Imperfect tacit collusion and asymmetric price transmission

Muhammed Bulutay,David Hales,Patrick Julius,Weiwei Tasch

doi:10.1016/j.jebo.2021.10.018

Abstract

We investigate asymmetric price transmission (APT) in laboratory experiments and find that imperfect tacit collusion is likely the cause in our otherwise frictionless markets. We vary the number of sellers across markets to evaluate the role competition plays in APT. We report similar magnitudes of asymmetry in markets with 3, 4, 6, and 10 sellers, but not in duopolies. Furthermore, sellers consistently set their prices above the best-response levels implied by their forecasts, particularly in periods following negative shocks. We interpret these pricing deviations as sellers’ intentions to collude, and note that they mechanically drive the pricing asymmetries we observe.

Highlights

The phenomenon of Asymmetric Price Transmission (APT), that is, that supplier prices rise quickly after positive input cost shocks, but fall relatively more slowly after -sized negative cost shocks, has been repeatedly documented in the literature such that we can rightly describe it as a stylized fact.1 while empirical evidence for the APT phenomenon is ample, identification of its causal forces is not settled
Our results point to the co-appearance of asymmetric price transmission and tacit collusion
The latter seems to be the result of strategic behavior, as our analysis of deviations from best-response action reveals. These findings are consistent with theories that cast tacit collusion as having a significant role in the emergence of APT, such as the trigger price model in Borenstein et al (1997)

Summary

Introduction

The phenomenon of Asymmetric Price Transmission (APT), that is, that supplier prices rise quickly after positive input cost shocks, but fall relatively more slowly after -sized negative cost shocks, has been repeatedly documented in the literature such that we can rightly describe it as a stylized fact. while empirical evidence for the APT phenomenon is ample, identification of its causal forces is not settled. Empirical studies of APT predominantly examine aggregate-level variables (e.g. inflation, concentration) proposed to be relevant in the theory literature The focus on such variables occurs because firm-level determinants are either not directly observable, or are not adequately measurable in panel data form. Rapidly raising prices in response to an upward cost shock poses no such threat to one’s competitors, and incurs no corresponding risk of retaliation Their arguments are sound, and consistent with a deep empirical literature finding correlations suggestive of tacit collusion, Borenstein et al (1997) conclude that they are unable to conclusively draw support for this hypothesis from their data. As no other empirical study of which we are aware has accomplished this either, we find motivation to turn to the laboratory to examine the role of tacit collusion in driving APT dynamics.

Objectives

Methods

Results

Conclusion