Article based on research by Anna Bayona
A study into the bidding behaviour of traders in wholesale commodity markets has important results for competition authorities, policy makers and researchers.
Anna Bayona, Assistant Professor at the Esade Department of Economics, Finance and Accounting, along with co-authors Jordi Brandts (lnstitut d'Anàlisi Econòmica and Barcelona GSE) and Xavier Vives (IESE Business School), investigated the link between information frictions and market power in a laboratory-based study. The results document causal relations between information frictions and market power, or the ability of a company to influence market price.
"We can connect our results to policy considerations for natural markets like the wholesale electricity markets or Treasury bond auctions," say the authors. "In our laboratory setting, collusion is excluded so our results suggest that high prices in natural markets point to the existence of some kind of collusion."
The findings led the study's authors to conclude that competition authorities may want to fine-tune their collusion screening processes by taking into account the information structure of the market, and especially cost correlation, in markets that are characterised by supply function competition.
The study was conducted in the form of a laboratory experiment, but in a setting related to real-world markets. The experiment examined competition in demand or supply schedules such as wholesale electricity markets, markets for pollution permits, or liquidity and treasury auctions.
"Understanding which factors contribute to the existence of market power is a central issue," say the researchers. "This is an area in which experimental economics can contribute. Specifically, in a laboratory setting we can vary different elements of the information available so that we can study the impact on market power."
When costs are interdependent, the supply function equilibrium predicts a higher degree of market power
The design of the study included a wide range of competitive environments where traders compete in scenarios with private information, including cost correlation (the statistic that measures how two costs move in relation to each other), and private information (information that is specific to each seller).
When costs are interdependent, the supply function equilibrium predicts a higher degree of market power than when costs are uncorrelated. When costs are correlated, the theory suggests that a seller will take a high price to mean the price of rivals is also high, and should compete less aggressively. But, as costs become more correlated, price becomes more prominent in the inference of costs and equilibrium supply functions become steeper, leading to greater market power.
"Understanding whether high observed market power in natural markets is due to an equilibrium or not is import for policy," explain the researchers. "A competition authority could mistakenly infer collusion from high margins in the wholesale electricity market, when generators were actually bidding non-cooperatively, considering the information conveyed by the price. Or, similarly, the Treasury may suspect collusion in a bond auction, when in fact, bidders are responding optimally to the incomplete information environment they face.
"It is therefore crucial to test whether the theoretical predictions are borne out in the laboratory, as mitigating market power is a primary concern of regulators."
In the study, "sellers" competed for several rounds, each being given a differing private signal that simulated naturally occurring environments. After all the decisions were made, a uniform market price was calculated, and each "seller" received detailed feedback about performance, market price, and the behaviour and performance of rivals in the same market.
Average market power is not statistically different between uncorrelated and interdependent costs
"Our central result is that average market power is not statistically different between uncorrelated and interdependent costs," say the researchers. "Consistent with the predictions of the supply function equilibrium, we first find that average market power in markets with uncorrelated costs is close to that expected by the supply function equilibrium. However, market power when costs are interdependent is far below that expected by the supply function equilibrium.
"This suggests that when costs are interdependent, participants are not sufficiently sophisticated with respect to the information structure and do not make inferences from the market price."
The experiment suggests additional questions for future research, both experimental and theoretical.
"An important extension of our work would be to allow bidders in the experiment to submit non-linear supply functions," say the study’s authors. "Additional future work could explore mechanisms by which participants learn how to improve information extraction from the price, for example by asking participants to come back to the laboratory a few days later as "experienced bidders," or by replicating the experiment with professional traders.
"Our experimental findings also call for the development of theoretical models that analyse market competition among participants who exhibit various degrees of sophistication in markets that are characterised by supply function competition and private information."
Original research article: Bayona, A., Brandts, J. & Vives, X. Information frictions and market power: a laboratory study, Games and Economic Behavior, 122, 354-369 (2020)
Join the Do Better community
Become a member and enjoy our free benefits. Get recommendations, receive personalised content in your inbox and save your favourite articles to read later.