From an April 11, 2006 article by Mark Hulbert at http://www.marketwatch.com/story/earnin ... erspective
According to data compiled by Yale Economics Prof. Robert Shiller, earnings per share on the S&P 500 grew at a 3.8% annualized rate between 1874 and 2004. In inflation-adjusted terms, the growth rate was 1.7% annualized.
...When I mention these numbers in seminars I give, I receive no end of protestations. Surely these numbers are too low, I'm told. How else could stocks have provided an 11% annualized nominal total return over the long term, and nearly a 7% annualized real return? The answers to their protests are instructive. There are two reasons why stock portfolios have grown faster than earnings, and neither provides much encouragement today.
The first is dividends. The S&P 500's average dividend yield since 1874, according to Prof. Shiller's data, is about 4.7%. In other words, nearly half of stocks' long-term total return has come from dividends. [Note. As of 9/14/2013, the dividend yield of the S&P 500 is 1.97%. Source: http://www.multpl.com/s-p-500-dividend-yield/
...The other factor besides dividends that accounts for why the stock market has risen faster than earnings is an expansion in the price/earnings ratio. This is hardly an earth-shattering insight, of course. It just means that if investors are willing to pay more for a dollar of earnings, the market can go up even when earnings do not. Since 1874, the S&P 500's P/E ratio has nearly doubled. Can we expect it to double again from today's levels? Anything is possible, of course. But that strikes me as highly unlikely.
One question the above article does not address is why or how did the S&P500 EPS grow annualized 1.7% points greater than inflation from 1874-2004 even after paying out average dividends per year of 4.7%? I think that part of the answer is an annualized increase in productivity over that period that is much higher than inflation as well as an annualized increase in the median or average wages that is significantly less than the annualized increase in productivity. But I haven't found productivity or wage figures going back nearly as far as 1874. However, I did find this at http://www.huffingtonpost.com/2011/03/1 ... 37814.html
Between 1979 and 2009, EPI [the Economic Policy Institute] says, U.S. productivity increased by 80 percent [non-annualized], while the hourly wage of the median worker has only gone up by 10.1 percent [non-annualized].
I feel very confused here:
- dividend yield is DPS/ price per share. Earnings or profits are what dividends are paid out of: the payout ratio measures the fraction of EPS paid out as DPS (DPS/ EPS)
- dividends are paid out of earnings per share. In the long run, DPS and EPS growth have to be equal. If DPS growth exceeds EPS growth (payout ratio = EPS/ DPS so falls to be below 1.0) then the market has to cut dividends, it is paying out finite retained earnings
(converse case, EPS growth exceeds DPS growth, then payout ratio falls constantly. That's theoretically possible, see Japan in 1990 (1% yield) or US in 2000 (2% yield). History suggests it does not last forever, although share buybacks and hostile takeovers (which cash out shareholders) provide possible alternative routes).
- productivity growth does not guarantee corporate earnings growth.
In a purely competitive industry, productivity growth does not convey higher productivity, long run, because it is duplicated by new entrants. In fact it is slow changing industries like tobacco that have produced the best shareholder returns.
-when we speak of PE doubling, we are also speaking of dividend yield halving (assuming constant payout ratio)
The formula Shiller is using is
return = current dividend yield + dividend growth + speculative return (fall in DY)
If we assume DPS growth = EPS growth long run, the middle term becomes earnings growth *per share*-- that's important in the age of the stock buyback.
So the reason stocks have returned as they have since the 19th century:
- yield started high-- probably 6-7%
- dividend aka earnings growth was good, a function of rising GDP (I think it's actually GDP per capita, but I cannot make a sophisticated argument for that, but my sense is corporates do well if individuals do well, in the long run, ie there is genuine progress)
- yield fell, in that the PE (the earnings yield, E/P, at a constant payout is just the dividend yield) of the market doubled over the period (at least)