Hacker News Reader

Yea, he built an amazing track record well before he started to buying businesses outright (and in dollar volume he's spent far more on public companies than private). His preference for buying companies instead of shares only came around later in his career, and solely due to the massive size of his portfolio that his success created.

He didn't get into insurance until about 20 years after he started investing. His greatest returns were before it.

Buffett/Graham believe in concentrated portfolio, not having too many positions in your portfolio, because if you have more than 20 or so positions its difficult to know any of them well. So this means when his portfolio grew in size, he started losing opportunities in the public markets because many companies were too small for him to buy enough of to have a reasonable position. So he ended up investing in larger cap companies, and buying smaller cap companies outright.

And his biggest flub wasn't textiles or airlines, it was maintaining Berkshire Hathaway as his investment vehicle. He gives a free ride (no fees) to investors on half his results, and because of the double portfolio size restricting his investment options, his returns are lower. He would have had a substantially higher returns if he just invested his own money, or if he had kept running an investment partnership and taking fees. He's cost himself at least $100B in personal wealth.

But despite his very infrequent mistakes, he's beat the market by well over 10% a year during his investment career, and there is almost no luck in that.

> Yea, he built an amazing track record well before he started to buying businesses outright

He himself has said that the easy days are long over. The actual stock market continues to form a closer and closer approximation of weak EMH.

> But despite his very infrequent mistakes, he's beat the market by well over 10% a year during his investment career, and there is almost no luck in that.

He beat the stock market, by and large, by not participating in it.

Owning insurance companies and buying companies outright is not a strategy I can pursue with my 401(k).

I like Warren Buffett. I've read the first 50 Berkshire Hathaway letters. I think both he and EMH proponents are "correct", but largely they talk past each other. Where they agree is that what works for Warren Buffett only really works for Warren Buffett. His literal advice to regular investors is to buy index funds.

I somewhat agree with you, but I don't think the market is necessarily converging to weak EMH. I think it's more fair to say that as the market changes over time, the skills required to reliably beat the market fundamentally change. It used to be that you could spot market inefficiencies with an MBA and a solid understanding of specific domains. Nowadays, you model the market as a physical system and look for obscure phenomenon.

On the one hand that sounds like what you're saying - the market is becoming more efficient. On the other hand, I propose that there is no evidence the market will become more efficient continually, and the new normal is due to the rise of software eating the industry.

I agree with you and disagree with other posters here.

It's physically impossible for the market to be efficient.*

Just try to use the efficiency for anything (set up any series of steps where one of the steps is "...because the market is efficient, therefore...") and whatever you've just set up just plain won't work.

Since it's impossible, instead of any version of the efficient market hypothesis people should just say "God's will" (as in God's will that prices should reflect available information), and it will show how silly they are.

* you can email me if you'd like a rigorous proof that it is impossible

It depends on what you mean by "efficient".

The weak EMH is pretty simple: there's money to be made in arbitrage, but the act of arbitrage spreads information. So a price quickly converges to reflect available information. The harder participants try to beat the market, the harder it becomes.

What's left is the lurch of prices upon truly novel information arriving. Stocks appear to move up and down at random, because future events are not fully predictable. If they were, prices would stabilise around the "true" price and remain stuck there.

A similar argument has been used to show that time travel has not and will never be invented. Arbitrage across time would be even more profitable than arbitrage across space.

> On the other hand, I propose that there is no evidence the market will become more efficient continually, and the new normal is due to the rise of software eating the industry.

Could you elaborate? I'm not sure I follow.

Sorry, I guess I wasn't clear. What I meant was that while finding alpha has indeed gotten harder over the past several decades, there are different candidate hypotheses to explore what will happen in the future. I interpreted your comment as a hypothesis that the market will continue to become more efficient as it essentially achieves weak EMH. However, I believe that you could consider the hyper-efficiency of recent times to be a natural byproduct of software and automation becoming ubiquitous.

Information gets cheaper, faster and more accessible over time but I believe that there's a diminishing return in how much more efficient the market can get without another huge jump in technology.

For example, HFT (specifically, the sort where you're actually thinking about the speed of light) is not categorically improving dramatically year after year anymore. The market is getting more efficient, sure, but it's getting more efficient very slowly, and would argue the rate of efficiency increases is decreasing. Without another categorical jump in technology (e.g. significant AI, quantum computing, etc) that dramatically improves data cleanliness or accessibility, I don't see the market getting significantly more efficient than it is now. And I don't agree that the market is yet at weak EMH.

It's not just Buffett though: https://www8.gsb.columbia.edu/articles/columbia-business/sup...

The argument here suggests that there may be choices to make that guarantee LOSS, not success. In that case, learning how to avoid the 'automatic loses' is important and the more people that learn it, the closer things revert to the power-law distribution.

I see no reason to conclude the hyper-winners are more than just luck. Perhaps all the '1% edge' associated with talent is simply avoiding bad mistakes that always deliver a loss. If everyone performs optimally, it becomes entirely and completely luck.

> ... If everyone performs optimally...

I guess it's a lucky thing that'll never happen? I seriously think the only skill involved in investing is a combination of humility and an admission that you're not smarter than everyone else. It keeps you medium conservative and provides solid returns.

It's impossible to empirically examine whether or not humility and reverse-Dunning-Krueger is important for trading acumen. On the other hand, it's demonstrable that domain knowledge and business acumen, or mathematically modeled market signals and statistical insight can result in reliably superlative returns.

Yeah, Buffet is actually not the best case to back up this argument with. RenTech/Jim Simons [0] is probably a better example of making money on pure skill.

[0] https://en.wikipedia.org/wiki/Renaissance_Technologies

Also:

"We find that the secret to Buffett’s success is his preference for cheap, safe, high-quality stocks combined with his consistent use of leverage to magnify returns while surviving the inevitable large absolute and relative drawdowns this entails. Indeed, we find that stocks with the characteristics favored by Buffett have done well in general, that Buffett applies about 1.6-to-1 leverage financed partly using insurance float with a low financing rate, and that leveraging safe stocks can largely explain Buffett’s performance."

https://www.aqr.com/library/working-papers/buffetts-alpha

This paper only starts in 1976, so it skips the first 20 years of Buffett's career when he used virtually no leverage and had his highest returns.

It also has problems with the classifications of his investments. It's a common problem for academic analysis of investing, if value investing works then I should be able to outperform solely buy purchasing stocks at low PE ratios, right? This doesn't work, so value investing "can't" work.

The truth is it's a skill. You have to be able to find businesses with long term competitive advantages selling at a substantial discount to their intrinsic value. You have to have the courage of your conviction through market cycles and not abandon your purchases even cheaper. You have to eliminate sources of bias in your analysis (Buffett won't look at it's price before he analyses a business so his valuation isn't affected). If you read Buffett there are a ton of things he does to limit bias in his job.

None of this can be academically measured.

Clif Asness is a successful investor. I wouldn't dismiss his findings so easily.

> this paper only starts in 1976

Explaining Buffett's performance since 1976 is an informative exercise.

It is, but it's similar to covering a start-up after their initial hockey stick growth. How to maintain growth is interesting, but definitely not the whole story.

Isn't in general, before the Internet , making money in business was much easier, hence qualititavely different, and maybe skill played a bigger role then ?

This is well covered by Charles Ellis: The stock market has changed dramatically since the 70s, when most money was invested with almost no information. Getting key information well ahead of the market was relatively easy. A lot of people made a lot of money then, until the amount of 'smart' money rapidly increased.

It's still possible to get better returns than average on the market: After all, most of congress does this. But the one way to do it is precisely like congress does, by having information that the market is unaware of.

This is what make people move money into venture capital: It's a way to invest into things most of the market is unaware of, and whee the number of players is small enough that it's still possible to get privileged information.

Buffet's current plan is only doable because he has access to more information, and has such a gigantic bundle of money that any effort to get an edge will be multiplies by investment size.

Where does the wealth of the richest people in the world come from? Tremendously risky bets, going heavily into a single, extremely successful venture that they had special access to. But what makes their ventures win, vs those of second players that did badly in the same industry? Non replicable things, distributed in a way you could call luck.

> He prefers to invest when it is actually possible to possess information or insights that have not been widely distributed to other potential buyers via public disclosure.

This is exactly the opposite of what he claims in his essays. One quote:

> The other way we differ from some arbitrage operations is that we participate only in transactions that have been publicly announced. We do not trade on rumors or try to guess takeover candidates. We just read the newspapers, think about a few of the big propositions, and go by our own sense of probabilities.

There are several others in "The Essays of Warren Buffett" where he also describes how he just uses public information.

Again, there is a distinction between buying private businesses outright and buying in the open market.

In the former case, Buffett solicits sellers to directly contact him. He makes this solicitation every year in his letter to Berkshire Hathaway shareholders.

In the latter case, Buffett recommends the Financial Times as a source of information.

Yep. If the market you're competing in is too efficient, find another market.