UBS: Furcht und Gier: Den Investitionsprozess von Emotionen befreien

Behavioral finance research has identified many ways investors allow emotion to get in the way of rational decisions to their financial disadvantage. In their studies, behavioral experts usuallyfocus on the foibles of the individual investor. Does this mean that trained professionals aren’t prone to dysfunctional behaviors, such as selling winners too soon, or holding on to a losing stock toolong in the irrational hope that it will reverse its losses, or that they never hesitate to add to a winner out of the misguided regret that they didn’t buy more in the first place? Many professionalinvestors believe that a formal investment process renders them immune from judgment-distorting emotions. That mistaken belief almost certainly detracts from performance. 

Quantifying the effect of behavioral biases 

In 2012, the US Intrinsic Value Equity team began a behavioral examination of investment decisions made over approximately the last 17 years in its flagship US Large Cap Equity portfolio. With the help of an outside consultant, it developed a custom behavioral attribution analysis of the portfolio decision-making process in each of three categories: buying, selling and sizing. The goal was to find quantitative evidence of behavioral biases and develop a process to minimize or eliminate them.
Grounded in 30 years of price-to-intrinsic value investing, the team’s core philosophy has always been to identify and capitalize on market mispricings caused by the unchecked emotions of other investors. For the first time, the analysis of the process would measure how human behavioral tendencies and emotions like fear and greed affected the portfolio’s performance.

From a naïve perspective, an easy conclusion was that everything was working fairly well. Over the long term, the team had outperformed by about 1.5%. They could see what had worked and what had not worked at the sector and stock selection level, but there wasn’t a clear message from a standard performance attribution identifying how we could improve. By digging in and analyzing, they found that while they were doing well, they could do much better if they improved in a couple of very specific areas. When they separated performance attribution into three areas, buy, sell and sizing, the picture started to getmuch more focused. The buy decision was very strong. That is where most of the alpha was derived. The sell decision added value, but there were clear opportunities for improvement. The position sizing detracted. This was clearly an area where there was alpha leakage.

Diagnosis: The buy

Buy decisions looked consistently strong. A winning buy is one initiating a position that ultimately outperforms its sector, factoring in subsequent additions, trims and the final sale. A losing buy underperforms under the same analysis. During the period under analysis, buy decisions consistently contributed to annual portfolio performance.

Diagnosis: The sell

Sell decisions provided solid proof that emotion influenced the investment process, and offered clear opportunities to improve performance. Sell performance was analyzed over two holding periods — those shorter than the average 22-month holding period and those held for a longer period. Sales of youngerpositions (trims and complete sells) are considered effective if after the sale the stock underperforms the portfolio. A stock that outperforms after it is sold suggests that the position was trimmed or liquidated before the full alpha was captured. Holding on to an older position is considered effective if it outperforms the portfolio from the day it becomes ‘seasoned’ at 23 months until it is sold. Underperformance suggests that these positions are being held beyond the information advantage ofthe buy ideas.

Der vollständige Beitrag als pdf-Dokument