Assessing the Transmission Mechanism of Federal Reserve Climate Stress Testing on Commercial Bank Lending Portfolios and Carbon-Intensive Asset Pricing

• Billy Elly

  Lautech

  Author

The intersection of climate transition risk and financial stability has emerged as a critical concern for central banks worldwide. While climate stress testing has gained prominence as a supervisory tool, the transmission channels through which such exercises influence commercial bank lending behavior and carbon-intensive asset pricing remain poorly understood. This study addresses this gap by developing an empirical framework to assess the causal mechanisms linking the Federal Reserve's pilot climate scenario analysis to bank portfolio reallocation decisions and the pricing of carbon-intensive assets. Drawing on the Federal Reserve's 2023 pilot climate scenario analysis with six major U.S. banks, combined with loan-level data from Y-14 filings, NGFS scenario projections, and proprietary emissions data, this research employs a difference-in-differences design combined with a Fama-French factor model augmented with an Emission Reduction Stress factor . The empirical results demonstrate that banks with stronger ex-ante climate risk management practices reduced lending to carbon-intensive sectors by 18.7% (p<0.01) following the stress test exercise, with an average transition risk impact of 30-100 basis points on probability of default for corporate and CRE loans . Furthermore, the carbon risk premium on "brown" assets increased by 89.4% in the post-announcement period, translating to a 50-basis-point rise in capital costs for major carbon-intensive corporations by 2030 under a Net Zero 2050 scenario . These findings provide the first quantitative evidence of climate stress test transmission through both the credit supply channel and the market valuation channel, with significant implications for macroprudential policy design and sustainable finance regulation.

• PDF

Copyright (c) 2026 Billy Elly (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.