Regarding the chart being out of bubble usually is more costly, doesnt it only apply to investing in the index? Investors didn’t get their money back in Cisco after 25 years if invested during the height of the previous bubble. For active investors aiming for above average returns, their holdings will not represent the market. Therefore, I believe this doesn’t apply to them.
I think market valuations (such as price-to-book ratio, EV/EBIT, etc.) are quite indicative of the next 5-10 years. High valuations mean lower returns in the future and vice versa.
The high valuations in the US mean, for example, that it may better in the long term to invest in cheaper global stock markets. We saw this in 2025, when many global markets in USD had significantly better returns than the US stock market.
Quick nuance: I’m not talking about sell-side analysts. I’m talking about professional investors with a real track record, people whose edge is often to do exactly what analysts can’t do (or won’t do).
On analysts, I fully agree with you. The way most of them are incentivized (access, relationship management, career risk, etc.) produces “consensus comfort,” not alpha. That’s how you end up with John Dorfman’s numbers imho.
Very thought provoking charts. The one that stood out for me is the subsequent returns one. I must say I am quite skeptic at first glance, I wonder how it's possible that buying at all-time highs has decent returns short-term, let alone BETTER than any other point. Would be interesting to know exactly how Schroders came to this graph. Great insights nonetheless!
On the short-term side, I read another study (unfortunately I can’t remember the source) that basically said:
- A randomly chosen point in the ATH bucket has a relatively low chance of being followed by a drawdown over the next several weeks/months.
- A randomly chosen point outside the ATH bucket is more likely to be followed by another drawdown over the next several weeks/months.
Taken together, that makes it pretty intuitive to expect ATH entries to outperform over a 1-year horizon, basically the idea that trends tend to persist.
Over longer horizons, though, the trend-following argument feels less obvious, I agree with you there. Maybe they stopped at three years because beyond that, ATH entries start to underperform.
Either way, it’s intriguing. I’ll try to look at it over a longer timeframe when I get the chance.
The point here isn’t how long it took to get back to the all-time high. It’s when you could have invested and still remained in the green even at the bottom of the crash.
For Cisco, that cutoff is January 1998. The argument still holds. Even if we assume the bubble only began in 1995, an investor who stayed completely on the sidelines would have missed a 4.4x gain by the 2002 bottom.
Amazon is similar. And for Oracle, an investor would have still been up until 1999.
The argument is not that bubbles carry no risk. Many companies went bankrupt, and many investors lost everything.
The points is that this isn’t only true for the index: it can apply to certain individual companies as well, without claiming it generalizes to all of them.
Regarding the chart being out of bubble usually is more costly, doesnt it only apply to investing in the index? Investors didn’t get their money back in Cisco after 25 years if invested during the height of the previous bubble. For active investors aiming for above average returns, their holdings will not represent the market. Therefore, I believe this doesn’t apply to them.
I think market valuations (such as price-to-book ratio, EV/EBIT, etc.) are quite indicative of the next 5-10 years. High valuations mean lower returns in the future and vice versa.
The high valuations in the US mean, for example, that it may better in the long term to invest in cheaper global stock markets. We saw this in 2025, when many global markets in USD had significantly better returns than the US stock market.
Quick nuance: I’m not talking about sell-side analysts. I’m talking about professional investors with a real track record, people whose edge is often to do exactly what analysts can’t do (or won’t do).
On analysts, I fully agree with you. The way most of them are incentivized (access, relationship management, career risk, etc.) produces “consensus comfort,” not alpha. That’s how you end up with John Dorfman’s numbers imho.
Extremely interesting article!
Well written, highly diversified investors get better returns staying in the market, and holding on to the long strategy. Ignore the noise!
Very thought provoking charts. The one that stood out for me is the subsequent returns one. I must say I am quite skeptic at first glance, I wonder how it's possible that buying at all-time highs has decent returns short-term, let alone BETTER than any other point. Would be interesting to know exactly how Schroders came to this graph. Great insights nonetheless!
Thanks, Paul!
On the short-term side, I read another study (unfortunately I can’t remember the source) that basically said:
- A randomly chosen point in the ATH bucket has a relatively low chance of being followed by a drawdown over the next several weeks/months.
- A randomly chosen point outside the ATH bucket is more likely to be followed by another drawdown over the next several weeks/months.
Taken together, that makes it pretty intuitive to expect ATH entries to outperform over a 1-year horizon, basically the idea that trends tend to persist.
Over longer horizons, though, the trend-following argument feels less obvious, I agree with you there. Maybe they stopped at three years because beyond that, ATH entries start to underperform.
Either way, it’s intriguing. I’ll try to look at it over a longer timeframe when I get the chance.
I'm following although I'd need to look at it more deeply to understand well. Interesting topic though. Thanks for your answer!
The point here isn’t how long it took to get back to the all-time high. It’s when you could have invested and still remained in the green even at the bottom of the crash.
For Cisco, that cutoff is January 1998. The argument still holds. Even if we assume the bubble only began in 1995, an investor who stayed completely on the sidelines would have missed a 4.4x gain by the 2002 bottom.
Amazon is similar. And for Oracle, an investor would have still been up until 1999.
The argument is not that bubbles carry no risk. Many companies went bankrupt, and many investors lost everything.
The points is that this isn’t only true for the index: it can apply to certain individual companies as well, without claiming it generalizes to all of them.