Guest Post by Benjamin Strong: Stock Market Valuation/Atlantic Monthly Catastrophic FAIL Department

A guest post by the late Benjamin Strong (1872-1928), sometime president of the Federal Reserve Bank of New York:

Back when I was alive, it was taken for granted that there was an equity return premium: that investments in stocks would, on average and over time, have significantly higher retuns than investments in bonds. You could see this by comparing the coupon yields of bonds with the earnings per share of stocks: earnings per share were higher.

Admittedly, only some of those earnings were paid out in dividends--the rest being retained and reinvested in the business. And even though the retained earnings had been earned and were the property of the shareholders, perhaps they would never see them. Perhaps the retained earnings would be wasted by feckless managers. Perhaps they were not really there at all in the first place and were just the results of financial manipulation--ENRON, anyone?

On the other hand, bond yields overstated real returns by the amount of inflation. Since the application of cutting-edge high-tech organic chemistry to the gold-bearing rocks of the Witwatersrand starting in the 1890s we have always had inflation. And stock earnings understate the returns to equity investments that companies derive by virtue of having established market positions in an economy growing in density and technology.

That there was an equity return premium was something that nobody doubted. The worry was that investments in common stocks were so risky as to be rank speculation: invest your money in a bond for two generations and it would still be there, but were you to invest your money in a stock there was a chance that you would be fabulously rich but the likelihood would be that the company had gone bankrupt and taken your investment with it.

The first major twentieth-century advance in investment science was made in the 1920s in the pages of the Atlantic Monthly, when Edgar Lawrence Smith made two points in an article he titled "Common Stocks as Long-Term Investments":

1. The power of diversification: even though each individual company was likely to go bankrupt and render its stock worthless over the generations, a diversified portflio of common stocks would provide insurance and would over the generations provide greater returns than bonds with no more risk.
2. The threat of inflation: in a post-WWI world where governments had learned that they could abandon the gold standard and survive and even prosper, even diversified investments in bonds provided no hedge against inflation, while a diversified portfolio of stocks did.

These two points established, to Lawrence Smith's and to my satisfaction, that common stocks were excellent long-term investments and not speculative at all--if your portfolio was properly diversified.

Indeed, those who starting in the 1920s invested in common stocks and had long enough time horizons for the central-limit theorem to make itself felt did indeed prosper mightily if they followed Edgar Lawrence's advice.

So now that I have been re-surrected or re-incarnated or whatever due to these alien technological marvels I do not understand, I eagerly opened its pages when I came across a new issue of the Atlantic Monthly: what new advances in investment science have been made now?

It understates the case to say that I am disappointed to find the Atlantic Monthly publishing an article by Megan McArdle denying that common stocks are sound long-term investments. And I am most disappointed of all because the article--"The Great Stock Market Myth: Why the Market's Rate of Return--and Your Nest Egg--May Never Recover"--contains no argument against the conclusions of Edgar Lawrence three generations ago.

Why does Megan McArdle believe that it is a "myth" that common stocks are sound long-term investments, promising likely real rates of return higher than bonds and of the order of magnitude of their earnings yields?

It is a mystery.

Why does the author of the article think that the idea that in the long run stocks are likely to outperform bonds on average by a substantial margin is a "myth", and that the stock market's rate of return "may never recover" from its depressed near-zero net real values of the past decade?

It is a mystery.

I cannot find a coherent argument anywhere in the article text.

There is a statement that:

Over the past decade, equity investing hasn’t offered much of a premium. The market went up (the Dow hit another record high in the middle of the decade). But then it went down again. In finance terminology, we experienced a lot of volatility—the major indexes have fluctuated a lot—but not much real growth...

But there was plenty of growth in earnings over the past decade. Returns over the past ten years have been low not because of slow growth in earnings: earnings were as much higher in the 2000s relative to the 1990s as they were in the 1990s relative to the 1980s. Common-stock returns over the past decade have been low because price/earnings ratios have collapsed from their late 1990s bubble values. These low returns over the past decade set the stage not for lower but for higher returns over the forseeable future.

There is a statement that:

[O]nce everyone believes that the stock market offers high returns for relatively little risk, that notion stops being true. And everyone apparently does believe just that—even after the 2008 crisis, the price-to-earnings ratio of the S&P 500 remains near the top of its average historical range...

But "near the top of its average historical range" means "about where it was in 1880, or 1900, or 1938, or 1962, or 1995--and since those dates average annual real returns have averaged 7.25%, 7.18%, 6.86%, 5.43%, and 5.94%, respectively. It's very hard to say that a price-to-earnings ratio "near the top of its average historical range" implies a likely future real return of less than the earnings yield of about 6% per year.

And there is a statement that:

[A] 2 percent return seems to be a real possibility—in fact, it’s a hair above the 1.8 percent that Smithers & Co., an asset-allocation consultancy, forecast for U.S. equities over the next decade...

As best as I can figure out from this marvelous internet of yours at http://www.smithers.co.uk/page.php?id=34, Andrew Smithers's 1.8% per year return forecast is not a long-term forecast. It is a medium-term forecast of likely returns over the next decade. It is driven by Smithers's expectation of a 40% fall in the price-earnings ratio to its long-run historical average over the next decade. After that, the real return on equities predicted by the logic of Smithers's q-theory model is something like 8% per year. Thus Smithers's long-run--fifty year, say--forecast rate of return is not 1.8% per year but instead something like a real return of 7% per year.

Regrettably, I had to do this digging myself. There is no explanation in the Atlantic Monthly article of who Smithers and Company are, or why anybody should trust their forecasts, or what the logic might be behind their forecasting a 1.8% per year real return for U.S. equities over the next decade.

And that is it. That is all the analytical backing behind the headline that "The Great Stock Market Myth: Why the Market's Rate of Return--and Your Nest Egg--Will Never Recover" that Megan McArdle's article contains.

At the moment what Robert Shiller calls the Graham Ratio--the ratio of U.S. stock index prices to a ten-year average of earnings--is 20.6. Since the center of gravity of a ten-year moving average is five years in the past, that indicates a current permanent cyclically-adjusted earnings yield of roughly 6%, which is almost exactly the current earnings yield on the S&P Composite.

The return on stocks is the dividend yield plus the capital gain. If the P/E ratio is stable, the capital gain is equal to the growth of earnings. This year's earnings are (a) paid out in dividends--that is the dividend yield--and (b) reinvested to increase future earnings. If the rate-of-return the company gets on its reinvested earnings is the same as the rate of return on the stock market, then the dividend plus the capital gain will together be equal to earnings. If the internal-external rate-of-return wedge is positive, then the dividend plus the capital gain will exceed the earnings yield; if the wedge is negative, returns will fall below earnings yields.

Thus the permanent earnings-to-price ratio--the permanent earnings yield--ought to give us something close to the expected long-run rate of return on the stock market. And that permanent earnings yield is the thing that is now something like 6% per year. That is very good indeed compared to 10-Year Treasury TIPS yields of 0.95%, and 30-Year Treasury TIPS yields of 1.80%.

Now even with Edgar Lawrence's insight into diversification stock market investments are risky because they are exposed to systemic risks. Perhaps the profits of American companies will collapse and never grow again. Perhaps American companies' reinvestments of retained earnings will be dissipated in prestige boondoggles due to failures of corporate control. Perhaps the American economy will enter a long-run decline and an American market position will be an eroding asset. Perhaps the earnings yield does vastly overstate likely long-run returns.

But if Megan McArdle believes that it does, isn't she under an obligation to her readers to offer an argument for why she believes that stock market returns in the future will grossly differ from earnings yields?

And isn't the Atlantic Monthly as an institution honor-bound to send the article back for another round of rewriting until such an argument is offered?

It is true that I do see much progress: electric toothbrushes, the internet, South Park, containerization, and the Green Revolution are all marvelous technological advances.

But I want my old Atlantic Monthly from the 1920s--the one that published Edgar Lawrence's "Common Stocks as Long-Term Investments"--back, not this new-fangled beast.