Felix Salmon

Revisiting the equity premium

By Felix Salmon
May 20, 2010

Allison Schrager takes up the question of the equity premium today:

The return equities generate in excess of the risk-free rate (which is normally short-term Treasuries), is often assumed to be between 5% to 8%. In my experience risk managers go silent when asked where exactly this number comes from.

Schrager herself isn’t much more exact, concluding that “for now it remains a difficult question”. But I think even so she’s a bit too optimistic about the outlook for equities:

A zero long-term equity premium assumes firms in most industries will not be very productive or profitable for decades.

I don’t think this is true at all. For one thing, stock prices tend to have some kind of productivity and/or profitability gains priced in to them: especially in the technology sector, it’s perfectly commonplace for a company to see rising profits which still disappoint the market so much that the stock price falls. Rising profitability or productivity do not by any means mean rising stock prices.

More to the point, improvements in productivity can end up either as returns to labor or as returns to capital; and returns to capital can end up either with bondholders or with stockholders. If productivity improvements end up flowing overwhelmingly to employees and to creditors, then there might well be very little left for shareholders, even if the company is becoming much more efficient over time.

Schrager then continues her argument with this:

Equities are inherently riskier than Treasuries. Equity prices must ultimately reflect and compensate investors for that risk or no one would hold them in their portfolio.

I’m not sure where that “must” comes from: maybe it’s some kind of corollary of the efficient markets hypothesis. Investors certainly hope that returns on equities will be commensurate with the risk that they’re taking. But there’s no rule saying that any given asset class will “ultimately reflect and compensate” those hopes. After all, if there were such a rule, then really there wouldn’t be any risk at all!

When I see people like David Merkel and Eric Falkenstein do the math and come to the conclusion that the equity premium is somewhere between 0% and 2%, I generally come away much more convinced than I am by the vague arguments of those who put it at 5% or higher — arguments which often boil down to “the future will be like the past, if you ignore the really bad bits of the past”. In any case, it’s pretty clear that the number of people with large stock-market investments is much greater than the number of people who really understand the full range of possible outcomes and are comfortable with how much they could lose in the markets if things go badly.

But I’m thinking that maybe the current bout of volatility is helping to bring that home, at least a little.

10 comments so far | RSS Comments RSS

I think that the risk free asset which is the US Treasury might have to undergo a definition change soon if the 100% sovereign debt and high fiscal deficits situation comes to a Greek Pass

Posted by AGreenInvestor | Report as abusive

This quote seems self-contradictory, “Equities are inherently riskier than Treasuries. Equity prices must ultimately reflect and compensate investors for that risk or no one would hold them in their portfolio.” [Schrager] The risk is partly observed in the volatility – varying prices. With the price varying, is the risk compensated properly all the time? Is the risk really moving, or is the rate at which the risk was compensated for going from under to over or somewhere between?

Sub-prime loans are inherently riskier than prime loans, however some sub-prime MBS’s were not yielding that much more than prime MBS’s (AAA rated tranches at least). You might even have a sub-prime loan with a teaser rate charging less of a premium than a really solid prime loan.

Option-ARMs further that point. An option-arm 100% financing no-doc loan is inherently riskier than a fixed rate prime loan fully-doc’d. Which yielded a better rate for the initial period? The extra risk was negatively compensated!

Posted by winstongator | Report as abusive

Assuming a positive equity premium is sort of like saying that equities are always a can’t miss investment. Isn’t the first rule of economics that there is no free lunch? Sometimes stocks are priced such that they are a good bet going forward, and sometimes not? Wouldn’t that have to translate into a variable equity premium, with that variation including trips into negative territory?

Posted by winstongator | Report as abusive

Brad says 4 per cent.

The U.S. Equity Return Premium:
Past, Present and Future
J. Bradford DeLong
Professor of Economics, U.C. Berkeley
Research Associate, NBER
Konstantin Magin
Visiting Assistant Professor, U.C. Berkeley

For more than a century, diversified long-horizon investments in
America’s stock market have consistently received much higher returns
than investors in bonds: a return gap averaging six percent per year.
This is what Rajnish Mehra and Edward Prescott (1985) labeled the
“equity premium puzzle”: the premium return on equities does not
seem to be matched by any obvious factor that makes the marginal
utility of wealth in states of the world when equities are cheap
sufficiently higher than in states of the world when equities are
valuable. The existence of this equity return premium has been known
for generations. More than eighty years ago financial analyst Edgar L.
Smith (1924) publicized the fact that long-horizon investors in
diversified equities got a very good deal relative to investors in debt:
consistently higher long-run average returns with no more risk. We
conclude that the equity premium puzzle has not been solved: it
remains a puzzle. And we conclude that we anticipate the equity return
premium to continue, albeit at a smaller level than in the past—perhaps
four percent per year.

Posted by maynardGkeynes | Report as abusive

Back when the market was truly stupid in the late ’90s, TIPS yielded 4%, indicating the market had priced in a new era where productivity was going to soar and real returns were going to be well over 4%. That was pretty crazy. By contrast, expecting private enterprise in a propertarian system designed and run by capitalists to deliver real returns below TIPS is almost equally crazy.

Current long-dated TIPS will deliver real returns a bit under 2%. As a result, if you believe equities are vaguely rationally priced, you should mark their expected real returns at greater than 2%. DeLong says 4%? That may be overoptimistic, but it’s in the discussion. Get 1% population growth, 0.5% immigration, excellent 2.5% productivity growth and no significant diminution in profit share, and you’re getting close. I’d be more comfortable with lower productivity growth myself, but I doubt we register a long-run 0%.

One aside: you’re confusing premiums over cash and over bonds. I like to use ‘real return’ instead of the former (since cash returns roughly equal inflation), reserving the phrase ‘equity premium’ for the expected long-term return premium over long treasuries.

Posted by wcw | Report as abusive

You can see historical equity premiums for 17 countries over 100 years as published by Professor Elroy Dimson.

http://faculty.london.edu/edimson/assets  /documents/Jacf1.pdf

The world averaged Geometric Mean equity premium of 4.6% and Arithmetic Mean equity premium of 5.9%

If it is believed there will be no equity premium then the value of equities should fall, increasing the earnings yield and increasing the equity premium. Though it is not guaranteed for any given year or decade it should over time persist. This has persisted for centuries.

Interesting point, growth is not necessary for high equity returns; in fact, I prefer low or no growth. Please read Jeremy Siegel’s Stocks for the Long Run.

Posted by cjkubx | Report as abusive

I feel that the equity premium is mostly a repercussion of the great depression. To a large extent, an entire generation that was too scared to invest set the stage for a major rebound which we call the equity premium.

In hindsight, it seems easy to say that in 2006-2008 these gains were finally sucked dry and the equity premium was at an end. Now, however, I believe this second “great recession” has set the stage for a new equity premium. The longer the market stagnates and the most fearful the public becomes, the more gains are bottled up for future investors who are willing to jump in.

I am personally tossing small chunks of cash into stink bid day-orders on blue chip stocks with high yields. This way I control my exposure to short term movements (and recent volatility means I can get discounts of 3-5%) and am paid 4-7% while I wait. These dividends continue to build my cash position.

I also hold a significant amount of fixed-income which I may back off in months or years.

Posted by RyanT | Report as abusive

Equity premium, sounds like hocus pocus to me. Schrager is almost correct…

Equities are inherently riskier than Treasuries. Equity prices must ultimately reflect and compensate investors for that risk or no one would hold them in their portfolio.

The reason you hold them in your portfolio is the earnings. Thus, the earnings should be higher than that of comparable investments. Using Shiller’s data set you can see from 1887 to 1960 the earnings yield averaged 4.53% spread over long term interest rates. Since then it has averaged -0.28%. Why did this happen; when did it become OK to accept a riskier asset class but with less of a stream of income? When it became a ponzi-esque game of the greater fool.

Posted by tcolemanuf | Report as abusive

WOW – from all of the research I have done, the equity risk premium varies with 10 year B and B- corporate bonds premiums compared to 10 year US treasuries for the US market. Please run the numbers and think hard on the subject as to why this seems true. (could it be that they are about as risky – hmmm)

Of course this describes the major index relationship and individual companies will vary depending on their characteristics. Large companies with great balance sheets will have a lower premium as will high growth companies until their growth rates falter.

I get such a big chuckle out of the “Fed Model” and the idiots that promote it. It is best used to measure over-valuation rather than under valuation. Stick to the 10 year B and B- corportes, making suitable adjustments for individual companies, and over the long run you will gain an understanding of valuations.

Posted by dirt2624 | Report as abusive

If you assume that companies would not be profitable for decades, would you then buy equities at a lower expected return to treasuries? Of course not.

Schrager also seems to confuse ex-ante and ex-post in her attempt at an article and to have a definition of “equity premium” that changes while she’s writing. No wonder her conclusion is so weak.

Posted by johnhhaskell | Report as abusive

Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/