Big Names, Big Market Calls

In the last two days there have been two rather spectacular predictions about the market.   While I am working on my own 2011 preview, I interrupt my regularly scheduled programming to consider the forecasts from these respected sources.

Two Extreme Viewpoints

There was breaking news yesterday about Yale Professor Robert Shiller's most recent market forecast.  He revealed his projected S&P 500 forecast for 2020:  1430   Yes, you heard that right.  Ten years from now -- virtually no gain from current levels -- an annualized growth rate of 1.3%.  This is an amazing call in an era when markets often exceed that percentage change in a single day, and when so-called "bullish strategists" are calling for a 10% market gain on the year.  Essentially, it is a forecast for a flat market.

If you believe Shiller, you should not invest in stocks -- or real estate, since he also does not like that sector. 

In sharp contrast, there was also breaking news today from Laszlo Birinyi, who also made a big forecast.  He sees the S&P 500 at 2854 by 2013.  This is much greater than the most bullish mainstream estimates.

What a difference!  At the time I am writing this, there is quite a difference in market coverage.  Seeking Alpha (where I am one of the original contributors) is typical.  There is no story on Birinyi, while Shiller is still a front-page feature, with supporting opinions wondering whether he might be too optimistic.  Wow!  There is also no featured article even suggesting that Shiller might be a touch too bearish.  This is such a dramatic difference between the two that it deservers a more thoughtful evaulation.

My Take

I do not agree with either forecast, and I have good reasons.

Shiller.  I start with a bias.  I like professors, I respect Yale, and I am impressed by Professor Shiller -- both his work and his personal appearances.  I am therefore surprised that I cannot find more useful advice from his analysis.  Here are three key points:

1. No one has ever made any money in real time by following Shiller, including Shiller.  He got out of the market in 1997.  I invite comment from anyone who has a track record of successful real-time trading based upon the Shiller method.
2. Shiller's approach does not successfully predict next year's earnings, the most interesting data for nearly every stock analyst and market forecaster.  I have an open challenge to anyone to show that Shiller's backward looking method is better than the one-year forward estimates by analysts at predicting next year's earnings.  This is a simple factual challenge with no takers so far.
3. Shiller's approach almost never signals a buy.  Maybe if interest rates get to 20% again we will see the P/E ratio's in the single digits that his backtesting shows to be a good time to have made a (theoretical) stock purchase.

To summarize:  The Shiller method provides no real guidance for investors.

Birinyi.  The Birinyi approach looks at history and measures various cycles and time frames.  As I have frequently noted,  there are not enough cycles to make such determinations, and circumstances have changed too much for such comparisons.  Instead, let us examine the Birinyi prediction in a different way.  After adjusting the assumptions a bit, let us consider the following:

• We will look at forward estimates of operating earnings (OK with me, but viewed as too bullish by many).
• We will start with S&P earnings of $100 for 2011 (very optimistic according to most).
• Let us apply earnings growth of 15% per year for three years, bringing us to earnings of $152 in 2014.  (This is also an estimate well beyond most expectations).
• Let us suppose that the P/E multiple expands to 15.  Oops, not enough.  16?  Try again.  It takes a multiple of 19 to reach the Birinyi target.  There have been multiples of 19 in the past, especially with very low interest rates, but it is not something that most would expect.

To summarize, the Birinyi target and time frame seem far too optimistic.  To be fair, I wonder why this approach is given less credibility than Prechter (Dow 1000?) or Shiller.  Readers interested in an actual analysis of past predictions should check out the Guru grades at CXO Advisory.  (Prechter 24%, Birinyi 51%).

The wise investor is aware of two things:

1. The long-term market trend in earnings and in price, and
2. His own needs and risk tolerance.

 

 

Comments

[oldprof]

qwsp -- You are absolutely right. These are not contrasting predictions for the same time frame. On a theoretical basis, both could be correct.

At least they provide a time frame, unlike many of those making dramatic calls.

I would still take "over" on Shiller and "under" on Birinyi in this little exercise on evaluating "big" calls. How about you?

Jeff

[qwsp]

Actually, you cannot compare these two predictions because they are 2 different time frames. Certainly Shiller understands that irrational exuberance can take the market higher over the next 2 years. Maybe Birinyi also suspects the index could trend down to Shiller’s number by 2020. Also note that Birinyi’s target for 2011 of 1333 is actually lower than the consensus of 1371. His real outlier is that 52% of the gains begins in July of 2012! You gotta give him props for this kind of precision! One final comment, Birinyi’s numbers are obviously actual print, but Shiller is probably referring to a real return of 1430 minus inflation but it’s not made clear in that sound bite.

[oldprof]

j'adoube -- Thanks for pointing this out. I hate mistakes like this. I apologize to Mr. Birinyi and to readers for the error.

Other bloggers work during the day, but I can only write at night. I wish that I did not make any mistakes, and found the ones that I did make. I have learned that many errors are corrected the next morning.

I feel especially bad about this one, since I have been a Ticker Sense participant for many years.

The first time I wrote about Maria I also spelled her name wrong. It was before she became famous, and she sent me an email correcting me!

Anyway, mea culpa -- spelling fixed.

Good job, J'adoube (and good name!)

Jeff

[j'adoube]

Dude, it's

Birinyi

not

Biryini.

Geesh.

[Rushabh]

Interesting piece. I wonder if you have read any of John Hussman's work? I see you state:

"I have an open challenge to anyone to show that Shiller's backward looking method is better than the one-year forward estimates by analysts at predicting next year's earnings. This is a simple factual challenge with no takers so far."

I am very sure if you read some of the stuff on Hussman's website the two of you would have a very intense conversation on historical market analysis vs. forward estimates and multiples.

Here is his website: http://www.hussmanfunds.com/

Great read either way. -Rushabh

[rich]

Hi Jeff -- I agree completely with your point that the Shiller PE isn't sufficient as a binary buy/sell signal used entirely in isolation. I definitely did not use it that way, as you point out.

I guess I interpreted your initial post as saying that the Shiller PE was useless. This is the idea I objected too, because I think it's very useful to have an idea of secular "expensivess" as an indicator of potential long term upside and downside (and thus, whether to look outside the S&P 500 for better risk-adjusted returns).

But you are absolutely right that by itself it is not useful as a buy/sell indicator. Sorry if I misunderstood the meaning of your post.

Rich

[oldprof]

rich -- I congratulate you on your successful decisions, but I am writing with investors in mind. You have used Shiller as one input, not really waiting for the official 'buy' signal.

I certainly agree that it is better to buy when the P/E ratio is low. I just prefer to look forward rather than backward, and I am comfortable with a one-year time frame. I also think that ignoring interest rates weakens the Shiller approach.

Please take a look at this comment from George Vrba, who writes that the Shiller approach is theoretically correct, but not practical. He has a very nice chart to illustrate his point: You might only get a buy signal every twenty years.

http://advisorperspectives.com/newsletters11/Letters_to_the_Editor-01042011.php

Thanks for your thoughtful comments.

Jeff

[rich]

Also, Jeff, as to your critique of Shiller's prediction. You are basically saying that it's hard to trade based on Shiller's valuation metric, and therefore that his 10 year forecast is invalid. But that doesn't follow. It is well documented that the Shiller PE is very predictive of long term (ie 10 year) returns. For near term returns, it has no predictive value (hence it being tough to "trade" on) -- but that does not invalidate a long term forecast based on the Shiller PE.

I would agree with your summary if you said that Shiller's work provides no real guidance for traders. For investors, who by definition should be most concerned with long term returns, I completely disagree.

[rich]

Per your invitation I will comment that I have had very good results using the Shiller PE as a key input into my investing approach. Not as a sole focus, but certainly as an input.

My serious investing timeline starts around 2004, so take that for what it's worth. But I kept very low US stock exposure in the mid-2000s, added in late 08/early 09 when the Shiller PE got back near/under fair value, and have scaled back on the way up. I guess we will see how that last part goes. I didn't go to cash, I've gone to other investments I thought had better value than US stocks (hence my statement that I don't look solely at the S&P500 Shiller PE, but use it to guide my US stock exposure). My returns have been tremendously market-beating and the Shiller PE has served me well... with the caveat that I am not judging my returns based on a quarter or a year, but over a multi-year timeline.

As for this comment by burt:

Schiller's data omit a slow structural change that has skewed the big caps: the increasing importance of non-US earnings. Many multinationals get most of their income from this. I believe that this is the reason for the trend upward in Schiller-like valuation indicators since 1950 or so. It would be instructive to get a series of US-only earnings and do the analysis on that.

What does that have to do with anything? The Shiller PE is about how much you pay for earnings, not where earnings come from. (Also, people might take your criticism more seriously if you could spell the guy's name right).

[oldprof]

DE -- This is a tricky methodological issue. You cannot test the proposition by looking for a correlation between "year-over-year changes in earnings and changes in stock prices."

You need to look at changes in expected earnings versus changes in stock prices. I have done this for various time periods and there is a relationship. It gets strained at times, and there is also a correlation with economic changes.

Zacks does a lot of work on this question as it pertains to individual stocks. They have plenty of data, and it seems pretty convincing.

And we also know that virtually any analyst talking about a stock speaks in terms of forward earnings.

So to answer your "who cares?" question, I think nearly everyone does, and they probably should.

Jeff

[DE]

"2.Shiller's approach does not successfully predict next year's earnings, the most interesting data for nearly every stock analyst and market forecaster. I have an open challenge to anyone to show that Shiller's backward looking method is better than the one-year forward estimates by analysts at predicting next year's earnings. This is a simple factual challenge with no takers so far."

Who cares about next years earnings? The link between near-term earnings and stock direction is tenuous. Outside of very large changes in earnings, there is essentially no correlation between year-over-year changes in earnings and changes in stock prices.

[A.N. Other]

Maybe the most accurate forecast is simply to take the current S&P 500 earnings yield and compound it out to 2020. That gives us about 6-7% per annum, assuming no earnings growth. Add in earnings growth at around 3% per annum and you get 9-10% returns for stocks.

Anyway, that is besides the point. Forecasting is a poor technique because the future is inherently difficult to predict. The decision to invest is far, far simpler. Simply look at the main investment alternatives. We have:

i) cash - yielding almost nothing, no growth
ii) bonds - yielding little, no growth
iii) stocks - earnings yield 6-7%, growth 3% per annum.

Wow, tough call - should I invest at 9-10% returns, or 0% for cash or 4% for Treasuries or 6% for corporate bonds? This one is obvious - you should have the vast majority of your investment in the stock market. Stocks are cheap, end of story.

[burt]

I just found your blog. Good work; I will put it into my reader.

Like you, I am amused by the pack of Schiller followers. They do simple analysis on his long term spreadsheet and come out bearish.

Schiller's data omit a slow structural change that has skewed the big caps: the increasing importance of non-US earnings. Many multinationals get most of their income from this. I believe that this is the reason for the trend upward in Schiller-like valuation indicators since 1950 or so. It would be instructive to get a series of US-only earnings and do the analysis on that.