We document that localized policies aimed at mitigating climate risk can have unintended consequences due to regulatory arbitrage by firms. Using a difference-in-differences framework to study the impact of the California cap-and-trade program with U.S. plant-level data, we show that financially constrained firms shift emissions and output from California to other states where they have similar plants that are underutilized. By contrast, unconstrained firms do not make such adjustments. Overall, unconstrained firms do not reduce their total emissions, whereas constrained firms increase their total emissions after the cap-and-trade rule, undermining the effectiveness of the policy.