The Natural Rate of Unemployment: Is it Useful, and if so, When?
Or, Sliding Down the Phillips Slide
I often get into debates about the ideas of Milton Friedman. I’m a bit of a defender of his legacy, so it makes sense that I should make a post evaluating the modern applications of his “Natural Rate of Unemployment” idea.
The idea has received much malign amongst critics of modern macroeconomics, as the idea is often misconstrued as implying that unemployment cannot be lowered via demand side action over the long-term. In this post, I would like to develop a restatement of the NRU and consistent advice for its policy applications.
A Nobel Idea
Milton Friedman introduced this term with his work on the Phillips Curve, and I believe it grew most popular after his 1976 Nobel Prize lecture. To him, the Natural Rate of Unemployment, is what you get when you cross adaptive expectations with the Phillips Curve. It’s the unemployment “Long Run Phillips Curve”. Essentially, Friedman believed that in the long term, price level rises could fail to induce extra employment, as workers would begin negotiating for wages on real, not nominal, terms. Therefore, only “real” factors could lower the long-run rate of unemployment.
What this means is that long-run price level increases will eventually fail as stimulus to employment. What this doesn’t mean is that demand side policies should not be attempted or that inflation cannot “grease the wheels of the labor market”, as James Tobin put it.
If inflation is constant, then workers will begin to supply labour merely based on the real wage. However, due to nominal wage rigidity, if the nominal wage is required to fall, inflation, even when expected, can still play a role in allowing the real wage to fall, preventing unemployment from rising. We find this in Dupraz, Nakamura, and Steinsson (2019).
The first thing to evaluate is what this theory means for inflation. We can assume that stable inflation in the long run is disregarded by workers when negotiating for wages. However, a variable inflation rate that still trends low will provide redistributive effects. As mentioned earlier, a stable or rising inflation rate in recessionary times will soften the blow to employment. Still, a variable and high inflation rate will incur serious welfare costs. From this, it appears that the best policy is a stable rate of inflation that rises during a recession.
However, there is a caveat-- I’m making the assumption here that demand-side policies aren’t shaping the natural rate. If expansionary fiscal and monetary policy is consistently implemented, then the economy will be stronger and actors will adapt to stronger importance. In effect, proper stabilization policy can be factored into expectations and lower the natural rate of unemployment. The natural rate of unemployment isn’t the real rate, it’s the adaptive rate.
Further Pan-alysis
Another concept to consider is the idea that the Natural Rate of Unemployment is a floor. If you look at the statistics, unemployment has jumped heavily over the rate, but very rarely under it. In effect, there is a plucking property (let us refer back to the earlier Dupraz, Nakamura, and Steinsson paper I mentioned) whereupon demand policy can force the floor down by strengthening the economy, and unemployment rising over the rate is a sign of poor demand policy. Since it’s hard to push under the floor, or capacity, the excess demand will be translated to inflation. This is where we get NAIRU from.
I dislike the Keynesian NAIRU (check the way it’s described here) because it gets the causation wrong. Old-school Monetarists believed that the Phillips Curve functioned with inflation expectations as the driver, and Market Monetarists believe the Phillips Curve functions with monetary/fiscal policy (read: nominal spending growth) as the driver. I personally favor growth of certain money supply divisia. Still, the point remains that inflation is not driven by low unemployment itself.
P.S. You may now ask, “what does the NRU matter for?” My answer is that it should be a guide to help measure if demand policy is strong enough, or perhaps too transitive. Strong, but transitive demand policy cannot lower the natural rate.