Micro-level evidence suggests that prices and wages are slow to adjust. This imperfect adjustment is a central feature of the monetary transmission mechanism in New Keynesian models. For example, given that prices are slow to adjust, a reduction in the nominal interest rate leads to a reduction in the real interest rate and therefore (in the baseline model) an increase in consumption and therefore output. Advocates of the NK framework often suggest that the framework can better explain the data than models with flexible prices. This, coupled with the fact that there is ample evidence of sticky prices at the micro-level, is often used to support the use of the NK framework.
But is micro-level evidence sufficient to support this transmission mechanism of monetary policy? No. If the NK transmission mechanism is correct, then we have to be able to explain the stickiness of the price level not simply the stickiness of individual prices. This position should not be controversial, but yet I think that it is under-appreciated and perhaps not well understood.
In his paper, “The Economics of Information”, George Stigler discussed price dispersion. One of Stigler’s assertions in the paper was that price dispersion results from “ignorance in the market.” In other words, as a consumer, I don’t know the price charged by every seller and therefore this opens the door to price dispersion — even among homogeneous goods. This insight was extended by Burdett and Judd, who showed that if one knows the price of every seller in the market, then the price will converge to the competitive price. If one knows only the price of one seller, then the price charged by each firm will converge to the monopoly price. If, however, the consumer knows the price charged by more than one firm with a positive probability, there is equilibrium price dispersion.
So why does this matter and what does it have to do with sticky prices? Well, as shown in a recent paper by Head, Liu, Menzio, and Wright, a model with equilibrium price dispersion can generate evidence of sticky prices even though money is neutral. To understand their insight, consider the following example. Suppose that there is a distribution of prices over which firms can maximize their profit. In other words, firms with a lower price are able to maximize profit by making up for the lower price with a larger volume of sales. In addition, suppose that this distribution of prices has support [p1, p2]. Since money is neutral, an increase in the money supply causes the distribution of prices to shift. Thus, imagine that an increase in the money supply causes the distribution of prices to shift such that the new distribution has support [p3, p4], where p1 < p3 < p2. This change suggests that the only firms that have to adjust their price following the change in the money supply are the ones that charge a price less than p3.
In the scenario described above, prices appear to be sticky. Only a fraction of firms adjust their prices. However, money is neutral. The implication is not necessarily that sticky prices are unimportant, but rather that the observation of "sticky prices" is not sufficient evidence for the transmission mechanism of monetary policy advocated by New Keynesians. This conclusion is also not new. Caplin and Spulber and Goloslov and Lucas, for example, obtain similar results.
What then is needed to reconcile the NK transmission mechanism? It would seem that it must be true that the price level rather than individual prices must be sticky. The NK model explicitly assumes this by using a representative firm in the case of Rotemberg pricing or by simply assuming that the price level is slow to adjust because only a fraction of firms can change their prices in the case of Calvo pricing.
Sticky prices are not a distinctly Keynesian idea and are not confined to New Keynesian analysis. The notion of sticky prices was evident in the writings of classical economists. This post is not meant to suggest that sticky prices are unimportant for the monetary transmission mechanism. Rather, the point is to emphasize that micro-level evidence is insufficient to argue that sticky prices have macroeconomic implications.