Searching for the Bottom: Valuation

What a difference a day makes. There are varying interpretations for today's rally.

• Oversold bear market bounce -- a popular view.
• News flow on mark-to-market and changes in the uptick rule -- we doubt it.
• A reaction to fundamentals -- Doug Kass's viewpoint, and therefore deserving respect.

The Valuation Question

In normal times stocks react to asset allocation decisions based upon relative valuations of the various asset classes.  This has certainly not been a normal time.  The climate of fear has generated redemptions from hedge funds and mutual funds.  Even companies with solid earnings have moved lower, as everything got sold.

The reason is that no one knows what is "cheap."  Even Warren Buffett, known for buying only when stocks were really cheap, is under fire for being wrong, or buying too soon, depending upon one's perspective.

Market Valuation

In our investor portfolios we are focusing on stocks that have been unfairly crushed with the overall market and those that will rebound most strongly if the economy avoids a Great Depression scenario.  There are many such choices.

For our current purposes, however, let us consider the overall market -- the S&P 500.  Valuation is a matter of a deceptively simple equation -- the P/E ratio.  One takes earnings and applies the appropriate multiple, the "correct" P/E ratio.  This tells you the fair value price.

The problem lies in the various interpretations of both earnings and the multiple.

Defining Earnings

Earnings can be defined in many ways, with a myriad of decisions.  Here is a sample:

• Reported or GAAP earnings, which include all of the financial write downs and various other "one-time" charges for severance pay.
• Operating earnings, attempting to exclude the one-time charges.
• Normalized earnings, where the analyst looks at a general trend over a long time period, attempting to smooth the business cycle.
• Weighted earnings, suggested by Prof. Jeremy Siegel, who believes that the reported S&P earnings over-emphasize the smaller cap stocks.
• Peak earnings, advocated by John Hussman, where one looks to past peaks in earnings and considers the trend in these peaks.
• Forward earnings, where one looks ahead instead of backward, trying to estimate the earnings power of the upcoming year.

These methods all yield different answers -- but that is not all. Some analysts augment their method by calling into question past earnings.  The historical financial earnings are suspect, since the companies booked false profits on derivatives now known to be over-valued.  John Hussman has reduced his old peaks because he now believes the profit margins to have been excessive.  If one chooses to normalize earnings, how far in the past must one analyze?

Depending upon your answers to these methodological questions, you can get a wide variety of past earnings results.  We cannot even measure the past.

The Multiple

After deciding upon S&P earnings, one must still select an appropriate multiple.  Once again, there is an array of choices:

• The historical multiple for the S&P 500.  This varies with the time frame selected.
• The trough multiple, thought to reflect the valuation when markets reflected economic distress.
• A multiple reflecting interest rates and/or inflation.  It should be higher when interest rates are lower.

Take your pick.

The Kicker

Every analyst agrees that whatever the choice of earnings and multiple, that is not necessarily the bottom.  Markets overshoot in both directions, so the "fair value" could still be 50% too high.

The Conclusions

The most bearish pundits take reported earnings and find the S&P 500 to be trading at a P/E ratio of 60 or so.  There is no observable bottom for these pundits.  They come up with a DJIA at 3000 and the S&P at 300, with no confidence that these values are a real bottom.  No wonder investors are frightened.

If one looks at forward earnings of $60 or so, there is still the question of the correct multiple.  If one applies the "trough multiple" of 10 or so, the fair value for the S&P is 600, and it could go much lower.  No wonder investors are frightened.

If one disputes forward earnings, expecting further reductions in estimates, you could forecast earnings at $40 and apply the trough multiple.  That is an S&P value of 400, and we all know that it could overshoot.  No wonder investors are frightened.

The Doug Kass Approach

Doug Kass made a big call last week, saying that the market would bottom within three days.  Tonight on Kudlow he called it a "generational bottom."

Since Doug has been so accurate in forecasting the threats to the market, he deserves special respect when he sees a bottom.  His approach is to take the book value of the S&P 500, about $5.60, look at the normalized earnings over a long period (seventy years), and apply the historical ROE of 12%.  This implies earnings of $67.  The multiple over the same period was about 15, suggesting a fair value of 1005.  He also notes past bottom multiples of 11.5 or 12, implying an S&P value of 787.

The beauty of his approach is that he avoids all of the top-down and bottoms-up arguments.  He also takes plenty of historical data through many business cycles for his normalized approach.

Simply put, there is no one else in the valuation debate who is dealing with the issue on these terms.

Many investors who have been skeptical about stocks, waiting for an entry point, are frozen with fear that markets could move lower.  Doug suggests that it is time to take advantage of the fear.

Our Take

At "A Dash" we have advocated looking at forward earnings rather than looking backwards.  This is how people evaluate individual stocks, and it will eventually show up in the market.

The "trough multiple" concept is currently used in error.  Past trough multiples came when interest rates were in double digits.  If the interest rate is 20%, the P/E multiple must be 5 just to get fair value.  With the current low interest rates, the multiple on forward earnings should be higher.  If one takes a corporate bond rate of 8% or so, you should still have a forward earnings multiple of 12.5, and that is building in plenty of risk.

Our main conclusion is the following:

Valuation methods are not helping the average investor.

That is because the pundit methods change with the winds.  If one is bearish, it is easy to find a method to justify it.  Or vice-versa.

Valuation should provide an anchor for the investor -- a reason to buy or to sell.  If Mr. Market is offering a good deal, take it.  If the deal gets better, shift a little more from cash or bonds.  That is what Mr. Buffett has done.

When Doug Kass and Warren Buffett are aligned, it is time to pay attention.  The valuation models have been no more help in finding a bottom than the economic or political methods from our prior articles in this series.

Comments

[oldprof]

A.N. I agree with you. At least a sense of value gives some idea of when to start buying. The valuation debate has played out pretty much as I suspected, especially with backward earnings.

And no, I never got to the final planned part of the series. Events moved too swiftly, and other ideas seemed more important.

Thanks for you good comments.

Jeff

[A.N.Other]

Hello Jeff,

I don't think that any serious student of valuation claims that it can "predict the bottom". All valuation does is tell you what earnings power or asset value you are getting for your money.

By the way, did you get around to that "technical bottom" piece? My view is that even if "market feel" or technical analysis can give good odds on calling a bottom, this is useless once the market is up 10% or so and the panic has cleared. So, even a great bottom picker is basically making 5-15% in a few days/weeks, once every 4 years at a major market low. That's not a great annualized return.

[oldprof]

Mike C - I have some notes that showed everyone looking to technical analysis, since nothing else was working. The TA guys -- at least several of them -- were saying that below a certain level there was no support. It was the last element of the bottom-searching problem.

The market started the rebound before I could do the last part of the series. I am not sure if I will post it.

Thanks for noticing! I have a few other hanging agenda items.

Jeff

[Mike C]

Searching for the Bottom: Technical

Is this still coming? Or did you shelve this?

[Account Deleted]

I bought my house a year ago and the discount on the morgage was only for a year so it is about to run out. My financial advisor has contacted me about re-mortgaging and I have a meeting on monday. She has told me to get a valuation before then if possible to see if the value has gone up as this would help get a better mortgage deal. How do I get one? I need a written copy of the valuation.

[Trevor @ Financial Nut]

I'm so impressed with our thorough and comprehensive your posts are. Keep it up. That's actually not very common. :)

[DaveinHackensack]

Mike and Mark,

I think there may be a simpler explanation for multiple compression, one that's independent of interest rates: the longer a secular bear market continues, the greater the number of investors who give up on stocks. So it takes lower P/Es (and higher dividend yields) to attract the fewer remaining equity investors.

In any case, independent of the valuation of the market as a whole, there are individual stocks that appear to be unambiguously cheap today, e.g., profitable companies trading for less than their net cash.

[Mike C]

@Dave, Mark

Yes. In fact, an ongoing debate/discussion (in a variety of places including this blog) is "how far back" does one have to go to make historical comparisons and is going "too far back" irrelevant? I think it can be argued persuasively that in fact the most appropriate comparison is in fact the 1930s/1940s which is the last time the U.S. really experienced this sort of deflationary/deleveraging contraction as a result of a massive bubble in credit/debt. Therefore, maybe we need to look way back for appropriate stock valuations rather then just the past 20 years.

In the spirit of sharing, a few valuation notes I came across the last 2 days:

http://marathonreturns.web.officelive.com/Documents/The%20Long%20Run,%20Andrew%20Smithers%20Interview,%20January%2030,%202009.pdf

http://www.investmentpostcards.com/wp-content/uploads/2009/03/reinvesting-when-terrified.pdf

Smithers is the Tobin Q guy. I am optimistic about future equity returns from today's level given that the guys who correctly identified the market as overvalued in 2006/2007 now believe it is undervalued.

[Mark A. Sadowski]

I've been confounded by the fact that most analysts don't consider the fact that the relationship between interest rates and P/E ratios was rather poor before the mid-1960's. The relationship seems to break down during periods of low inflation/deflation. In the Great Depression for instance the 10-year rates dropped to about 3.5% in mid-1932 and yet ten year average inflation adjusted P/E ratios fell below 6. That's why I've felt that the S&P 500 will drop below 400 before the end of this bear market. Thanks for the link to Katsenelson's PDF slides.

[DaveinHackensack]

"The "trough multiple" concept is currently used in error. Past trough multiples came when interest rates were in double digits. If the interest rate is 20%, the P/E multiple must be 5 just to get fair value. With the current low interest rates, the multiple on forward earnings should be higher. If one takes a corporate bond rate of 8% or so, you should still have a forward earnings multiple of 12.5, and that is building in plenty of risk."

P/E multiples don't show this correlation with interest rates if you go back more than a few decades, and thus aren't likely the cause of multiple compression, as Vitaliy Katsenelson demonstrated in his research on secular market trends. See, for example, his chart which accompanies this recent post of mine, "Stocks for the (Very) Long Run". In 1950, at the end of the secular bear market (or secular "range-bound" market, as Katsenelson calls it) that followed the Great Crash, the S&P traded at a trailing P/E of about 7x, and this was during a period of historically low interest rates.

[oldprof]

Mike C - I'm not sure how to improve on the conclusion I wrote. Valuation has been no more help in finding a bottom, this time, than a number of other methods.

I did not link to Barry's article on this occasion. I agree with his basic argument, that earnings are an unknown. I often referred to the Fed model result as a long-term sentiment indicator. If you actually look at data, you will see that bottoms up analysts were better than pundits or top down analysts for years.

I think we should be careful of drawing too many conclusions about methodology from what happened last year. A lot depended on the Lehman decision, as you can see from looking at any data series on stocks or the economy.

I want to emphasize that the point of this article is part of a series explaining why investors do not have their normal reasons to buy, and the bottom is therefore difficult to find.

As you may have noted, I have not written much about valuation since it has been pretty much irrelevant in the face of a lot of other factors. That will be the occasion to discuss the methods of your favorite analysts. There will be a day where the advantage of looking forward will be very clear.

Thanks for a (typically) thoughtful comment.

Jeff

[Mike C]

A few questions:

1. What is the purpose of valuation? To call/identify a specific bottom price? Or to identify an entry point where the long-term investor (7-10 years) is likely to achieve average to above-average returns (unlike the previous 10+ years)?

2. Regarding forward earnings and interest rates, what does the past 12-24 months tell us about the practical utility of these metrics?

Weren't stocks (as measured by the S&P 500) "cheap" based on forward earnings and interest rate comparisons at the Oct 2007 peak before losing 55% of their value?

Is it time to bury this model?
http://www.ritholtz.com/blog/2009/03/review-the-fed-models-flawed-valuation/

3. "Simply put, there is no one else in the valuation debate who is dealing with the issue on these terms."

Kass's method seems very reasonable and sensible to me, but I think the above statement is questionable. Essentially, he is doing the same thing as Hussman or Grantham which is to make some attempt to normalize for long-term trend earnings based on the business cycle so that you do not make the mistake of overweighting either the peak of the business cycle as many did in 2007 or make the mistake of overweighting the trough of the business cycle as many gloom and doomsters are doing now.

"Our main conclusion is the following:

Valuation methods are not helping the average investor."

I had been looking forward to this particular note on valuation, but I'm left scratching my head wondering what you are ultimately saying here. That valuation does NOT matter, or that any model is as good as any other model.

[asfdad]

No need to decide which of the three it is ... yet. Just need to respect the price action and the truth will out itself in a few. I think this thing has legs for at least a week or more.