The business cycle is dying
By Scott Sumner
I was almost going to write “dead”, but then thought that this would give readers the erroneous impression that I don’t expect any more recessions. Rather I believe the term ‘cycle’ is no longer descriptive.
In November 1982, the unemployment rate hit 10.8%, the highest rate since WWII. The US had experienced 8 recessions over the previous 37 years (9 recessions over the previous 38 years), and they seemed to be getting worse (1974 and 1981-82 were the two worst.) No one expected that the US would experience only 3 recessions over the next 37 years. What went right?
During the postwar period, monetary policy was under the influence of a rather simplistic Keynesian model. The problems of the 1970s and early 1980s led to a Keynesian/monetarist synthesis called “New Keynesianism”, which incorporated monetarist ideas such as the Natural Rate Hypothesis and the importance of the Fisher effect. It also incorporated Keynesian ideas such as interest rate targeting and policy activism. The Taylor Rule is a simple way of thinking about this synthesis.
As monetary policy became more effective, recessions became much less frequent. Importantly, this process is likely to continue. If we achieve America’s first ever soft landing over the next 2 or 3 years (as I anticipate), then I would expect only one recession during the next 37 years.
The economics profession has not yet fully absorbed the implications of improved monetary policy. Here’s Noah Smith:
It’s impressive how well the U.S. economy has held up during the past year. As early as 2018, leading indicators were suggesting a heightened risk of recession in 2019 or 2020. Then early this year the yield curve inverted, a traditional signal that recession is imminent (the inversion has since reversed, but this typically happens before growth actually goes negative). The trigger for a downturn wouldn’t be hard to identify — a slowing China, combined with President Donald Trump’s trade war. Already, countries such as Singapore and South Korea, which export lots of manufactured goods to China, are slowing down.
But despite all the pieces that seem to be in place for a recession, it hasn’t happened. This expansion is now the longest in postwar history, having recently surpassed the long boom of the 1990s.
That’s all quite reasonable, but I have a slightly different take. Smith is quite right that real shocks are often a “trigger” for recessions—recall the housing/banking crisis of 2007-08. But they are not the cause of recessions. Instead, recessions are caused by tight money policies that result in a sharp slowdown in NGDP growth. As monetary policy improves, real shocks become increasingly unlikely cause bad monetary policy, and hence less likely to trigger recessions. More likely, real shocks will trigger RGDP growth slowdowns, as in 2015-16.
But 2015-16 wasn’t even close to being a recession, as the unemployment rate continued to fall. America has never even had a mini-recession where the unemployment rate rose by only 1.0% to 2.0% and then starts falling, so a period where unemployment declines is radically different from a period where the unemployment rate rises by more than 2.0%.
Just as the bad period of the 1970s and early 1980s led to improvements in monetary policy, the Great Recession has led to further improvements. The Fed now understands that the natural rate of interest is in long-term decline. The past three recessions were partly caused by the Fed overestimating the natural rate of interest, and hence not cutting rates rapidly and deeply enough to prevent recession when trouble developed. They are less likely to make that mistake in the future. The Fed also has a better understanding that inflation shocks like 2007-08 are not worth worrying about in an environment where NGDP growth is slowing. They also understand the need to commit to something like level targeting in a slump. The one negative, of course, is that the Fed is more likely to face the zero bound problem going forward.
The reduction from 8 recessions in 37 years to only 3 recessions in 37 years is less impressive than it seems. Because of lackluster recoveries, the percentage of time with high unemployment has declined much less than you’d expect. But a further reduction to one recession in the next 37 years really would be a big deal. That would imply that the Fed has learned how to engineer soft landings, creating long periods of near full employment.
Given that we’ve never had a soft landing, why am I so optimistic that we are about to achieve the Holy Grail of macroeconomics? Because other similar economies such as Australia and the UK have learned how to achieve soft landings, and hence I see no reason why recessions won’t continue to become increasingly uncommon in the US.
In the 38 years after WWII, recessions occurred roughly every 4 years, on average. They were similarly frequent before WWII. If I’m right that we will have only one recession during the next 37 years, then the term ‘cycle’ will no longer be appropriate. In that sense, the business “cycle” is dying.
Right now, this is a “tentative prediction”, contingent on no recession over the next three years. If this soft landing occurs, then I’ll switch from a tentative to a confident prediction. I obviously won’t live long enough to see if my 37 year prediction comes true, but I hope and expect to live long enough to see if it is likely to come true, if soft landings become a normal part of the American macroeconomy.