Currently, a lot of attention is paid to the yield curve and how the Fed might react to recessionary risk. Claudia Sahm of the Federal Reserve takes a different tack, she has detailed a novel way to both forecast recessions and make them less severe. Her findings are helpful to investors too. She details a sound metric for tracking recessions early. It's to look for a +0.5% rise in the unemployment rate relative to the unemployment rate's past 12-month low. This metric has called each of the seven recessions back to 1970, never getting it wrong over that period. That's a good success rate, and comparable with the inverted yield curve. However, it's a slightly later signal than the yield curve but more robust in Sahm's view over the long-term.
Unemployment rising tends to coincide with the very start of a recession. It's a fairly early signal, but it tends to be triggered 1-4 months after the recession begins looking back over history. It's also a robust indicator, some have argued that the inverted yield curve we see today is the result of unique bond market circumstances. We could have a false positive from the yield curve. Whereas the unemployment rate is more directly tied to the immediate fate of the economy. Put simply, if the unemployment rate rises, then people lose their jobs and when people lose their jobs, they spend less, and dropping consumer spending typically causes recessions. Hence there's a fairly clear link from job losses to falling economic growth. By contrast the mechanism by which an inverted yield curve impacts the broader economy is less obvious, working through the financial markets in a less direct way. For example, there's now a risk that the yield curve is now so closely watched given its past success that the economy now reacts to it in a different way than previously.
Watch For 4.1% Unemployment
So essentially, given that the current 12-month low for unemployment is 3.6%, if, theoretically if we saw a 4.1% unemployment rate at some point within the next 12 months, then it's reasonable to assume on Sahm's metric that would be in a recession. Probably in the earlier stages of one. The chart below makes clear how this has worked historically.
Sahm isn't just trying to forecast recessions though. She's hoping to smooth them. Her main concept is to tie cash payments to people to when the unemployment spike hits. That could prove a rapid and effective way to combat recessions. If made automatic, it could boost consumer spending right when support is needed. Giving people money is also likely to have a more immediate and predictable impact than, say, the actions of the Fed in moving interest rates to manage the economy, which is less predictable and can have more of a lag. However, in order for her proposal to work, Sahm needs a clear recessionary indicator to tie the payments to, and it's telling that she's picked a spike in the unemployment rate.
How Close Could A Recession Be?
Unlike the yield curve, unemployment is not flashing any recessionary warning signal yet. In fact, in a sense, we're seeing the opposite. Unemployment is falling. However, past experience shows that a unemployment can start rising both quickly and sharply should a recession occur. In 2008 it took 7 months for inflation to rise over 0.5%. In 2000-2001 it took 6 months.
However, clearly, rising unemployment is an ominous signal especially for the stock market. The unemployment rate as of the most recent May 2019 data is clearly declining on a medium-term view and declining or closer to flat depending on how you interpret the shorter term picture. So we would need rising unemployment to become concerned about this indicator, something we don't have today. Yet, unemployment did rise in Q4 of last year, which also coincided with the stock market correction over the same period. Clearly the markets watch employment data closely. It's likely the stock market would decline well before unemployment rose materially.
It is therefore interesting to combine the yield curve and unemployment indicators. While the yield curve suggests there is a reasonable chance that a recession is here within a year, the labor market is suggesting that a recession is not happening now. So if a recession is coming, as the yield curve suggests may be possible, then it's likely still several months out based on what unemployment suggests.
