About

As carbon pricing faces political barriers, governments and large private buyers increasingly turn to green procurement policies, buying only from low-carbon suppliers to encourage clean production. This paper studies how Buy Clean initiatives reshape competition and clean investment in the U.S. cement industry. I develop a two-stage structural model in which plants choose whether to switch from coal to natural gas and then compete in procurement auctions. Counterfactual simulations show that Buy Clean induces plants to switch to cleaner production but also exacerbates market power distortions. Restricting public purchases to low-carbon suppliers weakens competition, increasing their markups on green sales. Shifting demand toward clean plants pushes their output closer to productive capacity, raising their marginal costs and increasing their bids to all buyers. Dirty plants’ markups also rise as they face less competition from clean rivals. A border-adjusted carbon tax that induces the same switching rate cuts more emissions at a lower welfare cost per ton of CO2 avoided.

Natural resource management is inherently shaped by uncertainty, yet most policy evaluation assumes linear preferences and objective probabilities. We examine how risk preferences and probability weighting shape resource demand and welfare measurement in recreational fisheries, a setting where anglers must commit to a trip without knowing their realized catch. Using discrete choice experiments from the 2018 U.S. West Coast Sport Fishing Surveys, we estimate angler preferences for six species across three fisheries under both Expected Utility and Rank-Dependent Utility frameworks. Combining a CARA utility function with probability weighting, we find that anglers are risk-averse toward catch variability and systematically overweight low-probability, high-reward outcomes, consistent with an inverse S-shaped weighting function. These distortions are stronger among experienced and specialized anglers. Incorporating these preferences into a bioeconomic model, we show that linear-utility models produce biased estimates of effort responses and welfare effects when the upper tail of the catch distribution changes, whether through regulations such as bag limits or through changes in fish stocks.

Organizations designing performance-pay systems must choose both reward size and which tasks to target. A threshold model predicts that incentives will have their largest effects on tasks with intermediate untreated completion, where more decisions are close to indifference. In a randomized experiment, clinicians treated standardized patients under contracts varying reporting and per-task bonuses. Consistent with the model, effects were largest at intermediate control-group completion and smaller at the extremes. Reporting alone raised completion by 8.2 percentage points; a high bonus produced a smaller additional gain, while a low bonus did not. This offers a practical criterion for task targeting.

This paper investigates how political connections affect the allocation of public procurement contracts, exploiting China’s anti-corruption campaign as a quasi-experimental shock to firms’ political connections. Using an original dataset scraped from procurement records and a difference-in-differences design, I find that contract values awarded to politically connected firms fall by 39% after the campaign. The decline is larger for firms connected to local officials and in more corrupt regions. Estimating firm-level total factor productivity, I find that the campaign improves the allocative efficiency of public procurement.