• Markets
• Economy
• Federal Reserve
• Housing
• Energy
• Digital Media

What Will the Stock Market Return over the next 10 Years?

I came across an interesting historical chart of the post '29 market, and posted it last night without comment. It was an accidental Rorschach test -- the permabulls took it as proof that I was (wrongheadly) insisting a depression was imminent, while the more bearish among you saw parallels to the present.

It was neither.

The following charts, however (courtesy of John Mauldin), do have implications. They look at the returns you can expect over the next 5-10 years, based upon valuation metrics:

>
Jeremy Grantham breaks down historic P/E ratios into quintiles,

Note that based upon the histroy of the stock market, we are now in the top 20% of historical valuations. Real returns for the indices when purchased at that price are zero.   

Mauldin writes:

"What kind of returns can you expect 10 years after these periods, on average? Interestingly, the first two quintiles, or cheapest periods, have identical returns: 11%. That means when stocks are cheap, you should get 10% over the next 10 years. The last, or most expensive period, sees a return over 10 years of 0%.

This basically squares with data from Professor Robert Shiller of Yale. To ignore it, they must show why "this time it's different." In order for someone to predict, as many do, that stocks will return 6 to 7% over the next 10 years, that person must show that values lie between the second and third quintiles--or that investors will somehow start putting more value on stocks, and thus a floor on the market.

The problem is that the average overrun of the trend in a secular bear market is 50%, which is why stocks get so undervalued. By that, I mean stock market valuations do not stop at the trend. They tend to drop much lower. For the bulls to be right, we would have to see something that has never happened before. Stocks would need to drop to values 25% higher than the long-term historical average and no further.

Which is about where we are today."

Of course, this kind of analysis asks the question as to what "typical" P/E is -- and what it should be:

"What is the average P/E? If you start from 1900, it is 14.8. Starting from 1920 it is 14.4, and starting from 1950 it is 16.8. So, average depends upon when you want to start counting. Using only the last 56 years, we are close to the average P/E. The Grantham chart suggests we should see a return of 8% a year for the next 10 years.

By the way, if you exclude the years 1997-2002, the average P/E ratio drops by more than a full point! Prior to 1997, the average P/E was 13.8."

Now for the histrocial comparo: I have previously made the case that the present market reminds me of the 1972-73 period.

Have a look at this chart by Ed Easterling at Crestmont:

Esterling notes:

"In 1972, P/Es were almost 18, the market was approaching and exceeding new highs, volatility was low, and the market was in the first half of a secular bear market ...what happened next is now history--if it happens again, it won't surprise the old sages..."

Note the bottom thrid of this chart: P/E multiples kept dropping and dropping. That's caused multiple compression (see this chart also), and we are about half way through a similar "classic secular cycle."  That makes the present rally "another market high on the way to lower valuations in the future."

I agree with Mauldin, who concludes his 5/5/06 letter with this bon mot: 

"Call me a bear, but I continue to believe we are going to get a chance to buy this market at a much lower level than today's close."

Saturday, May 06, 2006 | 07:47 AM | Permalink | Comments (43) | TrackBack (0)
add to de.li.cious | digg this! | add to technorati | email this post

Comments

The "average" PE is one of the most misused data points in the investment game. Over the long run the stock market PE has swung from over 20 to under 10 over something like a 20 year cycle. When you use a term like average it implies that there is some central tendency to return to that average or that it occurs more frequently then any other value. But if you look at the frequency distribution of the PE from 9 to 20 you find that that at any given time that there is about the same probability that the PE will be at any value in that range as another value. That is, the pe is between 9 and 10 about as frequently as the PE is between 14 and 15 or between 17 and 18. There is no central tendency for the PE to converge on a value between 14 and 15.

Posted by: spencer | May 6, 2006 8:20:32 AM