Photo by Andrew Neel on Unsplash
Hello,
Welcome to Known Unknowns—a newsletter that is bullish-ish about the future.
Stocks are expensive… maybe they should be
Are stock prices too high? I don’t know, and neither do you. I am too much of an Efficient Markets adherent to believe otherwise. Though, I am also Schumpeterian. So, yeah, we just may be in a new AI-powered economy where we will be much more productive, and the firms best poised to make money off this are worth a lot—and will be worth much more than what they are now. But until we figure many things out, and that could take decades (possibly many decades), there will be some froth along the way. We will make big bets on the wrong horses. Or some companies that seem poised for greatness will falter after a while. Or companies that seem unstoppable now may not exist then. Who the winners will be and what winning even looks like is unknowable right now.
Yes, if you own lots of NVIDIA, odds are you will lose money at some point, but you might come out ahead in the end—maybe. There is no doubt that stocks are expensive. But whether they are over-valued is never possible to know. It is even harder to know that when our old models are still not great at measuring the value of intangible capital. And they are even less able to account for the value of a new and unpredictable technology.
I think the Google Gemini case is revealing. First, it showed that AI may get smarter than we are, but what it thinks is based on the information we give it. And our information, or our desire to override that information, can be stupid or inaccurate. It is easy to believe what you want instead of what is. It turns out AI, which is supposed to be rational and better that we are, falls for the same stuff as long as humans are involved. At the very least, Gemini showed that AI technology has a lot further to go before we hand it the keys. Second, the whole incident shows even big companies that are leading the way might not be up to the task and will eventually stop being so important. I am not making a call on Alphabet—it might turn things around, but this mess-up seems pretty fundamental.
By the way, it is worth remembering that extreme stock market concentration is nothing new. In 1950, three companies made up 28% of the S&P 500 (the Magnificent Seven make up 29% today), and none of the top 10 companies from then are in the top 10 now.
I am bearish on AI ethicists, who were treated like oracles and seemed obsessed with keeping the technology from reflecting our terrible biases. They made some good points that are certainly worth keeping in mind, but now, it seems their comfort with bypassing accuracy for the greater good may be a bit of a problem. This also shows why, as rational as AI might be, its path will be bumpy because humans are involved.
More generally, when forming my outlook, I think it is important to remember there is a natural response to uncertainty. We want to turn it into measurable risk and remove the downside. And that is a good approach, but we often forget uncertainty is still uncertain (and that’s true about AI itself or if we use AI to take on another uncertain problem). Or, we get fixated on the wrong downside risk. And it seems with AI, prophets are obsessed with the downside they can anticipate and are willing to sacrifice the upside. Not only that, but they believe they can steer this potentially humanity-altering technology as they want. It doesn’t work that way.
Gemini is a good example of this: Google believed they could reduce the risk of bias by forgoing accuracy. This is only one example of the zeal to regulate (this time from within the company) this technology we don’t yet understand, which ultimately created new risks.
And we are only getting started; I foresee similar issues going forward. This instinct is stronger in Europe, which is one reason why I am ultimately bullish on American tech stocks. But even a Magnificent Seven stock fell for this (and odds are all of them will at some point). So going forward, here are two pieces of advice:
1. You can’t predict uncertainty.
2. There is no reward without risk—and even big, fast-growing stocks are risky.
6% does seem awfully high
Now that I am a homeowner, I already dread having to pay a broker 6% (!) of the sales price when I sell (if I do—with my mortgage rate, I am not sure I can ever afford to move). But maybe I won’t pay that much.
When I bought, the seller paid such a high fee because she covered her broker and mine. And because she paid my broker, I was probably a more needy client. I bought the second place I saw, but I needed to see about 50 other places first. Then the seller backed out—twice—drama my broker managed, and he kept the deal alive. Then he compiled a 2000-page (not kidding) co-op application and strategized how I’d present myself: “I may have unstable income, but I will never ever do a renovation,” I solemnly told my co-op board, as we practiced.
He earned his fee and then some. But if I had to pay him directly or pay for his time or per service, maybe things would be different. Manhattan real estate is not for the faint of heart, and buyers with variable incomes need extra help. I could not have done it all on my own.
But it did always strike me as odd that even as technology should have made realtors less necessary (they no longer have a monopoly on listings), we got more of them. But as it turns out, it was because of a market distortion that ensured sellers pay for both the buyers’ and sellers’ broker fees. In other countries, this isn’t the case, and broker services are unbundled and sometimes not used at all. This will probably be the future in America, too, now that a federal jury has found that this whole state of affairs is indeed the result of collusion.
We’ll still have brokers, and people like me will still need them. But there will be fewer brokers, and they will need to provide great and highly skilled services, sort of like travel agents today but for a very big purchase.
News we already knew
Crypto is a high beta asset class
Until next time, Pension Geeks!
Allison
The commission to trade stocks 50 years ago was about 1% per transaction. If you ask a Millennial trading stocks what a commission is, you would just get a blank stare because they wouldn't know what you are talking about. Mutual funds typically charged 1% or more in expense ratios with front-end or back-end loads of 3% or more. Charles Schwab and Vanguard upended those business models in the 1970s and 1980s and the Internet brought costs almost to zero.
The brokerage commission to sell a house was 6% 50 years ago. Today it is... 6%. 50 years ago, the brokers had to do a lot of work because everything was on paper, even MLS, and they had to physically research things. Homebuyers would have to look in the newspaper real estate listings to see what was for sale. Computerized MLS came along, so their research could now be done on a computer in 1975. That was and is a closed system, so you had to pay MLS fees as a broker to access the information.
The 6% commission is because the access to a single computer database contractually requires a standardized commission structure, which is where anti-trust comes in. We are probably going to watch the equivalent of the Ma Bell breakup and reformation of new ecosystems for buying and selling houses over the next decade. It is one of the last few "old way" pre-Internet consumer-facing business structures still standing, but it is falling.
Great. I’m hopeful that AI will do much more than help us make bad decisions faster.