10:1 is low.

Long-run stable P/Es could rise if the internal rate of return of the economy as a whole fell, so that a good safe investment was one that paid 2% instead of 4% after inflation. That could happen under circumstances like these:

- If the peak-oil doomers turned out to be right, and our economic growth actually does turn out to be contingent on continually increasing fossil-fuel consumption; or

- If much of what we think of today as "profit" was actually destructive extraction of natural resources (e.g. overfishing); or

- If some kind of sustained disaster makes profitability difficult (e.g. the aftermath of global thermonuclear war, widespread coastal flooding destroying coastal cities, widespread Farmville addiction, or the gradual collapse of the Westphalian state system in the face of decentralized guerrilla warfare); or

- If we shift to a less efficient way of allocating productive resources than transparent capital markets, to an even greater extent than currently (e.g. war and other forms of theft, taxation for the benefit of wealthy bankers, insider trading, central government planning for the benefit of the politically well-connected).

I consider these scenarios unlikely.

If, by contrast, we keep inventing and putting into practice ways to produce more and more value for less and less effort and natural resources, and knowhow becomes more easily accessible rather than less, then we can expect that the internal rate of return that stocks must compete with in order to get investment dollars will go up. Which means long-run P/E ratios will go down.

To make this concrete, suppose that in 2029, you have US$20 000 to invest. (I'm speaking in 2011 dollars here to avoid talking about inflation.)

In 2029, Apple has settled down to a share price of $100 with annual earnings of about $10 per share (a P/E of 10:1), and no particular expectation that that is more likely to go either up or down in the next few years. So you could buy 200 shares of Apple and get about $2000 a year out of it, with some risk that Intellectual Ventures Hummer Winblad will get greedy and sue Apple into bankruptcy two years from now.

Alternatively, you can buy solar panels and sell the power back to the grid at the going wholesale rate of $0.015/kWh. In 2029, silicon solar panels have finally been edged out of the market by quantum-dot solar panels, which have an energy payback time of 3 months in a sunny climate. Like silicon solar panels, they're made out of some of the most abundant materials on the planet, and their fabrication is fully automatic, so essentially all of their cost is profit, the cost of the risk capital invested in their manufacture, and the energy dissipated in their manufacture. The energy dissipated is $0.033 per average watt, $0.011 per peak watt, but because of the large investments involved and the rapid expansion of solar panel manufacturing, that's only 10% of the actual purchase price of $0.11 per peak watt.

So instead of buying the Apple stock, you can buy 180 peak kilowatts of solar panels, which will generate 60 kilowatts, averaged over day and night, winter and summer. Instead of earning you $2000 per year, this will earn you $7900 per year, and your only risks are that energy prices fall further or someone steals your solar panels.

Since your objective in this investment is to make money, you buy the solar panels, as does everybody else. People sell their Apple shares in order to carpet the Gobi with solar panels. Consequently Apple's share price falls. Eventually it reaches US$25 per share, at which point its P/E is 2.5:1, and it's competitive with the solar panels again.

As long as there are investments available with rates of return similar to those I've postulated for solar panels above, shares will tend toward that 2.5:1 P/E ratio. They aren't doing it now because there are only very limited investments available with such high rates of return: installing a more efficient furnace in your house, maybe, but how many houses do you have? Solar panels, though, and thorium extraction from seawater, and automating custom manufacturing --- those are scalable investments.

I like your general logic - but I disagree strongly with your prediction that buying solar panels will generate 40% return on investment (in year 1!) in 2029.

I don't; the energy price I used is about a third to a quarter of today's wholesale electrical price, and the solar-panel price is about a tenth of today's. I assume that those prices will drop rapidly. I think they will probably drop a lot further than that, but it's very difficult to imagine what will happen when some resource drops in cost by more than a factor of ten.

> I like your general logic - but I disagree strongly with your prediction that buying solar panels will generate 40% return on investment (in year 1!) in 2029.

I'd like to disclaim that prediction! It was a scenario, not a forecast.

It doesn't have to be solar panels specifically, but my point is that over time, we may develop capital goods whose internal rate of return is well over the 3% we've become accustomed to. (Solar panels are a plausible candidate because their production is already highly automated, they're made of dirt-cheap raw materials, and they produce energy.) If that happens, whether it's solar panels or automated moon factories, P/E ratios will drop --- at least until the new exponential takeoff hits some resource limit. In the case of solar panels, that will probably be land, until we construct a Dyson sphere.