Does investing “leader” overconfidence (self-attribution bias) transfer bad trading practices to other non-professional investors who participate in a social trading platform? In their March 2018 paper entitled “Self-Attribution Bias and Overconfidence Among Nonprofessional Traders”, Daniel Czaja and Florian Röder employ data from a large European social trading platform to examine: (1) how self-enhancement (attributing successes to self) and self-protection (attributing failures to external factors) components of self-attribution bias affect non-professional trading performance; and, (2) how social trading platforms transfer any such effects to other non-professional traders. The selected platform lets traders (leaders) manage and comment on virtual portfolios publicly. When enough other traders (followers) express interest in such a portfolio, a business partner of the platform offers a product that replicates its performance. After excluding portfolios managed by professional asset management firms, the authors perform content analysis on leader trading comments to measure the difference between first-person pronouns and third-person pronouns as indicators of self-enhancement and self-protection biases. They then relate leader bias to leader future performance and to inflows of associated investable portfolios from followers. Using daily transaction and performance data for 3,519 social trading portfolios managed by 2,010 European non-professional traders and available for investment for at least 360 days, including 45,623 leader comments, during June 2012 through November 2016, they find that:
- Self-enhancement bias indicates future underperformance, while self-protection bias does not. Specifically, average 360-day future raw returns for self-enhancement biased (unbiased) traders is -1.14% (+0.7%). Self-enhancement bias does not elevate average future return volatility.
- Self-enhancement biased traders on average exhibit higher trading frequencies, higher portfolio turnover and lower portfolio diversification than unbiased traders.
- Biased leaders attract significantly more funds than unbiased leaders, suggesting that leader overconfidence positively affects competence and trustworthiness perceived by followers. Specifically, leaders with self-enhancement bias accrue on average 23% (28%) higher increases in assets under management over the next 90 (180) days than unbiased leaders. On average, followers of biased leaders underperform followers of unbiased leaders.
In summary, evidence indicates that self-enhancement bias as exhibited by non-professional portfolio leaders on social trading platforms both underperform unbiased leaders and attract more funds from followers than unbiased leaders.
Cautions regarding findings include:
- A 4-year sample period is short in terms of variety of financial market conditions.
- Findings may not translate to other social trading platforms or to professional portfolio leaders.
