When Trading Becomes Social: How Social Trading Platforms Affect the Disposition Effect

Abstract

With a market size in the multibillion of dollars, social trading platforms have, over the last decade or so, been gaining a strong foothold in individual investment. Users on these platforms can observe traders’ detailed transactions over time. They can also ‘‘follow’’ anyone of those traders, just like with other social media platforms, investing own money in accordance with the strategies of their trader of choice. In turn, traders have access to an abundance of real-time data, including those that view and follow them. In this paper, we investigate the intriguing question as to whether the disposition effect bias those individual traders typically experience in offline investment setting – i.e., exhibiting risk aversion for gains and risk seeking for losses, manifests in the social trading platform context. We also examine the impact on the disposition effect of two social features of the platform: (i) views to the traders’ page, which reflects the visible allocation of platform users’ attention to individual traders and (ii) followers of the traders, which reflects the allocation of money to mirror the trading allocations of traders. We construct a dataset of 400 traders and 80,000 trades over the period between January 2019 and May 2020 that we collect from Zulutrade, a leading social trading platform. Using a survival analysis methodology, we evaluate whether gains vs. losses exhibit different hazard rates for to sell transactions, and whether and how views and followers interact with gains vs. losses in impacting those hazard rates. We find traders to exhibit the disposition effect in the social trading context, with the disposition effect amplified with an increase in views and reduced with an increase in followers. Our findings have important implications for the design of social trading platforms and more broadly for the literature on decision-making biases in online contexts characterized by social features.