The Washington PostDemocracy Dies in Darkness

The question isn’t when is the next recession coming; it’s what are we going to do about it.


First, I sat down in a restaurant the other night and the guy at the next table turns to me and says: “When’s the next recession? That yield curve’s looking scary!”

Second, I can hardly open the financial section of the paper without reading some economist declaring that the next recession will hit in 2020. (Economists polled by the Wall Street Journal have a rough consensus that the next downturn will arrive that year.)

I would take none of the above too seriously. The yield curve (I’ll explain below) isn’t scary. And although there’s a substantive reason for “Recession 2020!”— the fiscal stimulus that’s juicing the current growth patch we’re enjoying is scheduled to tail off that year — I suspect another reason people like saying that year is that it’s far away. You won’t remember their forecast if it’s wrong.

But this is anything but a don’t-worry-be-happy column (I tend not to write a lot of those). It’s a warning to recession watchers and, more so, to policymakers: Worry less about when the next recession is coming and worry a lot more about what we’re going to do about it.

Economists can’t tell you when the next downturn is coming, but a lot of the current chatter is because of the current expansion entering its 10th year (it started in June 2009), which ranks it among the longest. Expansions don’t die of old age: They’re murdered by bubbles, central-bank mistakes or some unforeseen shock to the economy’s supply (e.g., energy price spike, credit disruption) and/or demand slide (e.g., income/wealth losses). Also, trade wars.

The other motivation for the current angst is the flattening of the yield curve, meaning there’s less of a gap than usual between short-term and long-term interest rates. Those who swear by the curve point out, correctly, that its inversion (short rates higher than long rates) has proceeded many a downturn. But I’m a skeptic, and I’m not alone. For technical reasons (see data note if you must), the recession-is-coming intel in the curve has gotten wound up with other noise.

Take that noise out, and you get the pattern below (this copies some recent Federal Reserve work; see note). The blue line shows the result of an old workhorse model that uses the yield curve to predict recessions. Though it’s not that high, it’s tilting up at the end of the series. However, when you adjust the curve to get a clearer signal of the part that correlates with downturns, that lift at the end of the series goes away. Other, more recent Fed analysis shows a similar result.

Again, none of this argues a recession isn’t out there somewhere ahead. It is. We just don’t know where.

If that last graph was meant to reassure you, this next one is intended to freak you out. I call it: The Debt before the Downturn. It shows the public debt-to-GDP ratio at a bunch of recent economic peaks. If you average this value over the past six recessions, you get about 30 percent. For our purposes, that means that when the recession hit, there was a good bit of perceived fiscal space to offset the downturn with countercyclical policies. When the next downturn hits, this ratio will almost surely be well over twice its historical average.

Consider compelling new research showing that with such high debt/gross domestic product levels, Congress will do too little to push back against the downturn. Pointing to the elevated debt level, they’ll devote too few resources to programs like infrastructure or subsidized jobs that go beyond the automatic stabilizers, like unemployment insurance. As the authors of the study found: “The fall in GDP with fiscal space is just 1.4 percent. The fall in GDP following a crisis without fiscal space reaches a maximum of 8.1 percent.”

To be clear, even at debt/GDP levels much higher than what we have now, it still pays to apply the needed countercyclical policies. In fact, other recent research shows that to not do so can make the debt ratio worse! As I like to say, when you’re in a recession, the concern should always be: Is our deficit spending large enough to offset the recession, regardless of the debt level?

But given politics, what matters here is perceived, not actual, fiscal space. And the record shows that the pattern in the above figure is one wherein the perception will be of very limited fiscal space.

Next, a key automatic stabilizer isn’t ready for the next downturn. As of the end of 2017, most state unemployment-insurance trust funds failed to meet the minimum standard for recession preparedness. In fact, 11 of these states — California, Texas, Ohio, Indiana, West Virginia, Massachusetts, New York, Connecticut, Illinois, Kentucky and Pennsylvania — had less than half the amount recommended. California’s system is particularly weak, with $1 billion in debt to the federal government leftover from the last downturn (the state is expected to pay off the debt this year, but they then will need to build up the fund).

Ben Spielberg and I wrote a whole paper on what to do to get ready for the next recession. We recommend more responsive triggers to make sure the automatic stuff — unemployment insurance, food stamps — turns on and off more responsively to conditions on the ground. I’ve since argued for a Full Employment Fund that could ramp up to subsidized jobs where and when needed. But based on the last figure, one of the most important fiscal policy corrections is to replace the tax cuts with a plan that stabilizes and reduces the debt in periods of full employment, to have more (perceived) fiscal space when we need it.

Whether we have time to do all that before the next downturn, Keynes only knows.

Data note: The probabilities in Figure 2 come from the model and data in this excellent little Federal Reserve paper. The green line adjusts the slope of the yield curve (10-year minus 3-month rates) for the “term premium.” Longer-term interest rates, like the yield on the 10-year Treasury, can be broken up into the expectation of the average of future short-term rates and the term premium, or the extra yield investors require to lock up their money for the term of the loan. Since it’s thought to be the first part — expected rates — that correlates with a future downturn, it makes sense to net out the term premium from the model. When you do so, you get the flat line in the figure.

Correction: An earlier version said “Recessions don’t die of old age.” That has been changed to “Expansions…” H/T: Alexandra B from comments.