Really, Truly, Holistically, Stocks Are Expensive: John Authers

To get John Authers’ newsletter delivered directly to your inbox, sign up
here
.

A Feather in the CAPM

Armistice Day, or Veterans Day, is a great time to pause and take stock. The U.S. Treasury market closes for the day, so there’s no risk of hurly burley over inflation and monetary policy. It’s a good opportunity to get out of the short term and look at the very long term, through the lens of one of the most ambitious pieces of research I’ve ever seen. Jacques Cesar of the consultancy Oliver Wyman has spent the last five years working out an alternative to the capital asset pricing model, or CAPM, and he thinks he’s delivered.

The consultancy is now sharing the work as a series of white papers. Two are available so far, along with an outline of the research, and you can find them here.

This is a big deal. CAPM, which breaks down the return on equities into a risk-free rate, an equity risk premium, and a factor called “beta” to cover sensitivity to the market, has held sway for decades. It lies at the base of the architecture of asset allocation models and “efficient frontiers” that dominate investing. After the implosion of 2008 it was widely held that a new model was needed. But a decade on, the same ones still tend to be used, on premises that many now believe faulty.

Cesar, a veteran consultant who spent most of his career in the fast-moving consumer goods and energy sectors but is a newcomer to financial markets, can be seen as the Elon Musk of financial theory. The “why not?” attitude as he tries to answer questions that have eluded others is the theoretical finance equivalent of setting out to colonize Mars. Good for him. But there is also the issue of hubris; does this make sense?

Let’s be clear about my limitations. I’m not an academic, and I don’t have the time to go through the research in detail, check the numbers (which involve heroic assumptions about incomplete data series going back 150 years) or stress-test the conclusions. That’s a job for academics. But I can say that Cesar’s “Holistic Market Model” makes intuitive sense, and that it does offer believable answers to some difficult questions. Points of Return readers would be well advised to download the material, which is all free, and ponder it when they have time. For a summary, here goes:

Earnings

When arriving at a valuation of a company, we need to know its earnings — the flows of cash that it will produce. Accounting standards change over time. How seriously can we take earnings per share?

The Cesar approach was to move from GAAP EPS to what he calls “Buffett earnings.” This refers to the surplus that can be distributed to shareholders once all reinvestments needed to keep the business going have been made. This requires judgment calls, but probably better reflects the true economic earning power of a company than GAAP (if you get it right). Historic cost depreciation, used in GAAP, understates the amount that companies need to reinvest in inflationary times, and overstates it in deflationary times.

The result is to render earnings significantly lower than they appeared from 1950 through to the mid-1990s (not coincidentally a period when the market generally did well without a major speculative bubble), and slightly higher than they have appeared in recent years:

This means that when we take a cyclical price-earnings multiple with the new numbers, the market is revealed not to have been as extremely cheap as it appeared in the 1970s and 1980s, and not as expensive as it appears now. This accords with common sense. As stocks relative to earnings so recalculated are still as expensive as they were on the eve of the Great Crash in 1929, we still shouldn’t find this too reassuring:

Pre-pandemic, the Oliver Wyman model suggests that stock valuations were 1.6 times their historical average, while the widely used multiples published by Professor Robert Shiller show a ratio of 2 times. The historic average earnings yield, so calculated, has been 6.8%.

Margins

We also need to know about profit margins, which can be measured in several different ways, and which are also prey to unreliable historical data. The literature assumes margins are cyclical and mean-reverting. The Oliver Wyman team built estimates of sales per share for the S&P 500, only regularly computed since 1991, back to 1871, and then expressed margins as earnings divided by sales. The bottom line findings are significant: Margins before tax are near an 80-year high, the combined corporate and personal tax rate is near an 80-year low, and margins after all taxes are near a 100-year high. Margins have been drifting up more or less constantly since the beginning of the Clinton presidency. If they are to revert downwards, it will be politics rather than any internal market dynamics that does the job.

A Discount Rate

Once you have a reasonable version of earnings, you need a reasonable rate of interest with which to discount future streams. This also allows you to come up with the “equity risk premium” (the extra return for investing in stocks rather than something risk-free) that is at the heart of the CAPM.

The new model accepts the underlying mechanics of CAPM, but only after making significant changes to all the inputs. Rather than just plugging the 10-year Treasury yield into the formula, Cesar comes up with a Really Truly Risk-Free Rate, or RTRR, which is converted into real terms by subtracting inflation expectations, and which also removes what he calls a “Treasury Risk Premium” that reflects changes in the incentives to buy bonds. In recent years, regulations have pushed banks and pension funds toward buying Treasuries, and so demand for them doesn’t merely reflect the appeal of their being risk-free. This captures the way in which “financial repression,” or using regulation to force people to lend to the government at low rates, helps to prop up asset prices. All of these adjustments lead to a “True North ERP” which comes from subtracting the Really Truly Risk-Free Rate from the Buffett Earnings Yield:

The basic and critical fact of asset allocators’ lives for the last four decades remains unchanged. The RTRR, like the 10-year Treasury yield, has been steadily trending down for all of that time, thus enabling higher valuations for stocks.

Moving the True North ERP

Having done all this work, it was possible to map how the ERP moved over time. That then enabled the team to look for factors that could explain the observed changes. Cesar and his team believe they can account for almost all differences in the equity risk premium, and therefore all variance in the valuation put on stocks relative to bonds, with the following four factors:

• “Business cycle and sub-cyclical variations in economic and financial risk.” A quantitative risk-aversion indicator shows where market crashes will happen.

• “Inflation outside of a ‘Goldilocks’ zone.” Extremes of inflation and deflation cause problems for equities. While it stays within reasonable bounds, inflation has much less impact.

• “Intergenerational increases in risk aversion driven by long secular bear markets.” This captures flows in sentiment that show up in long-term valuation graphs; for example, people were turned off equities by the declines after the Great Depression. This is modeled using “reflexivity” — declines will make people want to make the market fall more — as championed by George Soros.

• “Somewhat imperfect risk arbitrage between equities and Treasury bonds.” This is like an error factor. The equity risk premium tends to be even higher than it should be when the RTRR is extremely low, as it is now.

The one thing missing, on which the team is still working, is a factor that would explain periods of speculative excess. That one should be fun.

Supply and Demand

Cesar also looked at a completely different approach, derived from Victorian economics, which looks at supply and demand for stocks. Model the moves in supply and demand successfully, and you will capture changes in share prices. Starting with GDP, they modeled demand for equities by taking into account the following three factors:

• “Average income per capita by tranche of income.” This factor captures both the impact of productivity growth — which will raise demand — and of inequality. When there are proportionately more rich people who have much money left over after spending to support their standard of living (as is true now), demand will be greater.
  

• “Inflation and unemployment.” This part of the model is inspired by the “misery index” — the sum of inflation and unemployment. They attack demand for equities.
  

• Natural equity/bond allocation. This is driven by secular shifts in the behavior of both direct investors and pension managers — a variable that captures flows to match liabilities and so on.
  

Put these together, and you can work out what proportion of GDP will go into equities. Predict GDP and you have an estimate for the stock market, which works very well. The model’s forecast is in the next chart; it correctly envisions that equity wealth will be higher in proportion to GDP now than at any time since 1929:

once more

Where next?

If the model is right, we can assess the chances for future equity growth by looking at whether the secular drivers can be sustained. Growth in productivity and demand from pension funds, both identified as important, will continue. The critical questions for the next decade or so are:

• Will the near-40-year decline in the RTRR (driven by Treasury yields) be reversed?

• Will inflation remain in the Goldilocks zone, or will it escape at one end or the other?

• Will secular changes in the regulatory regime put pressure on pre-tax margins and/or will changes in the tax regime put pressure on the combined corporate and personal tax rate? This question could also be reframed as whether the four-decade rise in inequality will end. (It would be bad for stocks if it did.)

If I had had to draw up three questions to work out where we’re going, they would be very similar to this. It’s impressive that they are the outputs from a big empirical model, so I’d take it seriously. And taking these questions together shows why the question of inflation is so important.

Problems?

The most obvious is a possibility of data-mining or “retro-fitting.” The researchers have worked hard to find factors that explain what actually happened, and they’ve found a great fit. Will this continue to be the case? Many new investment factors have beautiful back tests and cease to work as soon as anyone tries to put real money to work in real time. There are bound to be changes ahead, at best.

But was this exercise worth it? Yes. At present, many people are using a model that they know doesn’t work, because at least it gives them an answer. After five years of work, Oliver Wyman has produced a model that has a better chance of giving the correct answer.

Crypto-Trust

A brief reminder that we are discussing Why Trust Matters by Benjamin Ho in a live blog on the terminal on Wednesday next week, at 10 a.m. New York time. It’s a fascinating study of exactly why trust matters to economies and markets, and how it can be built. It also covers crypto-currencies, designed to circumvent any need to trust each other or a bank or a government by instead putting all trust in an algorithm. Ho’s views on this are worth reading.

I’ll be discussing it with him, and by my colleague Stacy-Marie Ishmael, who spearheads Bloomberg’s developing crypto coverage. If you have questions, please send them to the book club email: authersnotes@bloomberg.net.

Survival Tips

I’ve been writing this on what in Britain is known as Armistice Day, commemorating the end of the First World War on the 11th hour of the 11th day of the 11th month. For a towering work of art to emerge from it, try Benjamin Britten’s War Requiem, which sets to music the poetry of Wilfred Owen, who was killed in the last week of the war. Owen is a principal character in Pat Barker’s Regeneration Trilogy series of novels, which are wonderfully moving, and might make great reading over the Thanksgiving break. And for remarkable comedy set in the unpromising terrain of the trenches, try Blackadder Goes Forth.

Finally, for an American version of what is known here as Veterans Day, here once more is the view from my window:

They shall be remembered. Have a good weekend everyone.

To contact the author of this story:
John Authers at jauthers@bloomberg.net

To contact the editor responsible for this story:
Matthew Brooker at mbrooker1@bloomberg.net

© 2021 Bloomberg L.P. All rights reserved. Used with permission.