Selected research papers

Competition Policy in Light of Innovation, with David Rahman

Horizontal mergers impact welfare along several dimensions: (i) they increase deadweight losses arising from market power, (ii) positive synergies decrease costs post-merger, (iii) thanks to higher expected profits, increased market power as a result of mergers encourages innovation along the extensive margin, (iv) by altering the competitive landscape, mergers affect productivity improvements along the intensive margin, and (v) mergers decrease the deadweight loss from replication of research. In this paper, we construct and calibrate a dynamic general equilibrium model that incorporates these five dimensions where mergers affect welfare. We study the effects of different types of merger rules applied by a merger authority for all industries in the economy. Our model suggests that merger policy mainly affects welfare by changing the trajectory of productivity growth.

Symbiotic Competition and Intellectual Property (online appendix), with David Rahman

According to Nordhaus, the optimal life of a patent T* marries motivating innovation with the “embarrassment” of monopoly, given that imitators would copy unpatented inventions and thereby reduce incentives to innovate. We extend this argument through a general equilibrium growth model with follow-on innovations, knowledge spillovers, and imperfectly elastic labor supply. Our calibration to the US yields T* between 8 and 14 years, much lower than the current global standard of 20. Decomposition of T* suggests that knowledge spillovers are quantitatively as important as market power, each reducing the optimal life of a patent by about 20 years. We also find that optimal patent policy depends significantly on the distribution of spillovers across industries. 

Price Dispersion in Dynamic Competition

In product markets, substantial price dispersion exists for transactions of physically identical goods. Moreover, in these markets, incumbent firms sell at higher prices than entrants. This paper presents a theory of price formation under dynamic competition that explains these facts by assuming both that consumers have imperfect access to firms and that their degree of access depends on each firm's sales history. The model has a unique equilibrium that features randomized pricing strategies, with incumbents always posting higher prices than entrants. For a fixed underlying environment, the equilibrium converges to a stationary equilibrium over time. As firms' entry and exit rates approach zero, this stationary equilibrium converges to perfect competition.

Market Microstructure and Informational Efficiency: The Role of Intermediation, with Brian Albrecht

The competitive market is informationally efficient; people only need to know prices to implement a competitive allocation. However, the standard formulation of competitive markets assumes that prices are not set by strategic agents but by "supply and demand" and thus neglects the underlying role of market microstructure. We show that if prices are determined by strategic agents, then intermediation is necessary for markets to achieve informational efficiency. We study two specific market microstructures: a model where trade is intermediated by market-makers and a model of random matching and bargaining. First, we show that an economy where competition among market-makers determines prices can approximate the informational efficiency of the competitive model. Second, we show that as the complexity of the economy increases, matching markets require infinitely more information than the competitive market.

The Economics of Ancient Mediterranean Slavery, with Walter Scheidel

This paper investigates the economic aspects of slavery in the Ancient Greco-Roman world. Existing evidence reveals significant variation in the relative cost of slaves compared to unskilled wages: it appears that at different times and places, a typical slave could be purchased for prices equivalent to wages paid from 150 to 1000 days of unskilled labor. To explain this great disparity, we develop a principal-agent model that predicts the return on slaves relative to wages, which varies as a function of the prevalence of slavery in the labor force. This model implies that slavery may have increased aggregate labor productivity by reallocating workers from less productive to more productive regions within the Greco-Roman world.

Imperfect Competition as a Result of Unawareness

This paper develops a dynamic model of price competition where buyers have constrained consideration sets due to unawareness. There are two sellers: an incumbent, who is initially more well-known among buyers, and an entrant. Awareness is influenced by word-of-mouth: if more buyers choose to shop at a seller, unaware buyers are more likely to discover that seller. In the unique equilibrium, both sellers randomize their pricing strategies, but one seller posts higher expected prices than the other. I show that if the incumbent's present actions can change the future state of the market to a high enough degree, he has a strong incentive to undercut the entrant. Thus, this model provides microfoundations to the concept of "advantage denying" motive and relates it to the empirical finding that it takes time for a seller's demand to grow.

The Minimum Wage as an Instrument for Social Insurance, with Keyvan Eslami

We consider the welfare effects of a minimum wage policy from a public finance perspective. In a competitive economy without labor frictions, we show that the minimum wage is an effective redistribution tool that cannot be replicated by a tax and transfer system, reiterating previous results. However, by incorporating an explicit model of wage formation into this environment via a random matching and bargaining model, we show that this result no longer holds. In such a setting, a tax on entrepreneurs and lump-sum rebates to employed workers can achieve the exact same redistribution effects as increasing the minimum wage above the stationary equilibrium wage. In this case, the minimum wage must exist in an optimal redistribution policy only if it can be implemented at strictly lower costs than a tax and transfer system.