Competition and Efficiency in Congested Markets

Daron Acemoğlu, Asuman Ozdaglar

Abstract

We study the efficiency of oligopoly equilibria in congested markets. The motivating examples are the allocation of network flows in a communication network or of traffic in a transportation network. We show that increasing competition among oligopolists can reduce efficiency, measured as the difference between users' willingness to pay and delay costs. We characterize a tight bound of 5/6 on efficiency in pure strategy equilibria. This bound is tight even when the number of routes and oligopolists is arbitrarily large. We also study the efficiency properties of mixed strategy equilibria.

Introduction

In most economic markets, competition is the gold standard for achieving efficiency. However, in congested networks—such as transportation systems or data routing infrastructures—increasing competition can paradoxically lead to a decrease in social welfare. This phenomenon arises due to "congestion externalities," where an individual's use of a resource imposes delays on all other users. While previous research has focused on "selfish routing" in the absence of prices, this paper investigates the impact of profit-seeking providers who own network links and set prices. We explore whether competition between these providers mitigates or exacerbates the inefficiencies inherent in congested systems.

The Pricing-Congestion Model

The study employs a game-theoretic framework to model a two-stage pricing-congestion game. In the first stage, multiple service providers (oligopolists) set prices for the links they control to maximize their individual profits. In the second stage, a large population of users chooses routes based on the effective cost, which is the sum of the link price (toll) and the latency (delay).

A typical parallel-link network where users choose between routes with different latency functions $l_i(x)$.

Users are assumed to follow Wardrop’s principle, a Nash equilibrium for traffic where no user can reduce their cost by unilaterally switching routes. The model assumes users have a common reservation utility $R$, meaning they will only use the network if the effective cost is less than or equal to $R$.

The aggregate utility function $u(x)$ represents the total willingness to pay for $x$ units of traffic, where $d$ is the total demand and $R$ is the reservation utility per user.

Efficiency is measured by the Social Surplus, defined as the total willingness to pay minus the total delay costs.

The Paradox of Competition

A central finding of this research is the "Monopoly Paradox." In this model, a single monopolist owning the entire network achieves 100% efficiency. To maximize profit, the monopolist sets prices that function as "Pigovian taxes," forcing users to internalize the social cost of the congestion they create. This aligns private incentives with social optimality.

Conversely, when multiple providers compete (an oligopoly), this incentive vanishes. If one provider raises their price, traffic shifts to a competitor's link, increasing congestion there. This makes the competitor's route less attractive, effectively increasing the first provider's market power over their remaining users. This "lock-in" effect allows providers to distort prices away from the socially optimal levels.

Key Results and Efficiency Bounds

The paper establishes several critical results regarding the efficiency of these markets:

  • Monopoly Efficiency: A single provider reaches the social optimum by internalizing congestion costs.
  • Competition vs. Monopoly: Moving from a monopoly to a duopoly typically reduces social surplus due to the loss of the Pigovian pricing incentive.
  • The 5/6 Bound: For networks of parallel links with linear latency functions, the authors prove a tight efficiency bound of 5/6 (approximately 83.3%). This means that even in the worst-case competitive scenario, the market captures at least 5/6 of the maximum possible social surplus. This bound remains tight regardless of the number of competitors or routes.
  • Equilibrium Complexity: While pure strategy equilibria always exist for linear delays, they may disappear with highly convex latency functions. In such cases, mixed strategy equilibria can lead to significantly worse efficiency outcomes.

Limitations and Future Directions

The current model assumes homogeneous users with identical reservation utilities and delay sensitivities. Real-world scenarios often involve heterogeneous users with varying values of time and budgets. Additionally, while the 5/6 bound is proven for parallel links, extending these results to more complex mesh topologies remains an area for future research. Finally, the behavior of average efficiency in mixed strategy equilibria is conjectured to follow the 5/6 bound, but a formal proof remains an open challenge.

Summary updated

Abstract

We study the efficiency of oligopoly equilibria in congested markets.The motivating examples are the allocation of network flows in a communication network or of traffic in a transportation network.We show that increasing competition among oligopolists can reduce efficiency, measured as the difference between users' willingness to pay and delay costs.We characterize a tight bound of 5/6 on efficiency in pure strategy equilibria.This bound is tight even when the number of routes and oligopolists is arbitrarily large.We also study the efficiency properties of mixed strategy equilibria.
BibTeX
@misc{acemoglu2005competition,
  author = {Daron Acemoğlu and Asuman Ozdaglar},
  title = {{Competition and Efficiency in Congested Markets}},
  year = {2005},
  doi = {10.2139/ssrn.676662},
  howpublished = {SSRN},
}
Figure