Li Ting (Tina) Chiu

Marketplace lending platforms (MLPs) act as intermediaries between borrowers and investors, facilitating the creation of unsecured consumer loans. We examine how MLPs allocate loans between retail and institutional investors. Controlling for observable risk factors, including interest rates, we find that strategic allocation varies with origination volume. Specifically, institutional investors receive loans with 8.0% lower default rates when volume is low, while retail investors receive higher-quality loans when volume is high. This time-varying pattern reflects MLPs balancing quid pro quo incentives with the costs of regulatory scrutiny. We document additional preferential treatment among institutional investors, where balance-sheet lenders receive higher-quality loans than pass-through investors (such as hedge funds and investment banks) during periods of asset-backed securities accumulation. Using the 2015 San Bernardino shooting as an exogenous shock to regulatory intensity, we provide evidence that intermediary allocation strategies respond to regulatory pressure. Our findings reveal how emerging FinTech intermediaries navigate traditional financial incentives and regulatory oversight, with implications for the protection of retail investors and the design of financial markets.

This paper examines the effect of family-level product diversification on equity-common-flow risk hedging. We document that due to low correlations among common flows to various asset classes, product diversification reduces family-wide flow volatility. Consequently, fund managers of more diversified families hedge less against equity common flow shocks. This effect is absent for externally managed funds, suggesting the internal compensation structures influence managers’ hedging decisions. Furthermore, common flow risk premium is significantly higher for stocks held by funds in concentrated families than for stocks held by funds in diversified families. Our study underscores how family-level product diversity affects managers’ incentives and their portfolio decisions.

This paper examines the structural barriers behind the underrepresentation of women in the US equity mutual fund industry. Using data from 1992 to 2022, we document a Glass Cliff in appointments: female managers are disproportionately assigned to funds with poor historical performance. We also document a demand-side friction that we argue produces it: investors translate a female manager’s performance into capital less strongly than a male manager’s. The flow-performance relation flattens after a woman is appointed, and flattens most when she replaces a departing man, while pre-appointment sensitivity does not predict who is appointed. This muted response is not the efficient discounting of a noisier signal: past performance forecasts future performance equally well regardless of manager gender. Post-appointment performance is likewise indistinguishable across genders, leaving no skill gap for the pattern to reflect. A small modification of the Berk–Green (2004) model of delegated portfolio management adds a gender-dependent slope to the flow-performance relation. It rationalizes these facts jointly and shows why they are connected. Because a flatter response forfeits inflows at strong funds and cushions redemptions at weak ones, the model gives fund families an incentive to assign women to weaker funds. Taken together, our model and evidence support a view of the Glass Cliff as fund families’ response to gendered information processing by investors, rather than an independent supply-side friction.

This paper documents a structural shift in U.S. corporate disclosure following the availability of large language models (LLM) in 2022. Using MD&A filings from 2015–2025, we estimate a difference-in-differences specification pairing each firm’s original filing (treated) with its own GPT-generated summary (control). The result is a sharp, post-2022 expansion of the cognitive thinking process, rising 28–34% of a sample standard deviation relative to the machine baseline, an economically large shift in a sticky textual domain. This expansion is accompanied by a mirror-image collapse in authenticity and emotional tone. These findings suggest that firms inflate cognitive content to preserve managerial reasoning under algorithmic compression while neutralizing tonal features that matter only to human readers; they are reshaping disclosures not for human listeners but for an algorithmic audience.

This paper examines the effects of institutional bond ownership on the cost of debt. Empirical evidence shows that higher long-term institutional bond ownership lowers the cost of debt, whereas short-term bond ownership increases capital supply uncertainty and debt cost. These findings are robust to controlling for bond/investor characteristics and using investors’ funding structure as an instrumental variable. The results suggest that short-term bond investors’ capital supply uncertainty contributes to fragility problems in the corporate bond market. Conversely, long-term institutional bond ownership enhances corporate governance, thereby lowering the cost of debt.

Theory suggests that issuing callable bonds conveys private information and reduces information risk. We examine the impact of information risk on yield spreads for both straight and callable bonds. Our analysis reveals that as information risk or credit rating deteriorates, the difference in the sensitivity of changes in yield spreads to interest rates between straight and callable bonds shrinks. The results are similar when using equity volatility as an alternative proxy of default risk. Overall, there is evidence of a negative relation between call premium and default/information risk, which has not been explored in the literature.

We propose a dynamic search-based asset pricing model with collateral and counterparty risks to examine the role of these risks in the repo market and the secondary market for collateral. We show that collateral and counterparty risks increase expected loss in repo transactions, discourage trading activity, and reduce demand for collateral. Funding costs in the repo market and yield spreads of bond collateral increase with these risks and these effects magnify in times of stress.

This paper shows that firms with higher equity ownership by dual holders are less likely to make large and public acquisitions. We show that the relation between dual ownership and acquisition decisions is causal using investors’ reaching for yield preference as an external shock to dual ownership. Firms with high dual ownership are less likely to make high-risk, low-quality, and shareholder value-destroying acquisitions. Dual ownership is also associated with favorable market reactions to acquisition announcements and superior operating performance in the postacquisition period. Overall, these results indicate that dual ownership improves corporate investment decisions by mitigating the conflict of interest between shareholders and bondholders and managerial agency problems.

This paper investigates the impact of institutional herding on stock price formation, conditional on firms’ credit ratings, using 13F data from 1986 to 2019. In line with the current literature, we find herding intensity is driven by past returns consistent with momentum trading; however, we also find that herding is more sensitive to past returns for non-investment grade (NIG) stocks than investment grade (IG) stocks, resulting in a market bifurcation. We then examine the price impact of these trades and find that herding in NIG equities enhances price discovery. One plausible explanation is that information gradually diffuses within non-investment grade stocks, and herding behavior strengthens information discovery. Finally, we show both momentum-triggered herding and non-momentum-triggered herding contribute to price discovery among non-investment grade stocks.