The Sky Is (Probably) Not Falling: Cognitive Biases And Investing

Includes: IVV, SPY
by: Ploutos

From the local high close on March 26th to the close on May 18th the S&P 500 Index (SPY, IVV) fell 8.56% in just thirty-eight trading days. Quick market moves lower tend to send investors panicking towards the sidelines. By definition, the move lower has to be populated by more sellers than buyers. Question: How often has a move this rapid and of this magnitude occurred in the market? Answer: Often!

Since the S&P's most oft used gauge went to 500 constituents in 1957, half of the years have featured a market move as severe and as swift as the one we have just witnessed. On trading days where the S&P 500 has experienced the level of loss from March 26th - May 18th over a matched time frame, the next thirty-eight trading days have returned on average 2.3%. Annualizing this figure gives us a return of 16.4%, far greater than the average annual return of the S&P 500 since its advent. Investors who bought when the market had moved lower this quickly were rewarded with above average returns. The annualized standard deviation of the forward returns from these local troughs is roughly 21%, which is not too elevated from average variability, meaning that investors have not taken undue risk to generate alpha from buying at the dips.

The last four years have each featured a move of similar magnitude over a similar time interval. The graph below shows the steepest matched duration market declines and the subsequent market performance over a similar time frame. Markets have bounced back from all of these corrections.

Critics of this analysis might point to data-mining given that since I picked the worst decline in each year, then the next move must logically be higher. There were actually 149 trading days from 2008-2011 that saw a larger decline over the preceding thirty-eight session period than what we have seen over the last seven-plus weeks. Eighty of these points occurred in the dark days of 2008. The average forward return over a matched time frame was 1.95%, an annualized return of 13.7%.

Behaviorists believe that investors are plagued by recency bias, a cognitive bias that makes investors place more weight on recent trends or data than placing their current environment into an appropriate historical context. Nobel laureate Daniel Kahneman noted in he and Amos Tversky's 1974 paper, "Judgement Under Uncertainty: Heuristics and Biases", that test subjects expected that a sequence of events generated by a random process (e.g. short interval stock returns) will represent the essential characteristics of the process even when the sequence is very short. The pair also noted that subjects who were asked to express their beliefs about the range of a quantity (e.g. the Dow Jones Industrial Average on a given day) systematically provided too narrow of a confidence interval. Human behavior may make us focus too closely on recent events (recency bias) and understate the variability of outcomes (overconfidence bias).

Investors in this market should understand that their recent experience weighs more heavily and that the recent move we have seen is far from abnormal. On average, investors will be rewarded for purchasing on these frequent dips, profiting from those who evaluate recent trends and data flow too negatively.

Disclosure: I am long SPY.