US stocks were overpriced in January. Late April feels like a bubble.

The stock market tracks GDP in the long run. To show this, I have plotted real GDP and the S&P500 index(both from since 1930 below.


Two things are apparent from the chart above:

  1. The two lines run roughly parallel on a log-scale over this 90-year period. Their ratio is therefore a mean-reverting quantity. If you think about it, it would be strange for the lines to diverge permanently.
  2. The stock market line is a lot more volatile than the GDP line. This means that stock price can fall far below what GDP would suggest (generally in a crisis) and rise far above towards the end of a bull market. The higher volatility of stocks is well-studied, e.g. under the heading of  excess volatility.

The ratio of stock market value to GDP serves as a useful valuation metric. The market is cheap when this ratio is low, and expensive when it is high. According to Warren Buffett, the percentage of total market capitalization relative to GNP is “probably the best single measure of where valuations stand at any given moment.” You can track its value on this link from GuruFocus. I have reproduced today’s chart below:Ratio_TotalMktCap_GDP

The above ratio falls as stocks crash. As you can see from the chart above, it stood near 70% at the stock bottom following the dot-com crash, and just over 50% at the bottom following the 2008-09 crash. The ratio was under 75% through much of the 1970s, 80s, and 90s.

By contrast, the ratio “fell” from over 150% to under 120% in March and stands at 133% as I write this. Such high valuation is surprising because we are in a lockdown caused by a pandemic that will bankrupt a lot of businesses unless they are bailed out. The stock market, after falling precipitously in March, made a rapid recovery in April. It is widely speculated that the recovery is due to purchases of securities by the Federal Reserve. Quantitative Easing (or QE, as such purchases are called) is quite possibly a contributing factor. Tom McClellan has a compelling chart (reproduced below) showing that periods after the start of QE in large doses saw stocks rise rapidly. The chart also suggests that the slowdown or ending of QE coincides with weak stock market performance.


The following article describes this behavior in more detail. As for whether we are undergoing QE “in large doses”, see the following chart. Notice the spike on the right? That’s the latest round of QE.


If the Fed had not resorted to QE, the S&P500, which trades near 2860 today would probably trade closer to 2000. This is a rough guess based on earnings-per-share (EPS) projections and price-to-earnings (P/E) ratios in crises. Ed Yardeni expects 2021 earnings for the S&P500 to be 150, and applying a multiple of 13-14 gives us an index level of around 2000. An alternative explanation for the stock market’s April rise is that the market is pricing in a quick recovery and return to pre-crisis levels of profitability. This hypothesis seems less plausible because projections of GDP growth for 2020 and 2021 suggest that GDP growth will not be back to 2019 levels until 2022. I am using projections from the IMF, which can be found here. The path to recovery is uncertain with potential negative outcomes such as multiple waves of infections (which are likely, see this paper), and positive surprises such as cures and vaccines.

Another observation is narrow leadership in stocks. Even as companies at risk of bankruptcy continue to trade well below pre-crisis lows, the five largest stocks in the S&P500 have held their ground. The combined weight of the top five stocks has not been this high since 1980s. Urban Carmel posted a chart from Bianco Research showing how such concentration tends to coincide with market peaks. Caveat: the chart does not show “peak” concentration today. It shows rising concentration in the top five stocks, but the peak may be ahead of us.

Some of these observations, such as the scale of QE and the narrow market leadership, suggest a bubble. However, a bubble does not mean imminent collapse. The Federal Reserve, the Treasury, and all branches of US government have been taking unprecedented steps to prevent economic collapse. I wouldn’t be surprised if the Fed resorts to purchasing stocks, or to widespread debt forgiveness. A lot of critics will complain that such actions introduce moral hazard (encouraging bad behavior in layman’s terms), but 1. it is hard to stop the government, and 2. the government could argue that a business that saw its revenues collapse leading to crushing losses in a pandemic did not necessarily engage in bad behavior. The public stock of many businesses (such as retailers) is either worthless or worth something meaningful depending on how government chooses to act.

The focus of this article had been on the US economy and stock market. The US may succeed in propping up its economy/market, but some other countries may not be so fortunate. I suspect that we will see more businesses collapse worldwide as with each additional week of lockdown or decreased business, cash runs out. In times like this, having a high level of liquidity can be a boon for investors.

One final word. I am grateful for what I have today. Being able to work from home is a privilege. COVID-19 and the policy response to it is inflicting excruciating pain on millions worldwide. The critically under-resourced have very little say in policies that dictate their survival. I made a small donation here yesterday. Give what you can wherever your heart desires.

Warren Buffett vs. the bond king

I came across two news items today. In one, Warren Buffett said that US stocks are not expensive. In the other, Jeff Gundlach trashed the S&P500 (deprecating passive investing in  general) and suggested buying emerging markets instead. On the surface the opinions seem to be at odds, but the two gurus are providing pieces of a jigsaw that makes sense if we find and put together the missing pieces.

Buffett was responding to a question on the overvaluation of US stocks as measured by the “Market Cap to GDP ratio” and the “Cyclically Adjusted PE ratio”. His answer was that ratios are informative without being absolute. Further to Buffett’s point, early signs of a recession are conspicuously absent. Buffett no doubt sees this through the reports Berkshire Hathaway receives. Granted this could change quickly, but for now the economy is doing well. Buffett prefers stocks because their expected return is much higher than that of treasuries. Two points struck me as essential to complete the picture, yet not clearly stated:

  1. Buffett is a stock picker. I’m not surprised he is finding good value in US stocks, because there are enough companies whose stocks are cheap in relation to their future cash generation potential. However, the US stock market also has plenty of companies that will do far worse (many will go bankrupt) than treasuries. Investing in a stock market index may not be a good idea because one ends up buying the good and the bad.
  2. Which is what Gundlach (google “bond king”) was saying at the SOHN conference: Investing in a stock market index (such as the S&P500) is a bad idea. His point is that stock pickers (like Warren Buffett) will do much better than a market index. But he is also implicitly saying something else: that we are not constrained to invest inside the US alone. It’s not “do we invest in US stocks or US treasuries?”. There’s an entire universe of securities that we have access to. Gundlach sees good returns – even without stock picking – in emerging markets.

It would be fair to say that those with the talent for picking stocks globally will do quite well.

A 40% drop?

Barron’s magazine published an article in this week’s issue : A Skeptic’s View of the Trump Rally – Dreyfus Global Real Return manager sees stocks falling 40% or more once long-term trends reassert themselvesThe article expresses the view of fund manager Suzanne Hutchins. I agree with her, although I don’t necessarily see such a drop in stock prices as imminent.

The number 40% reminds me of a simple calculation I performed in my mind recently. The Shiller PE ratio (inverse of the earning yield) of the US stock market reverts to its mean in the “long-term”. The long-term mean for the Shiller PE ratio is 16.7. Right now, the Shiller PE is at ~28. A return to historical mean would imply a drop of (28-16.7)/28 = 40.4%. That’s pretty close to the 40% number in the article.


There are many other considerations though.

First, in the language of statistics, the distribution of the Shiller PE ratio isn’t stationary. It demonstrates trends lasting decades. Its mean is strongly influenced by interest rates among other factors. In a low interest rate regime, considerations of opportunity cost would induce the Shiller PE to remain higher (stocks would maintain a lower earning yield) than its long-term mean. This has happened in recent decades thanks to an extended low interest rate environment. So, one could question the premise that the Shiller PE would return to its long-term mean. It isn’t hard to see from the chart above that recent decades have supported average Shiller PE values above 20. A drop from 28 to 20 is a big drop, but closer to 30% instead of 40%.

Second, when the market crashes, it doesn’t crash to the mean of historic patterns, it usually overshoots, i.e.,  crashes further down, so a large crash can be expected.

Third, major market crashes begin in anticipation of recessions, and the market hits bottom often when the recession in underway. In a recession, earnings are lower than in boom time e.g. today. Doing the math, if the PE ratio were to go down to its historic mean of 16.7 during a recession when earnings were, say 25% lower than they are now, then we would see a market drop of 55% from present levels.

Finally, and worth repeating: while I see such a drop as probable, I don’t necessarily see it as imminent. In the last two major market crashes 1. post the dot-com bubble, and 2. post the housing bubble, the market crashes were predicted and anticipated by many, but the decline did not start until long after people started predicting it. Signs of economic stress were felt for years even though the market did not crash. The St. Louis Fed publishes a Financial Stress Index, which is a composite of several indicators of economic stress (below).


Looking at the chart above tells us that stress levels were above zero since 1994 even though the market, driven by euphoria, went to bizarre highs all the way to the end of the 1990s. And post the housing bubble, although stress started to spike up in late 2007, the big crash only happened in late 2008. Right now, the stress index reads negative (i.e. low stress), and not just barely negative but its value is easily in the bottom quartile of values since 1994. The economy is doing fine. That’s one reason why I don’t see a crash as imminent.

Of course, there’s always the risk of unforeseen events (e.g. natural or manmade catastrophes) or harebrained policy actions triggering market collapses prematurely.

Attractive global returns for the patient investor

We talked about earnings yield in Where should I invest?. When it comes to stocks and stock markets, the Shiller PE ratio (also called the CAPE ratio) is a good way to measure earnings yield:

Estimate of earnings yield = (100 / Shiller PE ratio) %

Higher the CAPE ratio, lower the earnings yield. For a long-term investor, higher earnings yield is obviously more attractive. In a number of works published in the 80s and 90s, e.g., Stock prices, earnings, and expected dividends, researchers illustrated that high stock market earnings yield is generally followed by good market performance and vice versa.

One point worth noting is that Shiller PE is not very good for predicting short-term returns or for timing market crashes. The predictive ability improves when the forecasting period is longer, e.g. 5 or 10 years. This has been demonstrated, e.g., in the paper Dividend yields and expected stock returnsA high Shiller PE (low earnings yield) generally means that a country’s stock market is overpriced, and investing in that country will probably lead to disappointing results eventually. However, it does not guarantee that the country’s stock market will do badly next month or next year. A richly valued stock market can continue to become more richly valued for several years, eventually turning into a stock bubble before it crashes. This happened in the late 90’s when the US stock market became heavily overpriced, a phenomenon attributed to irrational exuberance.

Investing in a basket of countries with low Shiller PE ratios generally leads to higher returns, a result demonstrated convincingly in a number of publications, e.g., the paper Global Value. Conversely, countries with high Shiller PE ratios exhibit poor returns over the subsequent 10-year period.

Shiller PEs and other country stock market valuation metrics are posted quarterly on Star Capital. They recently posted an update showing that performance of stocks in different countries corresponded well with their Shiller PE based predictions in 2016. Russia, the country with the lowest Shiller PE posted returns of 37% (in ruble) and 58% (in US dollar) in 2016. Russia doesn’t sound like a safe place to invest. And that’s generally true of countries with low Shiller PE ratios. CAPE ratios in the neighborhood of 5 signal extreme distress or significant risk factors. Risk factors include the risk that foreign investors may lose all their money or that a country’s currency may go to zero. However, such terrible outcomes are rare enough that investing in a diversified basket of cheap countries produces high returns with high probability in the subsequent 5-year period. While it is no doubt daunting to invest in such countries, I remind myself that historical studies take into account periods of far greater uncertainty than today, including World Wars where most countries were direct or indirect participants.

As of the day of this posting, the US stock market is quite expensive with a Shiller PE of around 28. This is just under double the historic mean and median of around 16. However, this doesn’t mean that a market crash is imminent. I’ve provided one reason above – that overvaluation is a poor market timing instrument. There’s also another reason having to do with opportunity cost, an idea labeled as the equity risk premium. The Shiller PE corresponds to a 10-year average earnings yield for stocks, and I find it helpful to compare this with the 10-year treasury rate. Since stocks carry much more risk than treasuries, it is reasonable to expect that the earnings yield of stocks will be higher than the treasury yield to compensate for the higher risk, the difference of the two being the equity risk premium. In 1929 before the great depression, and in the late 1990’s before the dot-com bust, the 10-year earnings yield of stocks  went lower than the 10-year treasury’s yield, signaling extreme overvaluation. The US stock market has an earnings yield of 100/28 or about 3.6% now. This is clearly still higher than the yield of the 10-year treasury, currently at 2.4%. The difference gives us a roughly calculated equity risk premium of 1.2%, low by historical standards but still positive. In other words, high stock valuation is being supported by low interest rates. If the Fed were to raise rates rapidly, stocks would start to look less attractive, but we’re not there yet.

A note for those who follow the SciPre portfolio: Many countries in the SciPre portfolio have low Shiller PE ratios. Their undervaluation is an important factor contributing towards their inclusion in the portfolio.