Today, let us briefly address sentiment -- what it is, what it means and how to use it in your everyday trading. There is no piece of market data that is more misused, misunderstood or misapplied than sentiment.
The spark for today’s diatribe was a State Street study of cash allocations in investment portfolios. The study found that investors worldwide had as much as 40 percent of their holdings in cash in 2014, up from 31 percent in 2012. In the U.S., cash rose to 36 percent in 2014 from 26 percent in 2012.
This led to lots of news media coverage (see ``Fear of Equities Drives More Investors to Cash'' and ``Investor distrust drives rise in cash holdings''). Huge cash holdings aren't what we typically think of when discussing investor complacency. The always-savvy Michael Santoli pointed to this as an example of bullish desperation.
I am less convinced. Most of the time, sentiment data contains a lot of noise and not a whole lot of actionable information.
We can break sentiment data into two distinct groups: survey and market. Survey data is what we get when we ask someone a question about related market issues. It has an anecdotal component in which someone describes what they are doing in the markets. Market sentiment is some measure of price based on actual buying and selling. This includes a wide variety of indicators ranging from put/call ratios, percentage of stocks higher than their 200-day moving average, money flow, etc.
How should the typical investor use sentiment? The short answer is rather rarely. Sentiment is more or less meaningless much of the time. Some have argued it is as high as 90 percent (but I am working on reducing my hyperbole).
The longer answer is much more complex and nuanced.
Sentiment has various definitions, depending upon whom you ask. The American Association of Individual Investors (AAII) does a weekly investor sentiment survey, and their definition is “the percentage of individual investors who are bullish, bearish, and neutral on the stock market for the next six months.”
My definition: “Sentiment is a measure of investor emotion that often correlates strongly to the investor’s current portfolio posture. When they are long equities, sentiment trends bullish; when they are less long, in cash, or short, it is bearish. Sentiment vacillates between various states, and on occasion can reach extreme levels of emotion, usually described as fear or greed.”
The basic concept behind measuring sentiment is that somehow it can give insight into future market movements. The emotions of investors usually reflect some sort of a balance between buyers and sellers. When one side of the market begins to become crowded, it should show up in various sentiment readings, allowing a savvy trade to take the other side of the market from the crowd.
Unfortunately, it doesn't quite work that way. As a trading tool, sentiment has its greatest value when it reaches extremes in markets and investor psychology. These are, by definition, rare events. Crashes come along infrequently (more so over the past 15 years; less so before). Market tops are novel events -- not that you could tell from reading certain websites, where every day is the top and tomorrow is a disaster.
Lately, I have been seeing a recommendation for a trade based on every twitch in AAII data. If anyone can point me to a study that shows the predictive value of this data, I would love to see it, for in my experience, it is mostly useless. If anyone is making money consistently off of the AAII sentiment surveys, I am unaware of it. If someone wishes to correct my ignorance, it would be much appreciated.
The value I place on sentiment data is that it informs us of people’s investment posture. We become bullish after we buy equities; once we dump stocks, we adopt the bearish view. Sentiment is a rationalization process, one that investors engage in to make themselves feel better about the risks they have assumed and the potential for loss. Sentiment reveals less about what people think, and more about what they just did.
Unlike Garrison Keillor’s Lake Wobegon, where all the women are strong, all the men are good looking, and all the children are above average, we must deal with a normal Gaussian bell-curve distribution. Hence, all of us will not be above average. So too, all of us can't be contrarians.
To contact the author of this article: Barry Ritholtz at email@example.com.
To contact the editor responsible for this article: James Greiff at firstname.lastname@example.org.