Effects and Design of Central Bank Corporate Credit Facilities

Abstract

This dissertation investigates the price effects, real effects, and design of the Federal Reserve’s Corporate Credit Facilities (CCFs), launched in 2020 to stabilize U.S. corporate bond markets during the COVID-19 pandemic.

Chapter 1, based on work with Jessica S. Li, estimates impact of the CCFs on corporate bond spreads. The CCFs included both direct support via cash bond purchases and indirect support via exchange-traded fund (ETF) purchases. We exploit bond-level ratings heterogeneity across firms to identify treatment effects from direct bond support. The March 23, 2020 announcement of the CCFs reduced spreads by 96 basis points (bps) for eligible issuers. To estimate the impact of the April 9 expansion, we leverage a quasi-natural experiment involving “Fallen Angel” firms—those initially eligible, then briefly ineligible, but reinstated during the expansion. Relative to a control group, we find a -126 bps treatment effect. Using a causal machine learning approach detailed in Chapter 2, we estimate that ineligible firms would have seen a -500 bps spread reduction had they received direct bond support on March 23.

Chapter 2 develops a novel two-step semi-parametric difference-in-differences (DiD) estimator for dynamic and heterogeneous treatment effects, allowing for flexible policy counterfactuals. Applying the method to firm-level outcomes, I analyze the real effects of the CCFs on cash holdings, leverage, payouts, and investment. The estimator produces results consistent with conventional panel and event study regressions but highlights important heterogeneity. Firms generally increased cash and leverage, while payout and investment initially declined. However, CCF-eligible firms began deleveraging by 2021 and accumulated less cash compared to ineligible peers. Despite increased shareholder payouts, eligible firms did not raise investment levels, suggesting that the CCFs failed to achieve their stated real economy goals. Counterfactual treatment effects suggest that broadening eligibility for direct bond support might have increased leverage and payouts, but evidence on investment gains is weak or inconclusive.

Chapter 3 provides a theoretical framework to explain these findings. I construct a dynamic capital structure model with investment in which firms cannot commit to a debt issuance policy. In the model, unsecured debt interventions accelerate borrowing but undermine the benefit of lower bond yields due to increased leverage. The proceeds are primarily distributed to shareholders rather than invested. In contrast, secured debt interventions support better investment outcomes because the collateral constraint on secured debt issuance induces commitment. Even for financially unconstrained firms, secured debt interventions yield more favorable dynamics, aligning firm incentives with the policy's intended real effects.

Together, these chapters demonstrate that while the CCFs were highly effective in reducing borrowing costs, their real effects were muted or misdirected, in part due to firms’ inability to commit to future financial policies. The findings underscore the importance of policy design—particularly the role of collateral and commitment—in determining the effectiveness of credit market interventions.