1. “The Information Content of Mandatory Human Capital Disclosures – Initial Evidence”, with Salman Arif and Johnny Yoon. 

Abstract: This paper examines the information content of newly-mandated qualitative Human Capital Disclosure (HCD) for equity and bond investors. Using hand-collected HCD data from 10-Ks, we first document that the amount of HCD is positively correlated with firms’ human capital management ratings. Moreover, the initial HCD is informative to investors: stock prices react positively to HCD, but bond prices react negatively to HCD. Further analyses show that the equity market reacts positively to both social- and operational-oriented HCD components, whereas the negative bond market reactions are driven by social-oriented HCD components. Finally, comparing the first-year and second-year HCD, we document high similarity scores between the two years and find no evidence that asset prices respond to second-year HCD. Collectively, our evidence suggests that principles-based mandatory HCD initially contains value-relevant information for both equity and bond investors, but this information content declines over time due to its sticky nature. More broadly, we provide some of the first evidence that mandatory ESG reporting can have opposing implications for shareholders and creditors.

2. “Federal Regulations on Fair Lending and Sexual Orientation Discrimination in the Mortgage Market”, with Scott Liao and Lulin Song. 

Abstract: This study examines whether including sexual orientation as a protected class in federal fair lending regulations mitigates discrimination against same-sex co-borrowers in home mortgages. Relative to mortgages in states with similar protections in state laws, we predict and find that the federal regulatory changes reduce the disparity in mortgage pricing between heterosexual and same-sex co-borrowers in states without state-level protections. We document that this reduction in pricing disparity is only significant for counties with higher gay or lesbian population and for lenders that are more in compliance with existing anti-discrimination laws protecting racial minorities. Supplemental analyses suggest that these results are not driven by changes in borrower credit risk or lender underwriting standards. Collectively, the evidence supports that anti-discrimination fair lending regulations at the federal level mitigate discrimination against same-sex co-borrowers in the mortgage market. Our study contributes to the DEI literature and carries important policy and societal implications.

3. “Unintended Consequences of Mandatory Carbon Disclosures: Evidence on Local Toxic Releases”, with Scott Liao, Divyesh Shan, and Lulin Song. 

Abstract: This study investigates the unintended consequences of the Greenhouse Gas Reporting Program (GHGRP), which mandates regulated facilities to report their greenhouse gas (GHG) emissions. We argue that this rule change may compel firms to allocate resources toward GHG mitigation at the expenses of toxic pollution control. Trading off GHG mitigation benefits versus environmental liabilities related to toxic pollutions, we predict and find that, compared to non-GHG-reporting firms, GHG-reporting firms experience an increase in toxic chemicals released by their facilities in low- relative to high-enforcement regions, following the implementation of the GHGRP. This result is more pronounced in communities with lowe  income or higher proportions of racial minorities, reinforcing enforcement stringency and legal liabilities as a key consideration in the trade-off. We also find decreased pollution abatement activities in low enforcement regions further supporting the notion of resource reallocation. Cross-sectionally, our finding is more pronounced when firms face higher financial constraints or investor pressures. These results cannot be attributed to production shifting from GHG-reporting to non-reporting facilities, nor technology-driven mechanical relation between GHG reduction efforts and chemical releases. Our collective evidence highlights the unintended spillover effect of GHG emission disclosure on toxic chemical releases, contributing to the debate about environmental disclosure mandate.

4. “Level 3 Fair Value Measurement and Systemic Risk”, with Scott Liao, Jacob Ott, and Ethan Yao. 

Abstract: Motivated by bank regulators’ use of level 3 fair value assets as an indicator for systemic risk, we examine whether level 3 fair value assets contribute to systemic risk buildup via the balance sheet liquidity channel and how financial reporting transparency mitigates the risk buildup. We first document a positive association between level 3 assets and banks’ contribution to systemic risk. While the economic magnitude of this association is modest on average, we find that level 3 assets’ contribution to systemic risk buildup is more pronounced for banks with high reliance on non-stable funding. Importantly, we find transparent disclosures of valuation inputs mitigate level 3 assets’ contribution to systemic risk. Finally, our results show that opaque level 3 fair value positively predicts future security impairment losses, and banks with security impairment losses reduce repo and uninsured deposit liabilities. Collectively, the evidence suggests that level 3 fair value measurement has macroprudential implications and opaque level 3 fair values may exacerbate liquidity shocks during economic downturns. This study contributes to the financial stability and fair value accounting literature by exploring the role of fair value measurement in banks’ liquidity management, which is distinct from the regulatory capital channel examined by prior research.

5. “Mortgage Companies’ Compliance with the State-Level Community Reinvestment Act -- Performance Management or Real Investment?”, with Scott Liao, Clark Shu, and Ethan Yao. 

AI Podcast

Abstract: This paper examines whether non-bank mortgage companies’ compliance with state-level Community Reinvestment Act (CRA) is a result of performance management or real investment. Leveraging the 2007 CRA rule change in Massachusetts (MA), we find that affected mortgage companies increase purchases, but not originations, of low-to-moderate-income (LMI) loans that qualify for CRA credit. In addition, inconsistent with the argument that loan purchase allows mortgage companies to gain understanding of unfamiliar areas and increase liquidity for the segments of loans with high information asymmetry, this increased loan purchase is concentrated in loans issued in the mortgage company’s familiar LMI neighborhoods and made to safer borrowers. We also document immediately reversed LMI loan purchase after mortgage companies receive satisfactory CRA ratings when they are unlikely to be examined again. Importantly, we detect no increase in mortgage originations in LMI tracts either by affected lenders or in aggregate, but find a reduction in mortgage rates in MA LMI areas for safer borrowers. These findings suggest that mortgage companies manage CRA performance by trading safer LMI loans. While such performance management does not increase mortgage supply to LMI areas, increased secondary-market demand for safer loans in LMI areas benefits these borrowers through lower interest rates.