Essays on the Industrial Organization of Newly Legalized Markets

Escudero, Manuel Matias

doi:https://doi.org/10.21985/n2-5xt2-dm57

Work

Essays on the Industrial Organization of Newly Legalized Markets

Public

Download PDF

Conventional methods in industrial organization assume that firms are strategically sophisticated and set prices as best responses to their competitive environment. In the first two chapters of this dissertation, I use a detailed dataset of retail and wholesale prices from the newly legalized cannabis industry in Washington state to show that firms are instead widely using fixed markup rules. Such pricing rules generate lower variable profits than those that would be observed under the traditional assumptions of full information and optimal behavior. In Chapter 1, I show that almost half of all units sold in this market at the retail level have prices that can be explained by simple rules of thumb, such as setting a 50% markup over price. I also document a robust convergence process toward this 50% markup rule, and show that the same process is observed across heterogeneous retailers, a diverse set of products, and many different competitive environments. Additionally, I find that the effect of entry shocks on average prices is economically negligible. While my observations adhere to the conventional wisdom in the retail industry, these findings cast doubt on the idea that firms learn to price to exploit the specific differences across products and markets. In Chapter 2, I assess how the observed pricing behavior in this market fares vis a vis fully-informed, optimal behavior. By estimating a discrete choice model of demand for differentiated cannabis products and the simulation of counterfactual scenarios, I show that the simple markup rules used by retailers lead them to capture about 75\% of the variable profits they could have obtained if they were pricing optimally in a full information Bertrand-Nash pricing game. Additionally, I find that, although reasonable in values, the marginal costs that rationalize observed prices and shares as a Bertrand-Nash equilibrium outcome differ from observed acquisition cost data, casting doubt on the match between modeling assumptions and retailers' true pricing decisions. Finally, in Chapter 3 I explore the relationship between product sampling and commercial links in the context of the upstream market for recreational cannabis in Washington state. In particular, this paper focuses on the early stages of this market to understand the role product samples have in shaping the network of commercial links at the upstream level. Because the sets of players on each side of the upstream market were completely new, product samples can be understood as a signaling device from which buyers are able to learn and update their priors regarding the idiosyncratic quality of a given seller. I analyze the importance of product samples as a means to establish profitable relationships between manufacturers and retailers. Using data from this market and an instrumental variables approach, I am able to document that manufacturers do indeed use product samples as a means to approach new retailers and that this marketing strategy is able to increase the likelihood of new commercial links that are, in expectation, larger in monetary value, longer lasting in time, and that involve more frequent trade.

Creator