The increasing focus on Environmental, Social, and Governance (ESG) factors has raised concerns about how negative ESG news propagates through supply chains and affects firm outcomes. This paper examines the impact of ESG-related supplier shocks on firms’ supply chain configurations and economic outcomes. We find that negative upstream environmental and social (E&S) news significantly increases the likelihood of supplier relationship termination, but this effect is contingent on input specificity. Despite reputational risks, terminations are less likely for firms reliant on specialized inputs, showing firms prioritize operational stability over ESG considerations. We document an asymmetric response to foreign versus domestic ESG shocks, highlighting the role of international supplier dependence. Finally, we show that supplier terminations lead to increased operating costs, lower productivity, and reduced markups, but these costs are mitigated when firms replace non-specialized or domestic inputs. Our findings underscore the complex trade-offs firms face in balancing ESG concerns with supply chain efficiency.

I present a theoretical framework that features contractionary productivity dispersion shock which is a result of the interaction between substitutability of supplied labor and demanded goods. I introduce information friction as a source of nominal rigidity to study the impact of the productivity dispersion shock on the conduct of monetary policy. In particular, I assume firms have incomplete information about the productivity dispersion when they set the price. I show that in the environment with nominal rigidity, replicating full-information flexible price equilibrium is always feasible and optimal, however, the optimal policy is not an inflation targeting policy. The optimal monetary policy is the policy which eliminates the dependence of the idiosyncratic nominal variables on the unknown productivity dispersion and as a result makes the information friction irrelevant.

Life Cycle Assessment (LCA) is essential for evaluating the environmental impact of sustainable activities in industry. Despite its importance, there exist challenges negatively impacting its deployment, particularly the time-consuming process of gathering inventory data. This research introduces a novel framework that leverages advanced text-based algorithms from Natural Language Processing (NLP), significantly enhancing the efficiency of data collection in LCA studies. Focusing on the inventory phase, the novelty of this research lies in its ability to reduce data collection time by an estimated 80%–90% compared to conventional methods and improve accuracy by directly extracting materials from bills of quantities (BoQs), which usually list all the construction materials. While our methodology shows promise, it faces challenges due to project complexity, particularly the need for consistent terminology between BoQ and reference databases, though future advancements in matching algorithms may enhance our approach’s efficiency. Real-world case studies demonstrate the framework’s effectiveness, offering flexibility across industries and system complexities.

This paper identifies Bayesian learning as a main driver of firms’ subjective uncertainty and demonstrates its consequences for investment behavior. Using the Survey of Business Uncertainty, which provides both executives’ forecasts and realized sales growth, I show that subjective uncertainty rises with forecast errors but displays diminishing sensitivity, a signature of learning mechanisms. A Bayesian normal-gamma learning model explains this pattern and outperforms GARCH in predicting firm-level uncertainty. Embedding learning uncertainty in a partial-equilibrium investment model yields three predictions: investment responses are weaker in volatile environments, weaker for large shocks than small shocks, and weaker for positive shocks than negative shocks. Together, these patterns generate asymmetric, S-shaped investment behavior. Compustat data confirm these predictions, with Bayesian uncertainty significantly dampening investment responses. This study formally links learning processes to subjective uncertainty and traces their real effects on firm dynamics, showing how belief updating, not just realized volatility, shapes corporate investment.

Automating data-driven processes in construction is critical for enhancing efficiency and reducing human intervention. Life Cycle Assessment (LCA) in construction relies on the seamless integration of construction and environmental databases; however, manual processes are time-consuming and error-prone. This paper presents a methodology that fully automates data integration using fuzzy string matching and Random Forest (RF)-based supervised learning. The approach ensures accuracy and efficiency while eliminating manual effort. When applied to the Andalusia Construction Cost Database (BCCA) and ecoinvent databases, the methodology achieved high accuracy and reduced manual intervention time by 90%. This scalable solution advances automation in construction by facilitating real-time decision-making, improving digital workflows, and supporting sustainable construction practices through AI-driven data management.