Conference Agenda

Large-scale infrastructure, such as flood-control dams, is a central strategy for climate adaptation. However, while these projects are engineered to protect downstream assets, the distributional consequences of their adaptation costs remain poorly understood. Using a spatial regression discontinuity design (RDD) on geocoded manufacturing firms across China, I document a stark asymmetry in adaptation outcomes. First, I confirm the physical efficacy of this infrastructure: dams successfully decouple downstream production from upstream rainfall shocks, effectively mitigating the hydrological propagation of extreme weather. However, this protection generates unequal economic returns. Firms immediately upstream exhibit 27–36% lower total factor productivity (TFP) growth compared to their downstream counterparts. I show that this gap is driven by the unequal burden of defensive adaptation. Lacking physical protection, upstream firms are incentivized to divert resources into non-productive buffers—specifically, excessive inventory and flood insurance—which decouple input intensification from output growth. Furthermore, these efficiency losses are amplified by cross-sectoral spillovers from agriculture-dependent supply chains. These findings highlight the hidden costs of adaptation borne by unprotected regions, suggesting that infrastructure-based solutions, while physically effective, can exacerbate spatial inequality by distorting firm-level resource allocation.