and more frequent flash crashes/price surges. Participants earn more, provided they combine robot and manual trading. Compared to evidence from archival data, we find significantly higher use of liquidity-taking robots. We attribute this to the inability, in the field, to identify the presence of liquidity takers when they happen not to trade.
economics,business, finance
What problem does this paper attempt to address?
The main issue this paper attempts to address is whether the introduction of automated trading algorithms (i.e., trading robots) in financial markets can reduce asset price bubbles and improve price quality. Specifically, the authors investigate the following aspects through an experiment: 1. **Reducing price bubbles**: The authors hypothesize that since trading robots require participants to formulate strategies in advance (similar to the "strategy method" in behavioral game theory), this will prompt market participants to consider future market price trends more cautiously, thereby reducing the frequency and scale of price bubbles. 2. **Improving price quality**: The authors further hypothesize that the use of trading robots will make market prices closer to fundamental values, thus improving the quality of prices. 3. **Impact of different commitment levels**: The authors designed three different experimental treatment conditions: (1) no commitment (participants can stop and replace robots at any time), (2) commitment (once a robot is deployed, it cannot be stopped), and (3) penalty (participants will be penalized for not deploying robots on time). By comparing the results under these treatment conditions, the authors aim to explore the impact of different commitment levels on price quality and market behavior. 4. **Combination of manual and robot trading**: The authors also study participants' choices between manual trading and robot trading and their impact on market performance. Overall, this paper aims to experimentally verify whether the introduction of trading robots can improve the price quality and reduce price bubbles in financial markets, and to explore the impact of different commitment levels on this effect.