2015 Series • No. 15–10
Research Department Working Papers
Price Dispersion and Inflation: New Facts and Theoretical Implications 
From a macroeconomic perspective, price rigidity is often perceived to be an important source of price dispersion, with significant implications for the dynamic properties of aggregate variables, welfare calculations, and the design of optimal policy. For instance, in standard New Keynesian models, the key cost of business cycles stems from the price dispersion resulting from firms' inability to adjust prices instantaneously. However, different macroeconomic models make conflicting predictions about the level of price dispersion, as well as about its dynamic properties and sensitivity to inflation. These contrasting predictions can help us to discriminate across alternative models. This paper examines the link between price dispersion and inflation, and the role of sales in this relationship.
Key Findings
- Models without sales fail to match empirical findings even for regular prices.
- The best match of empirical findings, for both posted and regular prices, comes from a Calvo model with sales, calibrated to match the observed frequency of sales. Intuitively, sales in this model serve as a channel of additional price flexibility that does not interfere with the frequency of regular price changes.
Implications
The finding that the Calvo model with sales matches the properties of regular prices better than a similar model without sales implies that sales have an important interaction with regular prices that is lost when sales are omitted. This implication is at odds with the conclusions drawn by Guimaraes and Sheedy (2011) and Kehoe and Midrigan (2014), who argue that sales have little impact on macroeconomic dynamics and that Calvo models without sales capture the salient features of the data sufficiently. In contrast, this paper's results suggest that in order to capture the comovement of regular price dispersion and inflation, it is necessary to incorporate sales into the macro models. Because in macroeconomic models the level and dynamics of price dispersion have direct implications for welfare, the cost of business cycles, and the optimal inflation rate, relying on models that are at odds with the empirical properties of price dispersion can lead to nontrivial mismeasurement of optimal policy actions.
Abstract
In macroeconomic models, the level of price dispersion—which is typically approximated through its relationship with inflation—is a central determinant of welfare, the cost of business cycles, the optimal rate of inflation, and the tradeoff between inflation and output stability. While the comovement of price dispersion and inflation implied by standard models is positive, I find that in the data, it is negative. This is due to transitory price changes (sales): if sales are removed from the data, the comovement of price dispersion and inflation turns positive. Nevertheless, I show that a wide variety of price-stickiness models that ignore sales cannot quantitatively match the comovement even for regular prices. Modeling sales explicitly helps to reconcile theory with the data: a Calvo model with sales quantitatively matches both the negative comovement for posted prices and the positive comovement for regular prices. Thus, sales play a pivotal role in explaining the properties of regular prices; models without sales—failing to reproduce fundamental properties of price dispersion—may significantly mismeasure welfare and its determinants.