Sentiment is extremely difficult to use as an indicator

One of my favorite pastimes is dissecting accepted Wall Street wisdom to see if it contains any value for investors or traders. Often, upon examination, the widely held beliefs turn out to be closer to magical thinking than financial acumen.

One of the more recent examples is the way some analysts use data on sentiment to determine how much an investor should allocate to equities. The problem is that the sentiment data is inconclusive and sometimes contradictory. There is no signal within the noisy data.

Sentiment is extremely difficult to use as an indicator because it is only rarely at the extreme readings needed to generate a reliable trading signal. Recall the March 2009 low, where every measure of sentiment was deep in the red. Or October 2002, by which time the Nasdaq Composite Index had fallen almost 80 percent from its high and everyone hated tech stocks. These extreme events are rare.

One of the analysts who has observed these phenomena for many decades is Laszlo Birinyi, formerly of Salomon Brothers, and now ofBirinyi Associates. In this weekend’s Masters in Business interview, Birinyi describes why so many sentiment measures are worthless. There is no usable signal in the American Association of Individual Investors bull/bear survey, Birinyi says. Many other such polls also lack a consistent methodology or are otherwise flawed. They are useless to investors, he says.

Lately, I have been seeing a spate of articles and blog posts that try to make the claim otherwise. Many of these are either inconsistent in their understanding of sentiment or misstate the significance.

Related Posts