In textbook economics, the market for a good is in equilibrium when its price “equates supply and demand.” The supply-demand approach is a useful tool, but for many important markets, this framework doesn’t do a good job because of frictions that make it hard for buyers to find suitable sellers and vice versa. Take, for example, the labor market. This is a market in which the “good” being sold (labor services) isn’t standardized, so the notion of a “market-clearing price” isn’t useful. Instead, it takes time and effort for workers to find the right job, and, similarly, employers have to expend time and effort to find the right worker. That is, the labor market is characterized by search and matching frictions. Most of the research of Jim Albrecht and Susan Vroman is focused on developing models to better understand how markets with search and matching frictions work. They have applied their models to several markets, in particular, the labor market and the housing market.
Their most recent published paper focuses on the housing market. Search theory is a natural tool to use to analyze this market: anyone who has bought or sold a house knows that it takes time and effort to find a suitable counterpart on the other side of the market. When a house is listed for sale, the seller posts an asking price. Sometimes houses sell at a price below the asking price, sometimes above, and often exactly at the asking price. What role does the asking price play in the housing market, and, more generally, how are sale prices determined in this market? In “Directed Search in the Housing Market,” published in the Review of Economic Dynamics earlier this year, Albrecht and Vroman, together with co-author Pieter Gautier, analyze these questions by constructing a model in which they assume that sellers have limited commitment to the posted asking price. Commitment is limited in the sense that if only one buyer makes a bona fide offer at the asking price, the seller is obliged to sell at that price. This commitment is typically written into contracts with sellers’ agents. However, if more than one buyer offers the asking price, the sale price can be bid above the posted level, and, of course, if the only the offers received are below the asking price, the seller is free to accept or reject the highest of these. In addition to helping understand the pricing patterns that we see in the data, the model explains how the asking price can signal seller “motivation,” that is., how eager the seller is to make a deal. Buyers observing a variety of asking prices will direct their search so that their expected benefit is the same regardless of the price. Buyers know that there is a tradeoff: a low asking price is appealing but it appeals to many buyers so that the chance of being the highest bidder is small whereas bidding on a house with a high asking price likely means paying more if the buyer has the winning bid but having a greater chance of winning. In equilibrium, Albrecht, Gautier and Vroman find that asking prices can indeed signal seller motivation and that prices draw more buyers to the more motivated sellers, an efficient outcome.
Albrecht and Vroman, together with co-author Bruno Decreuse, are currently working on a search-theoretic model of the labor market in which “phantom” job vacancies affect the rate at which the unemployed find jobs. Phantoms are ads for jobs that have been already filled but not yet removed from online job sites like Craigslist and Monster.com. The unemployed direct their search (decide which listings to pursue) taking the fact that older listings are more likely to be phantoms into account. Albrecht, Decreuse and Vroman show that workers over-apply to relatively new job listings – “over-apply” in the sense that if workers could coordinate their search activity, they would choose to direct more applications towards older listings. The authors also show that phantoms are quantitatively extremely important. In a calibrated version of their model, phantoms account for a substantial fraction of unemployment and an even larger percentage of search frictions.