Is The Stock Market Too Expensive Right Now?

|
Includes: CRF, DDM, DIA, DOG, DXD, EEH, EPS, EQL, FEX, FWDD, HUSV, IVV, IWL, IWM, JHML, JKD, OTPIX, PPLC, PPSC, PSQ, QID-OLD, QLD, QQEW, QQQ, QQQE, QQXT, RSP, RWL, RWM, RYARX, RYRSX, SBUS, SCAP, SCHX, SDOW, SDS, SFLA, SH, SMLL, SPDN, SPLX, SPSM, SPUU, SPXE, SPXL, SPXN, SPXS, SPXT, SPXU-OLD, SPXV, SPY, SQQQ, SRTY, SSO, SYE, TNA, TQQQ, TWM, TZA, UDOW, UDPIX, UPRO, URTY, USSD, USWD, UWM, VFINX, VOO, VTWO, VV
by: Abdulaziz A. Alnaim, CFA

Most of what has been written recently on the subject has been centered around the valuation of the US stock market because it appears - at first blush - to be the most expensive relative to both its own history and other international markets.

Before sharing my opinion on the subject, let me try to summarize the different positions on this issue. First, there is a puzzling death of public bulls on the US stock market today (Trump notwithstanding), which is odd given how well it has done and how high it appears to be relative to earnings. Around this time of the market cycle we would expect to see many predictions of a market that would keep going up forever.

With bulls mostly out of the picture, investors seem to have lined themselves into two schools of thought. The first claims that the stock market is insanely expensive and is in the middle of a big bubble that will soon pop, hurting everyone. The second school of thought, most recently including Warren Buffett, says that markets look expensive but that long-term interest rates are so low that holding stocks is preferable unless you expect rates to move up materially in the relatively near future. Let us call the first group the “Absolutists” and the second group the “Relativistic Pragmatists.” Now let us look at the evidence.

For the sake of the argument, we will focus on the US market as measured by the S&P 500. The Absolutists use two measures to argue their point: Robert Shiller’s cyclically-adjusted 10-years P/E average and the twelve-month unadjusted P/E Ratio. Here’s how the S&P 500 looks on both measures (using Shiller data for both). Both data sets go back to 1871.

Measure Current Median Current vs Median
Shiller “CAPE” 31.11 16.14 93%
12-month trailing P/E 25.42 14.67 73%

Source: Robert Shiller, since 1871

The numbers above definitely look scary! But I think they overestimate the over-valuation of the S&P 500 for the following reasons:

  1. The past ten years include the period of the Great Financial Crisis (“GFC”) of 2008-2009 and one could argue that earnings during that period were depressed to once-in-a-generation levels and are not comparable to other recessions
  2. Shiller calculates his medians over the period that extends back to 1871. The US economy and stock market have changed substantially since then. A more recent period would be more representative.

Let us see what happens if we do these adjustments. I’ve normalized earnings during the GFC to calculate the adjusted current CAPE (table below). Deciding which period to measure a median in is trickier, however. Jeremy Grantham has pointed out recently that the S&P 500 P/E multiple seems to have changed structurally since 1996 (see This Time Seems Very, Very Different, GMO Quarterly Letter, Q1 2017). This period predates the current ultra-low interest rate environment and may indicate a structural change, possibly due to a change in the mix of industries that constitute the S&P 500. I’ve calculated the mean over three periods (1945, 1985, 1996) just to see how it changes.

Current Median (1945) Median (1985) Median (1996)
Shiller CAPE 27.00 18.16 23.37 25.95
Trailing PE 24.27 16.82 19.66 21.64

Source: Mayar Capital estimates

You can see from the above that while the market looks elevated, it may not be as elevated as the first look at the published CAPE may lead one to believe (emphasis on the word “may”). Notice that we still haven’t brought the level of interest rates into the mix. We will get to that later. For now, let us dig deeper to see if we can find other ways of measuring the valuation of the S&P 500 in an absolute sense. Let us look at the following questions:

  1. Would the picture change if we used a normalized earnings measure instead of the GAAP earnings that Shiller uses?
  2. How do the different sector components of the S&P 500 look relative to their history?
  3. How does the S&P 500 look using other valuation metrics, including ones that adjust for leverage and margins such as EV/EBIT and EV/sales?

Let us look at P/E ratios calculated using earnings before extraordinary items. This removes the impact of one-off losses such as write-offs and discontinued operations. Note that you need to take such earnings with a grain of salt because companies classify too many things as “one-off,” but given that we are comparing this measure to its historical median a lot of that does cancel out. On this measure, the S&P 500 currently trades on a trailing P/E of 21.67x compared to a historical median (since 1990) of 18.14x. Not cheap. But not bubble territory either.

Looking at the valuation of the different sectors in the S&P 500, we discover that most sectors are not trading at multiples substantially different from their historical medians, with the exception of the Energy sector whose earnings are at a cyclical trough.

This leads me to think that the multiple of the overall market is being skewed upwards based on changes in the weights of the different sectors within the index. When sectors with a high median P/E (e.g., IT and healthcare) outperform ones with a lower median P/E (e.g., Financials and Energy) they become a bigger weight in the index and raise the P/E of the overall index even if the individual sector P/E ratios remain unchanged.

To adjust for that, let us look at the P/E ratio of the equally-weighted S&P 500 index. This index weighs its components equally instead of the normal practice of weighing them by their market capitalization. As you can see from the table below, this measure shows the S&P 500 to be pretty much at normal P/E levels.

Measure Current PE Median PE Current vs Median
S&P 500 Equally Weighted (since 1990) 21.36 19.73 8%

Source: Bloomberg, quarterly data, since 1990

Let us now look at other valuation metrics to adjust for margins and leverage. We will look at the enterprise value (equity and debt, “EV”) divided by earnings before interest and taxes (“EBIT”) and EV divided by sales. The first helps neutralize the effects of changes in the capital structure of the company, higher or lower debt across time, and the latter neutralizes the effects of changes in profit margins across time. You can see from the below that the market is somewhat above its historical median when the leverage is taken into account, with the caveat that with interest rate where they are, more debt does add a lot of value to the company by lowering its overall cost of capital, which warrants a high multiple to EBIT.

Source: Bloomberg, quarterly data, since 1990

I think a fair conclusion from the above discussion is that the S&P 500 overall looks to be trading at levels between slightly higher and substantially higher than its historical norms. So, while a deeper look into the numbers shows that the level of over-valuation is much lower than a quick glace would lead us to believe, there is strong evidence that it exists.

Now let us turn to the arguments of the group I called the “Relativistic Pragmatists.” They argue that P/E ratios should never be considered in a vacuum and one should always compare earnings yield (the inverse of the P/E ratio) to interest rates. Currently, long-term interest rates are extremely low (2.3% on 10-year Treasuries, below 1% in most of Europe and zero in Japan) and thus stocks are the better alternative, in their opinion. In the charts that follow, we compare the earnings yield on the S&P 500 to the yield on the 10-year treasury bond. You can see that by this measure, relative to bonds, stocks do seem to be a lot more attractive. Even comparing the dividend yield to the bond yield shows stocks as relatively more attractive than in most periods in the past.

But long-term rates are important not just because of the alternatives available to you today, but because in an efficient market they should be a good forecast of what future spot rates will be and thus which discount rates one should use in valuing companies’ future cash flows. A person with perfect foresight who knows that bond rates will remain low for the next 20 years, would happily pay even higher prices for companies than those offered by the stock market today.

graphs

There is a caveat to the above argument, which I’ve rarely seen mentioned by the Relativistic Pragmatists. The version of the future where rates remain extremely low must, by necessity, be one with much lower growth rates. So while lower interest rates do increase the value of discounted future cash flows, lower growth rates decrease the un-discounted future cash flows themselves and even though those two effects don’t fully offset each other, the impact of lower rates on valuation is not as big as some of the Relativistic Pragmatists would lead you to believe. Perhaps this is best illustrated with an example.

Company A generates $1.00 a share of Free Cash Flows (“FCF”) a year and is expected to grow FCF at an annualized rate of 5% over the next 10 years. After 10 years we assume it will have a terminal annual growth rate of 3.0%. The company has no debt. The rest of the assumptions are in the table below. Based on these assumptions, the intrinsic value of each share of Company A is $18.72.

Period Growth Rate 5.0%
Terminal Growth Rate 3.0%
Risk-Free Rate 4.5%
Equity Risk Premium 4.5%
Equity Discount Rate 9.0%
Estimated Intrinsic Value per share $18.72

Many Relativistic Pragmatists would argue that if long-term rates remain very low then the discount rate used in the above calculation would come down, increasing the estimated intrinsic value of the company. That is true. If we lower the Risk-Free Rate assumption by a mere 1.0% while leaving everything else the same, we get the following:

Period Growth Rate 5.0%
Terminal Growth Rate 3.0%
Risk-Free Rate 3.5%
Equity Risk Premium 4.5%
Equity Discount Rate 8.0%
Estimated Intrinsic Value per share $22.55

As you can see, lowering the Risk-Free Rate increases the intrinsic value by more than 20%. But as I’ve mentioned above, a future with a risk-free rate that’s 1% lower is also most likely to be a future where growth is also 1% lower. Adjusting the assumptions (see below table) brings the intrinsic value down to $20.16 per share, merely 7.7% higher than our original estimate.

So while the impact of a prolonged period of low rates is higher intrinsic values, the magnitude of that impact isn’t as big as some of the Relativistic Pragmatists think once we also adjust growth rates down to be consistent.

Period Growth Rate 5.0%
Terminal Growth Rate 2.0%
Risk-Free Rate 3.5%
Equity Risk Premium 4.5%
Equity Discount Rate 8.0%
Estimated Intrinsic Value per share $20.16

As you can see from the above evidence, I find it hard to take a very strong position one way or another on this topic. While my hunch is to lean on the side of the Absolutists, doing that would also imply accepting that long-term interest rates will rise substantially and soon, something that the bond market is not currently discounting and one that the experience of Japan over the past 20 years proves may not be the case. I am also willing to accept that I might be wrong on this, which may result in us missing out on some upside as valuations climb upwards from here and we continue to sell into that rise.

Instead of trying to pick a side in this debate, we will focus on doing our job and finding investment ideas that work in all of these scenarios. We believe that the recent difficulty that our investment process has encountered in finding new investment ideas is due to the valuations of different companies becoming too clustered around their median, coupled with exceptionally low volatility, the combination of which make stock-picking harder.

We don’t think this situation will last forever. Even in a prolonged low interest rate environment we expect something like what happened in Japan over the past 20 years, with global stocks moving sideways from here but with higher volatility. We believe our strategy will perform well in that environment.

First published in Mayar Fund’s Letter to Partners – September 2017 (portions have been redacted)