When is the Best Time to Invest?

Here at "A Dash" I often try to stimulate objective thought by taking the reader outside of the normal market discussion.  Sometimes when you see the point in a context where you do not already have an opinion, it is easier to bring the conclusion back to your investing.

Are you ready for some football?

On Sunday afternoon there was a great AFC matchup of the Jets and the Steelers.  The Jets returned the opening kickoff for a touchdown, a surprising and exciting start.  At 11:03 of the 1st Quarter, CBS displayed a graphic informing us that the Jets were 9-0 when scoring a TD.  Quite remarkable.  Scoring a TD is certainly good to do, but it is far from a guarantee of victory.  Apparently the Jets were 1-3 in games where they did not score a TD.  That is much easier to understand!

At the end of the 1st Quarter the score was Jets 7, Steelers 0.  But wait, Steeler fans!  CBS sent another message, informing us that Pittsburgh was 6-1 when the opponents scored first.  Doing a little subtraction in the standings shows the Steelers were 4-2 when they scored first, a smaller percentage.  Does this mean that it would be good strategy for Steeler Coach Mike Tomlin to allow an early score by the Jets?  Maybe he should try to allow only a field goal, or take an intentional safety, since we know that the Jets always win when they score a touchdown.  Hmm!

Even if you are not a football fan, you should be able to see the lack of logic in these arguments.

What is Wrong

The problem is that the CBS analysts are now able to look at hundreds of different variables, looking for something interesting to report.  The logic, or lack thereof, makes no difference.  My colleague Allen Russell lectured his college methods classes on the topic of statistical signficance in a very innovative fashion.  On a monitor in the corner of the classroom he was tracking results from a computer program.  The program generated a random data set with a few hundred variables.  It then looked for "statistically significant" relationships between each pair of variables.  When one was found, the monitor would flash, Significant! Significant!  (Allen was a great and innovative teacher, and I continue to benefit from his thoughtful comments and research).

(This story is from the era of slow computers.  Nowadays the monitor would be continuously flashing).

Back to the Investment World

The same type of analysis occurs all of the time in the investment world.  I am particularly disturbed by this approach from a powerful and influential voice.  First, the list of independent variables:

1) S&P 500 more than 8% above its 52 week (exponential) average
2) S&P 500 more than 50% above its 4-year low
3) Shiller P/E greater than 18
4) 10-year Treasury yield higher than 6 months earlier
5) Advisory bullishness > 47%, with bearishness < 27%

You should note that we actually have many more than five variables.  The statements include multiple conditions, each one of which can be massaged to a specific level.  Why not 10% above a simple moving average?  Why not 30% above a 2-year low?  Why not bullishness > 53%.  Etc.

This is a well known statistical problem.  There are too many degrees of freedom in the independent variables compared to the number of cases to be explained.  The resulting list of ten "bad times to invest" shows adverse effects within either a few weeks or within 21 months.  How long must we wait to know the answer?  Meanwhile, we are warned that right now is one of those bad times.

The Missing Variable

The missing variable is this one:

A 50% market decline within the prior three years.  Or how about: Negative Quarterly S&P earnings within the last three years.

Viewed in the context of the Great Recession, where stock prices declined to Depression levels and have only recently rebounded to where they were before the fall of Lehman in 2008, the factors are easier to understand.  A 50% rebound is still far from the highs.  As long as the economy is improving, the S&P is going to be above its moving average, and that might continue for a long time.  The Shiller P/E is going to be saddled with 2008 earnings for as long as this backward-looking method keeps it in the calculation.  The 10-year yield, which collapsed to levels reflecting expected deflation, has rebounded somewhat, but is still near historic lows.

If you actually think about the variables, nothing is a real surprise.  There is also no historical precedent, so there is no good basis for prediction.

An Alternative Approach

Last week I noted that a widely circulated email provided analysis that demonstrated the important difference between current conditions and the other nine cases on the "bad times" list.  I am delighted to report that Georg Vrba has now published those results.  I urge you to read his article.

Rather than trying to explain certain situations — incorrectly beginning with the dependent variable — Vrba uses a number of indicators that can be applied to any time period.  In fact, his own published work was completely independent of the list of "bad times" to invest.  He is taking a method and applying it to someone else's suggestion of timing.  This adds to the interest.

Two Final Notes

1. I hope that readers will understand that this is a debate on the merits.  The Vrba article is written with care and respect and I have tried to follow his lead.  In the academic world disagreements about methods and conclusion are commonplace and vigorous.  Somehow we are having trouble translating this honest engagement into the 21st century blogosphere.  Maybe we can all try harder to explain our intentions in 2011.  At the moment the discussion has been conducted in the pages of Advisor Perspectives.  This influential publication (while available to the public) is read mainly by my colleagues in the advisory business.  The articles and commentaries reflect this more sophisticated audience.
2. This topic is important.  I have limited time to write, and I choose my topics with relevance and importance in mind.  I currently have more than 50 items on my "blog agenda."  I selected today's subject because something like a "list of bad times" is seductively dangerous for investors.  I disagree strongly with the conclusion that this is a bad time to invest.  I think quite the opposite.  It is my responsibility to state that position, and to explain my reasoning.