In yesterday’s post on underemployment, I employed various concepts used by many in high places in contemporary economic analysis, including “natural rates” of un- and underemployment–i.e., the lowest rates consistent with stable inflation–the “equilibrium” Fed funds rate–the one consistent with stable growth that’s neither too hot nor too cold–and so on.
Some readers and Tweeters were unhappy with such usage, arguing in so many words that a) it is inherently conservative to adopt such constructs as in practice they generate biases against truly full employment and thus against the less advantaged, who benefit most from tight labor markets and are hurt most by slack ones, and b) they’re unobservable and impossible to pin down within a policy-useful margin of error, so why go there?
I acknowledged as much in yesterday’s post: “…there are good reasons not to truck in this whole “natural rate” business at all; there’s no reliable way to nail it down, it moves around, and economists invariably tend to pitch it too high, at great cost to those who depend on truly full employment.”
It’s also important in this corner of economic analysis to be wary of the assumption that there’s an equilibrium state for these variables. That is, even if a rate of unemployment or interest really is consistent with stable prices/growth at time t, that rate may well be out (or moving out) of said equilibrium at time t+1. Check out all the time variation in the “natural rate of interest” from this recent Fed paper (see Figure 5).
And the unobservable problem is huge.
But here’s the thing. When in Rome…That is, there is IMHO a real utility in this work to working within the model that’s in the minds of the key players pulling the key levers. First, if your goal is to nudge the course of the ship a bit, you’ve got to be able to use the same map and navigation system as the captain, even if you don’t wholly believe it.
Second, these people aren’t dopes by a long shot. My beef is more that they’re working in models driven by correlations more than causation and these models have been missing pretty big time for a pretty long time, often at great cost. But prices/wages do, in fact, correlate with slack/capacity utilization, though to different degrees over different times and for different reasons than the classical models often reflect. EG, bargaining power is a critical variable that can get lost in a lot of the macro analysis.
So I will continue to use all the tools and models that everyone else does, but I won’t be limited to them and I will employ them with caution and a dose of skepticism, not cynicism.