The Truth About Valuation

Untold Truths About Valuation  

In the second part of our series on various misunderstood macroeconomic relationships, we concluded that there is little correlation between economic growth and stock prices.  We think Benjamin Graham most accurately captured this phenomenon by explaining that, “In the short run the stock market is a voting machine.”  In other words, short term price movements are driven primarily by Fear (2008) and Greed (2009).  But Graham also taught us that a security’s true intrinsic value will ultimately be realized over time, “In the long run the stock market is a weighing machine.”  As a student of Graham, Warren Buffett gets it. “Price is what you pay.  Value is what you get.”  The price you pay is the single most important determinant of expected returns.  

Not unlike the unlimited combinations of flavors for Frappuccinos at Starbucks, there is no shortage of definitions for “value” today.  The most common factor is the P/E Ratio. At the most basic level, P/E is simply the price (P) of each share divided by the company’s earnings (E) per share for an individual company. Likewise, the P/E on the overall market is an aggregate of the prices and earnings of all the companies trading on the market.  The problem for investors is that because there are many ways to measure value, one can see what one wants to see.  Ned Davis recently provided us with an interesting recap of how this definition has changed over time.  

When the long-standing GAAP P/E ratios got overvalued, most on Wall Street shifted from GAAP earnings to so-called operating earnings, which added back into GAAP earnings a lot of the bad stuff that had been written off in GAAP earnings. When these operating P/E ratios got too high in the late 1990s, Wall Street shifted its focus to so-called “forward operating earnings” or predicted earnings. Sometime around 2000, they had to predict growth stock earnings three years into the future to argue that growth stocks were still reasonable. By 2007, nearly all of Wall Street had shifted to relative, forward predicted operating earnings, or the so-called Fed Model. This showed stocks deeply undervalued at the 2007 peak. Interestingly, nearly all the “Fed Models” shown only went back to around 1980. The reason is if one took them back further, they fell apart.  

Ned has a tremendous ability for calling it as he sees it.  Perhaps it’s simply the lack of investment banking biases at NDR.  Perhaps it’s something in the Florida orange juice, rather than the NY water.  Whatever the reason, we appreciate you Ned.  And we aspire to do the same.  It is in this regard that we’ll outline a few additional truths about valuation.  Specifically, we’ll discuss the relationship with inflation, the Fed Model, and forward “operating” earnings.  


Suffice it to say that the quality of “sell side” research has deteriorated over time, as predictably bullish biases can be proven false by anyone with a basic understanding of history and a hand held calculator.  One of our favorite Wall Street fairy tales is the popular delusion that high stock market values can be justified by low levels of inflation.  Let’s assume for a brief moment, that the intent of these strategists is not to mislead their clients in an effort to generate greater fees for their respective firms (we recognize we’re out on a limb here, but stick with us).  Making this leap of faith would then lead us to the conclusion that they are simply too lazy to do the work required.  It is true that history demonstrates a clear inverse relationship between trailing inflation and price-to-earnings multiples (i.e. lower inflation has been accompanied by higher valuations and vice-versa).  This is illustrated in the chart below.


Unfortunately, most strategists stop here when they should be focused on the implications for expected returns.  In other words, do low rates of inflation justify high valuations and lead to attractive long-term returns despite elevated price-to-earnings multiples?  The answer is no – they do not.  Rationalizing high valuations using current levels of inflation and/or interest rates actually destroys the predictive ability of normalized valuations.  

We will concede that market history validates Wall Street’s conclusion that low trailing inflation rates have been associated with high price-to-earnings multiples, and vice versa.  But, while low inflation may help explain why stocks are overpriced, it does not alter the long term consequences associated with that overpricing.  Simply put, high valuations produce low long-term returns, while low valuations generally produce attractive long-term returns.  There is no need to overcomplicate this.


The Fed Model  

The Fed Model has been embraced by Wall Street cheerleaders as a simple and “reliable” method for estimating stocks intrinsic value.  Note that simple is the key word in that last sentence, as it conveniently takes less time to calculate which provides strategists with much more time to shake their pom-poms.  In any event, the so called Fed Model is a straightforward comparison of earnings yields (the inverse of price-to-earnings multiples) and treasury yields.  When earnings yields are higher than treasury yields, so we are told, stocks are attractive and vice versa.  Sounds intuitive and is certainly easy to grasp.  Perhaps Wall Street had even hoped to inspire additional confidence by giving this particular measure of valuation such a strong endorsement.  Granted, they could have done worse (better luck next time Nike), but does this picture really promote confidence?  


According to work done by John Hussman, there is almost no relationship between the two data sets (earnings yields and Treasury yields) with the notable exception of the period since around 1980.  What an amazing coincidence that this is the only period that “the street” has chosen to examine.  And how convenient that the model clearly indicates stocks as being deeply undervalued for most of the decade since the market “bottomed” in 2002-2003. Unfortunately, (for investors) it turns out that the model persistently indicates that stocks are a bargain.  Although not depicted in the chart below, the Fed Model would have also implied that stocks were about as “undervalued” as they are today, just before the 1929 crash!!


Forward “Operating” Earnings  

In recent years, price-to-earnings multiples based on forward “operating” earnings have become the new standard on Wall Street.  As we have learned, most “improvements” on Wall Street are more accurately portrayed as “distortions.”  This one is no different.  By way of background, forward “operating” earnings are based upon consensus analyst estimates for next year’s earnings, excluding “extraordinary” and “non-recurring” charges (that often recur quite predictably) and a variety of additional information that “the street” chooses to ignore.  Operating earnings are not defined under Generally Accepted Accounting Principles (GAAP) and are inevitably higher than actual earnings.  This comes in quite handy for bullish “sell side” strategists as higher earnings translate into lower valuations.  

The sell side has conditioned investors to believe that “normal” price-to-earnings ratios have historically averaged about 15 and regularly use this as a basis for comparison with valuations based on forward “operating” earnings.  What these strategists shamelessly ignore, however, is that the average they regularly quote is based on trailing earnings! Astoundingly, by using valuations based on elevated earnings, “the street” is able to once again demonstrate how “cheap” stocks are based upon “below average” price-to-earnings multiples.  Unfortunately, because data on forward “operating” earnings estimates has only been compiled since the early 80s, there is no long term historical data to even determine the “normal” level of forward price-to-earnings ratios.  And heaven forbid analysts have to dust off their HP 12C and actually build models!  Thankfully, there are a few of us that still do, and John Hussman’s estimates indicate that the average ratio based on forward “operating” earnings would be closer to 12, rather than the often-quoted 15.  Of course, this is also a period dominated by high and rising valuations, so if we exclude the extremes reached in the tech and telecom bubble, the average falls closer to 10.  

The fun doesn’t stop there either – at least for those of us with some skin in the game.  In order for investors to rely upon the current level of earnings to estimate the intrinsic value of a long term asset, those earnings must represent a sustainable level of profits; otherwise, it is imperative to normalize profit margins which are naturally cyclical.  For anyone that is incredibly curious about today’s forward earnings expectations and the assumptions for profit margins, see William Hester’s recent piece here.  

For everyone else, we’ll continue with several metrics that are actually quite good predictors of future returns after the holidays.  Happy New Year!!