Expected Inflation and Welfare: The Role of Consumer Search
Read more at Chicago Booth Review
In standard macroeconomic models, the costs of inflation are tightly linked to the price dispersion of identical goods. Therefore, understanding how price dispersion empirically relates to inflation is crucial for welfare analysis. In this paper, I study the relationship between steady-state inflation and price dispersion for a cross section of U.S. retail products using scanner data. By comparing prices of items with the same barcode, my measure of relative price dispersion controls for product heterogeneity, overcoming an important challenge in the literature. I document a new fact: price dispersion of identical goods increases steeply around zero inflation and becomes flatter as inflation increases, displaying a 𝚼-shaped pattern. Current sticky-price models are inconsistent with this finding. I develop a menu-cost model with idiosyncratic productivity shocks and sequential consumer search that reproduces the new fact and exhibits realistic price-setting behavior. In the model, inflation-induced price dispersion increases shoppers' incentives to search for low prices and thus competition among retailers. The positive welfare-maximizing inflation rate optimally trades off the efficiency gains from lower markups and the resources spent on search.
Inflation and price dispersion
The figure shows the relationship between inflation and price dispersion in the cross section of retail products. The unit of observation is a product category × geographic market × year combination. The dots correspond to average price dispersion at each of a hundred inflation bins. Price dispersion is measured as the standard deviation of log prices across all sellers of an identical good in a given geographic market and year.
Work in Progress
Online Shopping Access and Retail Pricing Behavior
Online shopping has been expanding rapidly in the past decade. At the same time, recent studies suggest that prices of physical stores are becoming more flexible. What are the effects of online competition on the pricing behavior of traditional retailers? I start by providing empirical evidence from scanner data that consumers' access to online markets affects the pricing behavior of brick-and-mortar stores. By exploiting within-product variation across geographic markets, I show that a higher online spending share is associated with higher price flexibility, smaller price changes, less prevalent sales, and lower price dispersion within narrow categories of goods. I conjecture that online shopping makes it easier for consumers to compare goods and their prices across sellers. These changes in buyers' shopping behavior might affect the firm-level demand elasticity, therefore the pricing behavior of traditional stores. I find supporting evidence for this hypothesis: the effects of online shopping access are particularly strong in markets with a larger share of bargain hunters – buyers who shop online and are more likely to purchase a given product from different retail chains.
Online spending shares by category
The figure plots the online spending share of retail products, by two broad categories, over time. The solid line corresponds to products in non-grocery categories (e.g., cookware); the dashed line, to those in grocery categories (e.g., frozen goods).
Interactions between Shopping Behavior, Inflation, and Labor Markets
I document two novel observations on inflation and shopping behavior by employment status. First, as absolute inflation increases from zero, households where at least one head is not employed pay lower prices, and households where all heads are employed pay higher prices for the same basket of goods. Second, households with an unemployed head visit a larger number of distinct retailers than those in which all heads are employed. Moreover, this difference increases with absolute inflation. Motivated by the evidence, I develop a monetary model with sequential consumer search to study the role of inflation and shopping behavior in labor market fluctuations. In the model, consumers buy a homogeneous good and have incomplete information about the price distribution, which gives them incentives to search for the lowest prices. The higher is absolute inflation, the higher is price dispersion of identical goods across sellers, and, thus, the returns to search. Nevertheless, search activity is costly and reduces the time available to work, so the same aggregate conditions that increase incentives to search might decrease labor supply.
Prices paid by employment status, and inflation
The y-axis shows the average difference between the log price index of households where at least one head is not employed and of households where all heads are employed. The x-axis corresponds to annualized geographic market-level inflation. The dots display predicted values from a regression of the log price index on absolute inflation by employment status. Market and quarter dummies are included in the estimation, as well as various household-level controls. The standard errors used to compute the confidence intervals (bars) are two-way clustered by household and market × quarter pair.