The Fool's Game
Predicting unpredictable things isn't worth your time. Yet stock picking is about predicting. What do?
Before we start, I wanted to say thanks again for your interest, and for the support from many of you. I hope you enjoy the read as much as I enjoyed the write.
Rather begin by going into detail about our portfolio or an investment idea (rest assured that meatier stuff’s coming to your inbox Saturday), I wanted to go over one of the most fundamental features behind investing and decision-making: unpredictability. It matters. A lot.
Many business professionals trick ourselves into thinking we have a solid grasp on what’s going on with things like the economy and inflation. When prompted, we might opine on where we think the central bank is going to set interest rates in the next 12 months, say to spur growth early in the economic cycle, or to cool inflation later down the line.
Some of that’s social: I mean, what else are you supposed to say when asked, especially when you want to look useful, and also not look stupid? Some of this is also rooted in in the genuine belief we’re pretty sure we know what’s going to happen, even if it’s just because a lot of people — usually with information, credentials, and/or a confident tone of voice — told us so. The curse seems to be worse the smarter, more energetic, and more useful we think we are.
Is that confidence actually warranted?
Some time ago, Deutsche Bank Research published a chart1 that does a great job of illustrating the futility of trying to predict the unpredictable. The authors plotted short-term interest rates and interest rate forwards (we’ll get into what they are). The chart below is one replication of their work that’s being updated and shared publicly.
You don’t need to be an expert or know acronyms to understand what’s going on here. Just look: (1) there’s a blue line, and (2) there are some gray “hairy” dotted lines coming off the blue line. Dates across the bottom, interest rate percentages up the side.
The blue line: this is the short-term interest rate for borrowing or lending US dollars (such as commercial banks lending between themselves to facilitate the movement of money and credit). There’s a footnote2 for the intellectual masochists who want to know more, but all you really need to know is the blue line tells you where short-term USD interest rates have gone between 2001 and today.
The gray “hairs”: while blue shows where interest rates actually went, the gray hairs are the “forward curve”, showing where the market thought interest rates would go at the time. Again, the masochists among you can read about forward curves in this other footnote3.
Look at the first hair on the left side of the chart. The blue line tells us in June 2001, US short-term rates were almost 4%. Now follow the first hair that departs from the blue line. Sitting in June 2001, if we follow the first gray hair forward it tells us the market believed short rates would go up, reaching almost 6.5% by about December 2004 where it terminates. What actually happened? The blue line shows us that over the course of 2001 to mid-2004, the Fed cut rates and short-term interest rates fell to almost 1% instead. Many very smart people believed in 2001 that rates would rise a lot. Rates then fell a lot. You can follow the other hairs and notice there seems to be a pattern here: the gray hairs rarely predict where the blue line goes. A lot of smart people are doing a lot of predicting, yet it very much looks like wasted effort. Even if you wanted to try and get really sophisticated and academic and argue the market consensus — the gray hairs — represent the mean of some underlying probability distribution of expected interest rates, the blue line rarely tracks to that mean.
You can build the same chart for any other market with futures or forward contracts being traded between lots of smart people. For example, you can compare the spot price of oil over time to what oil futures contracts looked like before. You’ll find the same phenomenon: lots of smart people not getting it right.
Now, there are some reasons you’d take positions in these markets even if you can’t predict the future. You might be an oil producer who simply needs to fix some of their pricing to ensure stable cash flows to fund a project, and this market price seems fine enough to you, and enough of that will influence pricing. However, the chart still illustrates the peril of predicting something like the path of interest rates.
The takeaway also isn’t that no one can do it. There are probably a number of firms who have — and not just due to luck — managed to squeeze out a small edge betting on interest rate movements or consistently finding pockets of some rate markets where the odds are mispriced. The takeaway is: that person isn’t you. Or me. And that’s because it’s influenced by far too many variables and interactions between variables for us to understand, calculate, extrapolate, and consistently be right about.
Now we have to throw some groups under the bus. I am sorry.
Despite this, we may hear from the Chief Economist of ABC Corp or the Chief Market Strategist at a bank or the Markets commentator at a news source about interest rates and where they think they’ll go, and what the US Federal Reserve’s going to do. It all sounds interesting, and they’ll say it with conviction. This is “nice to know” stuff to get a sense of where the world’s at, but your interpretation should stop there. Clearly, this isn’t useful stuff if what you’re trying to do is make investment and planning decisions in a business, or build a portfolio to save and invest for the long term.
Here’s another easy way to think about it. If the Chief Market Strategist really could predict interest rates, why are they so happy telling you about it rather than running the world’s most successful interest rate derivatives hedge fund and laughing while they pull billions of free money out of the market from a laptop on the sun deck of their outrageously large yacht? And why aren’t there hundreds of them, all sunning themselves in outrageously large yachts drifting lazily off the coast of Italy? Might it be because it’s not so easy?
This isn’t to denigrate anyone. The point is this should put you at a crossroads if you’re a self-aware person responsible for making forward-looking decisions.
We can return to the comfort of pretending the future can be predicted accurately and consistently, even though deep down we know it’s futile. Or, we can get humbled pretty fast, then begin thinking about smarter ways to win in spite of the world not being a predictable place.
If it really is a fool’s game, what do? Can you at least try to position yourself to win regardless?
I think you can, or I’d have stopped being a professional investor ages ago. Let’s pull an example from my portfolio.
In 2018, I bought some General Motors stock and in 2020 it became my largest position when I bought more. (I look stupid today because I still own it despite where it went over 2021-2023, but that’s a topic for another time). We’re not going to go over the whole idea here and slap you with the Excel model, but we will point out how unpredictability applies and what we did about it.
GM assembles and sells autos. It’s a very cyclical business because cars are “big ticket” purchases, and are a “durable” good. In plain English, that means they cost a lot, GM only gets paid once up front when you buy, and you can keep re-using them without the need to buy another any time soon. In lean times, unlike toothpaste, you can easily choose to forego buying a new car. You can see how this is bad for the guy trying make a business out of selling new cars.
GM makes most of its money in the US. In a good year, American drivers buy ~17.5 million new vehicles. In recessionary times, they buy 12-16 million. When the world is ending like it was in 2009, they buy 9 or 10. When they do buy in bad times, they tend to buy smaller cars with fewer features, for obvious reasons. Those cars make less profit per car. All this means it’s not unusual for an automaker’s revenue to fall 50% and for margins to go from ~10% to zero (or to a loss). Worse yet, automakers have “negative working capital”, meaning the timing of making and selling a car is such that they get paid by the dealerships faster than they pay their suppliers and employees (plus some warranty and other costs). That’s great in good times because it’s a source of cash and means the business needs less capital to operate. In bad times, cash magically disappears into the ether at precisely the time you would prefer to have some extra cash, which you don’t need a business degree to know is not good.
How can you own this thing then if you can’t predict the economic cycle and yet this is an ultra-cyclical business that sets money on fire at exactly the wrong time? And how can you buy it in 2018, which was arguably the peak of the economic cycle? I mean, upon buying the stock, my mother chided, “Didn’t these guys go bankrupt? Why are you buying this for us?”
Well, two things matter.
First, the people running the company are not stupid. Mary Barra (and competitors’ CEOs) know this is a cyclical business. Since watching GM go bankrupt in 2009, the companies spent years structuring themselves in such a way that they’re now much more durable. They’re now loaded with liquidity and their cost structures are far more flexible than when GM went to zero.
Second, if you reverse-engineered what was implied at $37 a share, the stock market already thought that the business was going to go through a deep recession in the near-term. That meant you were probably going to do quite well if there was no recession, and if there instead was a really bad one coming, your capital at least wasn’t going to be impaired. At the same time, I thought the market was ignoring some other things about the business, so I got the exposure my thesis while at the same time I didn’t need to call the economic cycle because a bad recession was already priced in. A lot of possible futures could have played out, but we were buying in at the low end of that probability distribution because the market’s expectations were already so low.
That’s how we did it that time around. There are a ton of ways to get around unpredictability. However, you won’t figure out what to do until you first acknowledge that when something is driven by many forces you can’t predict, then trying to do just that is a fool’s game.
Chris
If you know of the original report, I’d love to have it and credit the analysts at DB.
LIBOR/SOFR is the benchmark overnight interest rate at which commercial banks and other financial markets participants will lend to each other overnight, in US dollars and collateralized by US Treasuries. They usually do this through repurchase agreements (“repos”; you sell your Treasuries to them in exchange for cash, under a repo wherein you promise to buy back the bonds for cash, usually tomorrow). Most floating rate loans are priced off this rate. For example, if a mid-sized company borrows from a bank or private credit lender, the loan’s interest rate is usually priced with SOFR being the reference rate. The rate then fluctuates with time. SOFR is directly influenced by the rate the Federal Reserve sets the Fed Funds rate at. So this is basically the short term interest rate for US dollars. If this is not enough and you want even more punishment, you can find the NY Federal Reserve’s overview of SOFR here: https://www.newyorkfed.org/markets/reference-rates/sofr
Sometimes you don’t want to buy or sell something today. You want it in the future. That’s where forward contracts or futures markets come in. The forward contract market for short-term interest rates is a market for participants who want to borrow or lend short-term cash at a defined time in the future, but not today. To enter into the contract, market participants need to decide on the interest rate they will borrow or lend at. Hence, the price set by short-term interest rate forward contracts is what the market thinks that interest rate will be at that time. You can graph out a “forward curve” for different times in the future — for the contracts that are priced for a week from now, the ones a month from now, the ones three months from now, the ones six months from now, and so on, and that’s what gets you those hairs on the chart.