The Forecasting Systems Letter
Jeffrey Mishlove, PhD, CTA

---

The argument is often made by bears, such as Richard Russell in his Dow Theory Letters, that the reported earnings yield of the S&P 500 stocks is far lower than the historical average.  The "earnings yield" is simply earnings as a percentage of the stock price.  Furthermore, Russell notes, previous bear markets did not come to an end until the earnings yield of stocks became extremely attractive.  To illustrate his argument that today's earnings yield is very low by historical standards, the following graph plots the S&P 500 earnings yield back to 1943:

Richard Russell, among others, fortifies his argument by pointing out that stockholders in the past were much more interested in dividends than in reported earnings.  As we have seen in recent months, reported earnings can be little more than the result of accounting tricks.  Or, for a variety of reasons, they can evaporate quickly.  But, dividends are actually paid to stockholders.  Therefore I am also including the following chart that plots the dividend yield of the S&P 500 stocks since 1943:

It seems quite clear to me from both charts that today's reported earnings and actual dividends are both well below historical average levels.  From these charts, it would seem that stock prices could well drop by 50% or more (or earnings and dividends grow by 100% or more) before one could reasonably argue that, by historical standards, stocks were fairly priced.

Many bulls, however, argue that stocks are fairly priced today -- and the run up in stock prices over the last week suggests that many people adhere to this view.  These people argue that we cannot look at earnings and dividend yields in isolation from alternative investment choices.  Rather, these yields must be viewed in comparison to interest rates.  With this perspective in mind, I prepared the following charts -- showing the earnings and dividend yields divided by interest rate of the 30 year, long term treasury bond (a standard and safe alternative investment to stocks).

At first it might seem that even this figure is low today as compared to historical standards.  But, if one were to consider only the period since about 1980, the argument could be made that -- given the low interest rates -- stock are a good investment, today.  As you might imagine the chart of dividend yields divided by the long bond rate, shown below, looks very similar:

Again, whether or not we consider today's levels to be reasonable depends upon how far back we wish to look.  It does not seem terribly unreasonable, I suppose to argue that two decades is quite sufficient.  By that standard, today's stock prices do not seem to be out of line.  What is important to note in both of the charts above is the spikes that appear in 1974 and again in 1978 -- at the end of the last serious bear markets.  If we continue to see bear market psychology repeat itself, we may need to see similar levels today (meaning greatly reduced stock prices) before the present bear market ends.

It worth asking, however, why it was the case in the 1940s and 1950s and even up to 1978 that stocks produced such better dividend and earnings yields (even taking interest rates into account) than they do today.  One obvious place to look would be the long bond rate itself.  It is charted below:

One interesting feature of the chart above is that it is asymmetrical.  Prior to 1960, the prime interest rate was much lower than it is today.  And, in comparison to that interest rate, stocks were a much more attractive investment.  It will be interesting to see what happens if interest rates do drop further in the future.  This could happen as a result of deflation.  It could also happen if investors decide that stocks are more attractive than bonds.  From the conventional perspective, one might expect to see stock prices rise further.  But, in the light of this historical picture, I would expect instead to see stock prices remain relatively stable (given their already high volatility), thus elevating the dividend yield in relationship to the prime rate.  

This Friday, I will be flying to Albuquerque, where I will be staying until Tuesday, October 29, for the International Conference on Altered States of Consciousness.  While I am there, I will be presenting a workshop titled, "Mind Over Matter: Psychokinetic Weather Control."  This presentation will largely be based on my book, The PK Man.   I will also be presenting another workshop on research in reincarnation.  These interests are rather far afield from my work in financial forecasting.  The common link, however, is my empirical focus.  Whether, I am looking at parapsychology or the markets, I attempt to reliable, observed facts upon which to build my interpretations.  It is interesting to observe the problems of repeatability and the controversies that exist in both disciplines.  Ultimately both fields represent different perspectives on the functioning of the human mind.

There are, incidentally, many fascinating stories about psychics, astrologers, mediums and mystics who have made and lost fortunes in the financial markets.  From time to time, I will be reporting on these.

I will have my computer with me -- as well as a high speed, internet connection.  Nevertheless, while I am traveling, it will be difficult to maintain the same level of attention to this website and Letter.  I expect things to resume as normal, when I return on October 30.

The neural network signal is down for Wednesday and up for Thursday.  I often find when this situation occurs that the market does not simply turn around clean at Wednesday's close -- but rather beforehand.  With such a signal, it would be typical to see the market on Wednesday down in the morning and then rebounding at the end of the day.  In fact, we have already seen this pattern for the last two days.

At first glance, it appears as if the neural signal has been out of synch with the market since I began tracking the results last week.  Therefore, why pay any attention to it at all?  Actually, the system had been highly accurate the previous week, calling both the ups and downs of the market, and generally accurate in the months before that.  In part, I suppose, it was the extreme accuracy of that signal that prompted me to initiate this public, record keeping process.

It would be quite ironic for the system to experience a dramatic "phase shift" reversal just as I start to post its results publicly.  Since the losses posted for the system thus far (-$13,250) are not larger than other losses experienced by the component systems during successful, out-of-sample testing, at this point, I still expect to see this system return to profitability.

Of course, neural network systems tend to deteriorate over time -- and we may be at a crucial turning point in the market just now.  While I have argued above that we are not yet at the end of the bear market, it is entirely possible that we will see a rally comparable to the one that began about a year ago and ran until March 2002.  It is hard to judge just how a neural network system will respond to such a possible shift in market psychology -- other than to say that the shift that began about a year ago was included in the training period for these neural network systems.  

BioComp Profit Neural Network S&P 500 Futures Contract "MegaSystem" Forecast:
Wednesday's close: Down from Tuesday's close
Thursday's close: Up from Wednesday's close

Nirvana OmniTrader Composite Technical Forecast for the S&P 500 Futures Contract:
Short-term, aggresssive strategy: Down

Volatility Breakout System, Daily, for S&P Futures Contract: 
Buy Stop Order for Wednesday at 899.00

View the Results of Previous Forecasts
Forecasting Systems Letter Archive
Return to ForecastingSystems.com