Abstract
We offer a parsimonious model to investigate how strategic wind producers sell energy under stochastic production constraints, where the extent of heterogeneity of wind energy availability varies according to wind farm locations. The main insight of our analysis is that increasing heterogeneity in resource availability improves social welfare, as a function of its effects both on improving diversification and on reducing withholding by firms. We show that this insight is quite robust for any concave and downward-sloping inverse demand function. The model is also used to analyze the effect of heterogeneity on firm profits and opportunities for collusion. Finally, we analyze the impacts of improving public information and weather forecasting; enhanced public forecasting increases welfare, but it is not always in the best interests of strategic producers.
Introduction: The Stochastic Nature of Wind Markets
Traditional economic models of electricity markets often assume production is deterministic, a characteristic of fossil-fuel generators. However, the rise of renewable energy introduces a new variable: stochasticity. In wind and solar markets, energy availability acts as a "private signal"—producers have a better sense of their own potential output than their competitors or regulators do.
This paper addresses a critical tension in market design: the relationship between geographic dispersion and market power. When wind farms are physically close, their outputs are highly correlated. This shared information allows strategic producers to anticipate their rivals' states and potentially withhold production to inflate prices. We investigate how the level of correlation (stochastic dependence) between producers affects strategic bidding and overall social welfare.
Modeling Strategic Competition
The authors model this competition as an incomplete information Cournot game with correlated types. Strategic wind producers, operating with negligible marginal costs, face either a "High" ($H$) or "Low" ($L$) production state. A key innovation of the model is the dispersion parameter $d$:
- Low $d$: Wind availability is highly correlated (farms are geographically "close").
- High $d$: Wind availability is independent (farms are "far").
In the derived symmetric Bayesian Nash Equilibrium (BNE), producers decide their output based on their own realized state and their beliefs about their competitor's state.
Empirical Motivation
To ground the theoretical model, the authors analyze hourly energy production data from the Midcontinent Independent System Operator (MISO) region (2014–2016). This data confirms that wind availability ranges from highly correlated to nearly independent depending on the geographic distance between sites.
As shown in the figure, the distributions of energy production vary significantly, justifying the use of discrete states ($H$ and $L$) to capture the underlying stochasticity of the resource.
Key Insights: Welfare and the Information Channel
The research highlights that heterogeneity affects welfare through an "information channel." In highly correlated markets, a producer in a "High" state can safely assume their rival is also in a "High" state, which encourages aggressive withholding. Increasing dispersion "muddies" this signal. When firms are less certain about their rivals' output, they are forced to compete more vigorously, leading to higher total output and lower prices.
Social welfare benefits from a "double-whammy":
- Wind Diversification (WD): The standard benefit of reducing the variance of total supply.
- Strategic Curtailment (SC): A reduction in the incentive for firms to strategically withhold production.
The Profit Paradox and Forecasting
While social welfare strictly increases with dispersion, producer profits do not follow the same trend. If a wind farm's "Low" state remains relatively productive, producers may actually prefer high correlation. This allows them to maintain high prices through behavior that mimics coordination, even without explicit collusion.
Furthermore, the paper explores the role of public weather forecasting. While better forecasting is always socially beneficial because it reduces uncertainty, it can actually harm producer profits. By removing the "informational advantage" that allows for strategic withholding, public forecasts can force producers into more competitive behavior.
Conclusion and Future Work
The findings suggest that private investment in wind energy may naturally lead to sub-optimal levels of resource heterogeneity, as producers might favor clustering to maintain market power. Policy interventions could be used to encourage geographic dispersion to maximize public benefit.
Future research could extend this model to include day-ahead markets, transmission congestion, and more continuous representations of wind availability to further refine our understanding of strategic behavior in renewable-heavy grids.
Summary updated