The paper I chose is: An Empirical Model of Search with Vertically Differentiated Products" by Wildenbeest. The goal of this paper is to simultaneously account for two sorts of price dispersion that are prominent in the data: price fluctuations over time within firm, and differences in average price levels across firms. Price fluctuations can be explained by firms playing mixed strategies in a game with search frictions, while differences in price levels can be explained by product differentiation; Wildenbeest notes that no one has yet combined the two. On the one hand, search models that generate price fluctuations have assumed homogeneous firms, precluding differences in average price levels. On the other hand, models in which prices are simply determined by differing product characteristics do not generate frequent price fluctuations over time. By creating a search model in which firms are heterogeneous, Wildenbeest is able to account for both facts at once. Search costs may be dramatically overestimated when heterogeneity is ignored, if heterogeneity truly is driving a large fraction of the price variation. Thus, with his model in hand, Wildenbeest is in a position to more accurately gauge the importance of search frictions versus heterogeneity. He takes this empirical question to supermarket pricing data. I don't want to go further on the empirics since that's not the part that sells the best.
Notably, the Wildenbeest model enables us to estimate search costs and heterogeneity over product quality using only pricing data. In the Wildenbeest model, consumers are identical other than their search costs, and products are vertically differentiated over quality. Firm quality is fixed over time, so by selecting p, the firm is essentially choosing the utility level, that a consumer obtains by buying their product. In Wildenbeest's equilibrium, firms are symmetrically randomizing over u according to some common density. By observing the distribution of prices over time, we can infer this distribution of utilities.
The paper motivated me to intoduce price differentiation into a search model with a mixed strategy equilibrium. One of the puzzles in search model is why consumers stop searching early. There have been many explanations proposed, for example advertisement, product differentiation etc. However, there are also many empirical evidence against these models. My proposition is that we could add quality-adjusted and risk-averse utility into the model. Similar to players stopping early in a gamble, consumers may also be motivated to stop early because of the search cost. The proposal include several parts. First, I need to extend the model. Second, prove that an equilibrium or mixed equilibrium exists under the setting. Third, take the model to a highly identical product market and test the results. Some of the empirical models that I can follow on examining search cost include De los Santos, Babur, Hortaçsu, and Wildenbeest (2012), which uses online purchase data, and Hortaçsu and Syverson (2004) which explores the highly identical mutual fund data.
Wildenbeest, Matthijs R. "An empirical model of search with vertically differentiated products." The RAND Journal of Economics 42.4 (2011): 729-757.
De los Santos, Babur, Ali Hortaçsu, and Matthijs R. Wildenbeest. "Testing models of consumer search using data on web browsing and purchasing behavior." American Economic Review 102.6 (2012): 2955-80.
Hortaçsu, Ali, and Chad Syverson. "Product differentiation, search costs, and competition in the mutual fund industry: A case study of S&P 500 index funds." The Quarterly Journal of Economics 119.2 (2004): 403-456.
The paper I chose is: An Empirical Model of Search with Vertically Differentiated Products" by Wildenbeest. The goal of this paper is to simultaneously account for two sorts of price dispersion that are prominent in the data: price fluctuations over time within firm, and differences in average price levels across firms. Price fluctuations can be explained by firms playing mixed strategies in a game with search frictions, while differences in price levels can be explained by product differentiation; Wildenbeest notes that no one has yet combined the two. On the one hand, search models that generate price fluctuations have assumed homogeneous firms, precluding differences in average price levels. On the other hand, models in which prices are simply determined by differing product characteristics do not generate frequent price fluctuations over time. By creating a search model in which firms are heterogeneous, Wildenbeest is able to account for both facts at once. Search costs may be dramatically overestimated when heterogeneity is ignored, if heterogeneity truly is driving a large fraction of the price variation. Thus, with his model in hand, Wildenbeest is in a position to more accurately gauge the importance of search frictions versus heterogeneity. He takes this empirical question to supermarket pricing data. I don't want to go further on the empirics since that's not the part that sells the best.
Notably, the Wildenbeest model enables us to estimate search costs and heterogeneity over product quality using only pricing data. In the Wildenbeest model, consumers are identical other than their search costs, and products are vertically differentiated over quality. Firm quality is fixed over time, so by selecting p, the firm is essentially choosing the utility level, that a consumer obtains by buying their product. In Wildenbeest's equilibrium, firms are symmetrically randomizing over u according to some common density. By observing the distribution of prices over time, we can infer this distribution of utilities.
The paper motivated me to intoduce price differentiation into a search model with a mixed strategy equilibrium. One of the puzzles in search model is why consumers stop searching early. There have been many explanations proposed, for example advertisement, product differentiation etc. However, there are also many empirical evidence against these models. My proposition is that we could add quality-adjusted and risk-averse utility into the model. Similar to players stopping early in a gamble, consumers may also be motivated to stop early because of the search cost. The proposal include several parts. First, I need to extend the model. Second, prove that an equilibrium or mixed equilibrium exists under the setting. Third, take the model to a highly identical product market and test the results. Some of the empirical models that I can follow on examining search cost include De los Santos, Babur, Hortaçsu, and Wildenbeest (2012), which uses online purchase data, and Hortaçsu and Syverson (2004) which explores the highly identical mutual fund data.