Wednesday, September 8, 2010

Investor Sentiment: New Thinking For An Old Tool

I have spent a fair amount of time in these blog writings on deciphering where and how investors are positioned with their capital.  We have the "dumb money" and the "smart money" and all sorts of flavors in between, and if we could avoid being the "dumb money" and understand what the "smart money" is really doing, my investing world would be a whole lot better.  Unfortunately, the market is never really that easy or simplistic.

Extreme data points - in any data series - often don't herald a reversion to the mean but may lead to new trends.  A current and recent example would be the flow of money into bond funds.  Over the past year it has been estimated that more money has gone into bonds funds then into the market during the entire internet bubble.  Therefore, this must be a bubble in bonds.  Too much money has gone into bond funds and it is only a matter of time before the bubble is popped.  Then the amount money in bond funds will go back to a more "normal" level - one that we have seen in the past; that it is, we will see a mean reversion in the flow of money into bond funds.

But consider for a minute that maybe these new "extreme" flows into bond funds are just a new trend that could possibly persist for a very long time such that these new and current extremes set a new level as to what is meant by extreme.  Believe me in the markets it happens all the time.

Another example that I can show you graphically comes from the Rydex asset data and flow of money into energy funds back in 2004.  Figure 1 is a weekly chart of the Oil Services Holders (symbol: OIH) and the data in the lower panel represents the amount of money in the Rydex Energy Services fund.  From 2001 to late 2003, the amount of assets in this fund rarely exceed $100 million.  Then all of a sudden this value hit a new extreme (point #1 on chart), and another (point #2) and another (point #3).  Prices didn't mean revert (as one would expect when too many investors jumped on board), they kept going higher.  The initial and new extreme (point #1) wasn't a sign that led to lower prices and mean reversion.  It was a sign of higher prices to come.

Figure 1. OIH v. Rydex Energy Services Fund (asset data)/ weekly

Is this a failure of sentiment as a useful indicator or such tools and monikers that measure investor sentiment such as the "smart money" and "dumb money"?  Or is this just an inability of "us" analysts and pundits to correctly interpret the information?  In this series of articles I will argue the later.  Most interpret extremes in sentiment as having one outcome - it must mean revert.  That is the "typical" notion, but there is nothing "typical" about the market, and we see extremes getting more extreme all the time.  So over the next couple of weeks, I plan to put forth a newer, better way of interpreting these extreme data points.  

Consider this new thinking for an old tool. 


Captain said...

Guy, I think your absolutely right on in this post.       What is typically considered a "failure" of sentiment analysis can often provide more valuable trading information than a "successful" sentiment trigger.
      My trading in the last decade has been based entirely on sentiment analysis. I learned early on (the hard way) that when sentiment analysis produces ineffective or insignificant trading opportunities, it's because of me and my personal beliefs about what the market "should" do, and not the fault of the trading system.
       Thank you for your excellent insights and when I start actively trading again I will absolutely become a subscriber to your service.

Thanks for all you do....Jeff

ryanmburke19 said...


I think another analogous example in technical analysis is what I would call a "range shift." With bound indicators like RSI, when a powerful bull market is underway, the range shifts such that pullbacks are contained by the 40-45 level on the monthly chart rather than by a more typical 25-30. When that level is finally breached, it usually means the bull market is over and not that you should be buying because the range shifts when the secular trend shifts. Hence what had been a bull signal consistently for years becomes a bear signal.

I have found that intermediate term advance-decline oscillators exhibit interesting patterns as well depending on the underlying trend. During the bear market from 07-09, the market always peaked without making breadth divergences; it seemingly always topped out when things looked best. On the other hand, tradeable bottoms were typically only formed when there were breadth divergences. After the initial pullback following the March 09 low, the oscillator reversed: the market plowed higher despite negative divergences, while pullbacks required only a small oversold reading in the oscillator and no divergences in order to rally.

Thus I think its important to listen to indicators, but that also includes paying especially close attention when they begin to behave in a manner contrary to recent action.