The Limitations Of Sentiment, Revisited

In a blog post in March of this year I discussed the limitations of sentiment as a market timing tool. I wrote that while it can be helpful to track the public’s sentiment and use it as a contrary indicator, there are three potential pitfalls associated with using sentiment to guide buying/selling decisions. Here are the pitfalls again:

The first is linked to the reality that sentiment generally follows price, which makes it a near certainty that the overall mood will be at an optimistic extreme near an important price top and a pessimistic extreme near an important price bottom. The problem is that while an important price extreme will always be associated with a sentiment extreme, a sentiment extreme doesn’t necessarily imply an important price extreme.

The second potential pitfall is that what constitutes a sentiment extreme will vary over time, meaning that there are no absolute benchmarks. Of particular relevance, what constitutes dangerous optimism in a bear market will often not be a problem in a bull market and what constitutes extreme fear/pessimism in a bull market will often not signal a good buying opportunity in a bear market.

The third relates to the fact that regardless of what sentiment surveys say, there will always be a lot of bears and a lot of bulls in any financial market. It must be this way otherwise there would be no trading and the market would cease to function. As a consequence, if a survey shows that almost all traders are bullish or that almost all traders are bearish then the survey must be dealing with only a small — and possibly not representative — segment of the overall market.

I went on to write that there was no better example of sentiment’s limitations as a market timing indicator than the US stock market’s performance over the past few years. To illustrate I included a chart from Yardeni.com showing the performance of the S&P500 Index (SPX) over the past 30 years with vertical red lines to indicate the weeks when the Investors Intelligence (II) Bull/Bear ratio was at least 3.0 (a bull/bear ratio of 3 or more suggests extreme optimism within the surveyed group). An updated version of the same chart is displayed below.