You have heard this story. I am a good investor! Why? Because I have a good memory of my past trades and those trades have all worked out. This is a nice narrative, but it is not true. Some recent researchers formed some structured experiments and found that most investors have a memory bias. (See "Investor Memory".)
The researchers tested for a memory effect or bias with investments. They found that subjects over-remember positive outcomes and under-remember negative returns. This pattern of remembering leads to overly optimistic beliefs and reinvestment. Any bad trades will soon be forgotten while those winners will be front and center in an investor's mind. It takes a special investor to focus memory on both good and bad decisions.
Other researchers have found that positive memory biases lead to overconfidence and more trading. If I think I am right, I should do more of the same. (See "Investor memory of past performance is positively biased and predicts overconfidence") However, by exposing investors to their bad trades and training them not to have a positive bias, this overconfidence and memory bias can be reduced.
A quantitative model will not have a memory bias. It will view positive and negative decisions the same. There is no extra training needed, no overconfidence, no belief bias, and no excessive trading. If you don't use a model, at least spend time reviewing bad trades to minimize a memory bias.
No comments:
Post a Comment