In fall 2011 we have again reached a trailing-ten-years multiple that exceeds 20. This is uncharacteristically high. Those who use current earnings or forward earnings will arrive at lower P/E ratios but from a historical perspective, those more optimistic P/E are also high.

I don’t care much for forward earnings because they’re always someone’s optimistic best guess. I don’t care for current quarterly earnings either since they are much too volatile to base longer-term decisions on.

P/E is roughly twice the expected annual growth rate. This varies a lot, but with an average dividend payout rate of 50% of earnings, I think the dividend stream is a much better proxy for the growth rate because it is stable. Take the inverse yield and double it. That’ll give you a “P/E” proxy which doesn’t fluctuate as much. I’m talking rough numbers here. Rough numbers.

Many stocks trade on a valuation model that’s roughly

Price = P/E * Earnings

So if P/E is 20 and earnings are $1, that makes for a stock worth $20.

If P/E is historically high and P/Es are historically bounded with no “new-era” forthcoming, it means that P/Es are likely to grow down. A more typical value for the low range is a P/E of 10. Historically it takes about two decades to go from high to low.

Thanks to central bank policies we have now spent a decade going nowhere. This means that the leverage must be halved in 10 years. Linearly, that would be going from 20 in 2011 to 19 in 2012 to 18 in 2013, etc. until we hit 10 in 2020.

This translates to a downwards pressure of 5-10% annual on stocks, a pressure that must be compensated for by increasing earnings. Now, 5-10% growth in the entire market is impossible, because the entire market is the GDP and the GDP grows very steadily at 3%.

This is the macro conclusion why stock prices will be at best flat for the next decade as increasing earnings make up for decreasing leverage.

This is also why I’m not betting on capital gains.

Implications: A portfolio that relies on selling a small fraction of shares annually to create an income will end up with fewer shares than it anticipates. In a market with declining multiples it is better to rely on dividends. Some of these could be used to buy stock at lower multiples but that would be at the discretion of the investor. A total return approach where all the earnings are retained for further earnings growth mostly makes sense in markets where multiples are rising along with earnings. Such a market is at least half a dozen years off.