Understanding when and how environmental objectives enter government decision-making is central to the design of effective regulation in decentralized systems. This paper studies how pollution becomes decision-relevant for local governments by modeling urban land supply choices through a revealed-preference framework. Using the staggered strengthening of environmental targets associated with China’s national air-quality governance reforms since 2012, we structurally recover the parameters governing local governments’ trade-offs between fiscal revenue and environmental quality. We find a sharp reweighting of policy objectives once environmental considerations become institutionally salient. On average, local governments are willing to forgo 6.37% of expected fiscal revenue from a land parcel to avoid a 1% increase in local PM2.5 concentrations. This implied valuation substantially exceeds existing estimates of residents’ private willingness to pay for cleaner air, disciplining the role of administrative decision-making relative to bottom-up market demand. Importantly, the incorporation of pollution into the administrative calculus is highly uneven across space. We document pronounced strategic responses in land allocation: the implied valuation of pollution is extremely high within 3 kilometers of monitoring stations but declines rapidly and becomes negligible beyond 6 kilometers. These spatial patterns suggest that while stronger top-down environmental targets can effectively alter local government behavior, they may also distort land allocation and generate new forms of spatial environmental inequality.