“The Intelligent Investor” by Benjamin Graham is a wonderful book and a profuse source of ideas. Some of them are fairly simple to implement for an individual investor (like 25-to-75-percents rule). Others are less obvious and can be easily misinterpreted. One good example is a principle of market sentiments:

The market is a pendulum that forever swings between unsustainable optimism (which makes stocks too expensive) and unjustified
pessimism (which makes them too cheap). The intelligent investor
is a realist who sells to optimists and buys from pessimists.

(citation taken from the preface notes by Jason Zweig, but the idea is repeatedly reiterated throughout the book)

While idea sounds reasonable and intuitive, it’s not clear how to use in practice. How to understand if investors are too pessimistic? And how much optimism is too much to be sustainable? How to measure the sentiment when half of the internet is screaming about upcoming collapse of financial systems and the other part is absolutely sure that line always goes up?

NAAIM Index

I thought that I found all answers when I first heard about US National Association Of Active Investment Managers and their very own NAAIM Exposure Index. What they do is basically gather data on active money managers’ exposure to equity market. Range of responses is from -200 (leveraged short) to 200 (leveraged long). Then data is averaged and usually results in a number somewhere between 0 and 150. And it looks like most of the dips correspond to market dips:

(Graphs taken from https://naaim.org/ on March 2nd 2025)

NAAIM vs S&P500

Obviously I could not resist the urge to actually try to predict the market using this index. By market I mean S&P500, which is not perfect, but very popular representation of it. Time frame of the research starts in July 2006 and ends in December 2024. And I started with plotting 1-year market deltas versus NAAIM Index values, as the simplest possible measure:

As you can see, there’s almost no correlation visible. The best approximation got from linnear regression has pretty low coefficient of determination:

R-squared: 0.004146148498059832
Mean Squared Error: 0.029878859482074673

For a good linear correlation it would be close to 1. In a pursuit of better results I tried plotting Compound Annual Growth rate instead of 1-year delta:

R-squared: 0.04058949439695614
Mean Squared Error: 0.0008681478096842283
Pearson correlation: -0.2014683458932354

Here correlation is more pronounced. Coefficient of determination is 10 times higher (however, still very far from 1). Pearson’s coefficient of correlation suggests weak negative relation between exposure index and future returns over the years.

Picking days

Theory mathematicians love good statistical results, applied mathematicians prefer cash. As some years offer low index values and high resturns, others show overexposure and expensive market. However outliers exists on the graphs, as well as good bying opportunities each and every year. Therefore I tried to identify those.

Relative to moving average

Hypothesis #1: Low/high values relative to NAAIM Index moving average offer better/worse results respectively.

R-squared: 0.0015039742018334579
Mean Squared Error: 0.0008942895444651768
Pearson correlation: -0.03878110624819058

Unfortunately, not.

Local one-side extremums

Hypothesis #2: Values of the index lower than previous X values provide better returns.

Green dots are local minimums in last 5 values (5 weeks in real time), red dots are local maximums. Difference exists, but can be considered negligible:

Mean (green): 0.1455158309195501
Mean (red): 0.14480137385856384

Conclusions

I spent quite a lot of time thinking about possible applications of this index-market relation, built a few trading strategies and ran backtesting on them. For now, they all resulted in a failure. Why? Here’s my thoughts about it:

• Delay: index is published once per week.
  • Exposure could easily change during that time
  • Market reflects any publicly available information faster
• Incompleteness: index is gathering information from active managers who decided to disclose it. Massive hedge funds and US congressmen/congresswomen most likely don’t do it
• Accessibility: all public and easily accessible information quickly stops providing any benefits, as other market participants also use it

And returning back to Graham, Zweig, and their advice: I still don’t know how to sell to optimist and buys from pessimists. Unfortunately.