The allocation of decision‐making power is a critical choice that organizations make to mitigate agency problems and information frictions. This paper investigates the role of delegation for organizations where the agency problem is both pervasive and has potentially high welfare consequences: state‐owned enterprises (SOEs). I use a natural experiment in India to uncover the causal effects of granting SOE managers more autonomy over strategic decisions. Managers meaningfully exercise this autonomy, which results in greater value added, but also a reduced emphasis on outcomes valued by the government, such as a reduction in worker amenities (employee housing), and an increase in markups. Returns to autonomy are higher for firms with higher baseline incentive conflict.

We estimate productivity gains of 13.5% from workplace soft skills training among Indian garment workers. Productivity gains are greater when trainees work on joint operations alongside other coworkers, consistent with gains being driven by improved teamwork and collaboration. Furthermore, untreated coworkers on the treated production lines also show increased productivity. These improvements in the teamwork substitute for managerial attention. Despite productivity gains and higher promotion probabilities among treated workers, there are no effects on wages or retention, consistent with frictions in this labor market. Consequently, the net return to the firm was large: 256% 8 months after program completion.

We study how managers mitigate the negative impacts of environmental shocks. Pairing productivity data from a garment firm with granular measures of air pollution, we show that productivity suffers as a result of pollution shocks but that managers respond by reallocating particularly sensitive workers to improve worker-to-task matches, thus mitigating team productivity losses. Responses are smaller for more inattentive managers; these same managers are also least able to mitigate productivity declines. These patterns are confirmed by leveraging variation in opportunities for reallocation and comparing how close managers of differing attentiveness can get to the simulated production frontier by reallocating workers.

We study the adoption of energy-efficient LED lighting in garment factories around Bangalore, India. Combining daily production line–level data with weather data, we estimate a negative, nonlinear productivity-temperature gradient. We find that LED lighting raises productivity on hot days. Using the firm's costs data, we estimate that the payback period for LED adoption is less than one-third the length after accounting for productivity co-benefits. The average factory in our data gains about $2,880 in power consumption savings and about $7,500 in productivity gains.