Robert Huebscher recently interviewed Jeremy Siegel at advisorperspectives.com and as we wonder about the state of the market moving into 2014 it might be worthwhile to pay some attention. Prof. Siegel is the Russell E. Palmer Professor of Finance at the Wharton School of the University of Pennsylvania and a senior investment strategy advisor to Wisdom Tree Funds. His book, Stocks for the Long Run, now in its fifth edition, is widely recognized as one of the best books on investing. A year ago today (Dec 10, 2012) he went on record as saying investors should expect a 25% return from the stock market in 2013, so his recent record in the prognostication business is reasonably solid. To make things interesting, he's also a close friend of Robert Shiller ("best friend" he tells us in the interview, and an invited guest at Shiller's Nobel award ceremony) with whom he tends to disagree on the state of the market.
So, with that impressive academic bio, what does Prof Siegel have to say? Is the market as overvalued as the disciples of the Shiller CAPE ratio would have us believe? Not at all, according to Siegel; just the opposite in fact. He considers it to be undervalued by 10% to 15%.
Now anyone even peripherally acquainted with academic discourse knows that anytime you hear two senior professors discuss a topic you're bound to have at least three opinions on that topic emerge. Which is fine in the shadows of the towers of academe, but "Hey guys, this is my money we're talking about. Let's have a little clarity, can we?"
Let's see if we can sort a bit of this out. I'm not presumptuous enough to be telling you who's right, but I am going to try to pick out where the issues are. (And, if you need a clue to my own biases, you might reflect for a few seconds on that leather-clad fellow in my photo in the upper left.)
Before I start let me give you a few teaser pull quotes from Prof. Siegel's interview:
- "The biggest myth in the market today is that this bull run is because of quantitative easing. I'm not going to deny that an easy Fed is helpful, but this market is really being driven by fundamentals."
- "If they looked at P/Es, earnings and interest rates, they would realize that the fundamentals support this bull run."
- "The CAPE methodology is brilliant and works. The data are the problem."
- "Interest rates are going to be higher, that's not going to depress the stock market because earnings are going to grow faster."
What does CAPE tell us and is it correct?
I'll start with the CAPE ratio argument. Observers have been drawing attention to it for some time to support the argument of an overvalued domestic market. The concept of CAPE - Cyclically Adjusted Price Earnings - has been around for a long time. It was first raised by no less an icon than Benjamin Graham who, with his co-author David Dodd, was likely the first to consider the value in smoothing company earnings over time to correct for volatility in earnings calculations. Graham and Dodd's classic Security Analysis was first published in 1934. Decades later, 2014 Nobel Laureate Robert Shiller revived the concept in his 2000 book Irrational Exuberance. The metric goes by many names, CAPE Ratio and Shiller PE Ratio among them. It is the ratio of current stock price to the 10-year moving average of earnings. Here's a couple of charts (from http://www.multpl.com/) comparing the S&P 500 PE Ratio and its Shiller PE ratio as of mid-morning 10 December 2013 aligned with one of Doug Short's superbly annotated charts showing CAPE and S&P's prices on a log scale (http://www.advisorperspectives.com/dshort/updates/PE-Ratios-and-Market-Valuation.php):
Let's notice a few things here.
First, Shiller's data series goes back to 1871. More data is better data, right? Yet, the business environment has changed many times since 1871 which is something that has always bothered me about the concept. It was interesting to read Siegel's take on this adding, as he does, some specifics on those changes. We'll get to that shortly.
Second, I've added blue lines for some of the peaks for CAPE where that peak has been a strong correlate to an impending market crash, plus a horizontal line showing the history of CAPE at today's value of ~25.
Third, every time CAPE ratio gets to levels in the mid 20s, the market turns down (see Short's annotated chart). Of course, there are other downturns that are not correlated with high CAPE ratios (1916). And there are times when CAPE ratio kept climbing well beyond 20s into the 30s and 40s before the market responded downward. Those two peaks, however, were followed by two of the hardest landings in market history.
Finally, present CAPE ratio is 25% above the 140-year exponential regression line for CAPE ratio and 55% above the 140-year median of 15.9, which would seem to indicate a considerably overvalued market.
The CAPE argument is an important leg supporting the "market is unsustainably overvalued" point of view, so anyone who takes a contrarian view to the conclusion really has to address this supporting leg.
So let's turn to Prof. Siegel's take on this. In the interview, he refers to a paper he presented in the fall (available here) where he argues that although the CAPE methodology is "brilliant and works" there is a problem in the data. How so? Well, his view flies in the face of the "more data is better data" mantra and involves the unquestioned use of the 140-year data set.
A problem is that the determinations of S&P 500 earnings over the last 15 to 20 years, particularly in recessions, were much different from what they were in prior years. Siegel contrasts recent data and historical data, comparing earnings to the business cycle. The results are sufficiently inconsistent to indicate to him that one is not dealing with the same series across the full spectrum. This has a lot to do with how changed accounting practices and forced write-downs of assets affect reported earnings. These accounting changes date to 1993 and 2001. Capsule version: Firms must write down an asset that loses value whether sold or not, but cannot write up values unless they are sold. Full details are available here in slides 14-15 from Prof Siegel's presentation to Q-Group. Have a look and notice the impact of the rules change on Time Warner's reporting of earnings following purchase of AOL for an example.
Can we sort out what happens if we try to adjust for these accounting changes? I've run up a chart tracking cyclically adjusted price earnings ratios separating out the periods following the 1993 and 2001 accounting changes. I've used a 5 year moving average instead of Shiller's 10 year increment. Two reasons, first the intervals since the accounting changes are too short for the 10 year increment to be at all meaningful (of course, using 5 years still gives us a too-short-to-be-reliable-series in this context) and second, some economists consider a window shorter than 10 year to have greater predictive value in this context. I've plotted the data with the same exponential regression Doug Short uses in the chart shown above and added shorter-term exponential regressions for the periods where earnings were determined with different accounting standards. I include the phase-in periods for those 5 years under different standards at the ends of the intervals, so the data are not as clean as one might like at their right ends. The chart on the top shows the whole data series, with standard CAPE values, going back to 1990 plus the values based on 5 year earnings moving averages color-coded for the times of different accounting practices. The lower chart is the same, zoomed in on the period from 1990 to December 2013. The only change is the color coding for the regressions to increase contrast to make them legible in the top chart.
Although there is insufficient data to demonstrate anything conclusive, it does show that by shortening the time scale for the regressions to periods that only include similar accounting practices and moving from a 10 to 5 year moving average for earnings, current values tend to converge to the long-term Shiller PE ratio regression line. Notice as well how the 5 year interval tends to moderate the peaks. Finally, notice that this does not present a picture of a vastly overvalued market relative to recent history. While recent history has been marked by some severe corrections, they followed periods when CAPE values reached into the 40s for Shiller PE (35 for the 5 yr values) and well above 25 for a sustained interval. By those measures 25 might appear less threatening.
A second problem Siegel finds with the standard implementation of Shiller PE to the S&P 500 is its aggregation bias. This arises as a consequence of treating the 500 companies of the S&P 500 as a single entity. That huge spike in PE in 2008 was primarily caused by massive write downs affecting earnings from AIG (AIG), Citigroup (C) and BankAmerica (BAC). Consider this: AIG had a weight of less than 0.2% in the S&P500 index at the time, but its losses alone more than wiped out the total profits of the 30 most profitable firms in the index, firms whose market values comprised almost half the index (see Siegel's slide 15 for details). Yet, that 65+ PE spike gets rolled into the CAPE 10 year moving average despite it's being largely driven by those three companies.
He makes an interesting comparison with corporate profits to illustrate how the PE series has changed. Up until 15 years ago the S&P 500 was more stable than NIPA corporate profit series. That has changed because of the changed accounting guidelines and it is now less stable than profits. That, he tells us is a primary indicator of the flaw in applying CAPE ratios across the series, and amounts to "the major problem."
Is Quantitative Easing Inflating Stock Prices?
Siegel considers the view that the bull run is a consequence of quantitative easing "the biggest myth in the market today." Fed policy certainly helps, but it's his view that fundamentals are driving the market with earning up more than 10% over last year even with very slow GDP growth. He expects GDP growth to be 3.5% next year and "wouldn't be surprised if GDP pushes 4%" to drive further upside in 2014.
Corporate Profits and Consumer Spending
Margins remain high compared to long-term averages but their sustainability does not cause him any real concern. There are several components here. One is that corporate leverage is very low. Two is that 80% of corporate debt is long term, locking in low interest rates, so even if interest rates rise exposure will be small. And, three, high-margin technology sector is an increasingly larger slice of S&P.
Siegel does not see looming cutbacks in consumer spending. Why not? Primarily because consumers are finding their financial situations more comfortable than in the recent past. With the real estate recovery, many fewer homeowners are underwater. There will be no tax increase (remember in 2013 everyone took a 2% increase in Social Security taxes), so that hit to the pocketbook is off the table. And consumers are feeling more confident about the status of their 401(k) positions. All this points to stronger consumer spending in 2014.
Interest Rate Will Rise But Not Enough to Depress the Stock Market
Although interest rates will certainly rise (Siegel predicts 100 basis points above today's level), earnings are going to grow faster due to a stronger economy. If so, indications are bullish rather than bearish for stocks.
He goes on to speculate on what the Fed will do and concludes that it is likely to begin the taper in March, not December or January.
The strongest argument for an overvalued market comes from looking at CAPE ratios. Siegel makes a case that this is a flawed argument. One on hand the long-term series is not a valid series due to a major change in the way earnings are calculated, and on the other, due to aggregation bias the metric is frequently skewed. Unfortunately, he does not present any evidence to show how those biases can be corrected for evaluating the data, and there is really too short a time span from the onset of the accounting changes to evaluate how market changes will correlate with CAPE under current practices.
He also bases his case on the strength of market fundamentals and a lack of concern on his part regarding the end of qualitative easing. Other observers may disagree here.