There’s an ongoing debate about whether or not the U.S. is approaching a recession. As an investor, this question is of utmost importance. It is precisely at these times when fortunes can be made and lost. There’s no shortage of pundits with strong opinions in both the affirmative and negative camps armed with plausible narratives and supporting data sets. How to decide which side to take? Applying some proven forecasting methods to historical data can help bring clarity to this question.
Forecasting is tricky business. It’s really hard to do well consistently, especially in investing. Fortunately for us, Philip Tetlock has made a study of forecasting. In the book Superforecasting: The Art and Science of Prediction (aka Superforecasting) he and coauthor Dan Gardner share their findings of a multi-year study aimed to discover the best forecasters, uncover their methods, and to determine if forecasting skills could improve. There are many great lessons conveyed in the book. We can thus apply them to our problem at hand: the question of whether or not the U.S. will enter a recession.
By running a series of geopolitical forecasting tournaments, Tetlock and Gardner discovered a group of elite forecasters and uncovered their best practices. Ironically, specialized knowledge played little role in their success. Rather, it was their approach.
Superforecasting had a profound impact on me. I took away a more concretized framework for dealing with predictions. Rather than feeling my way through a situation, Superforecasting gave me a method I could apply. Well-devised forecasts share four characteristics:
- They’re probabilistic
- They start with a base rate formed by an “outside view”
- They’re adjusted using the specifics of the “inside view”
- They’re updated as frequently as required by incoming facts, no matter how small the increments