I still read quite often how stocks are extremely expensive or badly overvalued. Such a determination can’t be done in isolation. It is necessary to place the valuation of stocks into context versus other asset prices. Stocks are certainly not expensive relative to Treasury bonds or real estate. In reviewing the evidence, we judge that stocks are fairly valued relative to inflation historically and still cheap relative to bonds.
The most common argument suggesting that stocks are overpriced comes from the Shiller CAPE (cyclically adjusted price earnings multiple). Alternatively, it is argued that price to sales, price to book, and stock market capitalization relative to GDP all suggest that stocks are quite expensive. It is also argued that stock price earnings multiples are high compared to the multiples of the past 10 or 20 years. Let review each argument.
Over the past 25 years, the Shiller CAPE has suggested stocks were cheap precisely once—in 2009—and otherwise implied that stocks are quite expensive. This is a horrendous track record. The average investment lifetime for a person is roughly 50 years. If one followed the Shiller CAPE, investors would have remained out of the market for very long stretches with only brief periods in which they were invested. And in the last 10 years, investors would have missed almost all of the longest bull market in history. Such a program would have been devastating to any hope to build wealth. The Shiller CAPE has been a woefully unsuccessful valuation tool.
Price to sales, price to book and stock market capitalization relative to GDP can be useful, but one must recognize their weaknesses. Price to book is a very valuable tool for certain industries where book value matters, including banks, insurance companies and most financial firms. We use it for those kinds of firms every day. It is useless or inappropriate for other industries, such as with technology firms. Facebook, Google, Microsoft, and many others generate revenues from business activities that have no meaningful link to the cost of building those activities. Any of those companies could have a book value of precisely zero and they would still be extremely valuable properties. If Metlife, Bank of America, or Avalon Bay Properties had a book value of zero, which would mean their assets were roughly equal to their liabilities, their book and equity value would both be close to zero. And with so much of the growth in the economy shifting into technology and highly scalable businesses, price to book is no longer a useful metric for the overall market, even if it retains its value in select industries. A similar argument applies to price to sales. When GM manufactures a car, its costs per vehicle include the value of all the inputs, including steel, aluminum, rubber and everything else, which are reflected in the selling price (plus some profit margin). When a user clicks on an ad on Google or Facebook, the incremental cost of manufacturing is zero, yet the click produces more revenue and profit, which can elevate the price-to-sales metric without limit.
One of the most frequently used arguments supporting the case for stocks being expensive uses the actual history of stock prices as the benchmark. Indeed, today’s market multiple of more than 19 is high compared to the multiple of the past 10 or 20 years. But that isn’t the best criteria. Price-earnings multiples vary inversely with inflation and interest rates. The theoretical derivation for this relationship is provided by every basic finance textbook, which is also borne out in the historical record. Price earnings multiples are low (high) when inflation and interest rates are high (low). Since inflation and interest rates are near historical lows, it is theoretically appropriate for multiples to be at historically high levels.
Oddly, current price-earnings multiples may be distorted. Most everyone uses the S&P 500 to make the needed calculations, which gives enormous weight to 6 companies: Amazon, Facebook, Apple, Google, Microsoft, and Netflix. The results change quite dramatically if we use the S&P 494. The price-earnings multiple for the 500 is just over 19.0 for 2020 expected earnings. But it is only around 15.3 for the S&P 494, which compares favorably to the average price-earnings multiple of the past 50 years of around 16.2. So a handful of companies have a huge impact on this valuation tool. Other than these six companies, the entire rest of the market is quite reasonably valued.
The distortion to valuation measures can be seen in numerous ways. Value Line publishes a table every week showing the 100 lowest price earnings multiples stocks it follows. In February 2017, the highest price earnings multiple in that table was 11.1. To make the same list now, the price-earnings multiples must be no higher than 8.4. In effect, most stocks have gotten materially cheaper over the past three years, even as a few of our largest companies have performed extremely well, notably the five mentioned above, which has pushed up the average multiple.
Comparing yields also implies stocks are attractively priced. The 10-year Treasury yield is around 1.6%, which is quite low relative to the 1.75% yield on S&P 500 ETF. Over the 10 year life of the Treasury bond, stocks will yield more, (are taxed more favorably), and dividends will grow along with the economy. Historically, the S&P has yielded less than the Treasury, except during periods of severe financial distress. So we see stocks as cheap relative to Treasury bonds. Indeed, we have made the point often that an investor buying a 10-year Treasury to earn 1.6% is locking in a negative rate of return after inflation and an even more negative return adjusting for taxes. Obviously, we recognize foreign money may feel differently given how low rates are in other parts of the world, specifically Germany and Japan, where Government bonds have a negative yield. Some high-rated European corporates can even issue bonds at negative yields, so these are unusual times. But is it unlikely that everything is overvalued simultaneously. There must be one relatively “cheap” asset. Relative to Treasury bonds, we conclude that stocks remain attractive.
The above shouldn’t be taken to suggest that there’s no place in a portfolio for corporate bonds and preferreds. They do provide a lower return than stocks, but they also entail less volatility and less risk. That’s a worthwhile tradeoff for many investors. By providing ballast to a portfolio, a carefully constructed bond allocation enables many investors to stomach the higher volatility of the stocks within a portfolio, even if Treasury bonds are expensive relative to their own history. The key is managing bonds correctly to reflect their own risk profile.