Inflation Epicycles
Many mainstream macroeconomists still argue that DSGE (Dynamic Stochastic General Equilibrium) models are the way to go to understand what is going in the economic world, provided enough fudges and stitches are added to baseline version. Although heavy consumers of DSGE, central bankers failed to anticipate the recent episode of inflation.
This, according to some, is not due to the inadequacy of the model but to the extraordinary circumstances, or to the difficulty of measuring the relevant indicators of tensions in labour markets, etc. I have heard the “extraordinary time” exception before in order to save the Black-Scholes theory from infamy: yes, Black-Scholes fails in rough times, but is OK in normal times! Strange arguments precisely when options are insurance contracts, and when monetary policy is most needed to prevent crises.
My main peeve about DSGE is the “E” for “Equilibrium”, to wit supply = demand and markets clear *at all times*, even when demand is itself time dependent (the D in DSGE).
However, in the absence of tension between demand and supply, it is hard to see how inflation can come about.
So here’s the fairy tale:
DSGE modelers assume that firms are not free to set prices as they wish. They face “menu costs” that prevent them updating prices when needed. Imagine now that for some reason demand increases; firms cannot increase prices right away and while they should keep production constant, market clearing imposes that they make losses by producing more, creating what economists call a positive “output gap”. Eventually, firms update prices to catch up, partly based on expectation of future inflation (since they will again be stuck with new prices for a while, they better anticipate) and partly based on observed “output gap”. This leads to the famous “New-Keynesian Phillips Curve” equation, which says that if the economy is running hot (beyond capacity), inflation will follow.
But note that there is never “not enough goods” in this story; in fact when firms can quickly update their prices, no output gap can ever appear. (Prices just reflect production costs in that case.) It is hard to reconcile this story with what actually happened: there *were* supply chain disruptions and firms did *not* feel restrained to increase prices, in the face of abundant savings stashed during COVID. The economy was nowhere near equilibrium!
DSGE does not allow for excess savings, supply-demand imbalances, inventories and stockouts, and does not say anything about the time needed to reach equilibrium after a shock. These are major deficiencies (among many others) that should become addressed, for example using Agent Based Models – that allow for all sorts on Non-Equilibrium (NE) effects – or to formulate a proper “DS*NE*” model.
