More digital ink has been spilled by traders, investors, bloggers, and charlatans, none of which are mutually exclusive, on what will cause the next recession. Some of these guys have been around long enough to have called 17 of the last 2 recessions. The reason so much time and effort is devoted to calling recessions is that they typically coincide with bear markets. If you can avoid a bear market while remaining invested during a bull market then you’ll eventually end up rich. The problem is that this is hard to do repeatedly, and the attempt is often hazardous to your wealth.
What follows is a guide to taking an economic indicator and incorrectly using it to try to time a recession. And at the very end: a very tiny amount of useful advice.
What Will Cause the Next Recession
The key to figuring out what will cause the next recession and the resulting bear market is to realize that whatever model you’re using can be wrong for any future recession. What may have worked well for past recessions may stop working for the next one. For example, industrial production used to be a go-to indicator for recession forecasting back in the ’50s and ’60s. Now, since manufacturing is such a small proportion of America’s economy, not so much.
The same thing happened with Dr. Copper. A lot of people tried to use the spot price of copper to forecast economic demand. While that never really worked in the first place, with the evolution of our economy, a lot of people switched to using semiconductor demand for the same purpose. The popping of the Bitcoin bubble created a “whoopsie” in that approach due to the collapse in demand from bitcoin miners without it remotely affecting the rest of the economy, but it hasn’t gone completely out of fashion yet.
If you want to use a perennial favorite, you need to focus on principles that haven’t changed since the ’50s and ’60s:
- The stock market is priced based on future earnings growth.
- Individuals are responsible for the incremental purchases necessary to realize that growth.
- To make those incremental purchases we need more people with enough money to make those purchases.
The answer: use the unemployment rate, which shows whether there are more or fewer people available to make those purchases.
This one is well-known, and from a 30,000-foot view it makes for a good graph and good talking points if someone asks you where you think the market is heading and you only have 30 seconds to pull up some data to make it sound like you know what you’re talking about.
Using Higher Unemployment to Time a Recession
One way the unemployment indicator is used to time a recession, and a bear market, is by looking at the level relative to the past year’s average. The reason this relative value is used instead of an absolute value is that the stock market is priced based on each marginal unit that the companies in the index can add to profits. It’s not based on the 20- or 50-year average, but on the next 12 months’ earnings. If unemployment goes up over the next 12 months then there won’t be enough people with enough money coming in to buy the products and services that will allow the companies to meet the investors’ forward earnings forecasts.
The first chart above is from the St. Louis FRED database for the unemployment rate with a 12-month moving average overlaid. The second chart is the monthly prices of the S&P 500. As you can see from the charts above, using the unemployment indicator with a 12-month moving average, and starting with the dotcom bubble, you would have avoided a 36 percent drawdown, briefly gotten back in during October 2002 for a month where you would have gained ~6 percent, then gotten back out again until October 2003 where you would have missed out on a ~14 percent gain.
Fast forward to June of 2007, and the indicator would have gotten you out at ~$1,500, which was about 3 months before the index tanked into the Great Financial Recession. You would then have used it to get back into the market at the beginning of May 2010 when the index was down to ~$1,125, avoiding a ~25 percent drawdown, and then gotten back in as the market rocketed back up.
And if you look at the chart from a 30,000-foot view, you won’t notice the upticks in unemployment above the 12-month average that occurred in November 2010, September 2016, and December 2018. If you had continued to follow the unemployment indicator you would have missed out on a total of ~10 percent of incremental gains. Of course, of the three instances you would have avoided one tiny loss, one ~3 percent gain, and one ~7 percent gain.
So nothing’s perfect.
But there’s a broader point here: few people take the time to look at the data for themselves. In fact, you probably read all those numbers above, squinted at the two charts I posted, and thought “Yeah, that looks about right.” What you need to do is go to the two sources of data I mentioned and ask why, if the data goes back to 1948 in the FRED database, I only gave examples over the past 24 years. The key to making a recession forecasting model work is to ignore the data that doesn’t fit your narrative. Kind of like how the yield curve inversion that happened in 2019 didn’t immediately lead to a recession — just ignore that one.
Dimmer Switch Portfolio Management
In all seriousness, economic data is useful. While it certainly isn’t a precise tool and it’s always fun to lampoon talking heads who try to turn it into one, the data will often show times of euphoria, panic, or relative strength or weakness in the underlying economy.
The best thing to do during times of euphoria is to take a look at your portfolio and figure out which would be impacted the most when the euphoria ends. Short volatility, junk debt, preferred shares, emerging markets, and anything with sky-high growth expectations or heavy debt loads are always the usual suspects. If any of these things are in your portfolio then take the time to trim them back or eliminate them altogether.
There’s no reason to sell all your equities and go to cash because of some economic indicator. Treat your portfolio like you would a dimmer switch instead of an on-off switch. Dial it down, dial it up, but never turn it completely on or off.