So the defense of GMO forecasts is that they were spectacularly accurate in 1998 and for a small window after the greatest financial crisis since the Great Depression. Wonderful. Now can we discuss the other 20 years?

No, the defence is very much simpler. When the market is unusual high they predict low forward returns, and then you always then see low forward returns, though the time frame for that to happen is a bit uncertain. When they predict high returns you always see high returns, though the time for that to happen is a bit uncertain.

It's not rocket science. Valuations do always mean revert. You don't know precisely what level they will mean revert to, nor precisely when. But you know with absolute certainty that if you pay a cyclically high price for a given stream of future earnings you'll get a much lower forward return than if you pay less for that exact same stream of future earnings. Sometimes prices are high, sometimes they're low, and they right there for all to see.

The 20 year average smoothed earnings yield is 60% higher than today's level, and the ten year average is 30% higher. We might see even more multiple expansion from these levels, but it's not sensible to expect it and it would be truly stupid to count on it.

Smoothed S%P 500 earnings yields are currently around 2.6%. (different smoothings will give different figures, but it's around that). It was about 16% in 1982, for a sense of scale. Your dividends come out of that, and neither the earnings nor the dividends are going to grow faster than real GDP give or take. Even if valuation levels never mean revert downwards at all, how much are you going to make starting from that situation? If not the earnings or the dividends or multiple expansion, where does the money come from?

Not from a deus ex machina. As a reminder, there is no known case in the history of humanity that ANY technological revolution (or anything else) has caused the trend productivity growth of the already most productive ("frontier") country to rise at more than about 2.8%/year. Figures in the 1-2% range are considered pretty good. If the "AI" revolution managed to get that up to 3% for a while it would be truly noteworthy, but I don't expect it to get to that level.

So, no, it's not different this time, at least not much. The most *rational* bullish case for US equities is the Kalecki-Levy corporate profit decomposition equation, not "it's different this time" because of [insert name of technology]. There is essentially no prospect of falling US deficits in the next decade, and the most noteworthy flip side of massive public sector deficit is massive corporate sector profits, which is the primary reason profits have soared to previously preposterous levels as a fraction of US GDP in the last 10-20 years. This macro outcome can have some side effects, though, including on the currency in question and sometimes debt crises.

Jim