Doyoung Park

We study how place-based environmental regulation affects production networks. Linking establishment locations and emissions with firm-to-firm networks and Clean Air Act nonattainment designations, we distinguish regulatory exposure through a firm's own locations, suppliers, and customers. Supplier-side exposure increases supplier disconnections and customer-side exposure raises customer disconnections, while severed links are rarely replaced. We then turn to emissions, where adjustment within continuing relationships appears. Supplier-side exposure is associated with lower emissions, whereas customer-side exposure leads to higher emissions, even after accounting for network turnover. We provide a simple framework with pollution offloading, input content, and margin compression to explain the asymmetry.

How do external policy shocks reshape local resource use when multiple activities compete for a shared factor? We study foreign tariffs and U.S. agricultural land use from 2005 to 2023 using county-level harvested acreage as a measure of realized land use. Fixed 2000-2004 export and crop shares map crop-specific tariff changes into county exposure, allowing us to separate crop reallocation from changes in total acreage. Higher tariff exposure shifts land toward exposed crops, while the scale response depends sharply on which crops carry the exposure. Because this response changes a county-wide acreage base, exposure concentrated in one crop can generate acreage changes elsewhere in the local portfolio. Relative to the 2004 tariff schedule, observed 2023 acreage is 29.9 million acres higher, with scale incidence concentrated in high-acreage, corn-oriented counties and the largest crop-level movements in corn and soybeans. Removing the tariffs remaining in 2023 yields a 22.9-million-acre difference under predominantly non-corn exposure, shifting scale incidence toward specialized non-corn counties while corn and soybeans retain substantial scale components. Across both tariff environments, changes in the acreage base are quantitatively first-order, implying substantial physical land-use incidence that crop reallocation alone misses.

Energy audit mandates seek to improve investment by requiring owners to identify efficiency opportunities. We study New York City's Local Law 87, which mandates audits, retro-commissioning, and reporting for large buildings while leaving most broader capital improvements to owners. Using a regression-adjusted local-projection difference-in-differences design that exploits predetermined filing deadlines, we find recorded filing rises by 47.8 percentage points, while permit activity, energy use, and emissions change little. Area-weighted estimates rule out reductions larger than 1.1 percent in energy use and 0.5 percent in emissions. Mandatory diagnosis can generate a strong administrative response without comparably large near-term physical adjustment.

We study the joint roles of global value chain (GVC) positioning (the "distance" of an industry from its upstream inputs and downstream consumers) and trade competitiveness (the effective prices received by an industry for its exports or paid on its imports in terms of foreign exchange) in determining the intensity of industries' emissions of air pollution. Using data on country-by-industry emissions and GVC positions, we locate each country–industry by its distance from both primary factors and final demand, using their joint position to proxy for longer, more fragmented value chains. We then estimate how these two-sided positions and real effective exchange rates (REER) relate to emission intensity for carbon dioxide, sulfur oxides, and nitrogen oxides. The results show that industries' emission intensities depend fundamentally on an industry's position in relation to both ends of the value chain (i.e., upstream suppliers versus downstream consumers). These relationships differ across upstream and downstream margins, pollutants, and sectors. Our findings underscore that global production networks shape not only where goods are made, but also where pollution is generated.

We study how trade policy shapes greenhouse gas emissions embodied in agricultural trade, linking structural gravity simulations to country–commodity farm-gate CH₄ and N₂O emissions. Existing tariff preferences raise embodied emissions by 37.9 megatonnes of CO₂-equivalent — about $3.4 billion annually at a recent carbon-border price — with livestock, especially beef, driving most of the increase. A zero-tariff benchmark raises global CH₄ and N₂O emissions by 17.7 megatonnes CO₂-equivalent, driven mainly by trade in crops, especially rice. The results illuminate how the global regime of tariff barriers shapes the environmental consequences of trade.

As U.S. wildfires grow more frequent and severe, the electricity sector has shifted toward solar and wind. Does intensified wildfire smoke, by reducing renewable efficiency, drive up emissions through greater reliance on non-renewables? Using daily precipitation as an instrument for wildfire-related PM2.5, we find that (a) pollution emissions rise significantly on smoke days; (b) the shift from renewable to non-renewable generation partly explains the increase, though elevated electricity demand also plays a key role; and (c) the non-renewable response comes mainly from higher generation levels rather than fuel-mix changes or intensified abatement. The results reveal overlooked environmental consequences of wildfire smoke.