Time-dependent or state-dependent pricing? Evidence from a large devaluation episode

The real effects of monetary policy depend on the reasons behind price stickiness. In models with time-dependent pricing, firms readjust prices at previously (and possibly endogenously) determined times. In constrast, with state-dependent price setting, prices are readjusted whenever they are far enough from their desired levels, so a monetary shock leads firms to adjust their prices, which dampens the real effects of monetary policy. This paper explores the distinct predictions of price-setting models on how the frequency and magnitude of price adjustment react to shocks in order to distinguish between models. Consider a positive shock to the desired prices of goods (the prices that would be charged in the absence of frictions). In the simplest state-dependent models (e.g., Caplin and Spulber (1987)), a firm raises the price of its good whenever the difference between the desired and current price hits a constant threshold. Hence a positive shock to desired prices raises the frequency of price changes but leaves the magnitude of price changes unaffected. In recent models of state-dependent price setting, shocks might have some effect on the magnitude of price changes, but the response of the frequency of price adjustment to shocks is a key feature of all these models.