An Analysis Inspired By Grantham And Hussman On Profit Margins

Includes: DIA, QQQ, SPY
by: Tom Armistead

With U.S. corporate profits as a percentage of GDP at a record high, well-known bears Jeremy Grantham and John Hussman have issued dire predictions. This article applies factual analysis to margins and market level, finding little support for a strong bear case based on margin reversion.

From GMO's Grantham, as presented on the Morningstar website:

Another breakfast thought is that the earnings level is abnormally high. And GMO is pretty academic, sometimes to its cost, and so we ruthlessly normalize earnings to very long-term averages. And that's what separates us from most of the data. People think the American market is very cheap, in general. A great majority think it's reasonably cheap, and we don't, because we want it to be priced to the normal earnings. And the bulls say, well, of course, there's been a paradigm shift. Profit margins are higher and always will be higher, because that's what bulls like to say. And we say, other things being equal, we'll always bet on the average--profit margins will come down.

From Hussman:

Presently, market cap is elevated because stocks seem reasonable as a multiple of recent earnings, but earnings themselves are at the highest share of GDP in history. Valuations are wicked once you normalize for profit margins. Given that stocks are very, very long-lived assets, it is the long-term stream of cash flows that matters most - not just next year's earnings. Stock valuations are not depressed as a share of the economy. Rather, they are elevated because they assume that the highest profit margins in history will be sustained indefinitely.

Looking at the Data

When investors say "the market," they most often mean the S&P 500 (NYSEARCA:SPY). Data on the earnings and sales of the index constituents is available going back to 2000, in the form of a spreadsheet from S&P, which is the source used for this study.

Briefly, analysis consisted of developing TTM as reported earnings and sales for each quarter, then dividing to get the profit margin. The margin, in turn, was charted against TTM as reported P/E. Here's the chart.

Click to enlarge

The fit for the regression is visually good, and sports an R2 of 0.95. Intuitively it makes sense: the market compensates for abnormally high or low margins in a way that implies reversion to the mean.

Working with the Regression Formula

Applying the formula from the chart above, at the most recent margin of 8.2%, the S&P 500 would be normally valued at 1,525, and a TTM as reported P/E of 17.2.

The average margin for the 12-year period is 6.38%. If margins revert to that level, a normal value for the index would be 1,465, somewhat above the current level, and less than the recent highs. The P/E would be 21.2.

If margins cratered to 1%, the norm would be 1,073, with a P/E of 99.9.

That's what the data shows. The period used, from 2000 to 2Q 2012, includes two recessions, one of which was extremely severe. An important factor is that it embraces a period of extremely easy money, which dates back to Greenspan's actions in the wake of 9/11, and is unlikely to abate within the next few years. There are no historical precedents, in the U.S.

The market is already priced for margin reversion, most likely triggered by a recession.

Doing a 5 Year Projection

Assuming 2% inflation, 2% real growth, and 1% value creation by share buybacks or acquisitions, nominal growth of revenue would be 5% per year. Applying that growth, but using the average margin of 6.38%, earnings in 2017 would be $87.45, which when applying the formula would imply an index level of 1,857 at a TTM as reported P/E of 21.2.

Buying at today's prices in the 1,450 area, and including an average dividend of 2%, an investor would receive annualized returns of 7.12%.

10 Year Treasury Yields

The 10 Year Treasury recently yielded 1.59%. The Fed model suggests that the earnings yield on equities should track the 10 year, with the addition of a risk premium. As a factual matter, when the 10 year is related to normalized earnings, such as Shiller's CAPE, a paradoxical relationship emerges.

When Treasury yields are less than 5%, CAPE decreases as the Treasury yield decreases, due to the flight to safety and consequent high-risk premiums. For the purposes of this article, I analyzed the relationship from January 2000 to December 2011, developing the following chart:

Click to enlarge

The linear regression has a respectable R2 of 0.73.

Applying the formula, 1.59% X 577 = 9.17 + 1.42 = 10.6. 10 Year average earnings for the S&P 500 (the E in CAPE) are approximately $65. $65 X 10.6 = 690. This formula telling us that the S&P 500 should be at 690.

If Treasuries were yielding 4.31%, the average for the period covered, the formula calls for CAPE to be 24.9, which would put the S&P 500 at 1,620.

Going back to the 5 year projection, we have $87.45 as normalized earnings in 2017. From there, we can multiply by projected CAPE to arrive at a projected index level. If GS10 is still at 1.59%, this line of reasoning calls for the index to stand at 926. If GS10 reverts to the mean of 4.31%, the index would be at 2,177, for an 8.52% annualized appreciation, to which the investor could add dividends.

Mean Reversion Applies to All parameters

The line of thinking here relies on a number of estimates or assumptions: inflation, real growth, profit margins, and Treasury yields. The point, after developing the relationships and doing the calculations, is that if mean reversion is applied to all parameters and projected forward by a reasonable period of time, very good returns are possible from the current market level, even if margins decrease by 22%, to the mean.

Caveats and Reservations

The R2 of 0.95 for the relationship between margins and P/E is appealing. However, if the actual values are compared to the formula, the variance ranges from +30.6% to -22.7%. Market values were extremely high early in the period, and generally undervalued since the Financial Crisis. There's a very wide range between the minimum and maximum variance.

Monetary policy is moving inexorably into uncharted waters. Experience has shown that anytime a variable gets out beyond the 95th percentile, strange things can happen, quickly. That's where monetary policy is now. Logically, investors should cave to the financial repression and go into equities.

In point of fact, investors have started paying up for stocks, contrary to the indications from the GS10 formula. That implies that the fear is going out of the market. Given the savage losses experienced in 2008 and early 2009, fear could return very easily.

Shiller's data goes back to 1871. I've selected a brief and very peculiar period for analysis, hardly sufficient to achieve the credibility that comes from a longer time frame. However, the most relevant data come from the period of easy money. Easy money is the wild card here.

A resurgence of inflation, to where the 10 year yields more than 5%, would be a game changer. This article relies on the assumption that any such reaction develops at a measured pace.

Investment Implications

Between August 20th and September 14th, I accumulated a hedge in the form of deep in the money puts on SPY. I closed it recently, with a fine IRR of 100%, influenced in part by thinking along the lines presented in this article.

The other half of the thinking was, by holding the proceeds in cash, I will be under no pressure to sell positions to meet expenses.

Investors who take a long-term view can create an analysis similar to those presented by Hussman and Grantham, developing an expected 7 or 10 year return from various market levels. The following chart from Hussman illustrates his method:

Click to enlarge

Hussman is saying that based on his projection of where the S&P 500 will be 10 years from now, annualized returns to that point won't be attractive, when risk is factored in.

I've developed a chart patterned after Hussman, which also includes a 7 year return, an interval favored by Grantham, and a 5 year return. I used the data relationships and estimates discussed in the article, including rigorous reversion to the mean.

Click to enlarge

The reason to show returns for the shorter period, is to illustrate the fact that the slope of the line gets steeper as the time period is shortened. If investors think they can get 10% a year for 5 years, the 10 year return is meaningless. As a practical matter, an investor with a 2-year time frame is very patient, by today's standards.

While there is ample reason to anticipate short-term volatility, a mean reversion analysis suggests that long-term rates of return will be very acceptable, compared to the alternatives in Treasuries or other safe haven investments. I'm investing on the basis that now is a good time to own quality stocks, the proviso being, that I'm using funds that won't be needed during the next two years.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: I have no position in SPY. I'm long numerous constituents.