> the data of a few decades. 1926-2020 is 95 years. Calling that "a few decades" is downplaying it. I'd say it is safe to take 95 years and from that extrapola

> the data of a few decades. 1926-2020 is 95 years. Calling that "a few decades" is downplaying it. I'd say it is safe to take 95 years and from that extrapolate a 30 year future with wide margins. In your analogy: looking at 95 people and saying "the average group size is between 2 and 5". > ... 90 data points and assumes the next 90 data points will be alike. Are we reading the same article? I don't see it predicting anything 90-points forward. Even the predictions for 30 points forward are very much on the safe side: merely a caution and establishment of a (well-known) generalism. A conclusion in that article: > don’t know if the next 30 years will be this kind to investors in U.S. stocks. You could make the case investors should reduce their expectations going forward. > But if those lower expectations turn out to be correct this makes thinking and acting for the long-term even more important than ever.