Stanley Druckenmiller is not a household name. Even if I told you he was the guy that came up with the idea to short the British pound in 1992, you’d likely be scratching your head. But if I added that his partner George Soros encouraged him to invest in that idea so heavily that it literally broke the Bank of England, you’d likely recognize the famous story. Druckenmiller and Soros earned over $1billion on that trade. Druckenmiller has one of the best track records in the business, compounding by 30% per year over his career. How did he do it?
In an interview (YouTube) with the University of Southern California, Druckenmiller explained his investment approach. He says “too many investors think in the present. The present is already in the price. You have to think out of the box and visualize 18 to 24 months from now what the world is going to be and what securities might trade at.” Is one of the world’s best investors really suggesting we turn into fortune tellers? I’ve written many times about why forecasting is one of the cardinal sins in investing, yet here’s a fellow who’s shown it’s not impossible. How did he do it?
Thinking about the future is tough stuff -- almost impenetrable. In their book Superforecasters, Philip Tetlock and Dan Gardner illustrate the techniques of the world’s best forecasters. By no means am I suggesting that we should all try to be Superforecasters (or put our life savings into it!), but it is instructive in helping form an opinion about the world. It’s for people that believe making an educated guess is better than making a random guess. The best news is that you don’t need a rocket science IQ to think about the future. Just a bit of curiosity and willingness to think in terms of probabilities, not certainties.
Generally all of the Superforecasters follow the same approach: break down the big hard question into smaller easier data-driven questions. By piecing together answers from the smaller easier questions, we can obtain a better informed estimate to help us answer the big question.
For example, let’s start with the big question: is the US stock market going to be higher or lower one year from now? There are as many answers to this one as there are investors, making it way too tough to consider this question on its own. So what are some smaller questions where we can find an answer?
Let’s start with this one: How many years has the US market increased? We know that on average the US market goes up about 65% of the time on an annual basis. So we can use 65% as our baseline probability that the market will rise. That’s way better than a random guess. Now to refine our 65% up or down, let’s think about some additional smaller questions:
- What is the current sentiment of investors? We can find survey data going back decades that tends to show investor sentiment is a contrarian signal. So if sentiment is low, we should boost our 65% estimate higher. But if sentiment is high, then we should reduce from the 65% baseline.
- What is the shape of the yield curve (for non-geek-speakers, the yield curve compares short term to long term interest rates)? Decades of data suggests a flatter yield curve tends to be worse for the stock market than a steeper yield curve. Same as above, use this info to boost or reduce from our baseline.
- How much cash are professional money managers holding as a percentage of their portfolios? We can find evidence that shows professional money managers tend to increase cash before positive stock market performance, a contrarian signal.
- Is the stock market already in an uptrend? Data shows that on an annual basis a rising market tends to keep rising.
The list is by no means exhaustive, but is a good place to start assembling a superforecast. You’ll notice also these are all questions that can be backed up by real data. Since the answers to these questions are knowable and have a testable track record, we can combine their individual probabilities to build a better estimate of our big question. Stay away from small questions where we don’t know an answer, or the answer is not measurable or verifiable.
So is the market going to be higher or lower a year from now? We’ll never know the answer for certain, but we might be able to get reasonably close.