@SteveRoth I can't speak for David, but I think the claim is that in a Ricardian regime, there's an equilibrium inflation rate. An increase in the interest rate will temporarily bring the inflation rate lower than that, but it will be transitory — the rate will return to the equilibrium. 1/

@SteveRoth In a non-Ricardian regime, the equilibrium inflation rate is not independent of the interest rate. 2/

@SteveRoth So if the inflation rate starts at an equilibrium of n%, increasing the interest rate will (cet par) lead to a transient decrease in the inflation rate to some rate <n% (same as before, the conventionally anticipated effect), but when the transient effect wears off, the new equilibrium at the new interest rate has inflation now at some rate *higher* than n%, rather than merely revering to n% as under a Ricardian regime. 3/

@SteveRoth So if a naive central bank targets n% by raising interest rates when actual inflation is above n%, they get a "sugar high" (or sugar low) of disinflation, it seems to work, but then inflation reverts to a value even higher than its starting point, then the idiot central bank does it again, etc. each tightening's respite is just a prelude to an even worse catastrophe. 4/

@SteveRoth So even though in the short run, raising interest rates works under both regimes, only under the conventional model of a Ricardian regime is interest rate policy straightforwardly good policy. The effect is transient in both cases, but under the Ricardian regime it does no long-term harm, while in the non-Ricardian (more realistic, at least on face) regime, the transient desired effect is just a prelude for an even worse situation. /fin