UC Berkeley

In two essays, I present my work that applies empirical method to real-world problems in quantitative marketing and economics. In the first chapter, I use structural econometrics to investigate airlines' dynamic price competition. I show how widely-used dynamic pricing techniques affect firms profits under a competitive equilibrium. The second chapter is a joint work with Zemin Zhong. We focus on how China's anticorruption effort impact its economy, measured by car consumption and new business registration.

Dynamic pricing is becoming a common practice in many industries, but its effect under competition is uncertain due to the potential for the Prisoner's dilemma. The paper studies profit and welfare implications of competitive dynamic pricing in the context of the airline industry. The paper develops a structural dynamic oligopoly model where firms compete in selling limited capacities when facing demand fluctuations. The supply and demand are jointly estimated using a unique daily-level data on airfares and capacity utilization. The identification leverages a natural experiment of carrier exit. The estimates show that air travel demand exhibits a large degree of temporal heterogeneity and stochastic variability. The counterfactuals show that the ability to perform dynamic pricing increases total welfare. In particular, (i) price discrimination (charging late-arriving consumers higher prices) softens competition in the late market and increases profits substantially and (ii) revenue management (pricing on remaining capacities) intensifies competition and does not increase profits.

Corruption could either benefit economic growth by “greasing the wheel,” or distort supply of public goods and create inefficiency. Empirically testing the impact of corruption is difficult due to its evasive nature. We take an alternative approach by investigating the economic impacts of anti-corruption policies. We focus on China's recent anti-corruption campaign, the largest of its kind in recent history. As an important initiative of this campaign, the Communist Party's Provincial Committees of Discipline Inspection (PCDI) send inspector teams to investigate municipal governments for potential corruption. The variation in their timing allows us to use a difference-in-difference design to identify their impact on local economy. Using two unique administrative datasets of vehicle and business registration, we find that PCDI visits have a negative impact on both car sales and new business entry. For vehicles, the effect is surprisingly uniform across different price tiers: Luxury brands exhibit a similar drop as domestic brands, suggesting corruption's impact permeates households across a wide income spectrum. Over time, the effect is strengthening: We observe a 2\% drop in the first three months of PCDI visit and a 10\% drop one year afterward. The especially large impact cannot be explained by the decline in government officials' consumption behavior, suggesting anti-corruption efforts also affect the private sector. We test the idea using business registration data, and we found PCDI visits indeed discourage new business registration. We validate our empirical strategy by showing that (1) the timing of PCDI visits cannot be predicted by observable county characteristics and (2) car registrations exhibit parallel pre-treatment trends. Our results suggest there may be a trade-off in anti-corruption and economic growth.