We last wrote about stock valuation in August, when we looked at three types of P/E multiples and argued that stocks were more stretched than you would think if you only relied on the simplest measure.
Since then, we’ve had the non-taper, non-Larry Summers Fed, non-Syria ultimatum, non-keeping your plan if you like it, and non-market bubble (according to Janet Yellen’s soon-to-be authoritative judgment). You might say we’ve had an unusual amount of non-sense.
After all that’s happened – and with the S&P 500 (SPY) about 7% higher – it seems time to update our research. We’ll look at the three P/Es again as part of a new analysis that ties in credit markets and the Fed and ends with a prediction about the next bear market.
We start with a chart that marks and categorizes 11 historical bears that will play a part in our conclusions:
These bears are slightly different to other lists you may have seen because we’re using Robert Shiller’s long data history, which shows average figures for each month without the daily detail. We’re also identifying all corrections of over 20% as bears, even if the market failed to make a new all-time high at the prior peak. “Mega-bears” are corrections of more than 40%, while “über-bears” are corrections of over 60%.
Now for the valuation history, starting with traditional P/Es based on trailing 4 quarter earnings:
Here are Shiller’s P/Es, which are based on 10 year trailing averages for earnings:
And here are P/Es derived from 40 years of trailing earnings but using trend lines instead of Shiller’s averages (click here for further detail):
Trailing P/Es are inferior to the other two measures because they don’t account for earnings cycles. As we wrote in August:
[I]t’s clear that changing perceptions about earnings explain a substantial portion of the market’s volatility. Just as investors can easily forget that P/E doesn’t rise forever, they sometimes forget that earnings don’t climb forever. And when earnings are unusually high, traditional P/E multiples fail to capture the full risk of a correction.
Moreover, earnings cycles are more pronounced in recent decades, due to the Fed’s increasing interventions. (See here for more on this topic.) Interventionist policies suggest an even stronger case for following the Shiller P/Es or trend earnings P/Es – not the more traditional measure – and we’ll come back to these results in a moment.
Turning to the credit markets, the next chart combines the Fed’s “Flow of Funds” data on nonfinancial private debt with an earlier series that isn’t exactly the same but captures credit trends, nonetheless:
Three time periods stand out as distinct debt or policy regimes:
1929 to 1946 (the Great Deleveraging). Private debt fell from a pre-Great Depression high of 163% of GNP in 1928 to 83% in 1947. Moreover, there was an even greater fall from the economic trough in 1932-33, when nominal GNP dropped to about half the 1929 amount and pushed private debt above 250% on a GNP ratio basis.
1946 to 1998 (the Long Re-Leveraging). Using the Flow of Funds data this time, nonfinancial private debt climbed from only 37% of GDP in 1945 to about 130% by 1998.
1998 to today (the Big Experiment). While re-leveraging continued through the housing boom, it was further supported during the Alan Greenspan and Ben Bernanke Feds by a new policy approach, which combines limited intervention in good times (leaving bubbles alone, for example) with ample stimulus at any whiff of volatility, deleveraging or deflation. The Greenspan/Bernanke “puts” have emboldened risk takers and pushed valuation measures upward, as shown in the P/E charts. For credit and asset markets, the puts mark a new regime that stands far apart from the old-fashioned approach of “taking the punch bowl away when the party gets going.” We’ll call the new regime a “Big Experiment.”
The Big Experiment began, arguably, with the Fed’s actions after the 1987 market crash and then strengthened progressively. But we chose a 1998 start date because it coincides with Greenspan’s aggressive response to the LTCM crisis and implicit support for the Internet bubble, while also marking the early stages of the housing and mortgage booms. By the end of the 1990s, the Fed had clearly crossed the Rubicon into a new era that David Stockman aptly calls “bubble finance.”
The three regimes’ relevance may not be immediately clear, but consider what happens when we sort all bear markets since 1929 by size:
It turns out that sorting by size is the same as sorting by our three regimes. Together with the P/E histories, the table gives us a few reasons to expect the next bear to be a big one. (Note that we’re separating the size of the bear from its timing, which we’ll discuss in a later post and doesn’t depend much on valuation.)
First, one thing we’ve learned about the Big Experiment is that stocks tend to fall a long way after the Fed loses its grip. The bear markets of 2000-03 and 2007-09 were more severe than all but the Great Deleveraging bears of the 1930s and 1940s.
Second, stocks are more expensive than at any of the Long Re-Leveraging peaks, which is the only period in which bear market losses fell short of 40%. In fact, the Shiller P/E is now higher than it was in any bull market before the Internet bubble except for the 1929 peak, while the trend earnings P/E has breached even the 1929 levels.
What’s more, it’s reasonable to expect big cycles to persist in today’s policy environment, because it relies so heavily on the Fed. When so much depends on a single, binary factor (in this case, faith in the FOMC’s support), corrections tend to be particularly severe after the factor reverses (when judgment swings from full faith to loss of faith). By comparison, a less manipulated market responds to a wider variety of inputs, most of which develop gradually.
Why so doom and gloomy?
If you disagree with these conclusions, you probably believe we’re headed for another long re-leveraging, thanks to the fall in private debt since the housing boom. The Fed’s policies, you may say, are merely cushioning the path to a lower debt burden and more balanced economy. It’s only a matter of time before the post-WW2 credit boom reignites. In Ray Dalio’s parlance, this scenario would be a “beautiful deleveraging.”
While the beautiful deleveraging is worth contemplating, it seems highly unlikely. Here are three reasons for skepticism:
- Private debt hasn’t fallen all that much. Nonfinancial private debt was 156% of GDP as of the latest data point, which is exactly where we were in the second quarter of 2006 as the housing boom was running out of steam.
- The Fed’s actions have barely registered with the all-important middle class. Not only is median household income 8% below its 2007 peak after adjusting for inflation, but it was still falling as of 2012 (the latest data point). The median household is now earning the same real income as it was in 1989 – 24 years ago! In addition, the jobs picture is abysmal by any honest assessment (which should include the composition of job gains, unemployed folks who’ve dropped out of the labor force and involuntary part-time workers).
- Just about everywhere you turn these days, you’re looking at another classic sign that credit and asset markets are getting out of hand. From record margin debt to a deluge of covenant-lite loans to 1990s-like enthusiasm for Internet companies, it’s hard not to see froth (regardless of your views on bubbles). When we compare these warning signs to the slow progress on debt reduction and no progress for the middle class, the financial economy is too far ahead of the real economy for the eventual outcome to be “beautiful.”
Put differently, the giant gap between the financial and real economies tells us that any normalization of the real economy will prove fleeting.
Think of it this way:
You’re a baseball player trying to break into the majors despite mediocre fielding skills, no foot speed, and a batting average that hovers around 250. Egged on by your friend, A-Rod, you think you can make it by using steroids and turning yourself into a power hitter. But it doesn’t work out as planned. After a year, you’re losing hair, your skull’s gotten bigger, there’s fatty tissue on your chest that wasn’t there before, and you’ve still only managed 18 home runs in a season. You finally accept that it’s not going to happen for you.
In the baseball scenario, steroids didn’t show enough payoff before the side effects told you enough was enough. And you can say pretty much the same thing about our economic scenario and monetary steroids. We’re seeing dubious benefits and fast developing side effects from the Fed’s actions, causing many observers (and ex-central bankers) to recommend a rethink of the Big Experiment. Yet, the experiment continues.
Getting back to our question about the next bear market, the Fed’s unshakable commitment to its approach – despite growing evidence that it may do more harm than good – is our last reason to expect the next bust to be another killer. When the bull finally runs out of steam, it’s likely to be March 2000 or October 2007 all over again.
Bonus result for Austrians
Although we started our analysis with the Great Depression (because of limited debt data and less familiarity with earlier markets), we sized up all of the bear markets in the Shiller database, which goes back to 1871.
These include a fourth regime: the classical gold standard from 1880 to 1914. There’s also an extra period from 1914 to 1921 that was marked by both World War 1 and the fact that the newly hatched Federal Reserve hadn’t yet established its so-called stabilization policies. I’ll call this period “transitional.”
Here’s the full list of bear markets, sorted by size:
Note that every one of the pre-Fed stabilization policy bears is less severe than:
- All of the Great Deleveraging bears
- The most extreme Long Re-Leveraging bear (the 43% drop in 1973-74, which would look much worse if you were to factor in inflation)
- All of the Big Experiment bears
In other words, our stock market history doesn’t reflect very well on the Fed.
(Click here for technical notes about this article and a few more charts.)