Dissertation: Three Chapters on Markets with Frictions

Market Liquidity and Inventory Cycles (Job Market Paper)

[paper]

Abstract

Using U.S. data, I document that inventory investment is pro-cyclical, the inventory-sales (I/S) ratio is counter-cyclical, and aggregate markup is pro-cyclical. However, existing models fail to fully explain these empirical patterns. To address this gap, I introduce a model where inventory holdings are driven by search friction and influenced by market liquidity, defined as the trade-off between markup and the speed of sales. In this model, sellers stock goods and post prices, while buyers select which sellers to visit. Due to a lack of coordination among buyers, sales are stochastic, often resulting in leftover inventory. This friction introduces a new incentive for sellers to hold inventory: to improve buyer experience by offering better terms of trade. However, this incentive leads to a negative externality. In the calibrated economy, sellers, on aggregate, overstock by 1.6%, equivalent to 0.2% of annual GDP, while the annual welfare cost amounts to 0.27% of GDP after accounting for distributional effects. Unlike the Walrasian framework, where prices are determined by market-clearing conditions, sellers in my model actively adjust both prices and quantities. This pricing flexibility allows the model to generate predictions that align with observed patterns of inventory cycles.

 

Price Commitment, Reputation and Bargaining in Bilateral Trade

Abstract

Market activities are mostly bilateral. A clear understanding of the pricing mechanisms has nontrivial policy implications. This paper considers a price-posting environment with the option of bargaining upon meeting. Sellers have reputation concerns about price commitment, while buyers take into account potential contingent pricing when bargaining is acceptable. Despite being ex-ante homogeneous, the sellers who accept bargaining post higher prices than other sellers who don’t accept bargaining. Whether we observe bargaining in equilibrium depends on the market conditions. Bargaining happens when (1) trade frequency is low, (2) the market is small, and (3) the cost of liquidity is low. These results explain the observed market behaviors and provide a rationale for model choices in different research contexts.

 

Frictional Markup

Abstract

Markup is neither a sufficient nor a necessary indicator of market impacts. In classical Bertrand competition, firms heavily impact each other, yet the markup is zero. Conversely, in a competitive search environment, no firm has impact on equilibrium, but everyone has a positive markup. To address the question of how much of the actual markup is attributable to market power and search friction, respectively, I model finite-agent competition with two types of search frictions: noisy search and directed search. The former contributes to market power, while the latter sustains markup in a competitive environment. These two types of frictions can be identified from price dispersion and sales volatility. I estimate the model using NielsenIQ data and conduct counterfactual analysis.

 

Work in progress

Pricing and Matching with Frictions and Heterogeneity, with Peter Norman and Randall Wright

Market Order and Interest Rate Parity