Yves Nosbusch


A question or comment? We would like to hear from you!


Print this article
Print this article






Overconfidence is a flaw that seems to be rooted in the human psyche. It is a fascinating phenomenon, both for its pervasiveness and for the consequences it may have. As such, it deserves close attention when we are faced with major decisions. One of the most frequently cited studies on the subject asked people to rate their driving skills. More than 90% of participants considered themselves to be better drivers then the median of the group. This is of course impossible by definition and the conclusion seems clear: many of the participants simply overestimated their skills.


Since this study was published, many similar examples have been documented in a wide variety of contexts. For example, when students just starting at university were asked to predict their ranking on graduation, the majority anticipated above-average results. Moreover, things do not necessarily improve with experience. Ulrike Malmendier from Berkeley and Geoffrey Tate from the University of North Carolina developed a particularly ingenious approach ̶ based on the non-exercise of company stock options ̶ to identify overconfidence among CEOs. Their measure of overconfidence turns out to be closely linked to poor performance in mergers and acquisitions, suggesting that CEOs who are overconfident about the future of their own company often pay too high a price in these deals.


A related source of mistakes is to overestimate either the accuracy of information or our own ability to interpret this information. A study by Terrance Odean from Berkeley found that the private investors of a discount broker ̶ a financial intermediary which does not supply investment advice ̶ tended to engage in excessive trading, in the sense that their portfolio performance would improve if they bought and sold securities less frequently. This underperformance cannot be explained solely by the additional transaction fees incurred by trading more frequently. The findings of Odean are coherent with the idea that investors overestimate not only the accuracy of their information, which leads them to trade too often, but also their own ability to interpret this information, which leads them to make repeated mistakes. This can explain the underperformance of their portfolios beyond the additional transaction fees incurred.


Another commonly observed phenomenon is “home bias”, i.e. the tendency for investors to favour their own country or region. As a result, many of them hold portfolios with a strong geographic concentration. There are many potential explanations for this, but one in particular deserves to be mentioned in the context of this article. This is the idea that people prefer local investments because they believe that they have better knowledge and understanding of companies that are close to them. However, numerous studies have shown that, in practice, the portfolio performance of local private investors is no better on average than that of investors without geographical links to the area. Home bias is therefore coherent with the idea that investors are overconfident in their ability to assess local companies: by selecting portfolios with a high proportion of local securities, they do not achieve higher returns on average. On the other hand they clearly suffer from a lack of diversification and they could reduce the volatility of their portfolios by using greater geographical diversification.


To err is human. That much we already knew. What is more interesting is that, in many cases, these mistakes exhibit repetitive patterns. This makes them predictable and thus, at least in part, rectifiable. If we hope to correct them, we must first recognise the situations in which they are likely to arise and be on our guard in such circumstances.


Yves Nosbusch

Chief Economist

BGL BNP Paribas

Published in the Luxemburger Wort (in French) on August 30th 2014