Wednesday, February 12, 2014

Trendline S&P 500 EPS

Chris Brightman has a new piece out that is worth reading. He is with Robert Arnott's firm, Research Affiliates, in Newport Beach, CA.

Brightman's point is that reported earnings per share for the S&P 500 composite are markedly above trend. He is likely to prove too pessimistic, at least if he means to suggest a "multi-decade period of zero or negative growth in real earnings per share," which he mentions at one point. However, the underlying logic of his piece is good.

What I love about the piece most, though, is the bonus feature. Brightman compares current EPS to a trend line that goes back to the start of 1871. But he then provides a widget that lets you set your own time frame for the comparison. In other words, we may all agree that we're over trend, but whether we are 30% or so above, or almost 60%, varies with the start date. The widget lets you see how the slope and position of the trend line differs depending on the dates through which you draw it.

Will earnings revert back to some version of the trend lines we can draw with the widget? That seems like a fairly good bet, in the long run at least. This is, in part, because the politics of the situation will ultimately demand it, as Brightman argues.

But when will reversion happen? No one knows.

Keep in mind that the measurement of inflation and the rules of accounting have changed over time, which can distort things a little and make the tool less precise than it looks. Increasing use, over time, of stock option compensation and share buybacks are also factors. Having dug into such issues from time to time, I do not think any of them is big enough to distort the basic picture. Sometimes they offset each other, as in the case of previously-unexpensed options and buybacks. Anyway, a degree of approximation goes with the territory.

As you use the widget, please note that: (1) it's the dotted line, not the solid line, that moves when you change the start and end dates, and (2) the "Distance to Long-Term Trend" always compares current EPS to the 1871-2013 trend, rather than to the shorter period you may pick. Just keep your eye on the dotted line you are drawing.

By the way, I have written before about profits being high as a share of GDP.

HT for the Brightman piece: Michael Brakebill.

POSTSCRIPT (2/28/14): If the real (inflation-adjusted) trend line growth in earnings per share of the S&P 500 is around 2%, as it is in the Cowles Commission data series, then Brightman is correct that it should take more than a decade for the trend line to catch up to where earnings are now. A catch up in nominal terms, however, would be a lot easier to achieve and would happen sooner.