Howard Marks on the uppside of bubbles [2026]

요약 보기

Marks는 버블의 상방 효과로 기술 발전 가속을 설명하며, AI가 버블인지는 AI의 미래를 알 수 없어 단정할 수 없다고 말한다. S&P500은 상위 7개 종목이 약 40%를 차지해 대표성이 낮다고 지적한다. 매그니피센트 7의 PE 30대는 비싸지만 과도하지는 않다고 보며, 나머지 493개 종목의 PE 약 20(역사적 평균 16)은 높다고 평가한다.

다듬은 글

※ 자동 자막 원문(다듬기 전)

01. I really liked one concept which which was novel to me and it's the concept of the upside of bubbles. Could you expand on that so our audience understands that a bit better? Well, I am not a technologist. And so to write a memo about whether AI could be a bubble, yes, I had to learn a lot. And uh my son Andrew who uh is a technologist and and and you know a forward thinker uh gave me a bunch of stuff to read.

02. So I I I read the books and I cite them in the memo. By the way, we're talking about memos. Uh David uh I I hope that your audience knows that they anybody who wants to can get the memos at oakreecap.com. Sign on click on insights. click on Howard Mark's memos and they can get all the memos they want and they're all free.

03. So uh some of the uh readings that Andrew gave me uh were from books about bubbles and they're cited in the memo so you can find them if you're interested. And what they basically said is that there is a everybody's worried about bubbles, but there's actually an upside to bubbles. And the upside is that the excitement that surrounds this technological innovation is so strong that money is committed and the the thing whatever it is gets built out very fast.

04. And and I I I found this very interesting and I I think it's correct. And what I say in the memo is that, you know, when you have these bubbles and you had a a bubble in railroads in the 1880s and uh radio in the 1920s and uh internet in uh 2000 or optical fiber in 98 etc. When you have these bubbles, if people remained objective and prudent and skeptical and demanding and risk averse, it would take a long time for these things to get built out.

05. But because of because of the enthusiasm or excitement that surrounds the bubble, they get built out much faster than they otherwise would have. So that's the upside of bubbles that they accelerate technological progress. But of course remember these books are written by technologists not investors.

06. So the technologist thinks about this and says oh bubbles are great because they accelerate progress. The investor says not so great. I could lose money. the the the technologist is not concerned about the fact that people could lose money because of the beneficial uh effect on society. But anyway, I do think that in some way that that is a beneficial effect of bubbles that they get stuff done fast.

07. A bubble is a burst of enthusiasm. Uh in 1996, Alan Greenspan, Fed chairman, uh coined the phrase irrational exuberance. And I would So what I say in the memo is that a bubble is irrational exuberance. And you look at AI today, you know, there's exuberance. The question is whether it's irrational. But I can't tell you that because I don't know how AI is going to end up.

08. AI conceivably, it's not impossible that AI will justify today's enthusiasm. So that's why I can't tell you whether it's a bubble or not. Great. Yeah. And speaking of bubbles, you also wrote about bubbles in January this year on your memo on bubble watch. And you I mean you made a point that the forward returns for the S&P based on probabilities and historical data are likely to be lower

09. than the the short-term historical run. So if you could rewe the S&P however you wanted, how would that change that outlook in your opinion or how would you approach that? I think that to be a representative index of 500 stocks, it's not good that seven of them are four 40% of the index. So, you probably would not weight the stocks based on their market value, which is what it is today,

10. right? Maybe you'd weight it on earnings or something like that. Or maybe you'd do equally weighted, but but if it it's not much of a 500 stock index is seven if seven of them are 40%. Right? 1.4% of the stocks are 40% of the index. That's not very representative. But other than that, I mean the interesting thing which I mentioned in another memo earlier this year is that the seven stocks, the so-called magnificent seven are overweighted but not insanely expensive.

11. I mean they have PE ratios which is the main metric for expensiveness valuation. They have PE ratios in the 30s and that's high but not crazy in my opinion. These are great companies and I would if I had to opine which I don't have to do because it's not my business but if I had to opine I would say that that the Nifty50 are not overpriced at 30 something times earnings.

12. When I came into the business as a young man at age 29 when I got out of Chicago, there was a bubble in what were called the nifty50, the 50 best and fastest growing companies in America. And they were selling at PE ratios between 60 and 90. So, so 30 doesn't seem very expensive to me. You see where I'm going?

13. and right and uh but what I think is high is the PE ratio on the other 493 companies which averages around 20 or 19. The historic average for the S&P is 16 and that includes the highest valuation highly represented overweighted stocks. So obviously the not highest valued, highest represented stocks, PE ratios might have averaged 14 or 15 and today they're close to 20.

14. So that seems very high to me. Anyway, that's kind of a sidetrack, but that's that's the way my mind works.