Hitchhiker’s Guide to the Equity Risk Premium

Deep Thought

In The Hitchhiker’s Guide to the Galaxy, a giant computer called Deep Thought is programmed to calculate the answer to the ultimate question of life, the universe and everything. Millions of years later, its circuits finally reveal the answer to the humans who have been agog with anticipation: 42.

That’s disappointing, and it then needs to embark on an even longer quest to work out what the question was. As the question turns out to have been, “What is 6 x 9?,” it turns out that the universe is based on a contradiction.

All of this is obviously both absurd and very funny. All of us have some sense what the ultimate question might be, and probably have some degree of an opinion as to our answer. None of us can possibly calculate it with precision. The mysteries of life are unknowable.

It’s hard not to think of Deep Thought as the Equity Risk Premium comes back into conversation. The ERP is the extra return you get from stocks compared to bonds in return for taking the extra risk. The higher the premium, the more attractive stocks will be, and vice versa. It lies at the center of the Capital Asset Pricing Model, which still guides much work on valuation. Business school students forever plug the ERP into their spreadsheet before they get started.

The problem? Like the ultimate question, the value of the ERP is unknowable in real time. It can only be known with hindsight, and history shows that it varies widely. That hasn’t stopped financial academics devoting lifetimes of research to calculating it. And the result of their research can be summarized as follows:

The problem? Like the ultimate question, the value of the ERP is unknowable in real time. It can only be known with hindsight, and history shows that it varies widely. That hasn’t stopped financial academics devoting lifetimes of research to calculating it. And the result of their research can be summarized as follows:

That chart was presented at a symposium of the biggest names in academic finance held by the CFA Institute in 2012, by Brett Hammond, then the chief investment strategist at TIAA-CREF. Some attendees said it was the most memorable slide they had ever seen. Quants can fire all the computer power and data they want at the issue, but at the end of the day, it’s hard to improve on a sensible round number: 4.

While we can’t identify the number in advance, we can know it in hindsight — just subtract the performances of bonds from stocks. If we look at 10-year periods, the ERP moves around a lot. In the decade after the CFA Institute’s first big colloquium on the subject in 2001, stocks did terribly and the premium turned out to be -1.1%. Over the following decade, when stocks went on a historic rampage, it was 11.1%. So says Hammond in the introduction to this year’s latest symposium on the ERP, which can be found here. It includes contributions from more or less everyone who has contributed to the debate, and it’s well worth some of your time. This is their spread of estimates, from 2001 and again this year:

So the great minds of finance can’t tell you by how much stocks will beat bonds over the next 10 years — it could be anywhere from 1% to more than 7%. A sensible median estimate is, you guessed it, 4%. However, the fact that the ERP moves around can be useful; even if it’s unknowable in real time, we can derive what the market is implicitly assuming as the equity risk premium by comparing the current real cost of debt and equity for companies, and subtracting the current real interest rate. That gives us the market’s aggregate assumption of what the ERP should be. This calculation shows that the ERP has tanked over the last year, as shown in this chart from Michel Lerner of the Credit Suisse HOLT team:

In more intuitive terms, if stock valuations have gone up, and the cost of borrowing for companies has gone up, while there is a better return available from government bonds, then the amount of extra return that investors expect in return for the risk of buying stocks must have declined. They’re buying stocks happily assuming there’ll be very little reward for taking that risk. A fall in the ERP on this scale is unusual. To quote Lerner, such a fall in 12 months “has only occurred seven times prior to now since 1983 and usually coincides with a large monetary event or an equity market shock/bubble.”

In straightforward absolute terms, without comparing to bond yields, equity valuations are again getting hard to ignore. In the 12 months to the end of August, the S&P 500 managed a price return of 14%. Of that, increasing price/earnings multiples accounted for 12.4 percentage points, according to a decomposition by Patrick Palfrey of Credit Suisse AG. Earnings themselves accounted for 1.4 percentage points. So valuations have boomed. This tells us nothing about timing, as any valuation can always become more extreme in the short term, but a lot about likely future returns. For valuations like this to make sense, investors must be assuming that bond yields are going to tank, which can only happen if inflation is not only beaten but totally trashed. To quote Lerner again, equity valuations are now at levels “consistent with the post Global Financial Crisis (GFC) era of loose monetary conditions where aggressive central bank intervention dulled the ups and downs of economic cycles.”

If inflation comes down and stays down, and rates aren’t “higher for longer,” that might be accurate. But the rising bond yields and rate expectations of the last few months suggest exactly the opposite.

In much longer context, this chart from Ian Harnett of Absolute Strategy Research offers the two most popularly cited long-term valuation metrics for US stocks — Shiller P/Es or “CAPEs” (for Cyclically Adjusted Price/Earnings multiples), which compare stocks to average earnings in real terms over the previous 10 years, and Tobin’s Q ratio, which compares stocks to the total replacement value of their underlying assets. Even if tech stocks, obviously generating excitement at present, are excluded, these measures suggest that stocks are historically expensive, at valuations comparable to the eve of the Great Crash in 1929 and of the GFC in 2007:

Over periods of 10 years or more, these indicators prove to be strongly predictive. The all-time record for the Shiller P/E in 2000 was followed by a terrible decade; the 50-year low it touched in the early 1980s was the signal for a fantastic bull market. A high Shiller P/E functions like a low implicit equity risk premium; it only makes sense if you think inflation will be tame, and hence bond yields can come down. This is explored in the CFA symposium; here is Harnett’s illustration, going back to 1910:

High valuations will have no effect on how the stock market performs next week or next month. But they do tell us that the market has baked in an assumption of calmly quiescent inflation, and that the losses if that doesn’t come true will be painful. As we await the latest installment of the US inflation saga on Wednesday, the stakes remain high. Sadly, there won’t be a simple answer.