AFA PhD Student Poster Session
Poster Session
Saturday, Jan. 3, 2026 7:00 AM - 6:00 PM (EST)
Sunday, Jan. 4, 2026 7:00 AM - 6:00 PM (EST)
Monday, Jan. 5, 2026 7:00 AM - 6:00 PM (EST)
- Chair: Sydney Ludvigson, New York University
AI and Demand-Based Option Listings
Abstract
This paper examines the impact of artificial intelligence on option market efficiency through Nasdaq's implementation of AI-driven option listings in August 2022. Prior to this innovation, over 80% of option listings experienced minimal trading volume, reflecting substantial demand uncertainty. Using a difference-in-differences approach with matched samples based on underlying stock characteristics, I find that AI-based option listings significantly increase trading volume by better aligning strike price selection with realized demand. The effect is particularly pronounced for deep in-the-money equity puts, which exhibit delta close to -1 and serve as synthetic short positions. Analysis of option Greeks reveals patterns consistent with improved targeting of hedging demand. I develop a characteristics-based model where a centralized exchange choose listings based on options Greeks to maximize expected trading volume. The model demonstrates that as the AI reduces prediction error, option listings become increasingly contingent on realized uncertainty in demand. While AI-driven listings improve market efficiency, the model reveals that exchanges' volume-maximizing objectives may diverge from social welfare maximization, with mixed implications for retail options trading. Overall, the findings suggest that AI-driven demand prediction can substantially improve market efficiency in settings with high demand uncertainty, with implications for other markets facing uncertain demand.AI Automation and Effort Allocation: Evidence from Sophisticated Investors
Abstract
How does AI reshape the distribution of effort across different tasks in this knowledge economy?Focusing on hedge funds as a high-powered information intermediary in financial markets, I
examine the role of AI technologies in affecting information acquisition behavior. I hypothesize
that AI reduces the costs of collecting machine-based information, thereby allowing hedge funds
to exert more effort on information acquisition that requires human interactions. Consistent with
my hypothesis, I find that hedge funds increase earnings call participation—at both the extensive
and intensive margins—after adopting automated SEC filings downloads. I further show that
automation-adopting hedge funds (AHFs) earn significantly higher abnormal returns from
attending earnings calls than non-adopters. Post-automation call attendance is also associated with
profitable stock trades, suggesting that human-interaction-based information is value-relevant.
Performance effects are more pronounced for hard-to-research stocks. Overall, this study
documents within-worker productivity effects of AI and identifies a novel “complement-via-
substitution” mechanism: AI augments high-skilled labor by facilitating the shift of effort from
automation-prone tasks to human-intensive ones.
AI Washing
Abstract
This paper investigates AI washing — the practice by which firms exaggerate or falsely claim their investments in artificial intelligence (AI). We leverage large language models to analyze earnings conference call transcripts (“AI talk”) and employee resume data (“AI walk”) from U.S. public firms between 2016 and 2024. Our analysis reveals that AI talk does not predict future AI walk, even over multi-year horizons. Substantive AI walk, rather than AI talk, positively predicts future AI patenting activity, in terms of both quantity and impact. Firms engaging in empty AI talk without corresponding AI walk generate fewer and lower-quality AI patents. Importantly, institutional investors appear to recognize this disconnect, allocating more capital to firms with higher AI walk. While AI talk is associated with short-term stock return gains, potentially motivating inflated disclosures, only AI walk is correlated with superior long-run stock performance. In addition, firms with high managerial incentives are significantly more likely to increase AI talk without a corresponding rise in walk, suggesting strategic hype. Overall, our findings highlight a critical disconnect between firms’ AI rhetoric and their substantive AI investments, revealing a misalignment between short-term market incentives and long-term value creation.Ample Reserves for Whom? The Role of Foreign Banks in U.S. Monetary Policy Implementation
Abstract
We document that foreign banks are the marginal holders of reserves in the U.S. ample-reserves system. Despite holding a minority share of U.S. banking assets, foreign banks hold nearly half of total reserves and exhibit highly elastic demand, responding aggressively to arbitrage opportunities between interests on reserves and funding rates. Using high-frequency shocks to reserve supply, we confirm that foreign banks absorb the bulk of adjustments to aggregate reserves fluctuations. We develop a tractable model that formalizes these findings and show that uncertainty originating from foreign banks, such as quarter-end window-dressing, raises the Fed's reserves supply required for reliable rate control.These results highlight the importance of institutional heterogeneity in reserves demand. In an ample-reserves framework, effective policy implementation depends not just on the aggregate supply of reserves, but on the type of the marginal holder.
Are Carbon Offset Projects an Effective Vehicle for Carbon Emission Reduction?
Abstract
This paper investigates the effectiveness of carbon offset projects in reducing carbon emissions, focusing specifically on REDD+ projects, which aim to reduce emissions from deforestation and have the highest carbon offset issuance. I build a novel dataset that includes all available REDD+ projects worldwide and high-resolution forest coverage data. Specifically, I define control and treatment regions based on grid locations within and outside project boundaries. Using a difference-in-differences design, I find that these projects have a nearly zero effect on slowing deforestation. Using subnational region data, I show that projects in regions with access to finance, higher education levels, greater trust, or more ethnic diversity tend to perform better, indicating that cultural and educational factors are crucial for the success of carbon offset projects.Artificial Intelligence, Opportunity, and Regulatory Uncertainty: Implications for Asset Pricing
Abstract
This paper studies two channels—opportunity and regulatory uncertainty—through whichArtificial Intelligence (AI) affects the stock prices and risk premia. On the one hand, advances
in AI present firms with opportunities, leading them to exhibit characteristics of growth
firms and earn lower expected returns. On the other hand, firms face increased regulatory
uncertainty in AI development, increasing their political risk exposure and resulting in higher
expected returns. Using conference call transcripts, I construct a firm-level measure of AI
Exposure that captures the level of attention analysts and managers devote to AI-related
topics at specific points in time. Empirically and theoretically, I show that these two channels
exert opposing effects: firms focused on opportunity earn a negative AI risk premium, while
those more affected by regulatory uncertainty earn a positive AI risk premium.
Asymmetric Labor Income Risk: Implications for Risk-Taking in Financial Markets
Abstract
How does the shape of labor income risk influence households' demand for risky assets? I exploit the Survey of Income and Program Participation (SIPP) to construct a state-dependent measure that interacts the variance of individuals' labor income growth with its skewness. Conditional on having the same income volatility, households facing right-skewed income shocks (opportunity risks) raise their equity share, whereas those subject to left‐skewed shocks (disaster risks) cut back sharply. This suggests that the counter-intuitive positive link between earnings volatility and stock holdings documented in the literature is driven almost entirely by the upper tail of the earnings distribution. These results show that higher-order moments of labor income risk shape portfolio choice, highlight the limitations of variance-based uncertainty metrics, and offer micro-level support for more robust portfolio choice models.Avoiding the Premium on the Premium? Self-funded Health Benefit Plans and Corporate Financial Decisions
Abstract
This paper examines how firms respond to frictions in the health insurance market by shifting toward self-funded Employer-Sponsored Insurance (ESI) arrangements in the U.S. Under self-insurance, employers avoid the high premiums charged by insurance carriers but take on greater financial risk associated with medical claims. I argue that this shift is driven by increasing insurer market concentration and obtaining insurance is costly. Exploiting national health insurer mergers and acquisitions (M&A) as a source of exogenous variation in insurer markups, I find that rising local health insurance premiums significantly increase the likelihood that firms adopt self-insured plans. Compared to fully insured firms, self-insured firms are better able to preserve shareholder value amid escalating healthcare costs. In response to higher expected medical claims, self-insured firms adjust corporate financial policies by increasing precautionary cash holdings and reducing capital expenditures.Awkward silence: Is manager hesitation informative?
Abstract
I study whether managers’ hesitation conveys information about market movements as well as analysts’ behavior. Hesitation is defined as the response time (RT) between analyst questions and managerial answers, measured using AI-based speaker diarization and transcript alignment over 7,300 earnings calls from S&P 500 firms (2017–2023). I find that longer RT is associated with lower contemporaneous and 1-quarter ahead cumulative abnormal returns. Quartile-split analysis confirms the association is due to abnormally long RTs. Analysts also revise earnings forecasts down and demonstrate increased uncertainty after they are faced with longer RTs. However, RT fails to predict future earning surprises, confirming prompt analysts’ response to hesitation information. This paper is the first to show that silence, specifically the managerial response time, can serve as an additional information channel. The findings contribute to the literature on information asymmetry and behavioral finance.Bailouts, Bank Regulation and Risk-Taking: A General-Equilibrium Exposition
Abstract
This paper provides a unified account of the role of bailout expectations and regulation in shaping the dynamics of banks' credit spreads and risk-taking incentives.I document that market-implied losses given default increased substantially following the Great Financial Crisis, became more volatile, and became more correlated with economy-wide and bank-specific fundamentals, reflecting diminished bailout expectations.
Using a dynamic general equilibrium of financial intermediation featuring default risk and time-varying bailout expectations, I show that lower bailout expectations substantially raised credit spreads at the onset of the crisis and that their importance diminished thereafter.
Banks endogenous deleveraging, together with higher capital requirements, explain around half of the recovery in credit spreads post-crisis.
These findings help explain how reduced bailout expectations and tighter regulation, by raising banks' cost of capital, have moved the banking sector away from risky asset markets.
Behavioral Cross-Selling: Evidence from Retail Credit Cards
Abstract
Why do some non-financial firms rely on revenue from consumer financial products? At several large U.S. retailers, direct revenues from credit card partnerships exceed total operating income. This paper proposes a theory of behavioral cross-selling, in which firms use their access to customers to cross-sell products that capitalize on behavioral biases, such as inattention or forgetfulness. We test our theory in the retail credit card market using data from a major credit bureau. Although retail cards account for only 17% of credit card balances in our sample, they generate 45% of missed minimum payments, triggering late fees. Among individuals with multiple cards, nearly half of missed payments on retail cards could have been avoided by reallocating excess payments from other cards in the same month, suggesting they cannot be fully explained by liquidity constraints. Consistent with the theory, firms in locations with more avoidable missed payments are more likely to offer retail cards and provide larger sign-up incentives. We discuss how behavioral cross-selling can help explain practices in industries such as airlines, auto dealerships, tax preparation services, and sports entertainment.Belief Distortions and Unemployment Fluctuations
Abstract
This paper shows that distorted beliefs about asset prices can amplify unemployment fluctuations through corporate hiring decisions. I decompose time-series variation in the aggregate job filling rate into expected cash flows and discount rates. Under subjective beliefs implied by survey forecasts, the job filling rate is driven by predictable errors in expected cash flows, while discount rates play a limited role. In contrast, rational expectations assign a dominant role to discount rates. A cross-sectional decomposition also shows that subjective beliefs overestimate the importance of cash flows. These patterns are consistent with a model of constant-gain learning about prices and cash flows. The learning model can generate a realistic amount of unemployment volatility, which is an improvement over a rational benchmark that underpredicts it by an order of magnitude.Beyond the Headlines: Measuring Monetary Policy Uncertainty from Bank Earnings Calls
Abstract
The role of financial institutions in the impact of monetary policy uncertainty on the economy is not fully understood. I construct an index of bank-level monetary policy uncertainty from U.S. bank earnings calls since 2002 and validate the measure with its correlation with past interest rate forecast errors and aggregate disagreement in the Survey of Professional Forecasters. SVAR evidence reveals that monetary policy uncertainty lowers real GDP and increases credit spreads. Looking at the cross-section, banks with high uncertainty charge higher interest rates in syndicated loans. The findings stress that banks beliefs impacts both lending conditions and business cyclefluctuations.
Board Diversity and the Career Progression of Women
Abstract
Over the past two decades, many countries have implemented board gender quotas to promote corporate gender diversity. While these mandates have successfully increased female representation at the executive level, their broader impact on women in entry-level and mid-level roles remains unclear. We examine this issue using the 2018 California board gender quota as a natural experiment. Leveraging Execu-Comp and a novel employer-employee dataset, we analyze its effects on women’s career progression across different organizational levels. Our findings show that the quota significantly increased the hiring and promotion of women within affected firms. Specifically, we observe a rise in both recruitment and promotion of women in mid-management positions, as well as an increase in hiring at the top management level and at the entry level. However, the quota also led to higher female turnover, particularly at the entry and top management levels. To explore potential mechanisms, we analyze employee ratings from Glassdoor, shedding light on workplace dynamics. Overall, our results suggest that board gender quotas can create more inclusiveworkplaces by improving women’s representation across organizational tiers.
Breaking the Bond: The Effect of Banker Turnover on Municipal Bonds
Abstract
This paper explores whether individual bankers add value to municipal borrowers and identifies how this value is generated. A key identification strategy exploits the quasi-exogenous shock from the 2021 Texas underwriter ban, which barred five of the largest banks from underwriting municipal bonds in the state and triggered widespread banker departures. Using novel data on banker moves, I show that affected municipalities follow their banker at twice the rate of unaffected peers. Instrumenting the follow decision in an IV–DiD framework, I find that following the banker reduces yield spreads by 36 basis points, fully offsetting the ban’s 4 basis-point spread increase. These improvements arise via three channels: an informational channel, where unrated issuers lacking public credit signals experience an extra 16 basis-point decline; a network channel, where institutional investors allocate $1 million more per quarter to the banker’s new bank; and a banker-quality channel, where following an MBA-educated banker generates an additional 22 basis-point spread reduction. Overall, relationship-specific human capital significantly shapes municipal underwriting outcomes.Can investor coalitions drive corporate climate action?
Abstract
This paper investigates the effectiveness of collaborative engagement in influencing corporate behaviour. Specifically, I examine the impact of Climate Action 100+, the world’s largest climate-related investor coalition, on targeted firms. To proxy the coalition’s engagement goals, I collect novel data on climate-related disclosure, sector-specific carbon intensities and emission reduction targets. I find that collaborative engagement influenced only target setting among a subgroup of firms selected on a discretionary basis. Surprisingly, the coalition’s scale does not amplify impact and there is no evidence of spillovers to non-target firms. These findings question whether investor coalitions drive corporate decarbonisation.Clearing the Murky Waters: The First Analyst Recommendations and Retail Trading Costs
Abstract
Retail trading has been on the rise for the past few decades, with the COVID-19pandemic accelerating this trend. However, there is a growing concern that stocks
with high retail interest often lack analyst coverage, leaving social media—a relatively
noisy source—as the primary information source for them. My research shows that the
first analyst recommendations are associated with lower investors’ trading costs, with
a more substantial reduction in effective spreads for orders executed by wholesalers
than the ones executed on exchanges. I investigate the underlying mechanism and find
that the decrease in transaction costs charged by wholesalers is likely due to reduced
effective spreads from the top two wholesalers—Citadel and Virtu—who lose market
share relative to exchanges following these recommendations. These findings suggest
that analyst recommendations provide valuable information that changes the trading
environment and potentially reduces the information rents that the largest wholesalers
can extract from retail traders.
Climate Disclosure: Theory and Evidence
Abstract
This paper investigates the real effects of climate disclosure requirements. I develop a model in which firms are financed by responsible investors who are heterogeneous in their aversion to holding polluting firms and are imperfectly informed about firms' externalities. I demonstrate that improving climate disclosure requirements has an ambiguous impact on welfare and pollution due to two different effects. First, it reduces the size of the dirty sector. Second, it also reshapes firms' shareholder base, resulting in the dirty firm being financed by investors who are, on average, less concerned about pollution. This latter effect undermines the dirty firm's incentives to adopt green technology. This challenges the conventional view that improving climate disclosure is beneficial and suggests that some level of greenwashing could be optimal. Using data on the staggered adoption of mandatory climate disclosure requirements across countries, I provide supportive empirical evidence of these mechanisms.Corporate Payouts and Local Job Creation
Abstract
Corporate payouts have surged sharply in recent decades, fueling ongoing debates over their economic consequences. This paper examines the impact of corporate dividend payouts on local job creation. Using the IRS county-level dividend income data and a shift-share instrument, I find that household dividend income is positively associated with local job creation, particularly among small and young firms. I validate this finding by implementing a difference-in-differences design to exploit the large special dividends in 2010Q4 and 2012Q4 before the dividend tax rate was scheduled to rise. The effects operate through two channels: consuming and depositing dividends, which in turn fuels small business lending. These findings suggest that capital flowing from mature firms to small businesses can stimulate economic growth, challenging recent calls to restrict corporate payouts.Credit Relationships and Dynamic Credit Constraints
Abstract
This paper presents microeconomic evidence from the U.S. syndicated loan market showing that as a credit relationship between a lender and a borrower strengthens, borrowing is more likely to be linked to a firm's earnings through loan covenants rather than physical assets as collateral. I rationalize this in a model with limited commitment and information asymmetry, in which heterogeneity in relationship status leads to heterogeneous borrowing constraints. In a credit relationship, access to earnings-based credit increases over time because of a learning mechanism. The lender learns about the borrower's private information through repeated interactions and so updates its belief. This leads to a dynamic borrowing constraint for the firm, with a switch from collateral-based to earnings-based constraints as the relationship develops. Empirically, I find that the use of loan covenant, which is often linked to earnings and increases credit supply by more than collateral use, increases as the lender-borrower relationship matures. Moreover, covenants tend to replace collateral requirements in a relationship. This provides direct evidence of a dynamic credit constraint in relationship lending, and demonstrates a new channel through which relationships increase credit supply by expanding access to earnings-based contracts. Finally, the effect of relationships on access to earnings-based credits is larger for smaller, typically more informationally opaque firms, underscoring the importance of the learning mechanism.DatedGPT: Preventing Lookahead Bias in Large Language Models with Time-Aware Pretraining
Abstract
Large language models (LLMs) exhibit significant lookahead bias when future-trained data contaminates historical predictions, violating temporal causality. We present DatedGPT, the largest time-aware model family (1.3B parameters, 12 models) eliminating this bias through strict temporal pretraining. Our framework trains separate GPT-3-XL-scale models from scratch on annually segmented CommonCrawl data (2013-2024), enforcing causal data boundaries. Benchmark evaluations show progressive NLU improvements while ablation studies confirm temporal integrity: Pre-2020 models produce near-zero probability for future events like "COVID-19"" and ""Joe Biden presidency"" in contextually appropriate prompts—surfacing only after relevant timelines. By preventing future information leakageDeviations of Covered Interest Parity, Dollar Funding Pressure, and Currency Risk Premia
Abstract
This paper investigates the relationship between deviations from Covered InterestRate Parity (CIP) and the cross-sectional variation in currency risk premia. Motivated by the currency hedging channel proposed by Liao & Zhang (2025), I use CIP violations as a measure of dollar funding pressure in postcrisis era, reflecting imbalances between excessive dollar hedging demand and constrained funding supply due to intermediation costs. Utilizing G10 currency data, I show that currencies with higher unconditional cross-currency basis values yield significantly higher excess returns, compensating investors for bearing greater dollar funding pressure risk. A tradable trading strategy that longs in currencies with high basis and shorts currencies with low basis, referred to as the "global cross-currency basis"" factor
Distorted by Design: Size-Dependent Guarantees and Capital Misallocation
Abstract
This paper studies the allocative and welfare consequences of government credit guarantees, focusing on Japan’s uniquely large and persistent Credit Guarantee Scheme (CGS). Using a quasi-experimental design based on a 1999 policy reform that raised the capital thresholds for SME eligibility, I show that newly eligible firms contracted in size relative to always-eligible peers, suggesting strategic adjustment to retain access to subsidized credit. I also document that banks with lower equity ratios issue a disproportionately high share of guaranteed loans, indicating that bank fragility amplifies distortive effects. To quantify these patterns, I develop and calibrate a general equilibrium model in which size-dependent guarantees generate endogenous bunching and misallocation. The model maps reduced-form estimates into a structural borrowing cost wedge and computes counterfactual welfare across sectors. Results show that the guarantee program reduced output by approximately 5% prior to the reform, and that raising the eligibility cutoff in 2000 improved welfare by roughly 1% of pre-reform output. These findings highlight the hidden costs of tying financial support to adjustable firm characteristics in settings with heterogeneous firm productivity and financial frictions.Do Firm Credit Constraints Impair Climate Policy?
Abstract
This paper shows that firm credit constraints impair climate policy. Empirically, firms with tighter credit constraints, measured by their distance-to-default, exhibit a relatively smaller emission reduction after a carbon tax increase. We incorporate this channel into a quantitative DSGE model with endogenous credit constraints and carbon taxes. Credit frictions reduce the optimal investment into emission abatement since shareholders are less likely to receive the payoff from such an investment. We find that carbon taxes consistent with net zero emissions are 24 dollars/ton of carbon larger in the presence of endogenous credit constraints than in an economy without such frictions.Do Sell-side Analyst Reports Have Investment Value?
Abstract
This paper documents investment value in analyst report text. Using 1.2 million reports from 2000–2023, I embed narratives with large language models and fit machine learning forecasts of 12-month returns. Portfolios formed on these text-based forecasts earn 1.04% per month, incremental to analysts’ numerical outputs and to a broad set of established factors and characteristic‑based predictors. Predictability is strongest after adverse contemporaneous news, and is amplified for reports authored by more skilled and experienced analysts. To interpret the mechanism, I apply a Shapley decomposition that attributes portfolio performance to distinct topics. The Strategic Outlook section, especially forward-looking fundamental assessments, contributes the largest share, and trading on this content alone earns 1.41% per month. Beyond providing direct evidence that analyst narratives contain value-relevant assessments that diffuse into price over time, this study illustrates how interpretable LLM-plus-ML pipelines can scale and augment human judgment in investment decisions.ESG Favoritism in Mutual Fund Families
Abstract
We investigate whether mutual fund families favor their ESG funds at the expense of their non-ESG siblings. We find that the net-of-style return spread of ESG compared to non-ESG funds within the fund family is significantly greater than the gap with non-ESG matches outside the family. The difference is around 2% per year, indicating sizable cross-fund subsidization that is mainly used to avoid underperformance of ESG funds. We link this difference in performance to fund and family characteristics and relate the observed effects to measures of environmental awareness and fund flows. Additionally, we investigate potentialmechanisms of implementing ESG favoritism.
Estimating Investor Demand Elasticity with Endogenous Firm Responses
Abstract
Estimates of investor demand elasticity are biased when firms endogenously respond to demand shocks by timing equity issuance, as observed price changes confound demand-driven price pressure with fundamental improvements. Empirically, firms experiencing mutual fund flow shocks exhibit both higher stock returns and increased equity issuance, supporting this channel. To address this identification problem, I develop a dynamic structural model capturing strategic interactions among firms, mutual funds, and residual investors. Estimating the model via indirect inference, I find a price elasticity of 2.4 for residual investors—higher than previous estimates that ignore firm responses. The higher elasticity, combined with endogenous firm-investor interactions that cause marginal price effects to diminish with shock size, implies more moderate capital misallocation effects.Exchange Rate Expectations and Currency Demand
Abstract
International exchange-traded funds (ETFs) often allow investors to choose between the currency-hedged and unhedged product. Relative holdings between these products reveal the currency demand and exchange rate expectation of investors. I find that investors’ allocation to currency-hedged versus unhedged ETFs varies with survey-based expectations and simple estimates of exchange rate expectations, like the forward discount, but less with dollar and carry factors. Portfolio-implied currency return expectations, extracted from ETF trading, become more dispersed during volatile periods in FX markets and predict future realized currency returns. This suggests that belief heterogeneity contributes to exchange rate volatility, yet aggregate investors’ trading is rational.Expansion of Informal Finance and Household Behavior
Abstract
The use of informal finance is pervasive. This study investigates how the expansion of informal finance—prompted by the 2019 Provisions on Evidence in Civil Proceedings, which clarified and broadened the admissibility of electronic evidence—affects household lending behavior. Using individual-level transaction data from a major Chinese FinTech platform, I find that the enhanced legal protections for informal lenders significantly increase the availability of informal credit within borrowers’ social networks. These protections also lead to more formalized lending transaction notes, characterized by greater detail and inclusion of financial terms such as interest rates and loan amounts, making them more closely resemble formal promissory notes. For informal borrowers, the reform also leads to higher rates of informal debt repayment, a shift from formal to informal financing source, and increased discretionary spending, particularly on vocational training and professional tools. Despite these positive effects for informal borrowers, the findings also suggest that informal lenders face increased pressure to accept lending requests from acquaintances, even after experiencing non-repayment from other informal borrowers, thereby heightening their exposure to credit risk. Two primary mechanisms underpin these effects: the legal reform reduces the reliance on social trust and lowers the threshold for repayment capability required to initiate informal loans.FHLB as Lender of First Resort: The Good, the Bad and the Ugly
Abstract
This paper investigates the dual role of Federal Home Loan Banks (FHLBs) duringbanking crises, focusing on their impact on bank survival, managerial incentives, and financial
stability. Using a bank run model, we compare three funding regimes: no external
lending, FHLB advances, and discount window borrowing. As major borrowers from
money market funds and the largest lenders in the fed funds market, FHLBs can act as “the
good, the bad, and the ugly,” depending on bank fundamentals. FHLB funding is beneficial
for well-capitalized banks, providing critical liquidity. However, for weaker banks, FHLB
advances may delay failure, benefiting bank managers and FHLBs at the expense of the
Deposit Insurance Fund and uninsured depositors. Our findings highlight that the policy
effectiveness of FHLBs hinges on the capital strength of recipient banks.
Financing Risk and Startup Growth
Abstract
This paper investigates how financing risk, the forward-looking expectation of limited future funding availability, shapes startup behavior. I develop a model of intertemporal investment under uncertainty and show that financing risk, distinct from traditional financial constraints, distorts investment, growth, and survival. I construct a novel text-based measure of financing risk derived from 4.1 million news articles. Exploiting exogenous variation from macroeconomic uncertainty shocks, I find that among the recently funded startups, financing risk causally reduces innovation, especially resource-intensive, exploratory, and novel types. Financing risk also slows employment growth, weakens product development, and increases failure rates. These findings highlight the broader role of anticipated future funding constraints in shaping startup growth in the absence of current financial constraints.Firm-level Political Risks and Newly Issued Corporate Debt Maturity
Abstract
I examine the impact of firm-level political risk on the maturity structure of newly issued corporate debt. Using Hassan et al.'s (2019) political risk measure, I find that heightened political uncertainty leads firms to issue shorter-maturity debt. This relation holds across multiple proxies for new debt maturity and various econometric techniques designed to alleviate endogeneity concerns. The effect is especially profound for smaller firms, those with lower credit ratings, and firms with patents nearing expiration. Path analysis confirms that cash flow volatility partially mediates the relationship between political risk and new debt maturity. Conversely, political bipartisanship, contributions/lobbying to the federal ruling party, and alignment with the federal government at the industry and state levels help moderate the relationship. Clustering analysis indicates a significant uptick in shorter-term debt issuance during Q1-Q3 of presidential election years. Collectively, these findings underscore the significance of firm-level political risks in shaping corporate debt maturity decisions.Trading Volume and Monetary Policy Surprises
Abstract
High-frequency identification of the causal effects of monetary policy relies on measuring monetarypolicy surprises–changes in interest rate futures prices in narrowwindows around FOMC
announcements. Constructing these surprises typically entails two key assumptions: a fixed
event window and fixed loadings over time that aggregate the set of interest rate futures contracts
into the resulting surprise measure. This paper relaxes both assumptions and introduces
the Volume-Based Monetary Policy Surprise (VBS), which uses abnormal trading volume to let
the market endogenously determine the relevant event windows, set of futures contracts, and
their respective loadings for each announcement. This leads to substantially larger estimated
effects of monetary policy. The announcement-specific event windows flexibly capture when
prices continue adjusting beyond conventional 30-minute windows while the announcementspecific
loadings naturally shift toward longer-dated contracts when the Federal Reserve relies
on forward guidance about future policy. The VBS doubles the estimated impact on Treasury
yields and equity markets and generates sizable impacts on macroeconomic aggregates.
Flooded but Thriving? The Uneven Economic Effects of Floods and Flood Risk
Abstract
I study how floods and chronic flood risk affect U.S. establishments and firms by combining high-resolution remote sensing flood data, FEMA flood maps, and establishment-level data. After floods, I find that establishments show growth in employment and sales, with recovery supported by insurance payouts and federal aid. Using a triple-difference and spatial regression discontinuity design around regulatory boundaries, I provide novel causal evidence that flood insurance plays a critical role in enhancing recovery, particularly for establishments subject to the mandatory National Flood Insurance Program (NFIP) purchase requirement. Federal aid, notably SBA disaster loans, also strengthens recovery by supporting small businesses and generating local spillovers. In contrast, chronic flood risk is associated with persistent declines in employment and business diversity, likely driven by higher insurance expenses and increased business exits. At the firm level, these patterns persist and the dynamics aggregate: firms with greater flood risk disclose these risks more proactively and reduce investment in physical assets, while markets react more negatively to floods for firms lacking prior exposure or disclosure. Overall, my findings highlight how insurance and risk communication enhance resilience to climate shocks, while underscoring the need for policies that mitigate long-term vulnerability without encouraging unsustainable development.Flow-Induced Demand Pressure from Option-Trading ETFs
Abstract
The assets under management of option-trading exchange-traded funds (ETFs) have grown more than 120-fold since 2018. This paper exploits option-trading ETFs to examine how flow-induced demand pressure and exogenous rollover trade demand pressure affect the implied volatility surface. I show that demand pressure from these ETFs significantly affects the implied volatility surface, with the magnitude of the effect varying with option characteristics—particularly moneyness and days to expiration—due to differences in option vega. In addition, liquidity frictions also explain the magnitude of impact. These findings suggest that flow-induced demand pressure plays an important role in shaping both the term structure and the moneyness curve of implied volatility.From Failure to Innovation: Strategic Knowledge Transfer in Corporate Venture Capital
Abstract
This paper investigates how corporate venture capital (CVC) investors strategically extract value from failed start-up investments and facilitate the reallocation of innovation resources and talent. I find that failed CVC-backed start-ups hold a higher proportion of patents in technological areas where their parent firms do not have prior patenting activity, introducing parent firms to novel technological domains despite financial losses. CVC parent firms are more likely to cite patents from failed start-ups when those patents introduce previously unexplored technological knowledge. Post-exit, parent firms integrate this knowledge by exploring adjacent or broader technological domains. They also frequently acquire patents from these start-ups and hire inventors with specialized expertise in these unfamiliar areas, effectively incorporating both codified and tacit knowledge into their innovation pipelines. Finally, failed start-ups tend to have higher technological overlap with the competitors of the CVC parent than those successful ventures, suggesting that parent firms may use these investments to gain insights into emerging competitive technologies and strengthen their strategic positioning.From Index Trackers to Risk Managers: The Expanding Role of Derivatives in ETFs
Abstract
Using regulatory data from the SEC’s N-PORT filings, we provide the first systematic study of derivative use by exchange-traded funds (ETFs). Nearly 60% of ETFs use derivatives, with greater derivative weight and exposure than mutual funds. Derivative use varies across ETF types: passive ETFs primarily use futures and forwards to reduce costs, while active ETFs rely on options strategies to improve risk profiles. Despite charging higher fees, active derivative-using ETFs attract more flows and exhibit reduced fee sensitivity. We show that these flows appear to be driven by superior downside protection, suggesting that investors value this benefit. Moreover, the extent of derivative reliance predicts both improved risk profiles and higher fees. Overall, our study highlights the strategic role of derivatives in ETF market competition.Green Neighbors, Greener Neighborhoods: Peer Effects in Green Home Investments
Abstract
Utilizing a nearest-neighbor research design, I find that households exposed to green neighbors within 0.1 miles are 1.6 times more likely to make their homes green within a year than unexposed households. The exposure also increases the likelihood of multi-property owners certifying their faraway secondary properties green, emphasizing that information from neighbors, not neighborhood characteristics alone, drives the effect. While financial benefits including green home prices, electricity savings, and regulatory incentives strengthen peer effects, pro-environmental preferences do not. An information-cost-based discrete choice model explains the findings and suggests that incorporating peer effect metrics in subsidies may accelerate green home investments.How Do Lender-Household Relationships Affect Mortgage Refinancing?
Abstract
What role do lenders play in household refinancing? This paper provides insights into this question using a unique dataset that tracks lender-household relationships for about 25 million mortgage loans. The empirical results show that an exogenous disruption to lender-household relationships substantially reduces a household's refinancing probability by 69.45%. In particular, rather than switching to new lenders, the probability that a household refinances with new lenders also declines by 51.91%. The disruption of relationships does not affect refinance loans’ interest rates, fees, or performance. The evidence is consistent with the channel that relationship lenders help households refinance by providing information about potential refinancing opportunities. The paper further develops a structural model in which relationships affect households’ awareness of refinancing opportunities and refinancing costs simultaneously.How Insurance Claim Delays Shift Mortgage Costs After Disasters?
Abstract
I examine how delays in homeowner insurance claim payments after natural disasters affect mortgage outcomes. Focusing on Hurricane Irma and leveraging Freddie Mac single-family loan data linked to insurer payment behavior, I show that areas exposed to slower-paying insurers—those paying less within the year of loss—experience higher rates of mortgage delinquency, forbearance, and loan modification. The results suggest that timely insurance payments are critical for household liquidity and help reduce the need for costly GSE interventions. Back-of-the-envelope estimates indicate substantial welfare losses when insurance fails to buffer post-disaster financial risk.How Mutual Fund Managers and Investors Respond to U.S.–China Geopolitical Risk
Abstract
I examine how mutual fund managers and investors respond to U.S.–China geopolitical risk, using a novel U.S.-China Geopolitical Risk Index (UC-GRI) based on White House press briefings and a flow- and price-adjusted Fund Manager Active Reallocation (FMAR) measure. When U.S.–China bilateral geopolitical risk rises, U.S. managers reduce allocations to Chinese equities, reallocate from Asia high- to low-risk regions, and shift domestic holdings toward firms with lower exposure to China. U.S. fund investors show strong home preference and retreat from sensitive regions. Institutional investors respond more selectively than retail investors. Chinese managers pull back from local markets but maintain U.S. exposure. Chinese fund investors exhibit limited adjustments. These responses are more pronounced after 2018. The results highlight the heterogeneity in geopolitical risk perception across countries and investor types.Import Competition and U.S. Firms: A Text-Based Approach
Abstract
This paper develops novel firm-level, text-based measures of import competition by linking the product descriptions in U.S. firms’ 10-K filings with those in shipping manifests for imported goods. These new measures capture cross-firm heterogeneity in exposure to international competition that is not reflected in industry-level import penetration metrics widely used in prior research. Using data from 2014 to 2022, I find that greater import competition from low-wage countries such as China is associated with lower sales and profitability, even after controlling for industry-level import competition. In response, firms increase cash holdings and reduce dividends, reflecting a shift toward more conservative financial policies. The analysis also shows that firms strategically adjust their product scope: they narrow focus and pursue vertical differentiation to compete against low-cost imports, while broadening product offerings in response to rising local market competition. These results highlight the strategic financial and operational responses U.S. firms adopt in the face of global trade pressure and demonstrate the value of firm-level, text-based metrics in analyzing the impact of international competition.Information and Innovation: Evidence From Railroad Expansion in the Nineteenth Century
Abstract
I study the causal effect of information flow on innovation by exploiting the staggered expansion of the U.S. railroad network between 1840 and 1900. Using an event study analysis surrounding the opening of the first railroad in each county, I find that counties where the first railroad was opened generate more patents over the next five years compared to those without one. These patents tend to be more important and impactful, though less novel. The effect is stronger in counties that are more connected to other counties through the railroad network. My findings suggest that improved information exchange efficiency enhances innovation. In my future work, I plan to leverage text-based patent similarities to further substantiate the role of information diffusion as a key channel facilitating innovation activity.Institutional Equity Investors and Corporate Debt Financing: A Cross-Market Perspective
Abstract
Institutional equity investors significantly influence corporate debt issuance due to their triple role as shareholders, equity market participants, and informational intermediaries to the debt market. Among U.S. non-financial public firms, a five-percentage-point increase in institutional equity ownership is associated with a 22% rise in debt issuance. Two novel mechanisms contribute to this relationship: a supply-side information channel where debt investors learn from equity investors, and a demand-side benchmarking channel that affects shareholders' effective risk aversion. Causal evidence is established using a regulatory shock to mutual fund disclosure and equity benchmarking intensity as an instrumental variable. A calibrated model featuring both mechanisms reveals that the information channel accounts for roughly one-fifth of the debt-equity holdings relationship, with the informational frictions in debt issuance representing about 1% of equity value. Additionally, a benchmarking-induced increase in institutional ownership comparable to Russell 2000 inclusion amplifies the relationship by 9%. These results highlight the importance of cross-market investor linkages and firm responses to investor characteristics in understanding financing outcomes.Institutional Investors and House Prices
Abstract
Institutional investors are playing an increasingly important role in residential real estate markets. This raises the possibility that their actions might drive aggregate market outcomes and may change how and which macro-financial shocks transmit to house prices. We show that a demand shock from institutional investors has a positive and persistent effect on aggregate euro area house price growth and mortgage lending volumes. Institutional investors also increase their purchase activity following a loosening of monetary policy. Exploiting regional heterogeneity across eight euro area countries, we show that institutional investors weaken the link between house price growth and local economic fundamentals but strengthen the sensitivity to monetary policy and financial market developments.Interdependence of Bank Run Risk and Interest Rate Risk
Abstract
study how strategising to mitigate liquidity risk in stress periods exposes banks to interest rate risk in normal times. Building on (Drechsler et al., 2021), I show that small banks in the bottom quintile are not able to perform interest sensitivity matching and hence, are exposed to interest rate risk. These banks are primarily funded by retail deposits which results in low interest expense beta. Despite being funded by retail deposits, I show that stress periods trigger a relative reallocation of deposits from small banks to large banks, exposing these banks to higher funding instability in stress periods. To mitigate the anticipated bank-run risk, small banks hold shorter-duration assets to maintain liquidity in stress periods. Holding shorter-duration assets results in increasing their interest income beta.As a consequence, they end up pairing low-interest expense beta with high-interest income beta leading to an interest sensitivity mismatch. I also conduct additional tests using the variation in banks’ presence on the reciprocal deposit network to show that since small banks on the network experience lower bank-run risk in stress periods, they can perform interest sensitivity matching to mitigate interest rate risk. These results demonstrate the interdependence of liquidity risk and interest risk management and emphasise the importance of the stability of the deposits in a bank’s ability to provide long-term credit.
Interpretable Systematic Risk around the Clock
Abstract
In this paper, I present the first comprehensive, around-the-clock analysis of systematic jump risk by combining high-frequency market data with contemporaneous news narratives identified as the underlying causes of market jumps. These narratives are retrieved and classified using a state-of-the-art open-source reasoning LLM. Decomposing market risk into interpretable jump categories reveals significant heterogeneity in risk premia, with macroeconomic news commanding the largest and most persistent premium. Leveraging this insight, I construct an annually rebalanced real-time strategy that hedges the most priced jump risk, achieving an out-of-sample Sharpe ratio of 0.84 and delivering significant alphas relative to standard factor models. The results highlight the value of around-the-clock analysis and LLM-based narrative understanding for identifying and managing priced risks in real time.Investment and Governance: Through the Lens of Sustainability
Abstract
Does investment inspire better governance? Using a global sample of 3,944 sustainable bonds, issued by public firms from 2013 to 2022, the causal generalized method of moments (GMM) estimates suggest that 1% increase in sustainable debt to total debt ratio improves the sustainable governance practices by 9%. To address potential simultaneity bias, we employ a method that utilizes the heteroskedasticity of structural shocks. Our findings also confirm that the standard panel regressions, even with fixed effects, may exaggerate effects due to simultaneity. Our findings remain consistent across different measures of sustainable governance using different databases and battery of other checks.Is Ownership Concentration Driving Markups in Multifamily Housing?
Abstract
An increasing portion of multifamily apartment units are owned by a small number of owners. At the same time, rents have increased substantial; the share of Americans paying more than 30% income in rent increased from 20% in 2000 to 50% in 2022. In this paper, I examine whether ownership concentration in multifamily housing is driving higher rents and markups. I start by exploring the time-series and cross-section of ownership concentration from 2010 to 2024 across US submarkets. Ownership concentration in these submarkets has increased over this period, and the top five owners own 31% of all units in the average submarket and 52% in the average submarket-by-class as of 2024. Then, using within-market-by-year variation in ownership market share for the same property over time, I show that a one standard deviation change in ownership market share is associated with a 2.7% increase in rents and a 5.3% increase in net operating profits while maintaining stable occupancy rates, operating expenses, and capital expenses. The effects are robust to including increasingly tight fixed effects. Further, the effects are increasing over time: the effects of ownership market share on both rent and profits have both more than doubled between 2010-12 and 2022-24. To tighten the causal interpretation of my results, (1) I use difference-in-differences around changes in ownership of a focal property and (2) I analyze the effects of an owner buying additional properties in the same submarket on their existing properties. Finally, I study the impacts of changes in ownership concentration on the demographics of renters to understand the distributional and racial impacts of my findings.Labor Shortage, Hiring and Asset Returns
Abstract
In this paper, I find labor hiring constraint matters for stock return, and is essential to explain the negative hiring-return relation, both in the cross section and in time series. To proxy for labor hiring constraint, I construct a firm-year level measure of labor shortage using textual analysis of firms' SEC fillings. Via portfolio sorting and predictive regression, I show that labor shortage predicts low stock return, and the negative relationship between firm's hiring rate and its future return is only significant for firms that discuss labor shortage in their filings. These patterns are consistent with predictions from a neoclassical framework with hiring adjustment cost. In addition, I document relations between labor shortage and firms' policy and operation dynamics. Firms of high growth are more likely to discuss labor shortage. Once they do, their current hiring and future investment rate drop, leverage and book to market ratio increase. These findings are robust to small firm bias, and alternative interpretations of labor shortage measure.Learning about Noise, Information Disclosure, and Market Stability
Abstract
How do rational investors’ learning about noise trading and firms’ information disclosure jointly affect financial market stability? We examine this question in a model in which a part of a risky asset’s payoff is publicly disclosed and rational investors can learn about noise traders’ demand for the asset. We demonstrate that reducing information disclosure can eliminate the multiple self-fulfilling equilibria caused by learning about noise. In an extended model with endogenous information disclosure, we show that the equilibrium expected asset price is increasing and discontinuous in the intensity of rational investors’ learning about noise. Our results predict that (i) more information disclosure can exacerbate the market fragility caused by high-frequency trading, and (ii) market crashes can occur during the process of regulating high-frequency trading. We also discuss the empirical relevance of our results.Machine learning in foreign exchange
Abstract
We apply machine learning to predict currency excess return cross-sectionally while addressing the black-box issue through interpretability techniques. First, neural networks (NN) significantly outperform traditional models in predicting currency excess returns including the random walk, highlighting the advantages of more flexible predictive functions. Second, NN-based portfolios achieve higher Sharpe ratios, underscoring their economic value. Third, both local and global interpretability techniques reveal that interactions between global macroeconomic factors and currency-specific characteristics are key drivers of FX risk premia.Managerial Attention to Financial Markets
Abstract
This paper introduces a novel measure of firm-level managerial attention to financial markets, constructed from earnings call transcripts spanning 98,010 firm-year observations (2007-2023). Firms whose managers devote greater attention to financial markets exhibit greater investment-price sensitivity, supporting the price feedback theory. Managerial attention also shapes financing choices: high-attention firms are likely to avoid equity issuance in favor of debt, consistent with the pecking order theory, but become more willing to issue equity when market conditions are favorable. I document persistent heterogeneity in attention across industries, likely driven by differences in information asymmetry their business activities present to external financial markets.Market's Time-Varying Sensitivity to Macro News and Monetary Policy: the Role of Attention
Abstract
Market sensitivity to macroeconomic announcements (e.g., payrolls, CPI) varies substantially over time. Focusing on interest rate responses, I document that sensitivity increases during periods of high inflation, rising unemployment, and when the Federal Reserve emphasizes macroeconomic conditions. To uncover the fundamental driver, I distinguish between two hypotheses—changes in investor attention and variations in the perceived data dependence of the monetary policy rule—using an empirical strategy based on their differing implications for the transmission of monetary policy. Consistent with the attention hypothesis, I find that the impact of FOMC announcements is amplified in periods when market sensitivity to macroeconomic announcements is high. A standard New Keynesian model with cognitive discounting is used to illustrate the differences between the two mechanisms. Overall, this paper highlights the role of attention in the effectiveness of monetary policy and macro dynamics.Memory and Generative AI
Abstract
This paper tries to understand Generative AI’s decisions under risk as an economic agent. Exploiting a novel experimental setting, we show that it uses memories to make decisions, even when the memories are not in the same decision domain. When displayed with irrelevant images with positive feelings, it becomes more risk-loving and will choose to invest more in stocks, and vice versa. Although emotional shocks strongly bias investment choices, they have minimal impact on GAI’s beliefs. This mechanism is further causally supported with a supervised fine-tuning technique known as knowledge injection that can edit the language model’s memories. Empirical results show that both domain-specific memories like financial news and non-domain-specific memories like Yelp restaurant reviews significantly affect GAI’s investment choices.Mental Barriers to Investing: Psychological Fixed Costs and Stock Market Participation
Abstract
This study examines how affective states, proxied by mental health, shape household stock market participation. Using rich panel data from the German Socio-Economic Panel (SOEP), we show that better mental health significantly increases the likelihood of stock ownership, while symptoms of depression and chronic worry reduce participation. To isolate causal effects, we exploit the COVID-19 pandemic as a natural experiment in a Difference-in-Differences Instrumental Variables (DiD-IV) design. Individuals with weaker pre-crisis social networks experienced larger declines in mental health, which we use to identify exogenous variation. Our results suggest that a one standard deviation increase in mental health predicts approximately 120,000 new stockholding households annually in Germany. We introduce a conceptual framework with three channels: external beliefs, internal beliefs, and preferences to explain how mental health shapes investment behavior. Our results strongly indicate that mental barriers are key constraints to stock market entry.Mitigating Moral Hazard in Delegated Investment through Recommendation Algorithms
Abstract
Digital platforms are increasingly serving as intermediaries in delegated investment, particularly by adopting recommendation algorithms that deliver personalized suggestions to a large user base. We develop a model to analyze the platform's investor-optimal algorithm design where investors with heterogeneous risk aversion contract with a portfolio manager based on recommendation status. Investors may have limited knowledge about their types, while the manager has risk-chasing incentives due to limited liability. We demonstrate that algorithms can mitigate managers' moral hazard in over risk-taking --- without affecting the contract --- by acting as both information gatekeepers and commitment devices, harnessing the scale of the user base. Optimal recommendation probabilities are non‑monotonic in historical returns. Unlike in consumption platforms, algorithms here extract noisy signals about the manager's actions from historical returns, reduce recommendations under ambiguous signals, and potentially compensate for clear signals, leading to an information rent paid by investors. We further discuss on algorithmic inequality, the joint design of algorithms and contracts, and comparisons to fund ranking systems. Our results emphasize the innovative role of recommendation algorithms as a digital financial service. Methodologically, we provide a general approach for algorithm design problems in function space with potentially non-monotonic solutions.Mixed Messages by Firms’ Managers: An Exploration into the Strategic Use of Tonal Inconsistency across Managerial Financial Communications
Abstract
Using conference call transcripts and contemporaneous press releases for the same quarterly earnings announcement, I document that inconsistent managerial tone across these communication channels predicts future negative abnormal returns and influences the speed of market price responsiveness to earnings disclosures. My empirical results demonstrate that such tonal inconsistencies forecast significantly slower market reactions and are robustly associated with subsequent fundamental adverse changes in firm performance and profitability. I further find that these inconsistencies foster market ambiguity, thereby delaying investor responses and facilitating strategic and opportunistic insider trading that appears to capitalize on this delayed price incorporation. Overall, the evidence aligns with managers strategically employing mixed messaging in financial disclosures to mislead investors about underlying firm fundamentals by manipulating market perceptions and exploiting informational frictions induced by ambiguity.Monetary Policy Shocks: A New Hope
Abstract
I develop a novel framework for computing monetary policy surprises by sys-tematically processing the entire set of regular Federal Reserve communications
using context-aware Large Language Models (LLMs). My approach analyzes the
complete institutional communication cycle—Beige Books, FOMC Minutes, and
policy Statements—as an integrated narrative system rather than isolated docu-
ments. The multi-agent architecture employs specialized LLMs to read and synthe-
size these full documents in sequence: extracting economic assessments from eight
annual Beige Books, analyzing internal deliberations from Minutes, incorporating
the full history of policy statements, and generating genuine surprises by comparing
ex-ante expectations with realized decisions. Crucially, I form these expectations
using only information available 2-3 weeks before each meeting, ensuring surprises
reflect the unexpected evolution of Fed thinking during the blackout period. The
resulting narrative surprises are as unpredictable as market-based measures (8-12%
R² on standard predictors) yet explain 61.5% of policy rate changes compared to 15-
17% for market surprises. This stark difference reveals that most monetary policy
”news” stems from how Fed thinking evolves between its documented communica-
tions and final decisions—evolution that cannot be predicted ex-ante but dramat-
ically moves policy when revealed. By processing the Fed’s complete regular com-
munication apparatus as an interconnected system, the framework demonstrates
how LLMs can measure the true information content of central bank transparency,
distinguishing predictable policy rules from genuine monetary shocks.
Housing Duration and Interest Rates: Evidence from Reaching-for-Income Investors
Abstract
In fixed-income markets, long-duration assets are more sensitive to interest rate changes, and this principle is commonly assumed to extend to other asset classes. I show that the opposite holds in housing markets: short-duration properties are more, not less, sensitive to interest rate changes. Using data from the American Community Survey, I find that a one-percentage-point cut in interest rates raises house prices by 1.86 percentage points over two years. However, housing markets with a duration one standard deviation below the mean experience an additional 0.71-percentage-point price increase. I argue that this inversion arises from a discount-rate channel driven by “reaching-for-income” investors. Short-duration properties offer higher rental yields. After rate cuts, income-seeking investors disproportionately target high-yield, short-duration properties for investment, prioritizing near-term income over long-term returns. This behavior pushes up prices and lowers discount rates in short-duration markets, generating a non-parallel shift in the term structure of housing discount rates. These findings highlight investor preferences as an important driver of heterogeneity in housing market responses to interest rate changes.Monetary Policy, Industry Leaders and Growth
Abstract
We provide novel evidence on heterogeneity in the transmission of monetary policy shocks to firms' financing conditions and investment, highlighting systematic differences between industry leaders and followers. Our key insight centers on differences in firms' profitability risk over the business cycle. Industry leaders - firms with larger market shares - generate profits that are relatively stable and less sensitive to aggregate economic fluctuations. When monetary policy tightens, aggregate risk is repriced, prompting investors to disproportionately raise the required rates of return for follower firms, whose profits are more cyclical and thus riskier. Following monetary tightening, industry leaders experience significantly smaller increases in financing costs, enabling them to sustain relatively higher growth expenditures in the form of physical investment and research and development. A stylized model featuring heterogeneous profit cyclicality rationalizes our empirical findings. Our results highlight previously unexplored distributional consequences of monetary policy, emphasizing its persistent effects on industry competition and firm investment dynamics.Mortgage structure, household saving and the wealth distribution
Abstract
Fixed amortization schedules in mortgage contracts force homeowners to save into illiquid home equity. I show that in an otherwise standard life-cycle model, homeowners rationally respond to mandatory repayments by cutting consumption and increasing precautionary saving in liquid assets. Consistent with this mechanism, which does not require any behavioral biases, I document in Euro area data that younger, poorer homeowners have much higher saving rates than their non-mortgage peers, and allocate a large share of saving to mortgage repayment. The exception is the Netherlands, where interest-only mortgages are common. This complements recent quasi-experimental evidence showing large effects of amortization requirements on saving. A quantitative version of the model reproduces these facts and shows that mandatory amortization increases both home equity and financial wealth accumulation, particularly up to age 40. Wealth-to-income ratios increase by close to a quarter for lower-income homeowners at age 40, while the impact for the highest-income households is minimal. These effects build up over time and have substantial implications for aggregate consumption and wealth: mandatory amortization dampens total wealth inequality, but increases consumption volatility and financial wealth inequality.Mutual Funds' Preemptive Response to Major Cyber Attacks
Abstract
This paper examines U.S. actively managed equity mutual funds’ trading behavior around cyberattack announcements. It documents that active funds preemptively trim high–cyber‑risk holdings and tilt into low–cyber‑risk stocks, particularly among funds that disclose cyber risk in their prospectuses. It shows that firms with stronger data‑privacy safeguards exhibit smaller tilts. It further demonstrates that these anticipatory reallocations generate significant subsequent net inflows driven by institutional investors. These findings extend the literatures on cyber‑risk spillovers, pre‑announcement trading, and mutual‑fund skill, and highlight active managers’ informational advantage in responding to emerging cybersecurity threats.Nonbank Lending and the Transmission of Monetary Policy: Evidence from U.S. Small Business Loans
Abstract
Nonbank lending has reshaped the original credit markets since the Global FinancialCrisis. This paper finds that nonbanks expand significantly when monetary policy
tightening. Using Uniform Commercial Code filings, I find that a one percent increase
in the Federal Funds rates is associated with a 1.4 percent rise in nonbank market
share. Nonbanks gain significantly more market share in counties where banks hold
greater market power in local deposit markets. These findings are robust to controls
for the bank capital channel and time-varying firm heterogeneity. The real impacts
suggest that the interest-rate-driven rise in nonbank lending does not promote local
economic development, as bankruptcy rates and unemployment rates both increases
significantly in more concentrated counties following interest rate hikes. This paper
highlights the importance of nonbanks in the transmission of monetary policy to small
business lending.
Open Banking and Competition in Banks and Fintech: Evidence from Mobile Apps
Abstract
This paper examines the impact of open banking adoption on competition and innovation in financial services sector. I construct a novel dataset by combining official open banking authorization records with historical Android app source code to track the integration of open banking among finance apps in the UK and EU. Linking this with high-frequency app performance data, and exploiting cross-country variation in authorization status within the same app, I provide causal evidence that access to consumer banking data boosts app performance. These effects are especially strong during the COVID-19 pandemic, particularly among lending and investment apps and fintech startups. I also find that banks experiencing greater competition from fintech apps, as measured by the similarity of their digital services, face declines in loan issuance, although their profitability remains stable. Overall, the results show that open banking intensifies competitive pressure and reshapes market dynamics in the mobile finance ecosystem.Open-source Generative AI and Firm Value: Evidence from the Release of DeepSeek
Abstract
In this paper, I exploit the release of DeepSeek in January 2025 as an exogenous shock to the availability of open-source Generative AI (GenAI) models and investigate its value implications for corporate GenAI adopters. I document that U.S. firms with high pre-event GenAI exposure earn an average cumulative abnormal return of 1.2% over the event window, relative to low-exposure firms. The effect is stronger for firms with tighter financial constraints and for firms with proprietary information concerns, consistent with the benefits of open-source GenAI in reducing adoption costs and mitigating privacy concerns. In the post-event period, high-exposure firms are more likely to articulate adoption plans and benefits of open-source models during conference calls, and to embed DeepSeek in their algorithms shared on GitHub. High-exposure firms also experience an upward revision of analyst forecasts and a more positive media tone. In contrast, GenAI providers and their hardware suppliers experience negative abnormal returns. Lastly, the baseline analysis for Chinese firms yields a larger market impact of the DeepSeek’s release. Overall, the findings indicate that open-source GenAI can further unlock the valuation potential of adopting GenAIOptimal Capital Deployment: Dry Powder and the Option to Invest
Abstract
Private equity (PE) firms operate in an environment where uninvested committed capital, or dry powder, influences investment pacing, pricing discipline, and fund performance. This study seeks to examine optimal fund investment in the context of the constraints implied by available dry powder with a focus on identifying the dynamic strategy that maximizes GP expected value. For example, the analysis seeks to examine if GPs undertake strategic actions, such as accelerating deployment or deterring competition in ways that increase GP value but not LP returns. In addition, the analysis seeks to evaluate potential spillovers of GP strategic behavior such as whether rising aggregate capital overhang inflates valuation multiples or allows dominant firms to restrict liquidity and limit new entry. By integrating fund-level cash flows with transaction pricing and investor composition, this study hopes to provide new evidence on how dry powder functions both as a constraint and a strategic tool, with implications for fund managers, investors, portfolio companies, and policymakers.Optimizing portfolio weights with machine learning
Abstract
A neural network that directly optimizes portfolio weights outperforms the conventional two-step approach of forecasting returns and then sorting stocks, achieving an out-of-sample Sharpe ratio of 4.06 from 1987 to 2021. Its output layer assigns a weight to each stock, and the loss function is defined directly on the portfolio’s return, with an additional quadratic return term to penalize volatility. The model can also optimize net‑of‑cost returns by incorporating transaction‑cost assumptions into the loss function, yielding a portfolio with lower turnover and more liquid stocks.Partisan Beliefs and Housing Decisions: Evidence from the U.S. Real Estate Market
Abstract
Partisanship influences both housing market expectations and home purchase decisions. Using survey data, we first document that individuals expect higher home price growth when their affiliated party controls the White House, and this tendency has increased significantly over time. We then construct a novel dataset linking voter registration with housing transactions in New York State from 1996 to 2023. We find that Republicans have an annual home-purchase probability that is 0.25% higher than Democrats. Moreover, homebuyers whose party affiliation aligns with the sitting president’s party are more likely to buy—a partisan-alignment effect equivalent to over 1.2 million additional home purchases nationwide each year. This effect is stronger among men, younger individuals, and consumption-oriented homebuyers. Moreover, we show that political affiliation moderates gender disparities: the gender gap in homebuying is narrower among Democrats and significantly wider among Republicans. These findings highlight the role of partisan bias in shaping major household financial decisions.Partisan Bias in Venture Capital Financing
Abstract
This study investigates the effects of political homophily between venture capital (VC) partners and company CEOs on investment decisions and outcomes. Using a comprehensive dataset of U.S. VC investments matched with political donation records from 2000 to 2021, we find that political similarity increases the likelihood of investment but negatively impacts exit performance, lowering IPO and M&A success rates and delaying exits. These findings support the in-group favoritism explanation. Shared partisanship promotes trust and collaboration but can lead to overconfidence and groupthink that deteriorates exit performance. Alignment with the broader political environment (e.g., the incumbent government or local political preferences) can mitigate these effects by enhancing legitimacy and access to resources. Our analysis of investment structure shows that politically aligned CEO-VC pairs favor early‐stage, first‐round, non‐syndicated deals, yet experience slower follow-on financing and longer intervals between rounds. These findings offer novel insights into how ideological alignment influences venture investment behavior and performance with implications for entrepreneurs, investors, and policymakers.Partisan Fed
Abstract
I show that political alignment between Federal Open Market Committee (FOMC) members and the incumbent U.S. President systematically influences monetary policy. I construct two novel, individual-level measures of political alignment for each member of the FOMC, based on their political campaign contributions and appointments to public roles. Using a Difference-in-Differences design around four presidential party transitions between 1992 and 2019, I find that a individual-level positive shift in political alignment with the sitting U.S. President leads FOMC members toward more expansionary policy preferences and more optimistic macroeconomic forecasts (over-forecasting GDP and under-forecasting inflation). At the committee level, a one-point increase in political alignment within the FOMC lowers the federal funds rate by approximately 25 basis points relative to the Federal Reserve staff’s benchmark recommendation. These politically driven rate decisions generate a partisan business cycle: periods of political alignment between the Fed and the executive lead to more frequent interest rate cuts, stimulating short-term gains in output, employment, and stock market performance, but contributing to higher inflation in the long run. Conversely, during periods of political misalignment the FOMC raises interest rates above the apolitical benchmark, resulting in short-run output contractions, but controlling long-run inflation.Payoff over Panorama: Mental Accounting and Asset Class Selection
Abstract
This study examines how mental accounting shapes the participation and asset class selection decisions of individuals. Using a unique question from the Dutch National Bank Household Survey, I identify individuals with mental accounting bias. This enables me to compare the financial decisions of individuals with and without this bias to determine its impact on investment behavior. My findings show that mental accounting is associated with 3.5% lower participation in risky markets, representing a 12% relative decrease. However, conditional on investing, individuals with mental accounting bias favor high-risk, high-return assets. Specifically, compared to individuals without this bias, they are 30% more likely to invest in cryptocurrencies but 20.7% and 22.7% less likely to invest in stocks and mutual funds, respectively. Moreover, among those who invest, individuals with mental accounting bias are most likely to exclusively hold cryptocurrencies, avoiding diversified portfolios. The influence of mental accounting extends to other investments, including real estate, bonds, options, and individual stock selection, and it is persistent over time. The findings align with Mental Accounting theory and help explain risk-taking behaviors beyond what risk and loss aversion alone can account for.Pension Reform and Asset Allocation: Evidence from Annuity Cuts to Retired Civil Servants in Taiwan
Abstract
How do households reallocate assets when faced with a sudden decrease in lifetime income? We examine retired civil servants’ asset allocation in response to a retroactive pension cut in Taiwan that sharply reduced their monthly annuity benefits. Specifically, the replacement ratio for these retirees was cut from the 75%-95% range to the 30%-62.5% range, representing a substantial and unexpected loss in guaranteed retirement income. Leveraging comprehensive administrative tax data and a difference-in-differences design, we compare asset allocation and financial behaviors of affected retirees to those of unaffected private-sector retirees. We find that, in response to the reform, civil servant retirees increased their stock-to-wealth ratios by 10%, stock holdings by 7%, exposures to systematic market risk by 13-15%. They also increased their exposures to other risks and exhibited a heightened disposition effect: the gap between realized gains and realized losses widened by 42%. These behavioral changes were accompanied by reductions in bank deposits and total wealth, as well as an increased likelihood of labor force re-entry. Our large-scale empirical evidence shows that households respond to cuts in annuitized income by shifting toward riskier financial positions, consistent with the predictions of prospect theory—that individuals become risk-seeking in the domain of losses.Political Affiliation and the Pricing of Climate Risk in Mortgages
Abstract
Using voter registration data of loan officers originating residential mortgages in coastal areas, I analyze whether climate change partisanship is reflected in mortgage lending.I find that Democratic loan officers charge higher loan spreads for mortgages on properties exposed to sea level rise (SLR) than Republican loan officers. The results hold after controlling for loan officer fixed factors. Partisan sorting is more pronounced for properties outside FEMA-designated flood zones, and for loan officers located nearer the coast or in communities with fewer climate change believers. These findings highlight how political ideology shapes the pricing of climate risks in mortgages.
Political Donations and Rent-Seeking: Evidence from the U.S. Health Insurance Industry
Abstract
In recent years, health insurers’ political contributions have increased substantially. Using health insurers’ financial, regulatory, and campaign donation data, I find evidence for potential pay-to-play that health insurers donate to state officials for public insurance contracts and other regulatory favors. Using a corruption index based on news coverage, I find that insurers donating more to state officials receive higher Medicare and Medicaid premiums from states exhibiting higher levels of pre-existing corruption. McDonnell v. United States increased the standards for corruption prosecution in 2016, which impacts high-corruption states more than low-corruption states. My difference-in-differences tests show an increase in health insurers’ contributions to politicians and their public insurance premiums from high-corruption states relative to those from low-corruption states after 2016. Insurers’ donations are associated with more lenient regulation and better financial performance, but lower health care affordability and more suicides among socio-politically disadvantaged groups. The evidence for pay-to-play is stronger for highly leveraged insurers and for donations to federal election candidates. My results suggest the health costs of political rent-seeking, which more effective pay-to-play laws may alleviate.Price path convexity and analyst recommendations
Abstract
We find that analysts’ stock recommendations are influenced by the shape of recent price paths. Analysts are more likely to issue downgrades following convex price trajectories. This pattern of downgrades is especially pronounced for stocks exhibiting strong past return momentum. The relationship weakens for firms with low information asymmetry and during periods of macroeconomic uncertainty and elevated investor sentiment. Moreover, stocks downgraded following convex price paths exhibit lower subsequent returns, suggesting that convexity is an informative price signal and that analysts respond rationally.Private Fund Capital Calls, Investor Portfolios, and Spillovers
Abstract
Institutional investors commit trillions of dollars to private funds that give fund managers discretion to make capital calls on short notice. Using novel data on insurers' private fund investments, this paper examines the implications of unexpected capital calls for investor portfolios and financial markets. I first document that investor-level unexpected capital calls are substantial. Despite this, investors do not appear to manage capital calls by increasing cash buffers ex-ante. Instead, when faced with unexpected calls, insurers rebalance their portfolios primarily by selling corporate bonds with high risk weights and unrealized gains. I demonstrate that this seemingly counterintuitive behavior is likely driven by concerns over regulatory capital. Moreover, these asset sales induced by unexpected calls spill over into the corporate bond market, leading to temporary price declines. The spillover effects are more pronounced when unexpected calls coincide with other adverse shocks. Overall, these findings suggest that unexpected capital calls pose significant challenges to investor portfolio management and may introduce new sources of financial fragility as private fund investments continue to grow.Real Effects of Bernanke–Kuttner: The Risk Channel of Monetary Policy on Corporate Investment
Abstract
The literature extensively documents that monetary policy announcements convey information that affects risk premia and investors’ risk perception, yet little is known about how these shifts influence real activity. Exploiting aggregate news shocks (decomposed into policy rate, growth, and risk components) identified from daily asset price changes across asset classes, I provide causal evidence that positive risk news released on FOMC announcement days—news that raises perceived risk and uncertainty about future cash flows—reduces subsequent corporate investment in tangible capital, and this effect is amplified by high debt burdens. Consistent with a flight‐to‐quality mechanism that raises external financing costs for firms with ex‐ante low credit quality, positive risk news leads high-debt-burden firms to (1) reduce investment more sharply; (2) curtail net borrowing; (3) accumulate larger cash buffers; and (4) concentrate investment cuts when their debt is short term—i.e., when rollover risk is highest. At the aggregate level, the investment response to risk‐news shocks is state‐dependent and strengthens with the share of high-rollover‐risk firms; nevertheless, the unconditional effect is statistically insignificant because such firms hold only a small share of the tangible capital stock and therefore, on average, contribute less to aggregate investment.Real Effects of Corporate Debt Collateral Eligibility
Abstract
This study examines the real effects of corporate bond collateral eligibility and its transmission through production networks. Specifically, we exploit the daily announcements of the European Central Bank (ECB) eligible collateral list to investigate how collateral eligibility spreads through business relationships. Our findings reveal that when firms' bonds are included in the ECB's eligible collateral list, these firms do not increase their own investment activities, but instead choose to expand their trade credit support to both upstream and downstream partners (suppliers and customers), thereby facilitating their partners' investment and employment growth. This study provides novel insights into understanding the real effects of corporate bond collateral eligibility and its transmission mechanisms through production networks.Refinancing Inequality and Implications on Monetary Policy
Abstract
I investigate the heterogeneous mortgage refinancing propensity across income groups and its effect on the refinancing channel of monetary policy. I document that low-income households engage significantly less in refinancing than high earners, a pattern referred to as “refinancing inequality”. This indicates that the refinancing channel does not effectively serve households most likely to be “wealthy hand-to-mouth.” Despite its importance, this pattern and its effect on the policy channel have been understudied in the literature. I demonstrate refinancing inequality over a longer time horizon with more robust measures than previous studies; the bottom quintile households exhibit less than half the probability of refinancing compared to the top quintile. The estimated potential savings through refinancing are significantly large particularly for low-income households, amounting to more than 10% of income. Furthermore, I show that households refinance substantially less in response to monetary policy shocks as their income decreases. These results suggest that the aggregate effect of expansionary policy could be larger if the refinancing frictions were mitigated.Regulation and Intermediation in Over-the-counter Markets
Abstract
This paper investigates how banking regulation affects the trading behavior of bank affiliated dealers in over-the-counter (OTC) financial markets. To micro-found the regulatory costs of holding inventories of OTC-traded assets, I develop a search model of an inter-dealer OTC market in which risk-averse dealers, facing idiosyncratic uncertainty in their endowment income and OTC asset returns, choose portfolios of risk-free and risky OTC assets subject to a regulatory balance sheet constraint. Acquiring additional inventories tightens regulatory requirements: dealers close to their constraint purchase assets only at sufficiently discounted prices to remain in compliance and avoid the penalty of lower utility from a violation. I solve the model numerically and show that inventory costs are state-dependent: tighter constraints lower the marginal value of OTC assets and increase trading costs. I examine stricter regulation by increasing the risk weight on risky OTC assets from 50% to 150%: stricter regulation raises holding costs and average spreadsRelief Beliefs: Effects of Anticipated Student Loan Forgiveness
Abstract
Political support for student loan forgiveness has been growing, particularly on the left, but evidence regarding its effects remains limited. We evaluate the immediate consumption response to President Biden's 2022 loan forgiveness announcement which promised debt relief of $10,000 to $20,000 for approximately 42 million borrowers. We find that retail stores located in counties with a 1 percentage point higher share of eligible student loan borrowers saw a persistent 0.1% increase in weekly sales. The positive spending response was absent in counties with high shares of financially delinquent households, consistent with delinquent borrowers being liquidity constrained and unable to smooth consumption. Novel data on debt relief eligibility and applications suggest that student loan borrowers anticipated relief they ultimately did not receive and adjusted their spending in response.Resistance and Arbitrage: International Trade in Brown Loans
Abstract
I develop a novel measure of carbon sensitivity in lending to assess reductionsin portfolio exposure to brown assets. Using syndicated loan data, I show that countries with greater resistance to brown lending, proxied by economic development, experience faster shifts in the sectoral composition of loan portfolios. The decarbonization is driven primarily by domestic credit reallocation. I find consistent evidence of risk transfers to less regulated lenders and foreign countries, indicating arbitrage and incomplete regulations. Furthermore, lenders’ climate risk-taking and transfer behaviors vary sharply by syndicate role, loan type, and specialization. The existence of international trade in brown loans has important implications for supervisory evaluation. Using the European Central Bank’s climate guide, I show that accounting for regulatory leakage reveals effects contrary to common wisdom.
Revisit labor market concentration under Work-From-Home: An Integration of Labor Market Boundaries
Abstract
Post-COVID work-from-home (WFH) policies have reduced geographic constraints, enabling workers to access remote job opportunities across regions. This shift has expanded local labor market boundaries and altered labor market concentration levels. In this study, I adjust a traditional concentration measurement to include remote job postings and analyze the impact of WFH on labor market concentration using the difference-in-differences (DID) approach. I find that excluding remote jobs significantly biases the concentration estimate. Additionally, I re-estimate the effect of concentration on wages using the two-stage least squares (2SLS) method, finding that concentration exerts downward pressure on wages for occupations with lower education requirements. In contrast, wages for occupations with higher education requirements increase with concentration. These results suggest supply-side shifts in the labor marketScale-Dependent Returns and the Interest Rate
Abstract
I revisit the relationship between household wealth and returns on wealth in the United States over the past 70 years. While recent studies find that wealthier households earn higher returns, I show that this pattern is specific to the post-1980 period. Before 1980, the relationship was reversed: returns declined at the top of the wealth distribution, and the bottom 90 percent earned higher returns than the top 10 percent. I attribute this reversal to differences in exposure to interest rate risk. Wealthier households hold longer-duration assets, such as stocks and private businesses, whose valuations (and hence returns) are more sensitive to changes in real interest rates. Rising real rates before 1980 depressed their returns, whereas the post-1980 decline in real rates boosted them. To explain why richer households hold longer-duration portfolios, I develop a model in which households choose asset duration to hedge income risk. Because their income is more correlated with short-term interest rates, wealthier households optimally select longer-duration (countercyclical) assets to offset this exposure.Sentiment-Induced Misvaluation and Predictability in Financial Markets
Abstract
The literature contends that the ability of certain empirical measures and surveys to forecast returns indicates that they capture investor sentiment. I argue that valid measures of investor sentiment must satisfy additional conditions beyond return predictability to imply sentiment-induced misvaluation. Specifically, both positive and negative sentiments should explain returns and volatility contemporaneously, and forecast returns. I show that several well-known sentiment indexes fail to fully meet these necessary conditions and introduce three new empirical indexes that perform better. These proposed measures demonstrate superior predictive power for returns both in-sample and out-of-sample, particularly over longer horizons; and survive the inclusion of non-sentiment variables known to predict returns. Further evidence shows that their robust forecasting ability extends to broader financial outcomes, including changes in flows to active equity mutual funds, the VIX, and credit spreads.Specialization in Non-Bank Lending
Abstract
Using SEC filings data on Business Development Company (BDC) investments, I document thatthese non-bank lenders specialize by concentrating their lending in selected industries, and
they are more concentrated than banks; however, they diversify over time. BDCs offer more
generous and flexible credit terms within their specialization sectors, leveraging informational
advantages from their industry focus. I also show that BDC industry specialization improves
loan performance within their focus industries. To identify drivers of specialization, I employ
changes in aggregate bank C&I lending standards from the Senior Loan Officer Opinion Survey
(SLOOS) As an instrument, it shows that tighter overall bank credit conditions make BDCs more
concentrated. I provide new evidence on the growth of these non-banks as they have become
substitutes for bank financing. These findings extend bank specialization theories to non-bank
intermediaries and highlight the recent growth in direct lending.
Specific Signals in a Noisy World: Idiosyncratic Forward-Looking Disclosures and Predictable Returns
Abstract
This paper examines how idiosyncratic (firm-specific) versus systematic (non-specific) forward-looking statements in corporate disclosures affect markets differently. Using natural language processing, I classify forward-looking statements from 10-K filings and develop neural network-based growth probability measures for each type. Analysis of U.S. public firms (1998-2022) reveals idiosyncratic forward-looking statements significantly outperform systematic ones in predicting future growth. A one standard deviation increase in idiosyncratic growth measure yields a 1.25% excess stock return at 180 days, while systematic information loses predictive power when orthogonalized against idiosyncratic content. Additionally, firms with higher idiosyncratic forward-looking growth experience reduced stock price volatility. Analysts respond positively to idiosyncratic—but not systematic—forward-looking information specifically during downward forecast revisions, consistent with managers selectively disclosing negative information. These findings demonstrate the superior information content of firm-specific forward-looking statements and document investor underreaction to idiosyncratic information, consistent with limited attention theories.Strategic Capital Deployment in Private Equity
Abstract
Private equity fund general partners (GPs) adjust their investment strategies in response to early returns rather than following a prespecified plan. Funds experiencing higher early returns in the fund life cycle subsequently shift away from riskier investments in later years and experience lower returns. Funds with early success also reduce exposure to higher risk sectors and increase their sector and geographical concentration as well as increase the fraction of the fund invested in later deals and hold onto these deals longer. In contrast, funds with low early returns do the opposite. Despite a lack of within fund persistence, funds with strong early returns still outperform in final fund-level returns and raise their next fund faster. These findings are consistent with GPs using early success in a fund to raise a next fund sooner and then turning their attention to this next fund.Stress-Tested: Municipal Bond ETFs During Market Turmoil
Abstract
Financial innovations like exchange-traded funds (ETFs) are often credited with enhancing market efficiency and liquidity, but these benefits may not extend to all asset classes. This paper examines why municipal bond ETFs (Muni ETFs) experienced large and persistent price deviations from net asset value (NAV) during the COVID-19 crisis. The creation and redemption mechanism failed to close price-NAV gaps, suggesting a breakdown in arbitrage activity. Using bond-level trade data, I find no evidence of that arbitrage participants remained active while small trades dominated sell activity for ETF-held bonds. In contrast, bonds held by municipal bond mutual funds experienced increased trading volume, consistent with flow-driven selling. These findings suggest that, while the Muni ETF mechanism did not trigger fire sales in the underlying market, it left most selling pressure in the ETF secondary market. The results highlight structural fragilities in ETF arbitrage mechanisms in illiquid markets, especially under stress.Subjective Beliefs and the Q-Theory of Investments
Abstract
I study the role of managers’ and investors’ subjective beliefs for corporate investments and valuation. I augment a q-model allowing jointly for managers and investorsto have non-rational expectations. The model highlights the importance of small dis-
tortions in expectations for investment decisions and valuation. The disagreement
between these two economic agents impacts Investment-Q regressions and Investment-
CashFlow sensitivity, even without other frictions. I construct a novel dataset on
managers’ and investors’ sales forecasts. I test the model predictions and estimate the
model to quantify the role of subjective beliefs. I find that managers are optimistic and
overinvest, while investors are pessimistic and underprice firms. Therefore, the documented gap between investments and valuation widens further. This requires higher
implied markups to reconcile this evidence. These results highlight the importance of
departing from rational expectations to evaluate the role of economic rents for corporate policies.
Subscription Revenue Models as Financing Models
Abstract
I explore the financial implications of recurring revenue models through a particular channel documented in recent literature: subscriber inattention and inertia. I propose that quasi-liability-like inflows of subscriptions, amplified by these documented sub- scriber behavioral biases, have first-order effects on capital structure and firm value by allowing firms to hold less precautionary cash than they otherwise would. Event stud- ies around two FTC actions–the 2023-2024 “Click-to-Cancel” rule, and the FTC’s 2024 lawsuit against Adobe–reveal large negative cumulative abnormal returns to subscrip- tion firms, particularly those most associated with cancellation frictions, representing about 2% of firm value. Using a newly constructed panel dataset identifying “sub- scription economy” firms based on 10-K keyword frequency from 2002–2023, I show that, controlling for firm characteristics, time fixed effects and industry fixed effects, recurring revenue-exposed firms are associated with 2.2% lower cash-to-assets and 5.6% higher investment. These results are not explained by cash flow volatility, and pro- vide suggestive evidence that the choice of adopting a subscription revenue model may function not only as pricing strategy, but as a shadow financing mechanism.Sustainability-linked Loans and Financial Benefits
Abstract
I develop a model in which firms can choose to issue either sustainability-linked loans (SLLs) or non-SLLs to finance investment. Firms issuing SLLs face a trade-off between the benefit of a lower interest rate and the additional costs required to meet SLL-related sustainability targets. Within this framework, the model predicts that SLL issuers exhibit lower default risk and experience positive stock returns following loan issuance. Using a sample of bank loans issued between 2016 and 2022, I find empirical evidence consistent with these financial benefits of SLL issuance. To address potential endogeneity concerns, I employ the EU Taxonomy as an instrument for SLL issuance and obtain consistent results.Symbolic Regressions: Opening the Black Box of Equity Premium Predictions
Abstract
This paper investigates equity premium predictability using Deep Symbolic Regression (DSR), a method that identifies sparse and interpretable functional forms in the data. Unlike traditional opaque machine learning models, DSR allows explicit capture of nonlinear economic relationships. The paper introduces a novel regularization parameter within the DSR methodology, ensuring robust model selection. Extensive simulations validate the effectiveness of this approach. Empirical analysis using monthly U.S. stock market data (1927–2021) demonstrates that DSR consistently outperforms benchmarks such as linear regression and random forests in forecasting accuracy. The findings highlight significant nonlinear dynamics in market returns, particularly during periods of economic stress, thereby providing a transparent and economically insightful framework for equity premium prediction.Take the Long View: Horizon Bias and Equity Term Premia on Earnings Days
Abstract
Horizon bias---the tendency for investors to be more optimistic over long horizons than short horizons---helps explain why a substantial share of the equity term premium is realized around earnings announcements. Retail investors, who are more prone to horizon bias, amplify mispricing by bidding up long-duration stocks ahead of earnings release. High arbitrage risk during these events deters rational speculators from correcting the mispricing until news arrival. Horizon bias also accounts for the announcement returns of value, profitability, and low-risk factors, as these strategies invest in firms with near-future cash flows. Furthermore, long-term optimism is more pronounced following waves of investor sentiment, leading to larger horizon bias. I show that active institutional investors time market sentiment, and their time-varying duration tilts are sufficient to explain the 8.7% spread of term premia across sentiment states.Tax Planning, Illiquidity, and Credit Risks: Evidence from DeFi Lending
Abstract
This study examines the link between tax-planning-induced illiquidity and credit risks in lending markets. Exploiting an exogenous tax shock imposed on cryptocurrency gains and millions of transactions in Decentralized Finance (DeFi) lending, we document that tax-motivated borrowing strategies to defer capital gains taxes significantly reduce market liquidity. This effect is pronounced among individuals borrowing in stablecoins (a way to monetize returns), those with higher loan-to-value ratios (more risk-averse towards new regulations and typically with larger taxable gains), those with high returns in the underlying asset (representing larger taxable gains), and those holding locked-in assets for over a year (i.e., converting high short-term to lower long-term capital gains tax rates). Using instrumental variable analysis, we provide a plausibly causal relation between tax-planning-induced illiquidity and increased credit risks. A standard deviation increase in tax-induced illiquidity leads to a more than twofold increase in the value of defaulted loans. Our results remain robust across a battery of checks, including analyses of subsamples of highly tax-sensitive borrowers, and align with well-documented tax awareness periods. Overall, our insights are relevant to market participants, assist in estimating revenue losses for tax authorities, and inform emerging policies on the tax treatment of digital assets.Tax Revenue Sensitivity and Bond Valuation
Abstract
How are municipal bond yield spreads exposed to property market value changes through the property tax channel? I examine the sensitivity of property tax revenue growth with respect to property market value growth for each local government. The sensitivity is estimated to be 0.43% on average, with variations observed across cities, counties, and school districts. The variation in sensitivity is driven by a local government's choice in property assessment values, tax rates, and tax exemptions. I further show that a large sensitivity increases municipal bond yield spreads by 23 basis points, referred to as the sensitivity premium. The sensitivity premium is pronounced at the bad state of the world, where property market values decline. Furthermore, municipal bond yield spreads are higher when cities, counties, and school districts face binding frictions such as a statewide cap on the growth of assessment values and infrequent reassessments.Technological Complexity, Valuation and Drives
Abstract
This paper measures industry-level technological complexity using the Network Diversity Score, which captures how varied the connections between knowledge components are. This approach outperforms other complexity metrics in key tests. The findings show that complexity rose steadily at both sector and industry levels until the Covid-19 pandemic, after which labor-intensive industries experienced a sharp decline. An excess complexity score is introduced to compare scores across industries and use it in our regression models. The results reveal that, in less competitive markets, complexity is driven by advancement in innovation and hurt by influence level, whereas, in more competitive markets, it’s purely driven by revenue.The Appraisal Mechanism: Spillover Effects of All-Cash Buyers on Local Housing Markets
Abstract
This paper examines the impact of growing all-cash home purchases on local home appraisals and transaction prices. Using micro data from 2018–2022 and a ring-based spatial design, I identify a new appraisal channel through which discounted cash sales can depress prices of nearby mortgage-financed homes. This spillover is stronger in low-growth, more affluent neighborhoods, whereas it is muted or offset in high-growth, less advantaged areas where it signals heightened housing demand. To validate the mechanism, I propose a novel algorithm that manually constructs comparable sales and show that appraisers predominantly rely on very local sales as comps, supporting the identification assumptions. A stylized model rationalizes these findings and indicates that these spillovers can reduce welfare by excluding mortgage-dependent buyers and eroding housing wealth in low-demand markets, but may improve affordability by lowering entry prices in hot markets.The Environmental Impact of Corporate Mergers
Abstract
This paper investigates the environmental consequences of corporate consolidation, focusing on how mergers influence firms' greenhouse gas emissions, green innovation, and vegetation health. I combine theoretical and empirical approaches to examine whether the potential reduction in competition resulting from mergers affects environmental performance. Theoretically, I model the trade-off between market power and emissions, showing that mergers can reduce emissions either by lowering output through increased pricing power or by incentivizing green innovation. Empirically, using firm-level data from 2006 to 2022, I implement a panel event study and a difference-in-differences (DiD) design that compares completed and cancelled mergers. I find that mergers lead to reductions in absolute emissions, a decline in green patenting, and no significant change in vegetation health—effects. These results are particularly pronounced in notifiable transactions subject to antitrust scrutiny. These findings suggest that corporate consolidation, often considered solely in the context of competition policy, may also have important implications for environmental policy.The Gift that Keeps on Giving to Investors: Does the Corporate Bond Market Learn from Politicians' Stock Trades?
Abstract
Using a comprehensive dataset of politicians’ stock transactions during 2012-2024, I provide evidence that their stock trading activities are incorporated by bond market investors, and the influence is observed in both the primary and secondary markets. Bonds issued by firms whose stocks were net purchased (sold) by members of Congress show lower (higher) offering yield spreads. The association is stronger when Democrats and Republicans trade in the same direction, and when more politicians are trading, but not when the dollar amount transacted is larger. In the secondary market, disclosed stock trades by politicians are followed by cumulative abnormal returns in the firm’s outstanding bonds. Cross-sectional tests reveal that these reactions are more pronounced for larger trades and smaller reporting delays, but they are not significantly correlated with changes in politicians’ holdings, party affiliation, or bond ratings.The Green Value of BigTech Credit
Abstract
This study identifies an incentive-compatible mechanism to foster individual environmental engagement. Leveraging a dataset of 100,000 randomly selected users from Ant Forest—a prominent personal carbon accounting platform embedded within Alipay, China's leading BigTech super-app—we provide causal evidence that individuals strategically engage in eco-friendly behaviors to expand their credit limits, especially when nearing borrowing constraints. These behaviors not only illustrate the green nudging potential of BigTech platforms but also generate value for the platform itself. By intentionally designing green actions to foster sustained engagement, the platform enables a mechanism that distinguishes intrinsically green and financially responsible individuals from others. Our structural model estimates that the implementation of a credit limit scheme tied to costly green behaviors yields an annual green value of $427.52 million. Moreover, this incentive-based approach outperforms traditional green mandates and subsidies in improving both consumer and overall societal welfare. Our findings highlight the role of an incentive-aligned approach, such as integrating personal carbon accounts into credit reporting frameworks, in addressing environmental challenges.The Human Edge Beyond Algorithms: Forecasting in Global Macro Shocks
Abstract
This paper examines the enduring value of human judgment in an era of increasinglypowerful AI. Focusing on over 200 macroeconomic shock episodes across 47 countries, I
investigate when and where analysts retain an advantage over algorithmic models. I use
machine learning trained on public data to construct benchmark forecasts for earnings
expectations and decompose the gap between human and machine forecasts into soft in-
formation, bias and noise. The results show that soft information, such as contextual and
non-public insights that are not captured in public data, significantly improves human
forecast accuracy, especially at the onset of macroeconomic shocks. This advantage is particularly evident in emerging markets, where limited disclosure constrains the learning
capacity of algorithms. As the shock progresses, however, the accuracy of human forecasts declines due to increasing bias and noise. These findings underscore the conditional
value of human input and the informational limits of automation under uncertainty.
The analysis also reveals substantial cross-country differences related to institutional
transparency, contributing to our understanding of belief formation, systemic resilience,
and the interaction between human and algorithmic decision-making in global financial
markets.
The Inequality Multiplier: Market Inelasticity and the Persistence of Wealth Inequality
Abstract
We study how recent changes in equity market macrostructure shape U.S. equity market capitalization, aggregate debt levels, and wealth inequality. Rising income inequality, combined with inelastic markets, drives persistent wealth disparities through asset price revaluation—a mechanism we term the “inequality multiplier.” Using a general equilibrium model, we identify two channels: (i) the equity investment channel, where wealthy households’ higher propensity to save amplifies equity price booms; and (ii) the borrowing channel, where increased indebtedness raises equity prices via rebalancing demand from financial intermediaries. Calibrating the model to U.S. data, we show that this multiplier makes wealth inequality self-perpetuating and drives a growing wedge between income and wealth inequality. The model replicates observed trends in equity prices, debt levels, and wealth concentration, revealing how asset market frictions drive inequality beyond existing explanations. The equity investment channel shapes long-run trends, while the borrowing channel explains short-run cycles in wealth inequality. Our findings link recent shifts in financial market structure to macroeconomic outcomes.The Intersection of Expected Returns
Abstract
A relatively small number of stocks plays a disproportionately large role in explaining the performance of 164 cross-sectional asset pricing anomalies. For instance, excluding the top 10% of stocks that are shared across the most anomaly portfolios for a given month reduces the average anomaly's return and alpha by approximately 40%. These stocks can be identified ex ante and used to form long-short portfolios that generate abnormal returns more than three times larger than that of the average anomaly portfolio. Consistent with prior research, I find evidence that biased investor expectations help explain the returns to these stocks, suggesting that a significant portion of the returns to the 164 anomalies can be attributed to mispricing. My results have implications for traditional asset pricing, behavioral finance, and for investors and practitioners in the factor investing space.The Intertwined Price Discovery Processes in Equity and Option Markets
Abstract
I analyze the linkage between the price discovery processes of the underlying equity and the option markets after option transactions. The price impacts in the two markets are heterogeneous and are connected through the implied volatility. This suggests that option trades contain two dimensions of information: price and volatility. Option trades associated with single-leg execution contain price information but not volatility information, while trades through the limit order book contain both price and volatility information. Furthermore, change in implied volatility is negatively related to underlying price impact, suggesting that part of the option trade information is hidden from the underlying market prices.The Nature of Deposits
Abstract
I unveil a novel role for demand deposits. In a dynamic model where banks are better informed about the quality of their loans and investors receive noisy signals about loan performance with a delay, demand deposits can be structured to correlate investors' withdrawal decisions with asset quality while maintaining the banks solvent. This explains why small and low-credit rating banks rely heavily on certificates of deposit as their main source of funding, while larger and safer banks on traditional savings deposits. The model also predicts that even though banks' liabilities are often homogeneous, the duration of their liability structure can be very heterogeneous and, in particular, that it is U-shaped in their credit rating. Finally, I also show that when the economy is populated by a large fraction of sleepy depositors, staggered intermediation or an FDIC insurance policy arise as natural candidates to solve the inefficiencies caused by their lack of information.The Price of Emissions: Carbon Risk in the European Equity Market
Abstract
We investigate whether carbon risk commands a return premium in the European equity market. Using data from the EU Emissions Trading System (EU ETS) between 2013 and 2024, we construct two complementary measures of carbon risk. The first is a forward-looking, market-based proxy derived from the one-year forward convenience yield in the carbon futures market, which reflects investor expectations about future allowance prices. The second is an accounting-based proxy calculated as the ratio of estimated carbon expenses to key income statement items for firms regulated under the EU ETS. Both measures reveal meaningful operational exposure to carbon expenses of companies. By 2023, firms in the top quartile of carbon expense intensity faced carbon expenses greater or equal to 12.4% of operating income. Our results highlight a divergence: the market‑based proxy is associated with a significant risk premium in the broad STOXX Europe 600, but within the subsample of firms directly regulated by the EU ETS, neither the market‑based nor the accounting‑based proxy is associated with a return premium.The Pricing of Geopolitical Tensions over a Century
Abstract
We study the asset pricing implications of geopolitical tensions using nearly 100 years of data. Leveraging widely adopted news-based geopolitical risk indices, we find that geopolitical threats (GPT) and acts (GPA) have markedly different effects. GPT aligns closely with investors’ risk perceptions from ratings and surveys and predicts long-run consumption disasters. It is priced across individual US stocks, equity anomalies, and international equity and bond indices. GPT also captures time variation in country level equity premia and firm investment. In contrast, GPA exhibits weaker and less stable links to risk perceptions, risk premia, and investment. We demonstrate that these findings are consistent with an Epstein-Zin investor facing time-varying disaster probabilities. Importantly, our results are incremental to news-based risk indices capturing war-related discourse, market volatility, economic and trade policy, and general macro-financial uncertainty. Overall, our findings underscore the importance of forward-looking measures like GPT for understanding how news-based uncertainty affects asset prices.The Quote Not Taken: The Effect of Market Structure on Liquidity Provision in Equity Markets
Abstract
Global equity markets have become increasingly fragmented, with off-exchange trading now accounting for the majority of volume in the U.S. This paper presents new evidence that off-exchange order flow predicts next-day on-exchange prices, creating an unusual return reversal that benefits off-exchange market makers. A long-short trading strategy based on publicly available flow data generates sizable abnormal returns. From 2014 to 2022, this short-term reversal cost off-exchange investors an estimated $193 million, nearly half of which materialized within the first five minutes of market open. While the presence of informed auctioneers mitigates the sensitivity of flow reversal returns to observable risk factors, it does not reduce the overall magnitude of these reversals. Unlike previously studied liquidity provision strategies, this reversal occurs intraday rather than overnight and remains robust to value-weighting.The Rise in Insurance Costs for Commercial Properties: Causes, Effects on Rents, and the Role of Owners
Abstract
Using a large and novel set of property-level operating statements for commercial properties, we analyze the causes behind the rise in insurance costs, its effects on rents and profits, and how property owners are managing it. First, we document a significant and persistent year-over-year increase in insurance costs across nearly all regions of the US over the last decade. 95% of individuals reside in counties where the increase in commercial insurance costs is at least double that of homeowner insurance costs, which is due in part to regulatory frictions in the homeowner insurance market. Second, we provide evidence of three primary drivers of rising insurance costs: heightened pricing of local risk, a substantial increase in reinsurance costs, and cross-subsidization across states. Third, we find that, on average, 67% of the rise in insurance costs is passed through to rents; however, the passthrough to rent is concave and decreasing over time. Finally, we examine the role of property owners and find that: 1) larger owners are able to maintain lower insurance costs, 2) the impact of owner size on insurance costs are more pronounced in high-risk areas, and 3) owners with a property portfolio that has lower average risk have lower insurance costs, even holding the risk of a given property fixed. The advantage of larger owners in managing insurance costs in risky areas has grown in recent years, resulting in properties in high-risk regions being increasingly owned by larger owners. Our findings suggest that commercial property insurance pricing is influenced by both localized risk factors and broader systematic risk exposure, with risks extending beyond individual properties through insurers and property owners' portfolios.The Rise of Single-Stock ETFs and More Volatile Stock Prices
Abstract
I investigate the novel instrument of single-stock ETFs. I propose that single-stock ETFs gain popularity because they satisfy retail investors’ demand for taking short-term leveraged long and short positions on popular stocks at a low cost. I show that flows, turnover ratios, and retail buy-sell imbalance of both long and short single-stock ETFs are positively related to the retail attention on the underlying stock. This suggests that single-stock ETFs are mainly used by retail investors to circumvent leverage and short-selling constraints. Furthermore, I find that following the launch of the first short single-stock ETF on a stock, the underlying stock experiences a significant and long-lasting increase in the idiosyncratic volatility.The Salience of Disaster: How Experience Outweighs Information in Pricing Earthquake Risk
Abstract
This paper investigates how salience influences decision-making in earthquakeprone real estate markets in Türkiye, focusing on two critical events: the 2018 revision to the national earthquake hazard map and the catastrophic 2023 earthquake that resulted in over 50,000 fatalities. Our findings indicate that while the updates to the hazard map have little effects on property values, the actual occurrence of a disaster significantly reduces home prices and increases insurance uptake in high-risk but physically unaffected areas. A one-standard-deviation increase in baseline seismic risk is associated with a 4% decline in home prices after the earthquake. Additionally, the data show that areas with strong social connections to disaster-stricken regions experience more pronounced declines in home sale prices, highlighting the role of personal relationships in amplifying risk perception. Overall, these results suggest that the salience of a vivid, catastrophic event is far more impactful in shaping economic behaviors than abstract, probabilistic information in high-risk scenarios.TheEffects of Relaxing Downpayment Constraint on the Mortgage and Housing Market
Abstract
This paper studies the effects of relaxing downpayment constraint on mortgage originations and house prices, using a geographic eligibility cutoff in a low-downpayment program introduced in United States by GSEs. A regression discontinuity design shows that relaxing downpayment constraint increased the share of high loan-to-value (LTV) mortgages by 8.2%, high LTV originations per capita by 10.5%, and house price growth by 12.7%. The effect on high LTV originations followed a bell-shaped relationship with the price-to-rent ratio, peaking in moderately priced housing markets and weakening in both low and high-cost areas. Price effects were weakest in areas with very high owned–rental market integration, due to substitution away from investor buyers whose rental demand is affected. These findings highlight how borrower characteristics, housing affordability, and housing-rental market integration shape the effects of mortgage policy interventions.Time-Varying Leverage Constraint, Safe Asset Demand and Dollar Exchange Rate
Abstract
I propose a two-country intermediary-based model with financial frictions to study the exchange rate movement and the dynamics of the U.S. external balance sheet. In the model, financial intermedairies are subject to time-varying leverage constraints which limit their ability to raise funds. The key asymmetry in the model is that U.S. Treasury bonds are considered to be safer and offering more liquidity than government bonds issued by foreign countries. Under the symmetric global financial shock, the model is capable to endogenously generate safe asset demand and provide micro-foundation for convenience yield that investors derive from holding US safe assets. In global recessions, the demand for U.S. safe assets increases, convenience yield becomes higher, leading to an appreciation of the dollar. Under the safe asset view, the seigniorage revenues from issuing bond that carry higher convenience yield raise the U.S. consumption share in recessions, despite the U.S. suffering portfolio losses from external positions. The model can also jointly explain the large and persistent CIP deviation due to the tightening of bank regulations after the GFC.Top Government Meetings in China
Abstract
In a “government centric” equilibrium like China, the central government performsfrequent and intensive interventions while investors prioritize information about government policies over macro-economic fundamentals. Consistently, top government meetings on economic policies are highly anticipated by the Chinese equity market. Just as the heightened uncertainty prior to FOMC meetings gives rise to the significant pre-FOMC drift in U.S. equity, we find significant pre-announcement returns prior to top government meetings in China. By contrast, we do not find significant pre-announcement returns before macro announcements in China, confirming the presence of a “government centric” economy with policy-driven markets.
Uncovering the Black Box of Startup Growth: How Startup Mergers drive Successful Exits
Abstract
How did Google, Facebook and Paypal grew from vague ideas to billion-dollar companies in just a few years? This paper opens up the black box of startup growthand reveals a previously unexplored mechanism for success: startup-to-startup mergers. Creating a novel dataset, I find that startups merge with each other as a growth strategy in preparation for a successful exit. This approach seems effective: merging startups are 3.2 times more likely to achieve a successful exit than non-merging peers.
VC Competition and Startup Financing Cost
Abstract
This paper examines how venture capital market concentration affects startup financing costs and investment patterns. We document two contrasting trends: increasing concentration among large VCs capturing greater market share, and rapid proliferation of micro VCs under $50 million. Using comprehensive data on 9,107 US startups and 21,642 VC investors from 2010-2024, we exploit cross-sectional and time-series variation in VC market concentration across sectors to identify causal effects on startup financing terms. Our empirical strategy employs multiple fixed effects specifications to address identification concerns. Results reveal that higher VC market power significantly increases startup financing costs, with high-power VCs extracting 3-5.5 percentage points more board control than less powerful counterparts, particularly in larger deals and syndicated rounds. Additionally, concentrated VCs systematically delay investment timing by 0.4-0.7 funding rounds, leveraging market power to cherry-pick promising startups at later, less risky stages. Effects are most pronounced in competitive, high-stakes scenarios. These findings suggest VC market concentration may harm startup innovation through higher financing costs and delayed capital provision, while micro VC growth provides important competitive counterbalance. Our results have policy implications for antitrust attention to VC market structure.When does ownership matter? Evidence from China
Abstract
We study the real effect of ownership activism by exploring a corporate charter amendment event in China, which significantly augmented the control rights of incumbent owners without altering ownership structures. Following the installment of enhanced control provisions, firms experienced increased profitability, productivity, and innovation. The performance improvements are mostpronounced in firms that adopted more substantial amendments and in those targeting critical provisions, such as decision-making processes and personnel management. Strengthening control rights also resulted in increased owner engagement in board functions. The improved corporate governance and disclosure quality of the affected firms support our proposed mechanisms. Our
findings highlight the crucial role of control rights, as opposed to mere ownership, in shaping corporate governance practices and firm performance.
When Elephants Walk: Large Investor, Information Advantage and the Fragility of the Asset Market
Abstract
How do investors' size and information precision drive asset price fluctuations and contribute to market fragility? When investors are large, they will trade cautiously. But when they trade, their price movement will make the price more informative. We adopt the rational expectation equilibrium model where the investors are large and behave as a price maker, showing that large investors are a double-edged sword to the market: they enhance price informativeness by injecting more fundamental information into the market, yet increase the asset market fragility to shocks due to their price impact concern, making the market illiquid and inelastic. We apply this model to examine the implications of the growing market share of passive funds over active funds. Our simulation reveals that a higher passive fund share improves liquidity, reducing liquidity shock fragility, but worsens the price informativeness. Depending on the size of passive investors, our analysis suggests two distinct policy strategies to improve market efficiency and reduce fragility. When the passive fund share is low, reducing informed investors' market concentration will improve price efficiency and reduce price fragility more effectively. When the passive fund share is high, enhancing information advantage for large informed investors will be more effective in improving price efficiency and reducing price fragility.When Models Fail: Evidence from Automated Underwriting in Auto Loan Markets
Abstract
While prior studies find the outperformance of automated over manual underwriting, I document significant heterogeneity in the adoption of automated underwriting practices across auto lenders and within the same lender. To explain this heterogeneity, I examine the performance of automated underwriting systems under conditions of heightened data uncertainty caused by the COVID-19 pandemic. Using a difference-in-differences design, I estimate the impact of this unprecedented shock on the performance of automated underwriting in the auto loan market. My findings show that the performance of automated underwriting, as measured by ex-post default rates, deteriorated substantially relative to human underwriters during this period. This effect is particularly pronounced among low-income borrowers. These results suggest the limitations of automated underwriting systems when faced with unprecedented shocks outside the scope of their historical training datasets, underscoring the continued relevance of human underwriters in addressing such challenges in the auto lending industry.When Roll-up Breaks: Serial Private Equity Acquisitions in the Hospital Industry
Abstract
This paper challenges the idea that private equity (PE) roll-ups consistently create value across acquisitions. Linking hospital-level operations and PE deal data, I classify targets as standalone, platform, or add-on. PE firms acquire financially strong hospitals as platforms, securing cheaper debt shortly after acquisition and achieving gradual improvements in profitability. In contrast, add-ons are highly leveraged at entry, but enter with healthy margins and focus on headcount reductions without further margin gains. These patterns suggest a two-stage strategy with uneven value realization and raise questions about the scalability of PE value creation in healthcare.When the Rules Change: The Uneven Effects of SEC's Accredited Investor Reform
Abstract
This paper examines how changes in investor protection regulation impact local angel financing and entrepreneurial activity. I exploit the SEC’s 2020 expansion of the accredited investor definition as a quasi-natural experiment, and find an increase in angel investor participation, investment volume, employment and early-stage firm formation. These effects are concentrated in sectors and regions with greater pre-reform exposure, as measured by ex ante investor sophistication. The regulatory shift also results in a reallocation effect in entrepreneurial financing behavior by reducing reliance on small business loans and mortgage-based capital. At the firm level, startups backed by newly accredited investors have shorter fundraising durations and improved firm survival. Although new entrants select higher relative valuation investments, they exhibit greater risk aversion. These findings underscore the trade-off between early-stage financing and investor protection, and highlight the regulatory challenges of equitable access to private capital markets.Why Do Firms Adopt ESG-linked Pay? The Role of Bank Financing
Abstract
This paper studies how debtholders influence firms' adoption of ESG-related metrics in executive compensation (i.e., ESG-linked pay). Leveraging exogenous variations in ESG regulations imposed on non-U.S. banks, we find that U.S. firms that have pre-existing lending relationships with these banks are more likely to adopt ESG-linked pay. This effect is stronger when borrowing firms have higher switching costs to new lenders, lower shareholder-manager coordination costs, and poorer ESG performances. We further show that borrowing firms adopt ESG-linked pay in response to firm value declines caused by heightened shareholder–debtholder conflicts following ESG regulations on banks, and such compensation scheme helps restore firm value and improve ESG performance. Taken together, our research demonstrates that ESG regulations in the banking sector can translate into compensation contracting outcomes because debtholders pass on regulatory costs to borrowing firms and reshape shareholders’ perceptions of firm value.Why Venture Later? Incentives, Learning, and Industry Allocation in VC Funds
Abstract
Why do private equity firms often delay investment in high-uncertainty sectors like deep tech, despite their potential for long-term gains? This paper examines how agency frictions shape cross-industry portfolio allocation decisions. I develop a dynamic model in which limited partners cannot observe fund managers’ effort to explore, and fund managers cannot fully monitor entrepreneurs’ experimental designs. These two layers of moral hazard reduce incentives for early exploration and distort capital commitments across funds. The model predicts that firms with high opportunity costs are more likely to specialize, while even low-cost firms underinvest in exploration when frictions are severe. Empirical evidence from matched fund-level data confirms these predictions, showing that higher moral hazard is associated with reduced exploratory investment in earlier funds and lower follow-on capital in subsequent funds. Ongoing work uses structural estimation to quantify the welfare implications of these frictions.JEL Classifications
- G0 - General