Below I list recent working papers in themes


Theme 1: The Interplay between Market and Society

Why do we need financial markets in the first place? To understand this broad question, I examine how market-based mechanisms affect the real economy and human society. I show that short selling can serve as an “invisible hand” to discipline firm incentives and gauge information flows (2015 RFS; 2015 JFE; 2016 JFE). 

The reverse question is how social considerations (e.g., ESG), institutions, and norms may influence firms and investors in financial markets. My research (2014 RFS) indicates that a country’s weak institutions can pose a fundamental challenge to its financial market. My other studies examine the impact of air pollution (2019 JFE), social trust (2020 JFQA), and social norms (2025 JFQA). 

My more recent projects aim to investigate how behavioral bias may affect different aspects of green finance (e.g., investors, regulators, etc.).

 

“Policy Uncertainty Reduces Green Investment,” with Mengyu Wang and Jeffery Wurgler, working paper, 2023. SSRN version. R&R at Journal of Financial Economics.

We find that Chinese firms reduce green investment as the uncertainty of supportive subsidies rises. This effect is identified from weather-driven fluctuations in air pollution that lead to fluctuations in subsidy allocations: Firms in cities where weather-driven subsidy uncertainty is high engage in less green R&D investment and hire fewer associated technical employees. More broadly, the results suggest that economic policy uncertainty may originate not only from political and macroeconomic shocks but also from behavioral mechanisms that link policy to salient recent conditions. 2023 NBER Chinese Economy Working Group Meeting.

 

“The Effect of Carbon Pricing on Firm Performance: Worldwide Evidence,” co-authored with Tinghua Duan and Frank Weikai Li, working paper 2024.

This paper examines the impacts of carbon pricing policy initiatives on the operating performance and market value of publicly listed firms around the world. Using a triple difference approach, we find a significant reduction in firm value upon the adoption of policy initiatives. This effect originates from both a discount rate channel and a more pronounced cash flow channel. 2024 ABFER.

 

“The Surprising Green Performance of Retail Investors: A New (Behavioral) Channel,” co-authored with Sumit Agarwal, Yanlin Bao, Pulak Ghosh, and Jian Zhang, working project 2024.

Contrary to the prevailing wisdom that green investors willingly accept lower returns for sustainable investment, our analysis of account-level data from a major Indian bank indicates the opposite. To explain this surprising observation, we hypothesize—and empirically verify—that green investments may help investors mitigate behavioral bias, such as the disposition effect and under-diversification. Tests utilizing abnormal temperatures as exogenous shocks support a causal interpretation.

  

Theme 2: FinTech, AI, and Big Data

This theme explores new developments in the market, such as social media, social networks, machine learning, and blockchain technologies. I am particularly interested in the recent AI progress and new developments that have a “blue ocean” flavor.

 

“Data Specialists and Market Efficiency,” with Massimo Massa and Yijun Zhou, working paper 2023. (SSRN version) R&R at Review of Financial Studies. 

In the age of big data, investors need to process ever more complicated, multidimensional data to decipher different aspects of a firm. How do investors deal with multidimensional data? Using trading concentration across news categories as a proxy for data specialization, we find more informed institutional investors tend to specialize in subsets of firm aspects (i.e., data specialists) instead of processing all available data. Such data specialization, however, may hamper market efficiency.  

 

“Predictive Crypto Crashes,” co-authored with Jennifer Li, and Li Liao, and Siyuan Yang, working paper 2023. (SSRN version coming soon)

Bubbles and crashes are hallmarks of cryptocurrencies. Can such bubbles/crashes be predicted in the cross-section of cryptocurrencies? We document that consecutive large price runups and price drawdowns, a leading asset pricing implication of slow-moving capital, can help predict crypto-crashes and negative returns. This effect may arise when investors reduce demand after experiencing inelasticity, leading to predictive price declines. Analysis using ICOs as a source of Ethereum blockchain congestion supports this mechanism. 2024 AsFA.

 

“A Tale of Two Zoos: Machine Learning Insights on Retail Investors” with Pulak Ghosh, Huahao, Lu, and Jian Zhang, working paper, 2023. (SSRN version coming soon)

We employ various machine learning models to analyze the returns for millions of retail investors in India. We observe that Neural Networks outperform other machine learning and OLS models in uniquely predicting both good and bad out-of-sample performance. Behavioral biases exert a more significant influence on their returns than holding-weighted firm characteristics. 2024 ABFER.

 

“Machine Learning as Arbitrage: Can Economics Help Explain AI?” with Huahao Lu and Matthew Spiegel, working project, 2024.

We show that economically motivated dynamic arbitrage portfolios (DAPs) can help explain the superior performance of Neural Networks (NNs). In predicting stock returns based on 153 firm characteristics (anomalies), DAPs rank anomalies similarly to neural networks in the cross-section. DAPs also account for approximately 87.9 bps monthly alphas of the high-minus-low portfolios selected by neural networks in the time series. Our results reveal three economic sources of neural network performance: a time-varying strategy analogous to dynamic arbitrage, a tendency to weight portfolios on unpublished anomalies, and exposure to microcaps.


  

Theme 3: Demand-side Frictions & Delegated Portfolio Management

This research theme considers the investor (demand) side of the financial market. I am particularly interested in how delegated portfolio management affects market efficiency. My early papers (2007 RoF; 2008 RFS; 2013 JFE) examine fund incentives and strategies of delegated portfolio management. More recently, I explore the economic grounds affecting the dual efficiency of security prices and delegated portfolio management, e.g., search frictions (2020 JF), social trust (2020 JFQA), leverage, and strategic complementarity.

 

“Hedge Fund Leverage, Delegated Portfolio Management, and Asset Prices,” with Charles Cao, Grant Farnsworth, and Yijun Zhou, working paper 2023. (SSRN coming soon)

We extend Berk and Green (2004) to model hedge funds as an institution with leverage advantage. Consistent with our model, we observe that: 1) higher leverage allows hedge funds to reap more economic rents (fees), 2) hedge funds simultaneously adjust leverage and holding beta during adverse funding conditions, and 3) a leverage-tightness factor constructed from hedge fund holding betas can significantly predict the cross-section of asset returns. The 14th Annual Hedge Fund Research Conference, 2024 CICF.


“Financial Intermediaries vs. Capital Allocation: The Forgotten Role of Mutual Funds,” coauthored with Massimo Massa and Yanbo Wang, working paper 2023. 

Recent evidence casts doubt on the efficacy of U.S. firms in allocating equity capital to more productive sectors. We document that equity mutual funds exhibit better allocational efficiency, which also implies a novel source of managerial skills. Our results suggest that financial intermediation helps the market achieve allocational efficiency. 2023 ABFER, 2023 CICF.


“Security Lending and Corporate Financing: The Case of the Bond Issuance,” co-authored with Jennie Bai and Massimo Massa, working paper updated in 2024. R&R at Management Science. (SSRN version)

The security lending market allows institutional investors to lend out their holding assets in exchange for cash collaterals, an important but understudied source of funding for corporate bond lenders. We find that this motivation can spill over to the traditional corporate bond market to influence bond issuance and prices. Our results highlight a “lender’s perspective” in digesting the real impacts of bond lending. 2018 CICF, 2019 AFA.

  

Theme 4: Supply-side Factors and Strategic Firms

This research theme investigates the firm (supply) side of the market, with a particular focus on how strategic firms may affect market efficiency and asset prices. For example, strategic business groups can alter firm risk via within-group asset allocations (2022 JFE), as has also been shown in a series of cases I developed when examining the revolution of corporate treasury operations in the new century. My recent paper on how banks use indirect networks to approach valuable firms (2024, MS) provides another example. This theme also dates back to my Ph.D. thesis, in which I explored “anomalies” associated with strategic firm investments that can generate non-normal cash flow distributions.

 

“Does the Community Reinvestment Act (CRA) Crowd Out Corporate Lending?,” with Massimo Massa, Ruichang Lu, and Wenlan Qian, ongoing project, 2023.

The CRA aims to promote mortgage lending by banks to mid to low-income borrowers. Could it crowd out corporate lending? Our findings suggest the opposite: relationship and investment-grade firms receive more loans from CRA-regulated banks. CRA-induced loans are larger and cheaper ex-ante but carry a moderately higher distress risk ex-post. Our results suggest that bank-level risk constraints may motivate banks to (over)supply credit to high-quality corporate clients to offset the CRA-induced risk from mortgage lending.


 “An Anatomy of Characteristics in Dynamic Trading,” with Matthew Spiegel and Tao Huang, working paper, 2021. SSRN version

We propose testing the joint and marginal power of characteristics in predicting stock returns via their contribution to optimal dynamic trading (e.g., Kyle 1985). Applying dynamic trading to a sample of 147 characteristics confirms their joint power in delivering significant out-of-sample returns (23% annually). However, most characteristics (88%) fail to supply independent information—removing them can further enhance returns to 33.7%. Our analysis also reveals a leading role of Fama-French-Carhart factors as informative characteristics. 2021 CICF.