Knowledge that Transforms

This paper examines the link between ownership concentration and corporate bond volatility. We show that more concentrated mutual fund ownership is associated with higher volatility of corporate bonds. This relation is stronger among more illiquid bonds, during periods of heightened bond market illiquidity, and among bonds held by corporate bond funds that invest in more illiquid bonds and experience higher or more correlated liquidity shocks. Using a sample of mutual fund mergers, we further show that increases in bond volatility are unlikely to be driven entirely by the endogenous ownership structure of corporate bonds. Our findings suggest that the concentrated ownership by corporate bond mutual funds provides another channel, apart from illiquidity, to help explain the excess volatility in corporate bonds.

In a novel econometric framework, we identify differences and significant asymmetries in financial markets' responses to the European Central Bank (ECB)'s various policy interventions during the eurozone crisis. Dollar liquidity interventions reduced stress in bond markets and improved economic sentiment, as reflected in higher equity prices. In contrast, the ECB's euro liquidity provisions and monetary stimulus measures delivered modest results. In both these cases, government bond spreads typically did decline but the risk of large spread increases and equity losses also increased. Only the Outright Monetary Transactions (OMT) intervention had a substantial expansionary effect. The results emphasize the importance of unambiguous monetary policy for driving market expectations.

Machine learning methods have achieved great success in modeling complex patterns in finance such as asset pricing and credit risk that enable them to outperform statistical models. In addition to the predictive accuracy of machine learning methods, the ability to interpret what a model has learned is crucial in the finance industry. We address this challenge by adapting interpretable machine learning to the context of corporate bond recovery rate modeling. In addition to the best performance, we show the value of interpretable machine learning by finding drivers of recovery rates and their relationship that cannot be discovered by the use of traditional machine learning methods. Our findings are financially meaningful and consistent with the findings in the existing credit risk literature.

Effective risk management in pension funds requires the use of appropriate long-term risk and performance indicators. However, Defined Contribution (DC) pension plans currently rely on short-term metrics that don't align with the retirement income goals of beneficiaries. Against this backdrop, we introduce a funding ratio measure for DC plans defined as the plan assets divided by accrued benefits derived from any given (defined) stream of contributions. The funding ratio's denominator is the present value of the total retirement income achievable if all contributions had been invested in fully amortizing fixed-income portfolios called retirement bonds. We also use the retirement bonds to introduce a class of target-income strategies that can effectively reduce income risk as retirement approaches, but also secure minimum funding ratio levels. These simple asset allocation rules strongly dominate the standard target-date fund strategies routinely used by DC plans, in terms of retirement outcomes for beneficiaries.

Using city-level data from China, we find that the overall size of finance does not affect the extent of resource misallocation. However, when finance is decomposed into different parts, we find that local government-driven finance backed by the revenues from land sales exacerbates the misallocation problem, while the remaining part of finance, which is more likely market-driven, significantly improves allocative efficiency. We use the instrumental variable approach to establish causality. Further evidence shows that local government-driven finance lowers allocative efficiency by facilitating the allocation of resources to the low-productivity state sector, but non-local government-driven finance reduces the extent of such distortions. Our analyses suggest that identifying the sources of financial assets and, more broadly, distinguishing between good finance and bad finance are critical to develop a socially desirable financial system.

We analyze the performance of marketplace lending using loan cash flow data from the largest platform, Lending Club. We find substantial risk-adjusted performance of about 35 basis points per month for the entire loan portfolio. Other loan portfolios grouped by risk category have similar risk-adjusted performance. We show that characteristics of the local bank sector for each loan, such as concentration of deposits and the presence of national banks, are related to the performance of loans. We conclude that marketplace lending has the potential to finance a growing share of the consumer credit market in the absence of a competitive response from the traditional incumbents.

This paper investigates the validity of banks' credit loss projections in the bi-annual EU-wide bank stress tests, which inform regulatory capital requirements. It finds that banks “re-optimized” their models in 2016 to bring down credit losses, exploiting flexibility in the stress test framework. Specifically, banks whose losses would have increased the most from 2014 to 2016 because of changes in the adverse scenario saw the largest decrease in projected losses thanks to model changes. Upon the release of the 2016 stress test results, stock prices and credit default swap spreads increased more for banks that achieved a greater reduction in credit losses through “re-optimization”, consistent with investors anticipating lower future capital requirements for these banks.

Regulations introduce substantial costs and constrain firms’ cost structures. This paper introduces a firm-specific measure of regulatory costs and explores its role in mergers and acquisitions (M&A) decisions; specifically, large (small) firms with high regulatory costs are likely to acquire (be acquired by) firms in the same industry. In contrast, regulatory costs do not have such an effect on cross-industry acquisitions. Herein, I introduce econometric techniques, including accounting for additional industry-specific variables, alternative regulatory cost measures, and difference-in-differences estimators, to address potential identification issues. Furthermore, regulatory costs drive acquisitions contributing to shareholders’ wealth; hence, regulatory cost burden is a crucial factor in M&A decisions and outcomes.

We examine the impact of employee-related Corporate Social Responsibility (ER-CSR) on pay disparity between top management and the average worker. Firms with higher ER-CSR ratings have a lower pay disparity and the effect is greatest when executives are paid the most. ER-CSR is associated with a lower ratio of top management's cash and long-term incentive compensation, relative to the average employee's pay. We find that the negative relation is driven by socially responsible firms paying their average employees more. Finally, we document that CSR activities related to employee relations and diversity are those leading to a significant pay disparity reduction.

We find that firms with CFOs who have extensive social connections within the finance industry hold less precautionary cash. CFO connections matter more than CEO connections, reflecting the preeminence of CFOs among C-level executives in cash management and negotiating access to corporate finance. Firms reduce the proportion of assets held in cash by seven percentage points in the two years following CFO turnover and the appointment of a CFO with finance industry connections. The stock market valuation of incremental cash holdings of firms with well-connected CFOs is lower than for other firms, consistent with investor recognition of CFO social capital as an alternative means to address constraints on external capital.