Predicting Recessions

The interest rate on 10-year Treasury bonds is now lower than that of 2-year Treasury bonds. This is the big one folks. The yield spread that’s supposed to matter has finally inverted.

If you’ve been keeping up with the financial news media or followed along on FinTwit, you already know what’s coming. The R word. Recessions almost always follow yield curve inversions. Of course by now, every analyst out there (including yours truly, circa June 2018) has studied previous inversions and has memorized some stat about an average of how many months and the percent of equity returns before recession hits. Still, I guess we’re nearing the end of this great bull market.

But wait! Maybe this time is different? Just last week, former Fed Chair Janet Yellen warned, “I would really urge on this occasion it may be a less good signal”. She’s not the only one downplaying the risk of negative term premiums. With a quick Google search, you can find similar quotes from Fed Presidents John WilliamsJames Bullard, Loretta Mester, Eric Rosengren, and Patrick Harker. Meanwhile, the Wall Street Journal’s Economic Forecasting Survey still favors continued expansion, though its implied 33.6% odds of a recession occurring over the next 12 months is at a multi-year high.

Here’s the rub though: people are terrible at predicting recessions.

The Business Cycle Dating Committee of the National Bureau of Economic Research (NBER) is tasked with identifying recessions in the United States. In their most recent announcement dated September 20, 2010, the Committee determined that the previous recession had ended in June 2009 – more than a year earlier. Similarly, they announced the December 2007 economic peak a full year late, on December 1, 2008. If you were investing in U.S. equities over the last couple recessions, this is how the press releases stack up.SPX and Business Cycle.PNG

To be fair, this group isn’t really in the prediction game. They only make a few announcements every decade, so they want to be sure they get things right. Surely economists in the business of making forecasts are a little more timely, right? Well, not by much.

The aforementioned Wall Street Journal Economic Forecasting Survey asks its participants each month for their estimation of the likelihood of a recession occurring over the next 12 months. The survey archive only covers the most recent recession, but the results don’t inspire much confidence. If the participants were omniscient, the responses would have all been at 100% in January of 2007. We certainly can’t expect that level of perfection, but by the time the recession began in December of that year, the survey suggested only a 38% chance of any downturn over the next 12 months. And though the odds quickly rose above 50% thereafter, by September 2008, the participants still placed a 40% weight on no recession at all for at least another year. Nine months after the recession began, these professionals still didn’t know it.WSJ Survey.png

But people aren’t only bad at recognizing when they’re already in recessions. They also have a tendency to see an imminent contraction when none exists. Countless reputations have been lost while trying to call the market top and predict economic meltdowns. Anecdotally, the last decade has widely been referred to as the most hated bull market of all time.

In either case, it’s not that the ones making predictions aren’t qualified to make such judgments. These are economists and analysts whose sole job is study such matters. They’re well-educated and extensively trained. The problem isn’t really that economists aren’t smart enough. The problem is that the economy is huge, there are tons of variables, and it’s downright hard, if not impossible, to accurately model its future.

For market participants, though, the most important takeaway should be to take the consensus opinions with a grain of salt. They aren’t accurate recession predictors, and even if they were, recessions aren’t perfect market-timing tools. Check out the first chart again. Stocks almost invariable peak before the recession starts, and by how long is difficult to guess. Calling the bottom is just as tricky. Perhaps we’re better off using economic data to more simply identify the risks and then focusing on whether the trend in prices confirms that assessment.

What do you think?

Nothing in this post or on this site is intended as a recommendation or an offer to buy or sell securities. Posts are meant for informational and entertainment purposes only. I or my affiliates may hold positions in securities mentioned in posts. Please see my Disclosure page for more information.

Comments are closed.

Up ↑