The Data Are Not Enough: A Causal Framework for Energy Storage

What is the causal effect of energy storage on grid stability? Current valuation methods cannot answer this question: they establish correlations, not causality. Because storage decisions depend on time-varying factors that also affect grid stability—weather, demand, market prices—conventional estimates are mathematically guaranteed to be biased, regardless of sample size or model complexity. This is not a data problem but an identification problem. Here we present a causal framework that resolves it, formalizing the minimal assumptions under which the effect becomes identifiable. Through simulations where ground truth is known, we show that conventional methods yield bias exceeding 100% while our framework recovers true effects with an average error of just 5.9%. Causal inference in energy storage is not merely a statistical challenge but a mathematical necessity: without explicit structural assumptions, all empirical estimates are uninterpretable. Our framework provides the foundation for evidence-based storage policy.