This paper presents a model of a monetary economy where firms face idiosyncratic productivity shocks and general inflation. Sellers can only change prices by incurring a real "menu cost." The model is calibrated using a new U.S. data set of individual prices from Klenow and Kryvtsov. The calibrated model predicts the effects of high inflation on the frequency of price changes, which align well with Israeli evidence from Lach and Tsiddon. The model is also used to conduct numerical experiments on the economy's response to credible and incredible disinflations and other shocks. The results show that monetary shocks do not induce large or persistent real responses. The model incorporates two types of shocks: aggregate inflation and idiosyncratic productivity shocks. It uses a discrete-time model to approximate a continuous-time model, allowing for the computation of optimal pricing behavior. The model predicts the fraction of prices changed per month, a key parameter in Calvo's (1983) model. The model is calibrated using data from the U.S. economy of 1988-97 and compared to Israeli inflation data from 1978-79 and 1981-82. The predictions of the model are surprisingly accurate. The model shows that idiosyncratic shocks account for most of the price adjustments in the U.S. More than 20 percent of all consumer prices are adjusted each month, and the average size of the change is much larger than the expected inflation between adjustments. Simulations confirm that idiosyncratic shocks are responsible for most of the price changes. When idiosyncratic shocks are shut down, the frequency of price adjustments is roughly unchanged in high inflationary environments but is much reduced when inflation is low. Even though idiosyncratic shocks cause most of the price adjustments, new prices reflect both firm-level and aggregate shocks. This makes it difficult to obtain a large degree of persistence for monetary shocks in the menu cost environment. The simulations also show this effect very clearly. Although the simulated economy fits observed pricing behavior well, it exhibits very small real effects of monetary instability. Simulated time series with realistic monetary variability have a variance of aggregate output that is ten times lower than that observed in the recent U.S. data. The estimated Phillips curve is also very flat: A one percentage point reduction in inflation rate depresses output by only 0.05 of one percent. Thus money appears nearly neutral: The effects of inflation shocks on aggregate production and employment, as we estimate them, are small and transient. Similar results hold for different policy experiments. For a credible disinflation—using an immediate reduction in the rate of wage inflation from 15 to one percent per quarter—there are essentially no real effects, consistent with Sargent’s (1986) findings. But even when the disinflation is carried out in a non-credible way, the resulting decline inThis paper presents a model of a monetary economy where firms face idiosyncratic productivity shocks and general inflation. Sellers can only change prices by incurring a real "menu cost." The model is calibrated using a new U.S. data set of individual prices from Klenow and Kryvtsov. The calibrated model predicts the effects of high inflation on the frequency of price changes, which align well with Israeli evidence from Lach and Tsiddon. The model is also used to conduct numerical experiments on the economy's response to credible and incredible disinflations and other shocks. The results show that monetary shocks do not induce large or persistent real responses. The model incorporates two types of shocks: aggregate inflation and idiosyncratic productivity shocks. It uses a discrete-time model to approximate a continuous-time model, allowing for the computation of optimal pricing behavior. The model predicts the fraction of prices changed per month, a key parameter in Calvo's (1983) model. The model is calibrated using data from the U.S. economy of 1988-97 and compared to Israeli inflation data from 1978-79 and 1981-82. The predictions of the model are surprisingly accurate. The model shows that idiosyncratic shocks account for most of the price adjustments in the U.S. More than 20 percent of all consumer prices are adjusted each month, and the average size of the change is much larger than the expected inflation between adjustments. Simulations confirm that idiosyncratic shocks are responsible for most of the price changes. When idiosyncratic shocks are shut down, the frequency of price adjustments is roughly unchanged in high inflationary environments but is much reduced when inflation is low. Even though idiosyncratic shocks cause most of the price adjustments, new prices reflect both firm-level and aggregate shocks. This makes it difficult to obtain a large degree of persistence for monetary shocks in the menu cost environment. The simulations also show this effect very clearly. Although the simulated economy fits observed pricing behavior well, it exhibits very small real effects of monetary instability. Simulated time series with realistic monetary variability have a variance of aggregate output that is ten times lower than that observed in the recent U.S. data. The estimated Phillips curve is also very flat: A one percentage point reduction in inflation rate depresses output by only 0.05 of one percent. Thus money appears nearly neutral: The effects of inflation shocks on aggregate production and employment, as we estimate them, are small and transient. Similar results hold for different policy experiments. For a credible disinflation—using an immediate reduction in the rate of wage inflation from 15 to one percent per quarter—there are essentially no real effects, consistent with Sargent’s (1986) findings. But even when the disinflation is carried out in a non-credible way, the resulting decline in