[Acknowledgment: Thanks to Edward Worsdell for helpful comments on an earlier draft of this post.]
Ever since the publication of the book Triumph of the Optimists: 101 Years of Global Investment Returns by Dimson, Marsh and Staunton (already 15 years ago!), it is basic knowledge that equities, as an asset class, outperform other asset classes in the long-run, in particular fixed income (both bills and bonds). As you can see in the following graph from the Credit Suisse Global Investment Returns 2017 edition, equities on a global basis (and not just US equities) have outperformed both bills and bonds by very wide margins – although this historical advantage seems to have faded away in the last 15 years:
From a theoretical perspective, however, the fact that equities should outperform both bills and bonds over the long-run has a long and distinguished lineage. It was Lawrence Smith, in 1924, in his classic book Common Stocks as Long-term Investments, who was the first to advance the idea (which for many was counterintuitive at the time) that stocks should deliver healthy returns for investors, with the proviso that you should hold them for several years – and over several business cycles. And as it happened, Keynes, who read Smith’s book as soon as it came out, was the first institutional investor (through the King’s College’s portfolio) to introduce stocks as a permanent, core asset in his portfolio allocation (and replacing the core role that real estate had played up to that point). And since then, the rest of the story is better known: from the Buffett’s investment success thanks to his patient approach at the helm of Berkshire Hathaway, to the books by Jeremy Siegel popularising the idea that equities, if bought at reasonable valuations, will do fine in the long-run.
The availability of data has also led to other minor, but also well-known, equity facts, such as the outperformance of small cap stocks over larger ones. The small cap outperformance is usually discussed in terms of average returns (as in Fama-French’s 1992 seminal paper), but often very little is said about the shape of the distribution of those returns. Hendrik Bessembinder has recently published a working paper provocatively titled “Do Stocks Outperform Treasury Bills”? that wants to address these issues. The working paper is a compulsory lecture for all stock investors, but especially for those investors who think that concentrated-holdings portfolios are the only path to long-term success.
The paper uses the CRSP database from the Chicago Booth School of Business, which comprises data from 1926 to 2015 for 25,782 stocks (a stock here is defined as a security associated to a PERMNO number, which tracks a specific security through its entire history regardless of name or capital structure changes). The results are on a monthly basis. There are several important conclusions that are worth discussing, but two of them especially emerge: i) the average life of a stock is around seven years (although this short average life is not only due to bankruptcies, but also to delistings, mergers and so on, it is still incredibly short), and ii) most of the stocks do not outperform (after taking into account reinvestment of dividends) even the one-month Treasury bill in a decade holding period (the paper also discusses other horizons with similar conclusions, I have simply selected the decade horizon for convenience).
The following table shows the distribution of stock returns for a decade holding period:
From a decade-holding-period point of view, as the author notes, “the most frequently-observed decade holding period return (rounded to the nearest 5%) is -100%”. The shape of the return distribution is positively skewed (14.2) and mean total (cumulative) returns exceed median returns by a large margin (118% vs. 14%). But more importantly, only 49.2% of the stocks outperformed the 1-month treasury bill, whereas only 37.4% outperformed the market return that could have obtained investing in a value-weighted stock index and only 33.9% outperformed an equal-weighted index. The main statistics are displayed in the following table:
From another perspective, the positive skewness also means that only a few stocks are responsible for the overall shareholders’ wealth creation. And here the results are worth taking a look at. The next table shows the 30 stocks that have created the largest amount of wealth since the inception of the database (i.e. 1926) until 2015. The stock that ranks first is Exxon, with a total dollar wealth creation of $939bn, followed by Apple and then GE. In terms of cumulative returns, Altria’s shareholders (Phillip Morris) have been the luckiest ones, with an impressive cumulative gross return of 202,9 million percent! – a result that already received a fair amount of attention in Siegel’s book The Future for Investors. But the really important thing to note here is that just the top 30 stocks were responsible for the 31.2% of the total stock market’s wealth creation since the inception of the index:
The same information can be gleaned from a cumulative distribution, which shows that out of the total universe of the stocks in the database (25,782 stocks), 50% of the total wealth creation came from just 86 stocks (0.33% of the total universe), whereas just 983 stocks were responsible for the total shareholders’ value creation. Beyond that there was some additional value creation (the graph tops at 120% and then declines), but the takeaway is that around 96% of stocks “collectively generated dollar gains that matched those that would have been earned had the invested capital earned the same rates as one-month U.S. Treasury bills” (p. 21).
The author ends the paper with a sobering conclusion for investors:
The results reported here also highlight the fact that poorly diversified portfolios occasionally deliver very large returns. As such, the results can justify a decision to not diversify by those investors who particularly value positive skewness in the distribution of possible investment returns, even in light of the knowledge that the undiversified portfolio will more likely underperform.
The “New Normal” in equity valuations? Some (theoretical) reasons why value investors are throwing in the towel
But this post is not about why stock investors, in general, should avoid concentrated portfolios. Rather, the post wants to contribute to an already lengthy debate that tries to address two more obtuse questions, namely: why are equity valuations currently so high? And, should we expect them to remain so lofty in the foreseeable future? Both questions have been addressed in this blog in the past, so it is high time for an update.
There have recently been two attempts at explaining why current stock market valuations are so high that deserve particularly close attention: a post written by the anonymous author of Philosophical Economics and Jeremy Grantham’s latest quarterly letter to GMO’s investors. It goes without saying that both of them are worth reading.
Philosophical Economics has been a long time advocate of the motto “higher (valuations) for longer”. All of his posts are carefully articulated, so it is no wonder that in his latest post Diversification, Adaptation and Stock Market Valuation he again discusses the issue of stock valuations from a slightly different angle. Actually, he starts from some of the Bessembinder’s findings to argue that indexing and the rise of vehicles to invest for passive investors have been one of the most powerful drivers for richer equity valuations over the last 30 years.
Imagine that you were an investor in 1926 deciding whether to invest in stocks or bonds. Obviously, you did not have a great deal of historical information at the time to make an informed decision (stock markets only begun to be active trading places in the latter part of the nineteenth century with the rise of railway stocks). Individual stocks seem to be obviously riskier, tied to the vagaries of individual companies, but does this riskiness compensate for the returns they deliver in the long-run? Take a look at the following graph:
In hindsight, the benefit of owning equities is clear. According to Philosophical Economics, this is one of the reasons why equities are currently so expensive by historical standards: investors know their historical performance and they factor this history into their portfolio decisions – and thus equities are overbought. Although equities seem riskier in the short-run, investors have now enough data to tell them that over the long-run equities are a no-brainer.
Even more, although you may think that equities as a whole are a good idea for the long-run, you still need a suitable vehicle to invest in them; in other words, as we saw earlier, you need some diversification. This is Philosophical Economics’ second point: once you have suitable investment vehicles to give investors access to passive strategies with a decent amount of diversification, the willingness of people to jump into the asset class is higher, as you can see in the following graph:
A corollary of this idea is that not only the stock market as a whole has become more expensive than it used to be, but also that the traditional small cap premium over the large cap stocks will tend to disappear, as investors notice that (in the long-run) they can snap up that 2% annualised premium without incurring additional risk (the previous graph shows the large performance gap between “small” and “equities” over the last 80 years).
On other hand, more striking is Grantham’s latest letter titled This Time Seems Very, Very Different. It is well known in the investment community that Grantham has always been a staunch advocate of mean-reversion principles when it comes to equity valuations. But in the letter, he admits that:
We value investors have bored momentum investors for decades by trotting out the axiom that the four most dangerous words are, “This time is different.” For 2017 I would like, however, to add to this warning: Conversely, it can be very dangerous indeed to assume that things are never different.
Why, according to him, could we expect higher valuations for longer? He argues that the main determinant for higher valuations are higher levels of corporate profitability. To support that assertion, he shows the evolution of NIPA non-financial corporate profit margins since 1970 and claims that one can see two very different periods, pre-1997 and post-1997:
But the most striking thing about Grantham’s analysis is not that one can argue the point that higher profit margins might not lead automatically to higher P/E ratios (although Grantham promises he will deal with this technicality in a future post), nor that the most relevant measure for valuations are returns on equity and not profit margins, but rather the reasons he offers for the persistence of higher profit margins, namely:
- Increased globalisation increasing the value of the brands (and the US has a wide universe of big brands).
- Steadily increasing corporate power over the last 40 years (although the reader should note that Grantham focuses in the differences between pre-1997 and post-1997 conditions, so maybe 40 years is not a very relevant period for this analysis).
- Reduced emphasis on growth relative to profitability.
- Increasing monopoly power.
- And lower real interest rates and higher levels of leverage since 1997.
For a house like GMO that has been promoting profit margin analysis based on the Levy-Kalecki profit equation in the past, such a list seems rather unstructured and arbitrary. A more compact macro analysis based on such an equation would say that profit margins have remained high thanks to higher-than-average levels in government deficits and lower-than-average levels in households’ savings. Also note that the third point Grantham has proposed (reduced emphasis on growth) is not a blessing for profit margins, but rather the opposite: investment creates profits, and there is abundant historical evidence that in those countries where the investment share in GDP is high (as in the Chinese case), profit margins are also bound to be high. Finally, the better-than-expected performance in the US balance of payments (mainly driven by a reduction in the oil and related products deficit) has also helped to lift overall profit margins in the last few years.
Other possible reasons for the “New Normal” in equity valuations: income inequality and tax havens
The post is getting a bit long, but I beg you to bear with me for a bit longer. So far, the debate on equity valuations has pitted two different views against each other: the “higher for longer” approach and why we should not expect the old good days in the foreseeable future (exemplified by Philosophical Economics) and the full mean-reverting perspective (exemplified by the GMO team and the “older” views of Grantham), which says that history is the best guide to predict equity valuations in the future.
I wrote some time ago that in this debate, virtue lies in the mean (no pun intended). As I explained in 2015:
Our opinion is that, yes, the US stock market is expensive, pretty expensive actually, but we are quite sceptical about the predictive value of historical averages for this kind of analysis.
Our readers already know that we believe that equilibrium valuations (defined as those valuations where an index is roughly properly valued) change over time and, in particular, that index valuations are not independent from the growth rate of the economy. In other words, it is non-sensical to defend that the historical average will remain immutable in the future regardless of future macroeconomic conditions.
So the way we should look at current (and future) equity returns is to expect higher average valuations than in the past, but also to consider current equity valuations as ludicrous. But to claim that once the market corrects we should expect a return for the old good days does not have any theoretical support. At the end of the day, stock market valuations are influenced by the macroeconomic environment, and there are many macroeconomic variables that are not mean-reverting. So the goal of a proper theoretical inquiry would be i) to identify which macroeconomic variables are crucial in determining equity valuations and then ii) to understand whether these macroeconomic variables are mean-reverting or not.
I have written on this topic in the past and the most I can say is that there are at least two crucial macroeconomic variables that everyone should take into account: the real growth rate of the economy and the marginal propensities to consume (out of income and out of wealth). What I found is:
i) lower real growth rates justify higher valuations, which from a microeconomic point of view is counterintuitive, because all the discounted-flow (dividends, free cash-flows, etc.) models tell you that lower growth rates lead to lower valuations (in an absolute and relative basis). If you still don’t believe those guys that say that we are living in a secular stagnating world, take a look at the following graph, depicting the US real GDP quarterly growth rate since 1960:
This secular lower growth rates have been coupled with higher equity valuation metrics. In a nutshell, the lack of investment opportunities in real assets (because of a lack of fixed capital investment) forces savings to chase financial assets, thus pushing valuations higher.
ii) lower marginal propensities to consume out of wealth drive higher valuations. Here your “microeconomic intuition” would lead you astray again. Although from a microeconomic point of view wealth is valuable as if you don’t consume it you can grow it through the capitalisation of returns, from a macroeconomic point of view wealth is valuable when it is consumed – the lower the rate at which wealth is consumed, the lower the overall returns it will yield. Measuring the evolution of the marginal propensity to consume out of wealth is a difficult endeavour (with a wide range of results depending on how you define wealth), so what I present here as a very rough proxy is the quarterly US household consumption divided by household total wealth (in a slightly different context, some results show that this ratio is a powerful predictor of future equity returns). You can see how the ratio has been trending downwards since the 1970s:
But if you think the previous measure is too rough, we have now plenty of evidence (here and here) that tells us that over the last few decades we have seen an unprecedented rise in income and wealth inequality. The higher the inequality, the lower the propensities to consume. Wealth inequality has been bolstered not only by the globalisation and financialisation of Western economies, but also by the rise of tax havens. From this point of view, tax havens have been a powerful driver for higher equity valuations through i) the lack of taxation of real economic activity and ii) wealth redistribution (maldistribution), which has pushed lower the amount of wealth that can actually be consumed.
So, from a broader perspective, the debate on current equity valuations cannot be separated from the debate on where we stand at the macroeconomic level. What my analysis reveals, however, is that many of the changes hailed by investors over the last three decades as good for equity returns (exemplified for instance by the concept of shareholder value orientation), aren’t worth a second thought. As long as we don’t see a broader change in macroeconomic conditions, we should not expect a brighter future for equity investors.