A Surfeit of Sadness

Let’s take a look at sentiment from the perspective of Drach Market Research, a small firm that was run by Bob Drach for decades out of the Florida panhandle before he passed away and was succeeded by his son. Bob used to be one of my favorite curmudgeonly contrarians when it came to sentiment. He was doing his own brand of quantitative research to prove his points of view well before computers hit the mainstream, and his son has picked up where he left off.

This week, Drach takes a look at the latest AAII sentiment survey, which reports that only 24.3% of investors have a bullish six-month stock outlook. That is an improvement from last week’s 21.1% (and a 20.0% reading on June 20). For reference, he reports, this number bottomed at 18.9% during the 2009 lows and averages close to 39%.

On a week-to-week basis, sentiment data is noisy, but it can be useful to study the extremes. For example, when bullishness is at or below today’s levels, Drach notes, the average future drawdown is less than half of the average drawdown across all market environments (-7% vs -19%). In other words, by the time people get this morose, a sizeable decline has typically already taken place.

The data is more telling when you have two consecutive weeks of bullishness below 25%. Drach says the last two times we saw back to back weeks at these levels were in July 2010 (a point from which stocks advanced and never looked back) and the week before the March 2009 lows. Based on this condition alone, Drach concludes, the probability that we are at permanent lows (meaning stocks only go higher from here) is 28%, vs. 11% for a standard market environment.

While this data is promising for bulls, he notes, there are some worrisome periods where a few weeks without optimism preceded major market declines — most recently in early 2008.

While this data is promising for bulls, there are worrisome periods where a few weeks without optimism preceded major market declines.

As a result, instead of only looking at two weeks, Drach turns his attention to 22 weeks. That is how long bullishness has been below average. There have only been two such streaks since the AAII survey began in 1987, he says. In each instance, the largest future decline was only -4%. At this stage in each of those streaks, however, investors were still more optimistic than they were today. One of these streaks extended beyond 22 weeks, eventually reaching levels as low as this week’s readings. When it did so, it marked a market low never to be revisited, he says.

Drach observes there are a few things that make the current sentiment environment different from most. One is that we are still near record highs. Generally, sentiment is driven by price action — and negative price action leads to negative sentiment. To get two weeks below 25% bullish, he reports, stocks have had to be down 15% on average.

But just because this time is a bit different does not mean investors are any better at predicting the future. In fact, when bullishness is this low this close to record highs, he says, the average future drawdown is only 3%. When posting back-to-back weeks below 25%, bulls this close to highs, the worst subsequent drawdown was -4%. Conclusion: Investors are not good at predicting changes in the trend.

One last point: A unique aspect of today’s sentiment is that while bullishness is low, bearishness is also not high. At present it is basically average. This tends to be favorable historically, Drach says. When both bullishness and bearishness measures are this low, by definition uncertainty is high. At these levels, there is a 25% chance we are at permanent lows, and a 90% chance that stocks will lose no more than 5% from these levels. The average future drawdown is only 2%.

Drach’s conclusion, based on the sentiment data: No matter how you look at it, the current level of caution out there should be seen as a positive for stocks. It’s a provocative perspective, but not unreasonable. I’ll keep an eye on this data for you as the summer and autumn progress.

Best wishes,

Jon Markman