Abstract
We revisit the linear Cournot model with uncertain demand that is studied in Lagerlöf (2006) and provide sufficient conditions for equilibrium uniqueness that complement the existing results. We show that if the distribution of the demand intercept has the decreasing mean residual demand (DMRD) or the increasing generalized failure rate (IGFR) property, then uniqueness of equilibrium is guaranteed. The DMRD condition implies log-concavity of the expected profits per unit of output without additional assumptions on the existence or the shape of the density of the demand intercept and, hence, answers in the affirmative the conjecture of Lagerlöf (2006) that such conditions may not be necessary.
Introduction
In industrial organization, the Cournot model serves as a fundamental framework for analyzing competition among firms choosing production quantities. While the deterministic linear Cournot model typically yields a unique Nash equilibrium, the introduction of demand uncertainty—where the demand intercept is a random variable—complicates the mathematical landscape. Such uncertainty can lead to a convex expected demand curve, potentially resulting in multiple stable equilibria. This multiplicity creates ambiguity for market participants and regulators alike.
The Problem of Multiplicity
When demand is stochastic, the "demand intercept" (the maximum price consumers are willing to pay) is no longer a fixed point but a distribution. Lagerlöf (2006) identified that this uncertainty could cause the expected price to decrease slowly as total supply increases, leading to non-concave profit functions and multiple equilibria. While Lagerlöf established initial conditions for uniqueness, they relied on restrictive assumptions, such as specific properties of the probability density at zero. This research aims to generalize these conditions, providing a more robust theoretical foundation for equilibrium uniqueness.
Methodology and Theoretical Framework
The study adopts a purely analytical approach, leveraging concepts from reliability theory and stochastic ordering. The focus is on the properties of the probability distribution of the demand intercept ($\alpha$). Two primary classes of distributions are analyzed:
- Decreasing Mean Residual Demand (DMRD): This property implies that if demand is known to exceed a threshold $x$, the expected additional demand beyond $x$ decreases as $x$ increases.
- Increasing Generalized Failure Rate (IGFR): This is a generalization of the hazard rate, often used to model failure probabilities in engineering, applied here to the distribution of demand.
By examining the First Order Conditions (FOC) of firms' profit-maximization problems, the authors define a function $\Lambda(x)$ whose roots correspond to equilibria. They prove that under DMRD or IGFR conditions, this function is monotonic, ensuring a single crossing of the zero-axis and thus a unique equilibrium.
Key Insights: The Hierarchy of Distributions
A significant contribution of this work is the mapping of relationships between various statistical properties. The authors demonstrate that the commonly used "Increasing Failure Rate" (IFR) property is a subset of both DMRD and IGFR.
As shown in the figure above, the results for DMRD and IGFR cover a much broader range of market scenarios than previous models restricted to IFR. Furthermore, in the special case of zero marginal costs (relevant for digital goods), the even broader Decreasing Generalized Mean Residual Demand (DGMRD) condition is sufficient for uniqueness.
Results and Implications
The paper presents two core theorems that resolve long-standing questions in the field:
- Theorem 3.2: Formally proves that the symmetric Nash equilibrium is unique if the demand intercept distribution is either DMRD or IGFR.
- Theorem 3.5: Shows that the DMRD condition alone guarantees the log-concavity of expected profits per unit. This result is particularly powerful because it holds even if a density function does not exist, confirming Lagerlöf’s conjecture that certain technical restrictions on density were unnecessary.
These findings suggest that equilibrium uniqueness is the standard outcome for most distributions used in economic modeling, as the IGFR class is exceptionally broad.
Conclusion
By applying reliability theory to Cournot competition, this research provides expansive sufficient conditions for equilibrium uniqueness under demand uncertainty. The identification of DMRD and IGFR as key properties simplifies the theoretical requirements and broadens the applicability of the model to real-world scenarios involving heavy-tailed distributions or non-standard densities. Future research may extend these unimodality conditions to asymmetric cost structures or more complex vertical market interactions.
Summary updated