Several recent articles have examined what seems to be a breakdown in the Phillips curve. Greg Ip explains: “Standard models of the economy are built on a simple relationship: When unemployment goes down, inflation eventually goes up. That relationship, dubbed the Phillips curve, has looked sickly for years.”
But there’s a simple reason to think that the Phillips curve is not always going to show up in the data. Imagine a central bank that is hitting its 2 percent inflation target every year, rigidly and perfectly. In that case, there isn’t going to be any relationship between unemployment and inflation. If unemployment is at 7 percent, inflation will be at 2 percent; if unemployment is at 4 percent, inflation will still be at 2 percent. On a graph that plots unemployment against inflation, there will be a straight line instead of a curve.
A graph that plots inflation against anything else will be a straight line too — if, again, the central bank is successfully targeting inflation.
That straight line can cause confusion, because it has the paradoxical effect of making a highly successful central bank look powerless. Increased demand for money balances pulls prices downward, so our perfect inflation-targeting central bank will have to match changes in that demand with changes in the money supply. Since it is doing so perfectly, however, it will look as though the money supply does not matter: Whether it grows fast or slow, inflation runs at 2 percent. It will look as though money demand does not matter either.
Monetary economist Nick Rowe has occasionally borrowed Milton Friedman’s analogy of the central bank as thermostat to illustrate the point. If everything is working well, we should not see any correlation between how hard the air conditioner is working and the temperature inside the house. From just looking at the data, we might get a completely mistaken understanding of the underlying causal relationships. We might think that the more power the air conditioner uses, the higher the outside temperature goes. We might think that the air conditioner has no effect on the temperature.
One reason monetary policy engenders so much confusion is that we are prone to making just this kind of mistake. A case in point: Several decades into a worldwide regime of inflation-targeting central banking, we’re still trying to figure out why our “standard model” of the relationship between unemployment and inflation has broken down.