I am a sceptic of thinking about low unemployment as a bad thing – which is what people who think about policy in terms of the Phillips curve do. Now a study by the Philly Fed is saying that the Phillips Curve is a poor forecasting tool. Will this have any meaningful impact on policy? Some thoughts below
Here’s the deal: a lot of people at the Fed think of low employment as something that stokes inflation and makes the Fed’s job harder. On the one hand, the Fed wants some inflation. Hence the meme on Amazon’s lowering prices at Whole Foods making the Fed’s job harder. But on the other hand, the Fed only wants moderate inflation – optimally 2%. That’s its target.
But a “very, very low unemployment rate” makes that target harder to achieve. At least that’s what NY Fed President Bill Dudley has said. And what he basically means is that the Fed wants to engineer the economy so that enough people remain unemployed to keep inflation from taking off and becoming entrenched. Let’s call this Fed-engineered unemployment or FEU (pronounced like the word FEW).
But what if the benefits of FEU were just a myth? For example, the Philly Fed paper says “forecasts from our Phillips curve models tend to be unconditionally inferior to those from our univariate forecasting models.” That’s Fedspeak for “if you want to forecast inflation accurately, don’t track two things as a trade-off to one another.”
Interestingly, the Philly Fed paper found that any trade-off that could actually be forecasted only happened when the economy was weak – meaning when the Fed would be cutting rates or holding steady. So the Fed’s own research now suggests that raising rates solely because unemployment was ‘too low’ makes absolutely no sense.
We can hypothesize why the Phillips Curve doesn’t add value in making policy. For example, some people point to technological change, globalization, and the waning clout of unions as factors. Irrespective, it is undeniably true that this Phillips Curve framework doesn’t help and that it helps least when the economy is firing on all cylinders.
I believe policymakers at the Fed should take this new information on and stop using the Phillips Curve to justify rate hikes. FEU is just bad policy.
What would that mean in the present context? First, it could mean the Fed would have to broaden its definition of inflation to include asset prices. Since asset prices are way up, rate hikes would still be justified. Second, the Fed could also decide that, because asset price inflation is more pronounced than consumer price inflation, rate policy was the wrong way to reduce accommodation. For example, Fed Chair Yellen has pointed to continued regulatory vigilance as important. Or third, it could also mean that the Fed would have to pause on raising rates until consumer price inflation actually began to increase.
My view: The Philadelphia Fed study is yet another hurdle to the Fed’s stated policy normalization plan. I continue to believe that balance sheet reduction will become a more important tool in the Fed’s policy framework. And we will just have to wait and see what that means for market liquidity and stability.