If you remain unconvinced by Bill Grossly Wrong on Stockmarket Returns, GMO have made the same argument more comprehensively in Reports of the Death of Equities Have Been Greatly Exaggerated. Read it, and if you still think GDP growth and stockmarket returns go hand in hand, please present some evidence.
The GMO piece deals with the main stumbling block quite well. Most disbelievers are confusing growth in corporate earnings with growth in stockmarket returns. Obviously, corporate earnings will grow in a growing economy and won’t in one that is stagnant. But that doesn’t translate into stockmarket returns:
Since 1929, [US] market capitalization has grown at 3.6% real, while corporate profits and GDP have grown slightly more slowly at 3.3%. The trouble is that none of this tells us much of anything about what the return will be to an actual equity investor.
Total corporate profits and total stock market capitalization have very little to do with earnings per share or the compound return to shareholders because new companies, stock issuance by current companies, stock buybacks, and merger and acquisition activity can all place a wedge between the aggregate numbers and per share numbers.
Just because corporate profits grow doesn’t mean sharemarket investors will do well. It all depends on how much additional capital has been invested to generate that growth. I said it in the last post and GMO mention it in their paper but it’s worth repeating again. The empirical evidence suggests that the relationship between GDP growth and stockmarket returns is an inverse one: historically, stockmarket returns have been worst in countries with the highest GDP growth.
My forecast of future stockmarket returns based on today’s earnings yield is simplistic in a number of ways. Firstly, I am assuming that the current level of earnings represents an appropriate base level. GMO’s analysis leads them to believe that current US profit margins are well above their long-term average and that, consequentially, corporate profits in the US are 30% above their sustainable level.
Several comments on my post also pointed out that, given that a large portion of corporate earnings in Australia are retained and not paid out as dividends, future returns will be as dependent on the return on that additional capital as they are on today’s earnings yield.
And, finally, GMO’s analysis suggests that historical actual returns have been consistently less than what would have been expected extrapolating a simple earnings yield:
If we stick with a corporate version of Hicksian income, where profit is the maximum amount a company could pay out to shareholders in a given period and maintain the same real earnings power, we might expect that the long-term return to shareholders would be the earnings yield of the market. This is theoretically very simple and appealing [and exactly what I did]. But when we do the math, it is difficult not to be a little disappointed on behalf of shareholders.
Since 1929, the average earnings yield on the S&P 500 has been 7.2%. The P/E of the market has also increased over the period from 13.8 to 15.8 on trailing net earnings. A naïve investor might therefore have expected to get a return of 7.4% above inflation, accounting for both the earnings yield and valuation shift. The actual return to the market since December 1929 has instead been 5.9% real. That’s 1.5% worse than one might have expected. What gives? The short answer is that earnings growth has been 1.7% real since 1929, while retained earnings have averaged 3.3% of market cap. That 3.3% could have been paid out as dividends, and if our earnings were truly economic profit that maintained the companies’ real earnings power, shareholders would have been able to pocket a dividend yield of 7.2% with flat real earnings. So, are corporations systematically flushing their retained earnings down the toilet? Possibly, but it’s also quite possible that earnings are simply overstated. Earnings are calculated not by economists, but accountants, and our guess is that if corporations had indeed paid out 100% of stated earnings, real earnings per share would have fallen significantly over time.
Those are US figures, but I don’t think Australian reported profits are any less susceptible to overstatement. There would seem grounds for knocking at least 1% off my 6% expected real return for Australian industrial shares.
Ex mining, I don’t think there are good grounds for adjusting the base level of profitability up or down in Australia. Nor do I think there is a good reason to expect returns on incremental capital to be lower than what we’ve experienced in the past (again I’m talking about industrial profits here – I expect most mining expenditure over the past few years to earn dreadful returns). We could, and probably will, quibble over those assumptions until the cows come home.
There are undoubtedly other rational arguments for expecting lower future stockmarket returns, too. The sole point of this and the previous post is that GDP growth isn’t one of them.