Menu costs, price dispersion, and aggregate fluctuations

I address the micro-foundations and aggregate consequences of sluggish price adjustment. A major theme of this investigation is the explicit modeling of the frictions that give rise to nominal rigidities; namely, fixed (menu) costs of adjusting prices. This dissertation consists of three chapters.; The first chapter asks whether nominal rigidities arise either (i) because of physical costs of price adjustment or (ii) because of information-gathering costs. If the state of the world is observed costlessly and physical adjustment costs are the source of price stickiness, firms are more likely to reprice in periods when aggregate and idiosyncratic shocks reinforce each other, thus triggering desired price changes in the same direction. I find strong support for this implication of the menu cost model using data for four-digit U.S. manufacturing industries and measures of technology shocks derived from production-function estimates.; Chapter 2 studies the ability of menu costs to generate output fluctuations. I use a dataset of scanner prices to document that (i) although the average magnitude of price changes is large, a substantial number of price changes are small in absolute value; (ii) the distribution of non-zero price changes has fat tails; and (iii) stores tend to adjust prices of goods in narrow product categories simultaneously. I extend the menu-cost model to a multi-product setting in which firms face cost complementarities in price adjustment. The model, because of its ability to replicate this additional set of micro-economic facts, can generate aggregate fluctuations much larger than those in standard menu-cost economies.; The final chapter is inspired by the pervasive positive relationship between the international relative price variability (RPV) and nominal exchange rate (NER) volatility. This relationship is interpreted as evidence that sticky prices, rather than trade frictions, are the source of the large international relative price movements. I show that menu-cost models predict a hump-shaped rather than monotonic relationship between RPV and NER volatility. The hump occurs at higher nominal exchange rate volatilities the less tradeable goods are. I use this insights to show that trade costs as large as 50 percent are necessary to reconcile menu-cost models with the data.