"Search Frictions, Network Effects and Spatial Competition: Taxis versus Uber" (Job Market Paper) Download
In this paper, I model the search and matching process among drivers and passengers of taxis and Uber cars in New York City to analyze the matching efficiency taking into account network effects and supply competition. Drivers make dynamic spatial search decisions to supply rides and passengers make static discrete choice decisions among taxi and Uber. I model network effects by adding demand and supply levels to both sides’ decisions in the form of a matching probability or waiting time. I use the nonstationary oblivious equilibrium concept to solve the model and analyze equilibrium frictions as mismatches between empty cars and waiting passengers. I show that network effects and supply competition, in addition to fixed fare and search costs, have extensive effects on frictions and welfare in three counterfactual scenarios. The first decreases Uber supply by 30%, the second improves traffic conditions and the third eliminates the Uber surge multiplier. I find that taxis’ pickups increase by 5.9% if traffic improves but do not increase significantly under supply regulation. Taxis’ profits increase by 1% under supply regulation and increase by 5.18% if traffic improves. Uber’s search friction increases after eliminating the surge multiplier or restricting supply. Consumer welfare decreases if Uber supply is restricted. Without network effects, search frictions and pickups will be underestimated.
"Vertical Relationship and Merger Effects in the US Beer Industry" Download In this paper, I study the MillerCoors joint venture of 2008 in the U.S. beer industry. In particular, I focus on impacts of the cost efficiency (in terms of shipping distance and production) and increased market power of this merger and importantly how they are affected by the vertical market structure. With vertical relationship, the upstream shock does not fully pass through to downstream retail price because both upstream and downstream firms will adjust markups to the shock. Thus, merger analysis in the beer industry depends on concentration of both upstream and downstream markets. I use random coefficient model to estimate demand for beer and price elasticities. In the supply side, I model the double marginalization problems of beer retailers and brewers by assuming linear pricing contracts between upstream and downstream firms. Cost saving of the merger is estimated by comparing pre- and post-merger implicit marginal costs. I simulate the two markups in the post-merger period without joint venture of Miller and Coors to quantify and disentangle the merger effects. I find that average cost saving of producing a 12 oz serving is 2 cents for Coors light and 1.6 cents for Miller lite. Cost saving through shipping distance is at most 7.4 cents for Coors and 2.2 cents for Miller. Market power of MillerCoors increases brewer's markup. However, retailer's markup decreases to mitigate the impact on retail price especially for more concentrated downstream markets.
Work in progress
"Network Externalities to Entry in Platform and Two-sided Markets"
"Network Effects and Dynamic Pricing in the U.S. Digital Game Market"