We show that household linkages, formed primarily of spousal employment ties, are important in explaining asset managers’ skills and their portfolio choices. Mutual fund managers with spouses that work at the executive and C-suite levels obtain monthly gross returns of up to 0.32% above asset managers with non-executive spouses. This effect is driven largely by managers’ quarterly stock trades in the industries where their spouses are employed. The spouse-industry stocks bought (sold) by executive-spouse linked managers outperform (underperform) those bought (sold) by the nonexecutive-spouse linked managers by 4.30% (4.33%) per quarter. These patterns suggest that spousal relationships facilitate fund managers’ comprehension of industry-level information. We further find that fund managers’ spouse-industry trades predict subsequent earnings surprises. Overall, our results indicate that the boundaries of information advantages in asset management extend beyond the managers themselves and highlight the importance of household linkages to information production in the asset management industry.

Growing climate risk is straining P\&C insurers' financial stability. State-level guaranty funds are designed to protect consumers facing insolvent insurers, yet their structure and effectiveness remain understudied. Unlike the pre-funded and centralized FDIC, these guaranty funds are state-run and rely on ex-post, risk-insensitive assessments of surviving insurers. While guaranty funds are intended to increase trust in the insurance system, in practice payouts can take years to reach homeowners, leaving them with unrepaired homes as they wait for claims to be paid. Crucially, we find that their structure significantly distorts insurance supply: solvent insurers exit states following insolvencies to avoid these ``tax" assessments and to escape inheriting concentrated exposures in affected counties. Furthermore, the post-funding mechanism degrades supply quality: fragile insurers, ignoring their own insolvency costs, underprice better-capitalized rivals for fully covered policies. Other explanations of exit, such as general industry-wide declines or climate risk management do not explain these exit and pricing patterns, though they can amplify the effects of the guaranty fund. These findings highlight how the current guaranty fund design induces moral hazard and amplifies pro-cyclical supply disruptions.

Standard credit derivative models typically assume exogenous corporate policies and rational expectations about systematic risk. We show that subjective beliefs about disaster risk, specifically their level, not just uncertainty, drive CDX (credit default swap index) rates through endogenous corporate responses. In a consumption-based model with Epstein-Zin preferences, firms optimally choose leverage and default boundaries while learning about disaster probabilities. When disaster risk beliefs rise, the model generates a feedback effect: higher perceived risk leads to higher optimal default boundaries, decreasing distance-to-default and raising leverage, which sustains elevated credit spreads even after uncertainty resolves. Estimating on CDX data from 2003-2022, we match both the level and time-variation of investment-grade spreads and leverage ratios. The model replicates the 1-to-10 year term structure of CDX spreads in out-of-sample tests.

Investments in electricity generation capacity incur substantial upfront costs and produce uncertain returns that can take years to materialize, making them highly dependent on external finance. We link loan-level regulatory data to power plant-level generation capacity and use two natural experiments to examine how financial frictions affect electricity producers' responses to taxes and subsidies. We find that increasing the cost of coal generation created spillovers through firms' internal capital markets and ultimately reduced investments in new solar capacity. We also find that subsidies designed to increase cash flows to early-stage projects disproportionately boosted solar capacity investments for financially constrained firms. These results highlight the importance of firms' financial frictions in understanding how policy can influence investments in electricity generation.

Why do some fund managers outperform others? I construct a novel dataset to study this question in private equity (PE), tracking the careers and deals of PE professionals who have deployed 97% of capital for U.S. PE firms over the past three decades. I find that about 40% of a PE partner's investment performance can be explained by which PE firm they worked at as a junior associate. Moreover, 60% of this early-career effect is accounted for by the identity of their primary mentor. Comparing deals led by former associates from the same firm or the same mentor provides evidence that these effects are driven by distinct approaches to capital structure and portfolio company management learned early in their careers. Instrumental variable estimates at the PE firm level document additional short-run effects of associate experiences: the loss of junior associates decreases deal-making activity. Overall, my results show that fund managers' long-term investing success is significantly related to where their investment careers begin.

We explore how monetary policy affects corporate investment decisions, and why there are long and variable lags in transmission. To do so, we hand-collect a firm-level dataset of U.S. investment plans and link it to managers’ cash-flow expectations. We first document new facts about investment plans. Firms' investment plans are persistent, and the initial plan explains most of the variation in realized investment. However, firms immediately update investment plans and earnings expectations in response to monetary policy shocks. We find a stronger effect on long-horizon plans than short-horizon plans. Plans for new and expansionary projects respond more to monetary policy than plans related to ongoing projects. These results help to explain the "long and variable" lags in monetary policy transmission. Regarding transmission mechanisms, we document a financing-cost channel, where policy-driven increases in borrowing costs reduce planned investment.

We use administrative data covering the universe of mortgage originations to individ- ual real estate investors in the United Kingdom to study financing outcomes following a large, unanticipated increase in interest rates. Post-shock, originations become more concentrated among specialist lenders, who exhibit lower rate pass-through for larger borrowers. To offset these smaller rate increases, they charge higher loan fees, thereby attenuating the impact of higher rates on interest-coverage ratios. High-frequency data on the menu of loans on offer show similar responses, indicating that specialist lenders adjust product design to target specific borrower types and, in doing so, reinforce market segmentation.

This paper documents substantial asset `reclassification' in the mutual fund industry, exceeding $450 billion in 2021. These reclassification events do not involve investor flows; instead, mutual fund assets are simply converted into twin investment vehicles, such as separate accounts or collective investment trusts. Analyzing the implications of asset reclassification for the mutual fund literature, we find that these events distort inferred mutual fund flows without reflecting actual asset movements at the investment product level. Failing to account for asset reclassification in flow-based regression analyses can lead to biased estimates, as it resembles a non-classical measurement error. We first analyze scenarios in which mutual fund flows serve as a dependent variable, focusing on flow-performance sensitivity. A regression utilizing reclassification-adjusted quarterly flows demonstrates a 40-100% greater flow-performance sensitivity for mutual funds with twin vehicles than one employing unadjusted flows. We then examine cases when flows serve as an independent variable, as in `smart money' tests, where measurement error artificially inflates true estimates.

Would the same person invest differently depending on the investment purpose? Using Danish administrative data linking parents and children, we compare adults who manage both a personal brokerage account and their child’s account. Exploiting within-family variation, we show that child-labeled accounts contain safer, more diversified portfolios, trade less, and have a weaker disposition effect, yielding higher Sharpe ratios despite lower expected returns. Parents' personal accounts take more systematic and idiosyncratic risk, tilting toward foreign and lottery stocks. These within-investor differences reconcile active behavior in brokerage data with the inertia documented in 401(k)s, suggesting that the gap reflects mental accounting.

Over the past two decades, many U.S. states have introduced tax incentives to at-

tract investment in data centers. We provide causal evidence that incentives double data center construction at the state-level. However, these effects are present only for large, high-power facilities used for cloud computing and, more recently, LLMs. Smaller data centers are mostly unaffected. The incentives primarily reduce operating costs and capital expenditures, and lead to annual savings in the millions of dollars for large facilities. Despite these substantial long-term subsidies, we find no clear evidence that data centers stimulate local growth in tech employment.

We show that stock market participants are optimists, while exit coincides with a deterioration in beliefs. Accordingly, belief-cycles drive participation-cycles in our model: Inexperienced cohorts enter with high leverage during episodes of market exuberance and low risk premia and exit during market downturns. In equilibrium, experienced cohorts typically remain participants throughout as the market clearing risk premium tends to gravitate towards their beliefs. These interactions generate procyclical participation, feedback between participation and experience-based learning, and welfare losses. Feeding the model with realized shocks reproduces fluctuations in participation, entry, and exit rates from multiple countries.

We test for and measure green waste: the misallocation of public subsidies for green investment projects. Our context is a major Norwegian program for green investment subsidies. We develop a model of subsidy allocation and apply it to detailed project-level data on carbon emissions and subsidy amounts for both marginal and inframarginal projects. We find that decision-makers could have achieved the same level of emission reductions at less than half the cost. To isolate the sources of this green waste, we use data on both ex-ante expected and ex-post realized emission reductions for each project. We find that decision-makers are able ex-ante to identify the projects with the highest ex-post emission reductions but unwilling to select them.

We conduct a randomized controlled trial to evaluate the impact of contract education and contract legal consulting on entrepreneurs’ outcomes in venture capital (VC) and seed funding. Entrepreneurs in the treatment group received targeted instruction in VC contract terms and startup financing agreements. We find that educated entrepreneurs are less likely to secure initial angel or seed funding and are more likely to pursue traditional employment rather than continue with their startup ventures. However, among those who do obtain an- gel and VC investment, treated entrepreneurs are more likely to receive subsequent funding rounds, achieve higher valuation in later stages, and experience a lower risk of involuntary removal as cofounders. These findings suggest that increased contract knowledge enables entrepreneurs to make more informed career choices and negotiate more favorable terms, albeit at the cost of reduced entry into the VC funding pipeline.

I document the role of a credit crunch in driving the housing market in the 1970s. Binding Regulation Q ceilings tightened funding and induced a credit crunch across the financial sector. I show that the crunch was particularly severe in housing because the primary mortgage lenders, savings and loan associations (S&Ls), lacked the funding flexibility banks had. Banks could substitute rate-capped retail deposits with wholesale funds exempt from the ceiling, whereas S&Ls could not. At the local level, a 1 pp tightening in the effective S&L ceiling is followed over the next year by a 4.7 pp drop in the mortgage growth rate and a 1.1 pp drop in the real house price growth rate. Effects through banks are muted. This mechanism operates alongside demand-side explanations and helps to explain the joint boom-bust patterns in prices and quantities in the housing market during this era.

How borrowers respond to future changes in the interest rate on their debt matters for the transmission of monetary policy and for financial stability. Combining data from a large bank, a letter RCT, and an online survey, we study this question in the context of the German mortgage market, where since 2022 borrowers have faced high interest rates when their rate fixation period ends. We find that various borrower actions substantially reduce the impact of higher rates on monthly payments. Survey responses indicate high awareness of the evolution of interest rates and corroborate a strong propensity to prepare for the rate reset, which we show experimentally is sensitive to the size of the rate increase and to the distance from reset. The letter intervention does not affect rate beliefs, consistent with high ex-ante informedness and selection into reading, but increases awareness of available options and refinancing propensities among borrowers who had not taken action until close to the reset date.

As stock market concentration has risen, regulatory limits on fund portfolio concentration have become increasingly binding, especially for large-cap growth funds. When funds approach these limits, they trim their largest holdings and reduce equity exposure. Funds perform worse when constrained. A constraint-based ownership measure predicts stock returns, particularly among the largest firms. These findings suggest that high market concentration can distort stock prices by limiting the ability of optimistic investors to scale their positions. Just like short-sale constraints can produce overpricing by limiting pessimistic investors' views, constraints on long positions can generate underpricing by suppressing optimists' views.

Using transaction-level data on Chinese retail investors across a boom–bust cycle, we study how salient realized outcomes shape the disposition effect. We measure experience as counts of large realized gains or losses and find a sharp asymmetry: sequences of large gains attenuate the disposition effect, whereas sequences of large losses amplify it. To account for these facts, we develop a disciplined memory-based recall model in which similarity-weighted retrieval of past outcomes guides sell decisions. The model reproduces both the baseline disposition effect and the documented asymmetries, highlighting the critical—and asymmetric—role of experience in shaping trading behavior.

We develop a new framework to study the joint dynamics of prices and trading volume in commercial real estate markets. We start from a micro-founded model where buyers and sellers differ in private valuations of building characteristics. We model listing and meeting probabilities based on investor and property traits. The model implies that transaction prices depend on both building features and participant identities—unlike traditional hedonic models. Using detailed transaction data and machine-learning methods, we find buyer and seller identities have a first-order effect on prices and transaction likelihood. We apply the framework to out-of-sample price prediction and counterfactual analyses on investment flows and prices. The approach extends to private equity, credit, and infrastructure markets.

We study a token-based platform where users choose between adopting the platform’s service and speculating on its native token. While both adopters and speculators earn the token’s capital gain, adopters also incur a participation cost to obtain a noisy service benefit enhanced by network effects. This tension between adoption and speculation can suppress participation, halt learning, and trigger collapse even when expected fundamentals are strong. We characterize the participation threshold, show how token-price volatility expands the collapse region, and identify the resulting learning traps. The model implies that inflation and adopter-targeted incentives support participation, whereas subsidies to speculators increase fragility.

Governments worldwide are increasingly adopting Renewable Portfolio Standards (RPS) to pursue net zero carbon emissions. Exploiting their staggered implementation in 32 U.S. states, we find that RPS explain higher yields and lower credit ratings on local debt securities. Consistent with an energy cost channel, RPS explain higher electricity prices and the effects strengthen when states implement particularly aggressive RPS. Effects weaken in areas with abundant renewable energy natural resources and when states introduce “clean energy” targets that permit a wider set of energy sources. Our findings shed light on the economic trade-offs facing policymakers as they administer the energy transition.

In this paper we explore whether the experience as a founder of a venture capital-backed startup influences the performance of founders who become venture capitalists (VCs). We find that nearly 10% of VCs were previously founders of a venture-backed startup. Having a successful exit, having gone to a top school, and being male increase the probability that a founder transitions into a venture capital career. Skill as a founder carries over to skill as a VC. Accounting for higher observable deal quality of successful founder-VCs (SFVCs), their investment success rates are 4 percentage points higher than professional VCs. On deals in the industry of their startup, SFVCs are more successful by 8 percentage points. Finally, we use two methodologies—a novel instrumental variable framework and a structural model—to explore what drives this superior performance: unobservable deal quality or post-investment value-add. While both mechanisms find some empirical support, the evidence in favor of unobservable deal quality is stronger.

We propose a general equilibrium model in which U.S. monetary policy affects global risk premia by revaluing the wealth of currency-mismatched intermediaries. Assuming their portfolios are mean-variance efficient, intermediaries must be short the dollar. A U.S. tightening thus erodes intermediaries' net worth and raises the global price of risk. We discipline this mechanism to rationalize the effects of U.S. monetary policy on international asset prices, and we study its real implications. In a future with higher dollar interest rates, as due to lower foreign demand for dollar-denominated assets, intermediaries' dollar funding and U.S. monetary policy spillovers are dampened.

Banks may play a more active role in corporate finance than previously documented. Banks with industry expertise steward firms to undertake value-creating opportunities in transition technologies or expansion (new to the firm) sectors. Bank financing can unlock neglected opportunities, especially when banks have specialization, and firms neglect opportunities because of short-termist discount rates. We set up an empirical two-way fixed effect identification approach, building off the dynamic comparables strategy in the spirit of Sun and Abraham (2021) to test these predictions empirically in U.S. administrative loan data. We leverage detailed regulatory risk measures and granular loan-level industry attribution in supervisory data that covers over 70% commercial lending. We find that firms are more likely to undertake transition and expansion loans with specialized banks. Furthermore, we show empirically that the effect of stewardship is stronger for short-termist firms, consistent with the idea that such firms would otherwise neglect value-creating opportunities (Kodak moments). We show the results are robust to limiting the sample to firms with similar transition opportunities due to the passage of the Inflation Reduction Act, limiting concerns about dynamic confounders. Consistent with stewardship emerging a consequence of bank specialization (e.g. Paravisini et al. (2023), Blickle et al. (2024)), we document that banks offer lower interest rates for these projects. Active bank stewarding moves the literature beyond the classic view of banks as mere credit providers.

The rise of sustainable investing challenges a core tenet of asset pricing: that stock prices primarily reflect discounted cash flows. Using earnings announcements as a key source of cash-flow news, we show that firms with high sustainable ownership are 45–58% less sensitive to earnings surprises. These dampened price reactions are accompanied by lower trading volume and no evidence of post-announcement convergence. The weaker response is not driven by smaller surprises, reduced informativeness, or alternative ownership structures. Within a flexible present-value framework, we attribute part of the dampening to lower persistence of earnings shocks, with the remainder explained by a 1–3% reduction in discount rates for high-sustainable-ownership firms. Overall, our findings demonstrate that sustainable investing alters the way markets incorporate fundamental cash-flow information into prices.

High asset price volatility alongside low portfolio flows reveals a fundamental trade-off between investor disagreement and price impact: When volatility is high but flows are small, investors must either largely agree with each other or be insensitive to price changes, implying large price impacts of small flows. We formalize this relationship in a price impact bound based on price volatility, flow volatility, and investor agreement. Applying our bound to U.S. equities yields large price impacts, implying that flows are central to understanding price dynamics. Our bounds align with event-study estimates while revealing novel patterns across horizons, assets, and aggregation levels.

Employer-sponsored health insurance is the predominant coverage source for the U.S. workforce, yet the rationale behind firms' provision of these benefits remains debated. We examine two primary motivations: as a retention mechanism reducing employee turnover, and as a human capital investment enhancing workforce productivity. To disentangle these perspectives, we exploit policy-induced shocks to labor mobility and track firms’ benefit adjustments. We employ a stacked difference-in-differences approach using novel datasets on health plan details, individual healthcare utilization, and state-level variations in non-compete agreement (NCA) enforceability from 2013 to 2020. Our results show that increased NCA enforceability leads firms to lower premiums primarily by shifting to High-Deductible Health Plans (HDHPs). This shift, in turn, boosts HDHP enrollment rates and significantly reshapes healthcare utilization among affected employees, leading to fewer preventive care visits but a greater incidence of severe, high-cost medical procedures. Overall, our findings support the retention cost perspective while also highlighting the unintended long-term consequences of such cost-saving strategies.

We study the macroeconomic implications of agency issues between shareholders and managers. Following \cite{jensen1986agency}, managers tend to over-invest (``empire building''), and do not minimize costs (the ``quiet life''), but are disciplined by the threat of takeover. We embed these frictions in a standard macroeconomic model with firm heterogeneity in productivity and factor adjustment costs. We assume that some firms are well-run (corresponding to a zero takeover cost), some are run by unconstrained managers (corresponding to an infinite takeover cost) and the rest are subject to a finite, nonzero takeover cost, and hence become endogenously well-run or poorly run depending on their capital and productivity. In equilibrium, agency frictions spill over from poorly-run firms to well-run firms: the poorly-run firms underprice the well-run, overuse inputs, leading to aggregate efficiency costs that are larger than the direct individual efficiency costs of the poorly-run firms. We argue that rising shareholder power (lower costs of takeover) in the U.S. since the 1970s has contributed to some important macroeconomic trends, such as lower investment and labor share, and higher profitability, payouts and Tobin's Q.

Meeting rising global electricity demand requires financing large-scale investments in power generation. While they produce the same electricity output as fossil fuel plants, renewable energy plants are financed differently: they typically rely on "project finance" and long-term Power Purchase Agreements (PPAs) that lock in electricity prices. We develop a framework that links power generation technology and financial structure in the presence of financial frictions. We calibrate it to micro-data to show how project finance can help mobilize private funds through a capital structure channel. Due to near-zero marginal costs, PPAs significantly reduce cash-flow risk. This allows greater use of bank debt relative to equity financing, which lowers discount rates and increases investment. This financing channel is technology-specific: project finance is much less attractive for fossil fuel projects with volatile input prices. However, we also highlight a downside of PPAs: the amplification of macroeconomic shocks, leading to higher investment volatility over the business cycle. We use our framework to explain the slowdown of renewable investment following the 2021-2022 inflationary episode and show that state-dependent investment subsidies indexed to cost shocks can help stabilize private investment.

We study how institutional investors manage private equity allocations beyond initial commitments. Using a novel method to identify secondary sales from reporting patterns, we show that public pension plans increasingly use secondary sales of private equity fund stakes to rebalance their portfolios when allocations exceed policy targets. Overallocated plans sell funds with higher asset values and tend to accept larger discounts. Even though secondary sales are costly, they serve as a key rebalancing mechanism under static allocation policies. Meanwhile, overallocated pension plans make only limited adjustments along other avenues, such as smaller new commitments and modest increases in target allocations.

We use a measure of financial constraint that distinguishes between a company’s

emphasis on equity versus debt financing to show that equity-focused constrained

firms endure larger declines in stock prices and implement deeper cuts in investments

when faced with contractionary monetary policy shocks. Equity-focused

constrained firms reduce equity issuance and are more reluctant to run down

cash holdings in response to tighter monetary policy. Contractionary shocks

reduce investor demand for the equity of constrained firms, increasing their cost of

capital. Our findings suggest that equity frictions are the main determinant of the

transmission of monetary policy to the corporate sector.

Understanding the nature of interconnections between banks is crucial to the knowledge of financial contagions and the appropriate policy response. We examine interconnections in the U.S. banking industry that stem from earnings information spillovers from local peers. Using a sample of nearly 8,800 commercial banks over three decades, we find that uninsured deposit flows respond strongly to the performance of peer banks – particularly when peers perform poorly but not when they perform well. Exploring the mechanisms, we find that the earnings information spillovers operate through both asset commonalities and panic-driven behaviors wherein depositors withdraw out of concerns about early withdrawals by other depositors. Our findings inform theories of contagion and their policy implications.

The persistence of mutual fund alphas is often viewed as evidence that some funds possess skill and that this skill persists. This interpretation is unwarranted when security factor lines are too flat. A sort on estimated alphas is a sort against betas: high-alpha funds are predominantly low-beta funds and vice versa. Thus, a strategy of investing in high-alpha funds benefits not from skill, but from a betting-against-multiple-betas effect. In-sample tests intended to separate luck from skill are similarly biased. We demonstrate this bias through simulations and show that applying a correction factor eliminates much of the apparent skill in mutual fund data.

With more than $40 trillion dollars under management, institutional investors of funds play an important role in the financial market. Using novel data on fund viewership, we are the first to examine how these investors allocate attention to specific institutional funds. Exploiting quasi-random variation in screen display features on a prominent institutional asset management platform as instruments, we provide causal evidence that their direct attention drives flows to institutional funds and exerts positive price pressure on their underlying stocks. Overall, our evidence suggests that even sophisticated institutional fund investors suffer from attention constraints, which have important asset pricing implications.

We develop a text-based measure of intangible investment intensity derived from firms’ 10-K filings. Our approach further classifies disclosure text into knowledge, customer, and organization capital. Firms with high intangible intensity are smaller, younger, and invest heavily in R&D and human capital, while the three subcomponents map cleanly to distinct economic firm types. Intangible intensity contains information about future profitability that is not captured by standard accounting measures. Managerial language thus embeds forward-looking signals about intangible investment that accounting data obscure.

This paper examines how AI planning, the core technology behind agentic AI systems that pursue long-horizon objectives by anticipating how current actions shape future payoff-relevant states, affects financial market stability. We develop a dynamic trading framework with positive-feedback investors, constrained arbitrageurs, and oligopolistic informed speculators who may coordinate intertemporally: trading aggressively in tandem to generate (negative) bubbles and subsequently unwinding their positions in a coordinated manner to extract profits. Such intertemporal coordination differs from traditional collusion because it faces two unique, fundamental obstacles: time inconsistency, as coordinated plans become incentive-incompatible once a large (negative) bubble has formed, and weak punishment, as deviations are difficult to penalize when no large (negative) bubble materializes. We characterize equilibria featuring coordinated creation of manipulative, exploitative (negative) bubbles. In simulation experiments, AI speculators trained via reinforcement-learning algorithms with explicit planning modules autonomously discover and implement intertemporal collusive trading strategies based on compounded price-trigger rules, coordinating without communication or shared intent. When feedback trading is strong, these AI-planning speculators dynamically converge on destabilizing strategies that create and exploit (negative) bubbles, manipulate feedback traders, and significantly amplify market fragility.

We study how state environmental review laws (SEPAs) shape local business activity. Using a stacked, border-pair event design around staggered SEPAs, we compare adjacent counties on either side of state lines. SEPA-side border counties experience slower post-adoption growth that cumulates into persistent level losses over two decades, with no mirror-image gains across the border and no broad underperformance in state interiors. Effects are strongest for SEPA-treated counties that border large metro areas, when SEPAs have substantive provisions, and when states devote more resources to environmental protection, and they fall disproportionately on pollution-intensive sectors and small firms. In sum, local environmental regulations can erode local agglomerations, adversely affecting local small business growth.

What are the general equilibrium effects of retail investor entry into private markets? The conventional debate focuses on direct effects: potential benefits (portfolio diversification, access to high-return assets) versus costs (excessive fees, inefficient capital allocation by retail investors). We identify a countervailing mechanism: retail investors fund experimentation with unproven general partners (GPs), enabling institutions to free-ride by poaching proven talent. We document that 34% of value created by institutional investments in private markets comes from GPs who started with retail capital. We formalize this stepping-stone mechanism by extending Berk and Green (2004) to venture capital with heterogeneous investors and uncertain manager skill. New GPs with moderate perceived ability initially raise retail capital to reveal their true skill; after observing performance, skilled GPs graduate to institutional capital while unsuccessful GPs exit. Retail investors pay for information production, but institutions capture benefits through superior manager selection and bargaining power. Our calibrated model shows that retail investor entry increases aggregate welfare when stepping-stone benefits exceed inefficient allocation costs. Policies restricting retail access may harm efficiency by restricting this talent discovery path.

We compute the cross-country transfers that result from unconventional monetary policy in the Eurozone. The ECB funds the expansion of its aggregate balance sheet mostly by issuing bank reserves and cash in core countries. The national central banks (NCBs) in periphery countries then borrow from the core NCBs at below-market rates to fund the asset purchases and bank lending. In addition, NCBs in the periphery lend more to their own banks at below-market rates. To compute the cross-country transfers, we compare the resulting cross-country distribution of NCB income to a counterfactual scenario without the ECB and without non-marketable intra-Eurozone debt. We document significant and persistent transfers from the core to the periphery.

In 2022, Fannie Mae and Freddie Mac waived a portion of their guarantee fees for mortgages to lower-income first-time homebuyers in order to reduce barriers to homeownership. We estimate that private intermediary mortgage lenders, on average, capture nearly half the value of the fee waiver rather than fully passing these waivers through to homebuyers. Moreover, despite Federal regulation prohibiting loan officers from benefiting financially from interest rate markups, we document significant variation across loan officers in pass-through of the waivers, consistent with loan officers continuing to have incentives and market power to charge higher prices. In particular, we find that loan officers with a track record of originating expensive mortgages pass through only a fraction of GSE fee waivers to borrowers.

We evaluate approaches to estimating demand elasticities in dynamic asset markets, both theoretically and empirically. We establish strict, necessary conditions that the dynamics of instrumented asset price variation must satisfy for valid identification. We illustrate these insights in a general equilibrium model of dynamic trade and derive the magnitude of biases that arise when these conditions are violated. Estimates based on static IO models are severely biased when the instrumented price variation is persistent or predictable. Empirically, we show that commonly used instruments yield elasticity estimates that are off by orders of magnitude, or even have the wrong sign. In contrast to standard multiplier calculations, our theory characterizes the dynamic asset market interventions required to sustain a given price path support process, with direct implications for policies such as Quantitative Easing (QE).

We construct a novel measure of partisan corporate speech using natural language processing techniques and use it to establish three stylized facts. First, the volume of partisan corporate speech has risen sharply between 2012 and 2022. Second, this increase has been disproportionately driven by companies adopting Democratic-leaning language, a trend that is widespread across industries, geographies, and CEO political affiliations. Third, partisan statements are followed by negative abnormal stock returns, with significant heterogeneity by shareholders’ alignment with the statement. We propose a theoretical framework and provide suggestive empirical evidence that these trends are driven by a shift in investor preferences.

This paper studies how insurance brokers affect product choices, premiums, and welfare in the employer-sponsored insurance market. We compile a novel database of contracting relationships among employers, brokers, and insurers in New York State. Exploiting variations in commission schedules, we document two market distortions: First, brokers exhibit traditional agency frictions, steering employers towards more financially lucrative products. Second, commission levels affect ex-ante insurer-broker networks and, in turn, insurers' competitive pressure, leading to anti-competitive distortions. We develop and estimate a structural model of employer insurance demand, insurer pricing, and formation of broker-insurer contracting networks. We use the model to quantify the steering-competition tradeoff: higher commissions exacerbate steering but may also broaden broker networks, increase insurer competition, and lower premiums. We also explore optimal commission and network regulations for insurance brokers.

This paper develops a complete-market production economy with heterogeneous beliefs about TFP growth.

Hiring occurs before TFP is known and is, therefore, risky (operational leverage).

The firm's discount factor depends on a wealth-weighted average of investors' beliefs.

Waves of optimism influence asset markets and percolate to labor hiring, tying together the equity premium, equity volatility, and labor volatility puzzles.

A taxonomy of belief systems shows that only extrapolative beliefs amplify the volatility of asset prices and hours and lead to risk build up.

Disciplined by survey data, the model matches asset-pricing and business-cycle moments, highlighting how heterogeneous beliefs can be a direct driver of aggregate fluctuations.

We show that uncoordinated climate policies, which expose firms to heterogeneous climate transition risk, induce supply-chain reconfiguration. Following the introduction of California’s cap-and-trade program, relationships involving regulated suppliers are more likely to terminate than otherwise comparable relationships with unregulated suppliers. Regulated suppliers experience operational disruptions, consistent with customers reallocating sourcing to reduce uncertainty about supplier reliability. These effects are concentrated among suppliers in competitive industries and producers of standardized inputs, where customer switching costs are low. The resulting reconfiguration is consistent with carbon leakage: customers that reallocate away from regulated suppliers shift toward supply chains with a higher carbon footprint.

Using proprietary data on U.S. venture capital (VC) fund distributions, I find that at least 56% of capital is returned to limited partners (LPs) in kind as publicly listed equity rather than cash. Distributions follow substantial price appreciation, reflecting both general partners’ (GPs) value-added activities and their ability to time exits. Distributed stocks experience a sharp, non-reversing price drop that modestly reduces LPs’ realized returns. Losses are smaller for distributions made by reputable VCs, not because they employ more LP-friendly distribution practices, but because they invest in companies with stronger post-distribution performance. The findings reveal a new agency friction: GPs report fund returns—and earn carried interest—using distribution marks that LPs cannot feasibly realize. However, a counterfactual analysis shows that prohibiting stock distributions would reduce LP returns even more, underscoring the trade-off LPs face.

Abstract: A significant fraction of measured surprise central bank interest rate hikes

is associated with simultaneous stock market run-ups and upward revisions in economic

growth forecasts. This evidence is often interpreted as contradicting the standard New Keynesian transmission mechanism and as indicating that the Fed possesses superior information

about economic fundamentals. We present a New Keynesian model in which investors are

uncertain about the Fed’s long-run monetary policy objectives and learn from observed Fed

actions. Our model does not assume that the Fed has superior information, but it nonetheless generates a “Fed information effect,” that is, a positive co-movement between interest

rate surprises, revisions in expected growth, and stock returns as an equilibrium outcome.

We show that the key predictions of our model are consistent with empirical evidence on

asset market responses to measured Fed information shocks.

We study how intermediaries transmit idiosyncratic funding cost shocks to consumers using a field experiment in the U.S. auto loan market. A subprime lender randomly varied the interest rate offered to dealers. Dealers could profit by marking up this rate to set the borrower’s APR. A 1 ppt increase in the offered rate reduced loan take-up by 14%, and the pass-through ratio to the borrower’s APR was 0.89. Dealers’ proposed margins were largely insensitive to the offered rate, revealing that sophisticated intermediaries transfer cost shocks mechanically, not strategically, highlighting a key micro-level feature of the transmission of interest rate changes.

We study how to estimate and interpret demand elasticities in dynamic settings where prices and flows are jointly determined. Because price changes necessarily alter expected returns or cash flows, there is no single context-free elasticity: measured elasticities depend on which expectations adjust. We formalize static, immediate, and dynamic elasticities and show that the dynamic elasticity equals the inverse of the equilibrium price multiplier. A tractable linear model delivers testable comparative statics linking price multipliers to risk, persistence, systematic exposure, and surprise. Empirically, we confirm these patterns and show that structural demand can be recovered from reduced-form price responses even under persistent shocks.

We construct novel datasets on AI innovation policies across 42 countries and

global AI patents to examine the impact of these policies on the productivity of AI

scientists. We find that scientists experience a decrease in productivity after their

country initiates AI-supporting policies. We argue this decline could be driven by a

shift in the nature of innovation being conducted; government support incentivizes

scientists to pursue more novel and exploratory inventions. While these projects

hold the potential to foster long-term growth, they are inherently characterized by

longer development timelines and a greater risk of failure, resulting in a temporary

decline in the average quantity and quality of innovations. This effect is potentially

compounded by the government’s inability to perfectly identify and fund the most

promising radical innovations, leading to a misallocation of resources. We develop a

general equilibrium framework that formalizes these dynamics. Our results highlight

the transitional risks associated with government support for innovation, particularly

as an economy navigates a new technological paradigm.

We investigate the role of financial advisors in shaping clients' asset allocation during retirement. Using data on more than 37,000 advised Canadian retirees, we document that advised retirees maintain high equity shares of 60--70 percent well into old age. This share of risky assets is roughly twice the level prescribed by common rules of thumb, target-date funds, or life-cycle models calibrated to non-advised portfolios. Conflicts of interest are unlikely to explain this risk-taking, as retired advisors hold similarly high equity shares in their own portfolios. We show that the observed portfolios are consistent with a standard life-cycle model featuring moderate risk aversion and modest financial wealth levels. Under a ``money doctor'' interpretation, advisor-induced beliefs about market returns can rationalize advisor fees of up to 140 basis points per year.

When the value of a pledgeable asset (or project) is uncertain, investors are tempted to examine it. The information cost is ultimately borne by the asset owner, reducing her financing capacity. A pecking order emerges. Debt generates a greater financing capacity than equity: unlike equity investors who own the asset directly, creditors own the asset only if the borrower defaults and, therefore, have weaker incentives to acquire information. The probabilistic asset ownership can be further diluted by introducing intermediaries between the borrower and the creditor, leading to a new theory of financial intermediation and credit chains. We demonstrate that the optimal financial architecture involves systematically sequencing multiple intermediaries with heterogeneous information costs and asset correlations, rationalizing the seemingly excessive complexity of intermediated credit flows.

We study commitments to reduce emissions by firms subject to the European Union Emission Trading System (EU ETS), the world’s largest cap-and-trade program. Commitments are associated with a drop in the number of carbon allowances surrendered, consistent with firms taking actions to reduce their emissions. However, firms subsequently increase their sales of allowances on the secondary market, transferring the right to pollute to others and potentially leaving aggregate emissions unchanged. They do not reduce emissions outside the EU or invest in green technologies. Despite this, firms benefit from commitments via higher ESG scores. Our findings underscore the importance of considering the interaction between carbon markets and private climate initiatives.

We examine how remote work affects firm productivity and employment outcomes. Using an instrumental variables design based on pre-pandemic occupational remote work suitability, we find that remote work increases the productivity of startups while reducing productivity for established firms. This effect operates partially through labor market scaling: remote startups experience substantially higher hiring rates, overcoming geographical hiring constraints for talent acquisition. Our findings suggest that remote work alleviates a key disadvantage that young firms face in hiring and expanding, while established firms experience challenges in adopting remote working practices.

We measure and analyze inflation uncertainty in the US. We construct a novel composite indicator of inflation uncertainty (CIU) from two components: a news-based measure derived from textual analysis of newspaper articles using large language models and a market-based measure that draws on prices of options on Exchange Traded Funds and commodities. Unlike survey- or inflation-option-based measures, our index is available in real time and extends back to 1926. CIU reveals that inflation uncertainty spiked during the Great Depression, World War II, the 1970s and 1980s, following the Global Financial Crisis, and in the post-pandemic period. We highlight the driving forces behind these fluctuations in uncertainty and analyze their economic consequences. Heightened inflation uncertainty is associated with higher prices of real assets—such as gold, silver, and housing—but with lower prices of nominal assets, including government bonds, corporate bonds, and equities. Moreover, we find that increases in inflation uncertainty are followed by declines in private investment and real economic activity.

Recent shifts in federal real estate policy, marked by lease cancellations and non-renewals, challenge the long-standing perception of government contracts as a secure and stable investment. We investigate how federal lease cancellations impact the pricing of government contract risk premia. Using unanticipated Department of Government Efficiency (DOGE) cancellations as a shock to commercial mortgage default risk, we find that first-loss CMBS bond tranches backed by loans tied to DOGE-notified leases experience a persistent 4% decline in price, with large, negative spillover effects to bond prices, delinquency rates, and rental cash flows tied to nearby private-tenant leases. These results reflect that early termination options were previously perceived by investors as a dormant clause. Applying arbitrage pricing models of commercial lease contingencies confirms the underpricing of risk associated with government tenants. Simulations of tail risks from early termination exposure result in aggregate property value losses in excess of $575 million for the Washington, D.C. securitized office market alone – well above total potential taxpayer savings from canceled lease payments.

Yes, by decreasing firm risk, not by increasing profitability, and with investors taking years to recognize the value created. We start, using novel AI patent data, by documenting significant corporate production of AI innovation as early as 1990. Then, we show that a signification motivation for a firm's AI production is the mutually reinforcing effects of the firm's innovation capacity (exogenous R&D stock) and its labor inputs' AI exposure (both the firm's own and its customers'). We use the interaction of these two effects to instrument for AI production. We find that producing AI creates firm value through a large, permanent decrease in risk (cash flow and stock return, systematic and idiosyncratic). Further evidence suggests that AI lowers physical capital intensity and increases bargaining power for producing firms. The initial market reaction to AI patent announcements is economically small, but abnormal stock returns thereafter are significantly positive (about 5% per year) for (only) roughly three years, suggesting initial undervaluation followed by gradual correction. We find no evidence of investor learning, except during the past five years. We empirically distinguish producing AI innovation versus AI adoption, automation, general technology, and other potential confounds.

Unfunded fiscal expansions are a significant source of risk premia in Treasury markets when central banks and governments subsidize taxpayers at the expense of devaluing bondholders' returns through inflationary finance. We formalize this redistribution mechanism in a two-agent model that features limited asset market participation and monetary-fiscal interactions. A fiscally-led policy regime reduces bond returns in response to surprise fiscal expansions, while the redistribution effects from asset holders to hand-to-mouth agents increase the pricing kernel, making government bonds a risky asset. This fiscal redistribution risk mechanism generates sizable nominal term premia in a calibrated New Keynesian framework with partially-funded spending.

Income shocks are widely studied, but expense shocks are usually neglected in modern consumption-savings models. We use a survey linked to credit bureau records to estimate expense shocks’ frequency, size, and financial implications. Expense shocks are more frequent and larger than income shocks. Very large expense shocks are several times more common than very large income drops. When we account for this additional risk in a standard model, our estimates imply: (1) many costly decisions such as loan defaults are driven not by strategic considerations but by expense shocks, (2) smooth consumption implies a lack of liquidity, not an abundance of it, (3) expense shocks explain most precautionary wealth, and (4) expense shocks are more important than income shocks for studying the marginal propensity to consume. Our results and publicly available data allow researchers to account for expense shocks.

The asset management industry is increasingly shifting toward tailored portfolios, fund proliferation, and decentralization of stewardship – trends partly driven by growing heterogeneity in investor preferences. While these developments better align investment products with investor demands, they also reshape ownership structures, potentially leading to more dispersed ownership and weaker managerial oversight. We develop a framework to evaluate these trade-offs and show that fund proliferation does not necessarily weaken governance: Stronger incentives for asset managers and concentrated portfolios of specialized funds can offset these effects, especially when investor preferences are intense. However, strong investor preferences can also induce asset managers to compete on a new margin—granting investors control by decentralizing stewardship and adopting pass-through voting—without internalizing the associated governance costs.

We study how location and household characteristics determine homeowners insurance pricing. Using 70 million policies linked to mortgages and property-level disaster risk, we show that credit scores impact premiums as much as disaster risk. Leveraging a temporary ban on credit-based pricing in Washington State, we find that the direct use of credit information explains most of this pricing variation. Insurance premiums act as a “second credit channel” and are as important as mortgage rates in driving housing cost variation across credit scores. We discuss mechanisms behind the role of credit information and their implications for housing affordability and exposure to disasters.

We examine how the expansion of work-from-home (WFH) arrangements affects entrepreneurial entry, leveraging the natural experiment provided by the COVID-19 pandemic and related stay-at-home mandates. Pandemic-driven widespread adoption of remote work leads to an overall increase in entrepreneurial entry, but also reveals a significant substitution effect. Areas with higher ex ante telework potential experienced notably smaller increases in new businesses. We propose and test a conceptual framework that emphasizes the substitution between employer-provided flexibility and entrepreneurship’s traditional advantage in offering autonomy, finding empirical evidence consistent with the model. Survey evidence further confirms that employer-provided flexibility reduces entrepreneurial intent, especially among those primarily motivated by autonomy or flexible schedules. These findings highlight a critical policy tradeoff: while greater flexibility in traditional employment enhances job satisfaction and work-life balance, it may simultaneously dampen entrepreneurial dynamism, particularly in small-business sectors essential for inclusive economic growth.

I document a large increase in intraday market reactions to Consumer Price Index (CPI) news during the 2021--2023 inflation surge, while reactions to other macroeconomic news announcements remain comparatively unchanged. An investor attention measure---constructed from Bloomberg Terminal coverage in the days leading up to the CPI release---can robustly explain the rise in sensitivity to CPI news across asset prices. While the attention measure shares common variation with a variety of factors, none of the alternatives considered can eliminate its explanatory power. Based on this evidence, I construct a similar measure for Federal Open Market Committee (FOMC) announcements. Higher investor attention predicts greater market volatility following FOMC announcements, and further evidence indicates that this increase reflects an amplification of market sensitivity to FOMC news. Overall, the findings point to substantial time variation in markets' sensitivity to macroeconomic news, a larger role for endogenous attention in macro-finance, and potential challenges in measuring monetary policy through financial markets.

We construct a new historical corporate bond database spanning 128 calendar years to address longstanding data limitations hampering corporate bond research. By hand-collecting monthly corporate bond quotes from three archival print sources, we complement existing datasets and create an extensive database dating back to 1895, comprising nearly 110,000 unique bonds and 8 million observations. Leveraging this expanded sample, we find that the lack of priced risks

in corporate bonds documented by recent studies stems from their reliance on short samples. With greater statistical power, we show that prominent bond and stock factors as well as several nontraded macroeconomic factors are significantly priced with theoretically consistent signs. At the same time, the predictive power of corporate bond spreads for real activity is largely robust in the longer sample, except when pre-war data are included. Our database, covering major economic episodes like the Great Depression, not only helps validate previous empirical findings but aims to facilitate further research by serving as a CRSP counterpart for corporate bonds.

While technological innovations typically increase wage inequality by favoring skilled workers, we show that instant payment systems instead reduce it. Using matched employer–employee administrative data, we study Brazil’s rollout of Pix. We implement a triple-difference design that exploits pre-treatment variation in mobile penetration across municipalities, the differential exposure of small versus large establishments, and the timing of Pix’s introduction. We find that wages in small establishments rise significantly relative to those in large establishments after Pix. These wage gains are concentrated in cash-intensive sectors (retail and services), with no effects in wholesale or manufacturing. Crucially, wage inequality declines, driven entirely by wage gains among workers in the lower half of the distribution. Our evidence points to increased small-business labor demand and firm entry, amplified by local labor market frictions. Overall, our findings show that technology’s impact on inequality hinges on the specific frictions it relaxes and the types of firms it disproportionately benefits.

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 political appointments to public roles. Using a stacked Difference-in-Differences design around presidential party transitions from 1990, 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). The results also hold when examining historical FOMC votes starting from 1936. At the committee level, a one-point increase in political alignment of 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 real gdp, employment, and stock market, 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.

529 college savings plans have become more prevalent than student loans among undergraduates, yet their educational impacts remain largely unexplored. This paper examines how wealth gains in 529 plans shape educational opportunities using a shift-share IV approach that exploits variation in target-date fund designs within these plans. Contrary to existing literature suggesting minimal effects of household wealth on college attendance, I find that 529 wealth gains are remarkably effective at increasing four-year college attendance, matching the impact of targeted grant aid per $1,000. These wealth gains also reduce student loan borrowing and boost private K-12 school enrollment. However, 529 wealth gains accrue disproportionately to upper-income households, exacerbating the four-year college attendance gap between students from the top and bottom income quartiles by 16% in 2023.

This paper develops probability pricing, extending cash flow pricing to quantify the willingness-to-pay for changes in probabilities. We show that the value of any marginal change in probabilities can be expressed as a standard asset-pricing formula with hypothetical cash flows derived from changes in the survival function. This equivalence between probability and cash flow valuation allows us to construct hedging strategies and systematically decompose individual and aggregate willingness-to-pay. Four applications examine the valuation of changes in the distribution of aggregate consumption, the efficiency effects of varying performance precision in principal-agent problems, and the welfare implications of public and private information.

Insurers of last resort are becoming critical due to climate-related stress on private markets. Using policy-level data, we show such insurers pass costs of climate disasters through both premiums and claim rejections. Following disasters, premiums rise in affected and unaffected areas, while rejections increase in unaffected areas. Spillovers depend on price sensitivity: less sensitive pay higher premiums, more sensitive face higher rejections. Premiums (rejections) increase during financially constrained (unconstrained) periods. Households respond by increasing insurance against disasters while accepting smaller risks. Welfare analysis shows rejections are critical in redistributing costs. Disasters force even government-backed insurers to act like profit maximizers.

We find that U.S. export restrictions weaken the innovation incentives of U.S. firms that supply sanctioned foreign entities. Export restrictions prompt targeted foreign entities to accelerate their innovation efforts, ostensibly with increased support from their governments—including weakening enforcement of U.S. intellectual property rights (IPR). Weaker IPR reduces U.S. suppliers’ ability to appropriate returns from R&D, leading to an 11% decline in R&D expenditures and a 17% reduction in R&D-related hiring. Post-sanctions, U.S. suppliers adjust their IP protection strategy: patent filings (which require detailed disclosure) decline by 22%, while mentions of trade secrets in regulatory filings rise by 41%. These effects are stronger when sanctioned entities are likely to reverse-engineer their suppliers’ technology and weaker when domestic competition necessitates U.S. firms to innovate. The impact is most pronounced in patent-intensive industries and for suppliers who hold patents in sanctioned countries. Finally, R&D employees who leave exposed suppliers experience adverse labor market outcomes. Our findings suggest that export controls may unintentionally fuel foreign innovation and IP appropriation, prompting U.S. firms to scale back innovation and favor secrecy over patenting.

The US government’s money demand is large and volatile and takes the form of the Treasury General Account (TGA) at the Federal Reserve. I study the drivers of TGA volatility and the options available to the Federal Reserve for adjusting its balance sheet in response to this volatility. Using interest rate control, control of the Fed’s overall policy stance, and communication as criteria, I argue that a policy of adjusting the Fed’s holdings of Treasury bills (or other short-maturity assets) performs better than the current ample reserves policy of letting the supply of reserves and overnight reverse repo facility balances adjust passively to changes in the TGA.

Corporate bond payoffs are intrinsically non-linear, making it difficult for traditional factor models to explain the cross-section of bond returns. We combine a reduced-form affine term structure model with mean-variance portfolio allocation to propose a default-adjusted CAPM.

Our framework introduces credit betas: a bond-specific measure of default risk easily derivable from bond analytics that maps directly to each bond’s loading on priced systematic default risk. This approach allows us to perform risk adjustments without estimating bond-specific default probabilities, a notoriously difficult task. The adjustment implies that cross-sectional returns increase with credit betas and decrease with duration. Using US corporate bond data, we construct a strategy that shorts duration while going long credit beta, delivering a marketorthogonal Sharpe ratio of 1.1, roughly 2.5 times the duration-hedged market return.

How does AI-generated personal financial advice from Large Language Models (LLMs) compare to economists’ normative models? We develop and implement a method to evaluate generative AI financial advice by simulating thousands of life cycle paths for consumption, saving, and portfolio choices under realistic income, employment, and asset return scenarios. Our approach compares these LLM-generated paths against optimal choices from a standard life cycle model and can parsimoniously summarize LLM advice across prompts and models by estimating structural time and risk preferences. Applying our method to OpenAI’s GPT-5 mini, we find that the advice qualitatively aligns with standard life cycle theory but deviates systematically in four key ways: (i) recommended consumption and saving paths imply unrealistically high patience, with estimated intertemporal discount factors well above one; (ii) recommended choices often reflect simple heuristics, such as round savings rates, fixed-percentage withdrawal rules in retirement, and common asset allocation rules-of-thumb; (iii) LLM recommendations exhibit substantially more inertia in portfolio rebalancing and moderately less consumption-smoothing in unemployment than our normative benchmark; and (iv) holding all else constant, recommendations vary systematically with demographics (e.g., recommending lower equity shares for women) and between repeated identical queries.

We introduce immigration into an overlapping-generations growth model. A planner chooses the rate of immigration to maximize the utility of a representative native consumer in the steady state, subject to providing each immigrant with a threshold level of utility. We derive a sufficient condition for the optimal immigration rate to be positive, identifying the components of the marginal cost and marginal benefit of immigration. The planner's optimal steady state can be implemented in a competitive economy using a fiscal package consisting of an entry fee for immigrants, a consumption tax, and government bonds.

We use survey responses to hypothetical scenarios linked with administrative 401(k) data to investigate the efficiency and equity implications of employer 401(k) match formulas. We find that (i) survey responses can be used to accurately predict saving responses to both observed and hypothetical plans, (ii) saving is inelastic to the match rate, (iii) non-elective contributions do not crowd out saving. These patterns imply that (iv) plans combining lower match rates with non-elective contributions generate higher savings and more equitable match distributions, and (v) many existing plans—including safe-harbor formulas—are strictly dominated along both

We study a dynamic problem of selling data without commitment to a budget-constrained receiver. The sender has access to a data-generating process, informative about a fundamental state, and can sell it either as granular observations (raw data) or summary statistics (information). Properly designed, such statistics ensure the residual uncertainty declines predictably

along with the receiver’s budget, supporting efficiency under gradual information sales. In contrast, selling raw data poses a risk that future observations increase residual uncertainty, exceeding the receiver’s remaining budget. Consequently, selling data is inefficient if the fundamental is discrete and requires excess budget if the fundamental is non-Gaussian.

El Nino cycle is a slow-moving global climate phenomenon that hits multiple countries over time in relatively predictable patterns, affecting economic growth and international trade patterns across countries. Examining over different El Nino cycles, we discover a striking pattern of cross-sectional predictability in foreign exchange spot and excess returns. Currencies that appreciated (depreciated) under previous El Nino cycles tend to appreciate (depreciate) when a new El Nino cycle hits. This cross-sectional predictive information arises from the heterogeneous effects of El Nino on countries’ business cycle conditions, resulting in heterogeneous exposures of currencies to El Nino cycles.

Industry size standards that determine eligibility for U.S. small business support have vastly increased over the past decade. We exploit quasi-random variation in the implementation of size standard increases to study the effects on small firms, resource allocation, and industry outcomes using administrative data from the Census Bureau. Following size standard increases, revenues decline for an industry’s smallest firms. These effects stem from procurement contracts being reallocated from smaller to larger firms. Consequently, small firms are more likely to exit and industries patent less. Our findings highlight the broad economic effects of changing eligibility for government small business support.

Employing data from Polymarket, a blockchain-based prediction market where users trade on Federal Reserve rate decisions and scenarios related to central bank independence, we construct a hawk–dove score for wallets and link beliefs to monetary policy expectations. Users who believe President Trump will fire Fed Chair Powell, and who expect stronger political pressure on the central bank, hold more dovish views and expect lower short-term rates than other users. They also expect higher long-term Treasury yields and higher inflation, consistent with reduced policy credibility. The findings indicate that political events affect expectations through perceived threats to central bank independence.

The recent surge in retail option trading has sparked concerns about leverage- and lottery-seeking. We study why retail investors trade options over stocks leveraging a novel trader-level dataset of 2,415,323 parent trades. Options constitute about one-third of all trades and concentrate in just ten underlyings, especially the S&P 500 index. We document sub-hour median holding periods and minimal combined stock-option positions. Investors use options primarily to access high-priced underlyings as the median option’s underlying is 31 times more expensive than the median stock. Contrary to leverage-seeking concerns, they achieve only two-to-one realized leverage by sizing option trades six times smaller than stock trades. Contrary to lottery-seeking concerns, option trades’ profits are distributed nearly symmetrically, not positively skewed. We leverage our sample’s heterogeneity to confirm core findings across investor types, brokers, and within traders. Overall, rather than a new extreme, retail option trading reflects familiar stock speculation.

Credit bureaus face significant frictions in collecting consumer medical debt liabilities data, which spurred an intense ongoing policy debate. Leveraging novel healthcare costs proxies based on Medicare spending data, we evaluate the impact of hidden medical liabilities on consumer credit scoring and access to credit. We document that the traditional creditworthiness measures underestimate the ex-post default for consumers residing in higher healthcare costs markets. Consumers in high-healthcare-cost CBSAs are 37.6% more likely to default than those in low healthcare-cost CBSAs. These effects are more pronounced among higher risk consumers, those with low credit scores and high DTIs. Lenders internalize these biases and impose higher mortgage rejection rates in high-healthcare-cost CBSAs, particularly for riskier applicants. These effects intensify following a policy shift that partially removed medical liabilities from credit reports without affecting consumer balance sheets. Our findings suggest that limiting the flow of medical liabilities data undermines the predictive accuracy of standard credit metrics, impairs the information value of credit bureau outputs, and leads to less efficient credit allocation.

In a cap-and-trade system, emitters face transition risk, to the extent that emission caps and permit prices are volatile. We show, theoretically, and empirically for the EU Emissions Trading System, that (i) emitters hedge with emission permits futures bought from financials, (ii) financial constraints limit hedging, in particular by limiting and delaying emitters' purchases of permits in the spot market, implying (iii) permit prices are below the prices of replicating derivatives portfolios. Moreover, we show theoretically that constrained Pareto optima are implemented in equilibrium with cap-and-trade systems, in which the variance of emission caps is set lower than in the unconstrained case.

Even very straightforward policy changes in the provision of consumer financial services can fail to impact the vulnerable households they would have been expected to benefit. In this study, we examine policy changes from large U.S. banks between 2017 and 2022, which eliminated non-sufficient funds (NSF) fees and relaxed overdraft policies. Using individual transaction-level data, we find that the elimination of NSF fees, not surprisingly, resulted in immediate reductions in NSF charges across the income distribution. However, relaxing overdraft policies resulted in reductions in overdraft fees only for wealthier households, along the dimensions of income and liquidity, and only those enjoyed subsequent declines in late fees, interest payments, account maintenance fees, and the use of alternative financial services, such as payday loans. Our results thus suggest that the policy changes were not substantial enough to significantly reduce the financial stress of the more vulnerable households. Finally, as our setting features multiple treatments and variation in treatment intensities, we theoretically motivate and empirically implement a new stacked event study estimator closely related to de \cite{deChaisemartin2024} to address the biases arising from staggered DID specifications.

We document new empirical facts and develop a theory of private equity continuation vehicles (CVs), in which general partners (GPs) transfer portfolio companies from an existing fund to a new vehicle they continue to manage. CVs can improve efficiency by extending the holding period of high-potential firms, but they also allow GPs to extract rents by exploiting their informational advantage and intermediary position between legacy and new limited partners (LPs). CV formation may involve two informational frictions: adverse selection, where new LPs overpay for low-quality assets, and inverse selection, where they underpay for high-quality assets. These frictions intensify when fewer legacy LPs roll over their stakes.

The GP’s coinvestment and carry, and whether the legacy fund is in or out of the money, determine whether CVs arise and how they perform. The model links LP liquidity needs, GP coinvestment, and contract design to CV performance and distribution of value across investors.

We study the economics of homeowners' property insurance by examining how contract design balances the trade-off between incentive alignment and risk sharing. Using granular contract-level property insurance data merged with property-level disaster risk for millions of U.S. households, we develop and structurally estimate a model in which insurers optimally determine contract terms given property risk and household risk preferences. The estimates provide, to our knowledge, the first large-scale contract-level structural measures of risk aversion, risk premia, and the cost of moral hazard, allowing us to quantify how disaster risk is allocated between insurers and households. We find that the cost of moral hazard is small, yet the very mechanism used to mitigate it substantially increases households' exposure to disaster risk: contract design leaves policyholders exposed to roughly 29 percent of total expected losses. This residual exposure is most pronounced for low-FICO households and for properties with the greatest tail risk. Counterfactuals indicate that mandating full insurance would lead to substantial market exit, increasing household vulnerability. We further show that insurers' financial constraints are systematically correlated with the riskiness of underwritten properties and with household characteristics.

Despite rapid advances in information technology (IT), corporate investment in new technologies remains slow and uneven across firms. This paper provides new empirical evidence that skill mismatch—the misalignment between firms’ skill requirements and workers’ capabilities—acts as a central friction in the diffusion process. Using large language models to construct granular measures of skill demand and supply from matched data on firm job postings and worker resumes, we test the predictions of a model in which skill mismatch evolves endogenously with firms’ investments in different IT vintages. We document a U-shaped pattern: skill mismatch is highest when new technologies are introduced, declines as firms and workers adjust, and rises again as technologies become obsolete. These gaps are substantial not only for technical skills but also for IT-complementary nontechnical skills such as managerial and organizational capabilities. We further show that firms invest more in intangible capital and exhibit lower measured productivity early in the technology lifecycle when mismatch is greatest. The findings illustrate how skill mismatch plays a critical role in both the timing and the realized gains from corporate investment in new technologies.

We propose a novel approach to sovereign crises. Using seven decades of data for 40 economies, we show that as governments devote more revenue towards debt service, growth slows. If fiscal surpluses do not adjust, inflationary pressures build: deficits financed through money creation or excessive borrowing undermine price stability and erode confidence. Developed economies raise primary surpluses almost twice as much as emerging economies in response to higher interest burdens. As a result, developed economies experience slower growth with disinflation, while emerging economies face smaller output losses but much higher inflation. Debt burdens predict hyperinflation and disasters in consumption and output.

We show that the standard Granger causality test for assessing informational efficiency between financial markets is misspecified in the presence of market-microstructure noise, a pervasive feature of financial data. Although the test remains statistically valid, its economic interpretation is flawed: predictability from microstructure noise is misread as information flow. We propose a new test robust to such noise and apply it to credit markets, overturning established results. The corporate bond market, not the CDS market, leads in price discovery; there is no evidence of insider trading in CDS; and bond transactions contain more timely information than quotes.

Robo-advisors typically make investment recommendations based on surveys of clients’ preferences. Consequently, biases induced by the surveys would be embedded in those recommendations and hence may influence investors’ financial decisions. We examine this hypothesis through two studies. First, we administer a nationwide survey of experienced U.S. investors and find that the order of listed choices significantly influences their responses to risk-tolerance questions commonly used by robo-advisors. Second, we collaborate with a leading robo-advisor in China to conduct a preregistered RCT and demonstrate that the order effect significantly alters both the robo-advisor’s risk assessments and its clients’ investment decisions, raising new ethical concerns in the FinTech era.

This paper studies the role of bailout expectations in shaping the dynamics of bank credit spreads and the implications for bank risk-taking behavior.

I propose a dynamic model of financial intermediation with bank default and time-varying bailout probabilities.

Credit spreads are driven by both fundamental risk and bailout expectations.

These two forces have contrasting implications for the joint comovement of credit spreads and default probabilities.

Combining the model with US bank credit default swap spreads and option-implied default probabilities,

I indirectly infer the relative importance of fundamentals and bailout expectations as drivers of spreads.

I find that 28 basis points out of the 34-basis-point rise in credit spreads after 2010 are due to lower perceived bailout probabilities, and that the remainder reflects weaker fundamentals and is partly offset by tighter capital requirements.

Finally, I use the model to measure the effect of lower bailout expectations and tighter regulation on the expected returns of bank assets and the cost of bank credit.

Abstracting from lower bailout expectations overstates the importance of regulatory tightening by a factor of two.

We examine the response of individual borrowing to changes in liquid wealth, exploiting a quasi-natural experiment in Chile. During the COVID-19 pandemic, the government temporarily allowed partial withdrawals from otherwise illiquid pension accounts. The policy’s nonlinear withdrawal rules generate several kinks, which we use to estimate the elasticity of borrowing with respect to liquid wealth through a regression kink design. We find substantial debt repayment among the predominantly low-income, young, and female population, particularly for individuals with higher debt-to-income ratios within that population. We interpret these findings through a model in which the marginal cost of debt increases with borrowing.

We analyze the adoption of clean technology by heterogeneous firms subject to financing constraints. In the model, capital goods differ in terms of their energy needs and age. In equilibrium, cleaner and newer capital requires more financial resources. Therefore, financial constraints induce an endogenous pattern in clean technology adoption: Financially constrained, smaller firms optimally invest in dirtier and older capital than unconstrained, larger firms. The model is consistent with the empirical patterns of technology adoption we document using data on commercial shipping fleets. We use a calibrated version of our model to simulate the aggregate transition dynamics to cleaner technology.

We provide the first analysis of firms' exposure to market expectations of future weather volatility. Using option contracts on temperature indices, we estimate the weather option implied volatility (WIVOL) and document it captures risks of future temperature oscillations, increasing with climate uncertainty about physical events and regulatory policies. Local weather risk shocks worsen operating performance and increase uncertainty about fundamentals, indicating incomplete hedging. The magnitude of these effects suggests that, while managers acknowledge weather risk, they seem to blame it on the weather rather than implementing risk management policies. In asset prices, firms with more negative return exposure to WIVOL trade at a discount and earn higher future returns, consistent with a compensation for weather volatility risk. Only exposures to local, not foreign, WIVOL predict returns, revealing a local channel for pricing uncertainty shocks.

We examine the tradeoff between fertility and entrepreneurship using China’s twochild

policy. Households exposed to the reform were more likely to have additional

children but less likely to engage in entrepreneurial activity. This tradeoff is driven

by heightened effective risk aversion: childbirth raises subsistence needs and amplifies

sensitivity to income volatility, discouraging risky occupational choices. The effect

is stronger among households with only daughters, higher consumption requirements,

tighter financial constraints, greater income volatility, and weaker intra-household risk

sharing. Our findings highlight how household characteristics factor in the effectiveness

of pro-natalist policies.

Using a large cross-section of corporate bond returns around the world, we construct a novel global credit factor that prices international corporate bonds in both the time-series and the cross-section. We estimate the global credit factor as a function of both the VIX and U.~S. credit spreads, and show that incorporating information from nonlinearities and from interactions between the two is important for the forecasting performance of the global credit factor. In the cross-section, riskier bonds and bonds of issuers in riskier countries have a higher loading on the global credit factor. Periods of tight credit conditions correspond not only to larger probabilities of negative bond returns and higher future bond return volatility but, importantly, to a persistent deterioration in firms' credit risk and increases in credit spreads, especially for riskier firms. This suggests that increases in the global price of credit risk have real consequences for corporate bond borrowers.

We study fairness-aware credit modeling with interpretable machine learning in U.S. mortgage markets. We introduce an in-processing training objective that augments the classification error with a differentiable equalized-odds penalty, ensuring parity in error rates across protected groups. Our fairness-aware machine learning model provides inherent interpretability, allowing for the decomposition of model outputs into feature contributions. Using HMDA data, we find that the fairness-regularized model substantially narrows TPR/FPR disparities and shifts importance from proxy-like variables toward core underwriting determinants, yielding transparent, regulator-aligned decisions. The resulting models constitute realistic, interpretable less-discriminatory alternatives to standard credit scoring rules, showing that substantial reductions in disparate treatment can be achieved at minimal cost. Quasi-experimental tests around underwriting thresholds and lender-level comparisons confirm that fairness regularization lowers minority shortfalls at the margin and reduces within-lender disparities.

Open-source development is an unconventional contract where firms can attract volunteer developers as inexpensive labor. Using firm-task-developer level panel data, I show that firms attract more open-source labor when facing an adverse venture capital shock, which generates spillovers that weaken the volunteer labor market. The collapse of a crypto startup in 2022 reduced availability of VC funding for other startups. Treated firms hired 6.1% fewer employees and relied on 10.5% more contributions from non-employees. Marginal contributors were of lower quality and future employers were 9.7% less likely, relative to the unconditional mean, to hire volunteers contributing to treated startups.

This paper resolves a long-standing zero-beta rate puzzle—the empirical finding that estimated zero-beta rates remain persistently high across factor models. I show that this apparent robustness arises from pervasive model misspecification rather than reflecting a genuinely high risk-free rate. When a factor model fails to perfectly price assets, the zero-beta rate is no longer uniquely identified, and the conventional estimator—based on the minimum-variance zero-beta portfolio—converges toward the mean return of the global minimum-variance portfolio as model misspecification increases. To quantify this mechanism, I introduce a new investment-based measure of model misspecification: the maximum Sharpe ratio attainable by zero-investment, zero-beta portfolios. This measure captures the economic magnitude of pricing errors and links model misspecification to empirically observable investment opportunities. Studying a comprehensive set of classical and modern factor models, I find substantial misspecification, explaining why all models yield similarly elevated zero-beta rates. Simulation analyses confirm that realistic degrees of misspecification can fully reproduce the empirical magnitude of the puzzle even when the true risk-free rate is low.

The extant evidence is mixed on whether companies respond to tax incentives when making capital structure decisions. We argue that it is important to examine capital structure choices in the context of external financing needs, such as when funding corporate investment. We examine 99 years of panel data for US public companies and find that firms use more debt to fund investment when corporate income tax rates are high, consistent with the tax hypothesis. We study, for the first time, the two largest tax hikes of the last century (one in the 1940s, one in the 1950s). Consistent with tax incentives, we find that firms increase leverage in response, and these effects are concentrated among firms with the highest needs for investment funding. A battery of tests shows that these results are unlikely to be driven by changes in economic conditions or investment opportunities associated with the coincident war efforts.

Stock markets in emerging economies often suffer from severe illiquidity. We show that public policy can improve trading conditions by incentivizing designated market makers (DMMs). Exploiting a Chilean tax reform, we identify a causal link between DMM adoption and sharply narrower bid–ask spreads. A novel analysis of intraday spread distributions shows that half of the improvement stem from spillovers to other liquidity providers. These spillovers reflect a virtuous cycle: tighter spreads attract higher trading volume, lowering average operating costs of liquidity suppliers and enabling further spread tightening. Economies of scale in liquidity provision seem crucial in illiquid markets.

We study whether AI-extracted facial features from borrowers’ photos can serve as a scalable proxy for “soft” information missing from traditional credit models, such as conscientiousness, patience, and self-control. These traits influence financial behavior but are rarely captured in administrative data. Linking LinkedIn photos and employment and education records to voter registration and Experian data for over one million U.S. borrowers, we find that facial embeddings add significant predictive power for default risk beyond standard observables such as as credit scores, gender, and race. The incremental value is largest for younger, lower-income, and thin-file borrowers, where traditional credit scoring technology is least informative. A separate model mapping facial images to perceived Big Five personality traits reveals personality as one mechanism through which images proxy for soft information. These results suggest that facial embeddings capture stable behavioral traits absent from standard credit data, as well as perceived attributes that may influence how individuals are treated by others. While such models offer new insight into the role of personality and soft information in credit markets, their use in screening raises important concerns about fairness, privacy, and autonomy.

A key function of government-sponsored enterprises (GSEs) is insuring mortgage default risk, yet little is known about whether these guarantees are fairly priced. Using a replicating portfolio and prices of reinsured mortgage default risk to value the cash flows from these guarantees, I show GSEs shifted from providing a 20-bps subsidy pre-GFC to earning risk-adjusted profits of 30-bps post-GFC. Using a regression discontinuity design, I document that this higher pricing is passed to borrowers through higher conforming mortgage rates. Following this increase, banks reduce but still securitize most mortgages through GSEs, implying lenders value GSEs for reasons beyond subsidized funding.

Shocks to U.S. monetary policy significantly affect sovereign credit risk, and financial intermediaries play an important role in this transmission. Using market-based measures of intermediary stress and regulatory data on dealer-country-level sovereign CDS positions, we show that a decline in risk-taking capacity by intermediaries prior to an FOMC announcement significantly amplifies the sensitivity of sovereign spreads to U.S. monetary policy shocks. To explain these findings, we develop a general equilibrium asset pricing model of the world economy which features occasionally-binding intermediary borrowing constraints that amplify movements in sovereign default risk and risk premia. Spillovers from U.S. intermediary financing conditions to foreign output are crucial to match the sign and magnitude of monetary policy effects observed in the data.

This paper examines how learning and competition shape brokerage commission fees in a plain-vanilla Chinese setting. We document three facts: substantial price dispersion across investors, higher fees for naïve investors, and widening dispersion over time. To explain these patterns, we develop a dynamic model in which brokers compete for naïve and sophisticated clients while learning types through repeated interactions. The model shows that an incumbent’s information advantage can sustain persistent market power—even over sophisticated investors—and that additional broker entry can backfire by raising markups on naïve investors and widening dispersion. Exploiting branch-level variation in rival presence, we provide causal evidence consistent with these predictions. Quantitatively, although the average number of brokers per city more than doubled over the sample, intensified competition slightly offset—rather than accelerated—the secular decline in fees.

Blockchain technology holds the promise of transforming our financial system but a key question lingers regarding whether this technology can ensure secure settlement. We develop an equilibrium model to study that question with regard to the most prominent blockchain type, a Proof-of-Stake (PoS) blockchain. We demonstrate that blockchain security increases with the productivity of the blockchain. Moreover, we show that there exist reasonable conditions under which a PoS blockchain is secure against arbitrarily large incentives to disrupt settlement.

We embed the budget constraints of the private, public, and external sectors within the aggregate budget constraint of the economy to examine whether valuation ratios in one sector forecast real returns and cash-flow growth in others. Exploiting the cross-sector restrictions implied by the aggregate constraint, we show that fluctuations in the government surplus-to-debt ratio robustly predict equity returns. The magnitude of this cross-sector predictability is on par with the own-sector predictability associated with the dividend–price ratio. We then develop a model in which distortionary taxes generate these patterns and use the cross-sector forecasts to calibrate the implied size of the tax distortions.

Textbook corporate-finance theory assumes that managers maximize the NPV (net present value) of expected future equity payouts. However, in practice, the people running large public companies often seem more concerned with increasing EPS (earnings per share). Perhaps this is a mistake. Or maybe EPS growth is a good second-best proxy for value creation. Whatever the reason, we show that the simplest possible EPS-maximizing model predicts important financing decisions, such as leverage, new issuance, share repurchases, and cash holdings. The principle of EPS maximization leads to a novel microfoundation for value and growth. Managers with an earnings yield above the riskfree rate view equity as expensive (value stocks) and adopt one set of EPS-maximizing policies. Those below this threshold see equity as cheap (growth stocks) and take a different approach to maximizing EPS. We examine the data and find strong empirical support for our model’s key predictions.

We use detailed data on fixed-income benchmark indexes in Canada and the United States to provide systematic evidence of how benchmarking shapes corporate bond ownership and prices. Funds hold sparse portfolios, and index weights strongly influence which bonds active and passive funds select. We rationalize these patterns in a model with benchmarked managers who face portfolio management costs, which predicts which assets managers optimally include in their portfolios. In the model, a bond's price increases with its benchmarking intensity (BMI)—a measure of the amount of fund capital benchmarked against the bond—while portfolio sparsity attenuates this price impact for excluded bonds. Exploiting discontinuities in benchmarked assets around bond maturity cutoffs, we show that increases in bonds’ BMIs lead to reductions in yield spreads and increases in fund ownership—but only for bonds predicted to enter sparse portfolios.

Broker-dealers do not participate directly in equity markets in large quantities; instead, they participate indirectly by lending to hedge funds via their prime brokerage divisions. We show that shocks to broker-dealer financial health affect their credit supply; however, hedge funds are typically able to diversify away these shocks. This is consistent with a high ability to substitute borrowing to non-distressed broker-dealers. This ability is not unlimited: when the shock to broker-dealer health is sufficiently broad and spills over to non-affected broker-dealers, it triggers hedge fund equity sell-offs. We show that such a broad broker-dealer shock occurred in the first quarter of 2016 when several European broker-dealers became distressed. These sales affected equity prices, generating abnormally low returns for the four months immediately proceeding the shock. Moreover, we estimate a price multiplier of (at least) three, meaning that if hedge funds, in net, sold off 1\% of the total shares outstanding, the stock price fell by (at least) 3\%. Overall, our results indicate that broker-dealer health matters for equity prices under conditions of broad distress.

We use establishment-level data to examine the relation between corporate taxes and default risk. Using a border discontinuity design, we document that higher corporate tax rates significantly increase default risk, particularly for geographically concentrated firms, with sizable spillovers across establishments. We further investigate \textit{how} taxes affect establishment-level default risk by exploiting the interest limitation rule introduced by the 2017 Tax Cuts and Jobs Act. Using a difference-in-differences design, we find that affected firms experience a decline in default risk. This occurs because the reduced tax advantage of debt induces firms to deleverage, thus reducing default risk through capital structure adjustments.

This paper develops a new method to extract exchange rate expectations from investment positions. I use relative allocations between otherwise identical exchange-traded funds (ETFs) offered with and without a currency hedge to measure investors’ pure currency demand and infer a distribution of currency return expectations. These portfolioimplied expectations predict future exchange rates more accurately than survey-based expectations or expectations derived from macroeconomic models or currency pricing factors. Dispersion in portfolio-implied expectations accounts for 27% of exchange rate volatility, consistent with models of heterogeneous beliefs.

We study the role of institutional investors’ subjective risk premia in explaining vari-

ation in their subjective expected returns (both over time and across investors). Our

analysis uses long-term Capital Market Assumptions from asset managers and invest-

ment consultants from 1987 to 2022. Perceived market risk premia account for most of

the countercyclicality and overall time variation in subjective expected returns, with

the remainder driven by alphas (perceived mispricing). The risk premia effect stems

almost entirely from time variation in perceived risk quantities rather than risk price

(risk aversion). Additionally, market risk premia explain most of the expected return

disagreement, but here alphas play a significant role, and risk price and risk quantities

contribute roughly equally to the risk premia effect. These results provide benchmark

moments that asset pricing models should match to be consistent with institutional

investors’ beliefs.

This paper shows that the contractual features governing bank lending to nonbank lenders (NBLs) are a key source of financial stability. I document that credit lines account for 90% of bank funding to NBLs in the syndicated loan market. NBLs use credit lines to manage investment and liquidity shocks, while banks’ liquidity advantage makes them natural insurers. To study the financial stability and welfare implications of this arrangement, I develop a quantitative model with endogenous credit limits and fees between banks and NBLs. Credit lines generate insurance revenues for banks that fund loan origination, reinforcing deposit creation. Credit lines’ contingent features make them cheaper but riskier than cash, and safer but more costly than loans. Quantitatively, credit lines raise welfare by 0.02% relative to cash and 1.83% relative to loans. Yet, deposit insurance induces banks to extend limits beyond the social optimum, making moderate tightening of capital and off-balance-sheet requirements welfare-improving. In contrast, partial guarantees to NBL debt weaken banks’ relative liquidity advantage, constrain credit line supply, and reduce welfare, underscoring the importance of bank credit lines to NBLs for financial stability.

We propose a new methodology to estimate the equity term structure. Instead of using

realized returns of dividend assets, we generalize the implied cost of capital approach and imply the term structure of ex-ante expected returns from the cross-section of observed stock prices and projected firm-level cash flows. Using US data for 1980-2024, we find an unconditionally upward sloping term structure of risk premia with rich cross-sectional patterns in the size, value and credit risk dimensions. Strikingly, value firms and speculative-grade firms have flat or even downward-sloping term structures. We also detect that the term structure flattens out in recessions.

We explore how biased investors affect the market for real assets and estimate the resulting efficiency losses. Investors subject to non-proportional thinking ask (too) high merger premia to sell low-price targets and offer (too) low merger premia to buy high-price targets. As a result, M&A premia are lower for high-price targets and both low- and high-price firms are less likely to be acquired than firms in the middle of the price distribution. We test these predictions using a large sample of M&A transactions. We also quantify the value lost because positive-synergy deals do not happen due to non-proportional thinking. Our structural estimation suggests that investors' mistakes reduce the frequency of M&A transactions by about 8% and the value created by the M&A market by about 6%.

The NYSE Trade and Quote (TAQ) dataset, used throughout finance research and securities regulation, is generated by a Securities Information Processor (SIP), which aggregates quotes and trades from all U.S. stock exchanges at a central location. We show that the SIP systematically reports events out of sequential order: quote changes that occur after a trade are frequently reported as occurring before the trade. The source of the issue is that trades and quotes are recorded with variable latency (due to, e.g., geography) and are extremely clustered in time. The result is a look-ahead bias: the prevailing SIP quotes, ubiquitous in signing trades and measuring spreads, incorporate price impact from the trade itself. We document that the look-ahead bias leads to incorrect trade signing and downward-biased effective spreads and price impact. The errors are extreme for the large fraction of trades with high reporting latency: approximately 20% of trades are incorrectly signed, and effective spreads and price impact are understated by more than 40%. We propose a signing methodology based on exchange rules that yields 100% accuracy for a significant majority of volume and, over all trades, outperforms existing methodologies.

This paper examines the role of repo lending between counterparties affiliated with the same bank holding company (BHC). Using confidential transaction-level data, we find that Treasury repo rates between affiliated entities are significantly higher than those between unaffiliated parties. This affiliation premium is more pronounced when dealers face tighter balance sheet constraints, suggesting that regulatory capital requirements raise the cost of external borrowing and enhance the value of internal funding. During a temporary period of regulatory relief when leverage requirements were relaxed, the affiliation premium nearly disappeared, only to re-emerge once the regulations were reinstated. Our findings underscore a unique competitive advantage for dealers within large BHCs: the ability to access internal liquidity from affiliated banks. This internal liquidity channel also facilitates the distribution of liquidity from banks with high level of reserves to the rest of financial system.

We document a chilling effect of climate regulations on cross-border corporate investment. Specifically, countries enacting climate laws subsequently experience significant declines in in-bound mergers and acquisitions by foreign firms. Completed transactions after law enactment feature lower premiums and less post-merger performance improvement, and deals initiated pre-enactment are more likely to be withdrawn. The chilling effect is amplified when acquirers are from countries less concerned about climate change and when climate regulations provide no subsidies. Key mechanisms underlying the documented effect include economic burdens and geopolitically-driven selective enforcement. Our findings suggest that climate laws can disrupt global capital and resource reallocation and highlight a hitherto overlooked cost of climate regulations.

Deviations from covered interest rate parity (CIP) have been attributed to limits to arbitrage, but trading volumes surge during periods of no-arbitrage violations. We show that such distortions are caused by constraints on non-U.S. agents’ access to wholesale U.S. dollar markets, and thus reflect a premium for obtaining unencumbered synthetic dollar funding. Three main findings emerge. First, non-U.S. banks substitute secured USD borrowing with foreign exchange swaps to comply with regulatory requirements. Second, the cross-currency basis reflects the shadow costs imposed on non-U.S. agents for borrowing in U.S. wholesale segments. Third, U.S. dealers extract rents on dollar provision; the costs are ultimately passed on to the real economy, with non-U.S. customers paying an additional $10.4 billion in hedging costs. Our study highlights how intermediary constraints impact interconnected money markets and drive segmentation in global dollar funding.

I study how equity analysts form subjective beliefs about firms' earnings using their own written text from over 1.1 million equity research reports. Using large language models, I identify the topics discussed by analysts and represent topic-level information using textual embeddings. I introduce a novel text-instrumented Coibion-Gorodnichenko regression to examine analysts' misreactions to specific information. Using this new procedure, I find pervasive underreaction in short-term earnings forecasts across topics, whereas overreaction in long-term forecasts is concentrated in qualitative, intangible topics rather than quantitative, statistical ones. Revisions driven by qualitative information in long-term earnings forecasts strongly predict future stock returns. Finally, I use textual data to investigate the mechanisms underlying the documented misreactions. The empirical results suggest that overconfidence is an important driver of the overreaction to qualitative information, while herding appears to be important in explaining the overall underreaction seen in short-term forecasts.

We study competition and pass-through in the market for retail deposits. Federal funds rate increases are associated with only minimal increases in deposit rates. Leveraging new data on offers mailed by banks to households, we show that federal funds rate increases are strongly associated with changes in mail volumes, sign-up bonuses, and offers targeting new customers. These margins and not deposit rates are the primary ways in which interest rate changes affect the deposit market. We rationalize the use of these margins and not deposit rates in a simple model with active and sleepy depositors. Our model implies that the marginal value of a new depositor is insensitive to interest rates and that depositor heterogeneity and adverse selection, as opposed to market power, are the primary reasons why banks offer far less than the full present value of future rate spreads to depositors.

This study documents a novel finding that fundamental volatility measured from financial statement variables negatively predicts stock returns, an effect that is strikingly pervasive (over a large number of accounting variables) and persistent (predicting returns over at least two years). A long-short portfolio based on fundamental volatility generates a significantly positive monthly return of 0.93% and significant alphas under a variety of factor models including the Fama-French five-factor model and the q-factor model. We propose a production-based asset pricing explanation to this high-volatility/low-return phenomenon. In the model, due to diminishing marginal return to inputs, firms with higher operating volatility have lower expected profitability and lower expected returns. Further empirical analyses confirm two key predictions of the model. First, the effect of fundamental volatility on stock return is mainly through the profitability channel. Second, the negative effect of fundamental volatility on both profitability and return is stronger for firms facing more severe diminishing marginal returns in their operations.

Net interest margin (NIM) measures net interest income per dollar of assets and remains remarkably stable for most US banks despite large fluctuations in the federal funds rate. I show that an important reason for this stability is changes in credit spreads on new assets. To derive the credit spreads, I construct a bank-specific risk-free benchmark rate for assets and infer the implied spread as the difference between the interest income rate and this benchmark. For most banks, the risk-free benchmark responds more to policy rate changes than interest expenses. Consequently, without changes in credit spreads, a 100 bp decline in the policy rate would compress average NIM by about 28 bp and expose banks to interest rate risk. NIM stability implies that most banks add new assets to their balance sheets at higher credit spreads when the policy rate falls and at lower spreads when it rises. Increases in implied credit spreads are associated with greater exposure to credit and duration risk, consistent with banks reaching for yield. The results provide a novel asset-side explanation of how NIM stability is achieved.

We construct the first nanosecond-level European Best Bid and Offer (EBBO) by consolidating direct-feed quotes from ten major venues for Eurostoxx50 constituents between 2020 and 2024. The EBBO reveals that consolidation halves quoted spreads and nearly triples displayed depth, yet only 79–86% of marketable orders execute at the best available price. Geographic latency, routing frictions, and incomplete connectivity generate welfare losses of more than €50 million per year. These results quantify some of the costs of fragmentation and demonstrate how a consolidated European benchmark could enhance execution quality, transparency, and competition—directly informing the design of the Markets in Financial Instruments Directive III’s (MiFID III) Consolidated Tape.

We develop CovenantAI, an artificial intelligence-powered covenant monitoring methodology, to examine whether institutional investors strategically trade around covenant violations in leveraged loan markets. Documenting a persistent decline in loan prices during the 100 days preceding violations—with pronounced drops 20 days prior—we find cumulative abnormal returns of -0.84% during the [-20,-1] event window. Price effects are most severe for loans amended post-violation or remaining in technical default. Covenant violations significantly increase downgrade and bankruptcy probabilities, particularly among non-investment-grade loans held by Collateralized Loan Obligations (CLOs). We document substantial cross-sectional heterogeneity in CLO constraints driven by overcollateralization ratios and CCC-rated loan holdings. Loans predominantly owned by constrained CLOs exhibit steeper pre-violation price declines and significantly more negative abnormal returns. Our evidence demonstrates that constrained institutional investors preemptively divest loan positions in anticipation of covenant violations, with trading intensity reflecting both violation severity and investor-specific portfolio constraints.

Do tax subsidies prompt investors to take on risk? We address this question within a framework in which venture capitalists (VCs) combine outside funding with incentive-based compensation and examine a policy that cut capital gains taxes on startup investments. Using bunching, regression discontinuity, triple-differences, and matching designs that exploit industry eligibility, investment vintage, and holding-period requirements, our study analyzes data from 158 thousand investor-firm pairings over two decades. We first identify strategic investment timing, with subsidies prompting clustering at tax-eligible holding-period thresholds. We then document strategic capital allocation, with firms just below eligibility thresholds receiving more funding than those just above. Most notably, when and where tax subsidies apply, VCs shift their project selection toward riskier ventures: they invest more into pre-commercial stage startups, become more likely to provide startups with their initial capital, become more likely to invest in industries in which they have no prior experience, and more in firms with pre-existing debt, while becoming less likely to co-syndicate their investments. These portfolio firms eventually show higher failure rates. On the flip side, the increased risk-taking yields salient return outcomes: tax-subsidized VC-backed ventures attain higher valuations at exit and are more likely to reach "unicorn status." None of these patterns are observed for comparable non-VC investors receiving the same tax subsidies. Our study is the first to show that tax policy can shift entrepreneurial financing toward riskier, more experimental, valuable ventures, with outcomes shaped by investor organizational structure and incentives.

The relative restrictiveness of a central bank’s supply of money predicts the raw and risk-adjusted returns of its currency—both next month and at least three years into the future. Archived data, known by currency traders at the time, estimates central bank restrictiveness as a scaling of the residual from out-of-sample panel regressions of M1 on macroeconomic variables tied to domestic and international transaction requirements. Carry’s ability to forecast currency returns is subsumed by the central bank restrictiveness signal, which also forecasts inflation.

AI agents can emulate the beliefs of human survey respondents. We leverage this idea at scale to examine how investor disagreement emerges in response to firm news and to measure such disagreement at high frequency. We endow a local large language model (Llama 3) with over 200 demographically representative investor personas and elicit their sentiment toward S&P 500 firm-specific news headlines from 2010–2025. LLM personas disagree in economically meaningful ways: disagreement is largest for social and governance-related news, and smallest for hard news tied to firm fundamentals. The measure aligns with human survey evidence and traditional uncertainty measures, yet it is distinct from social-media disagreement. Turning to market outcomes, LLM-derived disagreement is strongly associated with same-day and next-day abnormal trading volume. Our main findings are stable across the model’s pre- and post-training windows.

We propose a novel mechanism for why bank portfolios are opaque: banks main-

tain opacity to boost profits, even at depositors’ expense. We argue that banks choose

opacity to secure cheaper long-term funding while balancing insolvency and illiquidity;

and this trade-off determines the portfolio’s optimal opacity. We show that opacity re-

duces deposit rates but, by forcing depositors to rely on noisy solvency signals and act

cautiously, raises the likelihood of early withdrawals when interest rates are high. Con-

sequently, in high-rate environments, banks strategically tilt toward excessive opacity

to further cut deposit rates, at the cost of more frequent early failure.

Machine learning (ML) methods are highly flexible, but their ability to approximate the true data-generating process is fundamentally constrained by finite samples. We characterize a universal lower bound, the Limits-to-Learning Gap (LLG), quantifying the unavoidable discrepancy between a model’s empirical fit and the population benchmark. Recovering the true population R2 , therefore, requires correcting observed predictive performance by this bound. Using a broad set of variables, including excess returns, yields, credit spreads, and valuation ratios, we find that the implied LLGs are large. This indicates that standard ML approaches can substantially understate true predictability in financial data. We also derive LLG-based refinements to the classic Hansen and Jagannathan (1991) bounds, analyze implications for parameter learning in general-equilibrium settings, and show that the LLG provides a natural mechanism for generating excess volatility.

We reconstruct the debt ownership structure of 10,136 firms in 52 countries over 2002-2021. Contrary to prior literature, debt ownership is most concentrated in civil-law countries and most dispersed in common-law countries. This result is driven by junior arm’s-length debt dispersion. Where investor protection is stronger, dispersed arm’s-length debt coexists with concentrated bank debt. In weaker investor protection environments firms rely more on short-maturity and USD-denominated borrowing. These patterns are most pronounced among small-to-medium firms, whereas the largest exhibit dispersed debt ownership by international investors almost irrespective of country of incorporation. Our results support legal origin and financial contracting theories.

For bilateral bargaining under complete information in network markets, we show that holdup is eliminated when contracts are atomically settled with smart contracts. Applications include over-the-counter trading networks, purchases requiring third-party financing, and other multi-agent markets in which pairs of agents bargain bilaterally over terms that could potentially affect their outside-option values in negotiations with other firms. Under this protocol, the terms of a contract proposed to any firm can be given a “greenlight” by that firm, which automatically converts those terms into a binding contract if and only if the terms proposed to all other firms are also assigned greenlights and thus likewise converted to contracts. We analyze tree-based networks, including hub-spoke and chain networks. Under mild technical conditions, the unique outcome of any Perfect Bayesian Equilibrium is immediate agreement on socially efficient contracts that generate equal expected gains for every firm.

Institutional investors commit trillions of dollars to private funds. These commitments give fund managers discretion to call capital on short notice, effectively making investors their liquidity providers. Using novel data on insurers’ $370 billion private fund investments, this paper studies the risk of unexpected capital calls. Specifically, I examine the portfolio implications of capital call shocks and the resulting spillovers to public asset markets. I show that capital calls are difficult to predict and that unexpected calls are substantial. Nevertheless, I find no evidence that insurers build liquidity buffers ex ante. Instead, they adjust their portfolios only ex post, primarily by selling risky corporate bonds. These portfolio decisions are driven by regulatory capital considerations. Moreover, capital-call-induced corporate bond sales cause negative price impacts, especially for bonds with high risk weights. These spillover effects are amplified when capital call shocks are concentrated or coincide with other episodes of market stress. Counterfactual stress tests reveal significant aggregate losses under extreme scenarios. Overall, the findings highlight the liquidity risk embedded in private fund commitments and its implications for financial fragility.

We study how carbon prices affect labor market outcomes by exploiting a policy change in the EU Emissions Trading System that led to a sharp rise in permit prices. Using population-wide employer-employee matched data from the Netherlands and a matched difference-in-differences design, we find no adverse aggregate effects on employment or wages. However, the distributional effects are sizable: workers in firms with large permit surpluses experience wage gains, as do STEM-educated workers---especially those with stronger outside options. Plants employing more STEM workers achieve larger reductions in energy costs, highlighting the role of skills in facilitating the transition to low-carbon technologies. Our results illustrate that distributional effects of carbon pricing depend on market design and worker skills.

This paper documents sharp segmentation in cross-border financial operations across firms, with those in the top 1st percentile of the size distribution exhibiting a different pattern of internationalization than all others. Using administrative micro-level data covering the quasi-universe of France’s external positions from 2014 to 2018, we show that granular firms display no gravity, currency, or institutional biases and simultaneously engage in foreign investment and foreign funding, often with the same country. In contrast, bottom 99th firms exhibit pronounced biases and internationalize almost exclusively by raising equity capital from a single foreign country while holding no foreign assets. Large fixed costs of financial internationalization—substantially higher for foreign investing than for foreign borrowing—rationalize this segmentation and affect exchange rates. We show that exogenously identified idiosyncratic shocks to granular agents trigger capital flows that affect bilateral exchange rate with destination countries.

I study the relationship between sell-side analysts’ forecast bias and stock returns by comparing three forecast families—price targets (PTG), earnings per share (EPS), and long-term growth (LTG)—to ex-ante machine-learning (ML) benchmarks. Bias in return expectations co-moves positively with bias in cash-flow expectations, suggesting a common behavioral source of belief distortions. Analysts’ deviations from ML benchmarks across all forecast families predict returns in a systematic and nonlinear manner. Across horizons, forecast bias predicts returns positively in the short term but negatively in the long term, with LTG-based return-predictability regressions exhibiting similar patterns. Across the sign of bias, optimistic (pessimistic) forecast bias positively (negatively) predict returns. Finally, I document a common tendency to anchor expectations on consensus forecasts. An asset-pricing model with asymmetric information and positively skewed fundamental shocks generates these dynamics: analysts anchor on consensus and misperceive the informativeness of signals, producing the observed bias–return relation.

Firms facing serious litigation can use bankruptcy as a negotiating tactic. We develop a dynamic capital structure model in which firms choose negligence levels that boost short-run profits but create litigation risk. The model reveals substantial risk-shifting: highly leveraged firms engage in excessive negligence because limited liability allows them to shift expected litigation costs to creditors and tort victims through bankruptcy. After estimating the model, we evaluate four policy regimes corresponding to recent legal debates. Legalizing Texas Two-Step bankruptcies, which shield firms from litigation with minimal consequences, would produce catastrophic increases in negligence and harm victims despite eliminating bankruptcy costs. Protecting tort victims in bankruptcy with either superpriority or enhanced bargaining rights improves outcomes modestly. Comprehensive reform combining both of these protections dominates: negligence declines, victim recovery triples, and firm equity values increase, as both productive investment and improved deterrence benefit shareholders.

We derive new entropy and moment bounds for the stochastic discount factor (SDF). Our results generalize existing bounds which exploit risk-adjusted measures of investment opportunities---such as Sharpe ratios or expected log returns---that are maximized in the cross-section, across assets. By contrast, we can fix a single asset and optimally exploit information in its true and risk-neutral return distributions. Applying the framework to the S&P 500 index, we find that the $\theta$th SDF moment grows extremely rapidly when $\theta > 1$, and appears to diverge to infinity before $\theta=2$. But entropy measures and the $\theta$th moments with $\theta \in (0,1)$ are well-behaved theoretically and empirically, and can be related to measures of market risk aversion and of the attractiveness of investment opportunities.

Valuation models lie at the core of both financial theory and practice, yet we lack systematic evidence on how professionals value assets, which models perform best, and why. To make progress on these questions, we analyze valuation models in 1.1 million equity analyst reports. While, on average, simpler multiples-based models generate more accurate forecasts than more complex discounted cash flow (DCF) models, this masks important heterogeneity: skilled analysts produce superior forecasts with DCF models, especially for hard-to-value firms, underscoring the importance of expertise when employing complex models. To establish that model-specific expertise matters, we exploit a quasi-exogenous shock that forced some analysts to switch valuation models, and show that their forecast accuracy subsequently declines relative to analysts with established experience using the new approach. This highlights a fundamental trade-off between simplicity and sophistication in valuation, where optimal method choice depends on analyst characteristics, such as skill. Finally, given their unconditional superior performance, we study how analysts determine multiples. Analysts use historical, current, and peer-based reference points to contextualize their choice of multiples, but they do not use these benchmarks mechanically when determining their prices. Moreover, we show that sensitivity analyses have become increasingly common, bull and bear scenarios are asymmetric and account for greater downside risk, and their inclusion is associated with more conservative forecasts.

Prior studies of retail trading costs focus on stock market orders, largely overlooking the role of limit orders and the boom in retail options trading. We show that these underexplored factors are critical for assessing overall retail trading costs. First, nonmarketable limit orders account for 17% of stock volume but more than 34% of option volume in Form 606 reports by retail brokers. Second, using trader-level data on 451,299 option and 607,922 stock retail trades, we estimate effective spreads using actual trade direction to average 1.07% for options, much higher than 0.07% for stock trades. But these realized option costs are 60% lower than the 2.59% conventional effective spreads because retail investors commonly use limit orders to reduce trading costs. These patterns hold across trader styles. Finally, individual trader order choices rather than broker defaults drive execution quality, with trader performance persisting over time.

Increasing climate risk is making property insurance unaffordable and unavailable. I study a novel Australian policy response: government-provided, mandatory, risk-based reinsurance for cyclone damage in home insurance. Public reinsurance reduces premiums by 21\% and increases insurance availability by 11\%. These gains are not a subsidy but arise from eliminating large pre-existing markups in private reinsurance and catastrophe bond markets, flowing primarily to insurers most constrained by tail-risk exposure. The markup reduction stems from neutralizing the high premium for spatially-correlated and ambiguous risk, with increased competition providing additional benefits. This demonstrates that insurance market dysfunction originates from frictions in tail-risk reinsurance markets, and that targeted, cost-neutral interventions in these upstream markets can restore affordability and availability in home insurance.

Do creditor rights amplify or mitigate the macroeconomic consequences of credit cycles? Using a panel of 39 countries from 1978 to 2019, we show that credit expansions in economies with strong creditor protection are followed by smaller output losses, lower stocks of non-performing loans, and a greater reallocation of credit away from risky borrowers. Firm-level evidence from the staggered adoption of antirecharacterization laws across U.S. states shows that well-protected creditors cut credit to risky firms with poor growth prospects, while easing credit constraints for productive firms. Our findings suggest that stronger creditor rights can enhance macroeconomic stability by facilitating a more efficient reallocation of capital.

Using contract-level data on U.S. bond funds’ currency forward positions from 2010–2023, we document that foreign investors dedicated to emerging markets (EM) bear substantially greater currency risk than their bond holdings suggest. On average, funds amplify their EM currency exposure by 14% through net long position in forwards linked to bond positions, and even more when including forwards without corresponding bond investments. This forward usage pattern is strongly related to the degree of capital control imposed by currency issuers, underpinned by a triangular relationship between capital flow restrictions, funds’ bond portfolio deviations from local currency benchmark weights, and net forward purchases. Meanwhile, for funds that sell local currencies forward overall, bond portfolio weights strongly predict forward sales, indicating an inelastic hedging demand for currency forwards. Informed by the empirical analysis, an equilibrium model featuring investor heterogeneity, capital flow restrictions and forward market segmentation is able to rationalize empirical properties of the forward premia of EM currencies.

The U.S. stock market return during the first month of a quarter positively predicts the second month’s return, which in turn negatively predicts the first month’s return of the next quarter. This pattern arises because investors fail to fully recognize that earnings announced in the second month of a quarter are inherently similar to those announced in the first month, leading them to overreact to predictably repetitive earnings news. A model formalizing this form of correlation neglect yields additional predictions for survey data and for both the time-series and cross-section of returns, all of which are borne out in the data. These results provide evidence of correlation neglect even among sophisticated, financially incentivized decision-makers, underscoring its importance as a behavioral phenomenon.

The market value of deposit insurance changes over time and across banks as the value of the underlying put option changes, but the premium banks pay for the insurance does not adjust to completely capture this variation. Consequently, their incentive to supply insured deposits changes with the change in subsidy they enjoy from deposit insurance. Factors that increase the subsidy, such as asset risk, move the supply curve outward. Consistent with this idea, we show that the supply of insured deposits increase when banks become riskier. Our findings uncover a novel channel of deposit supply, with implications for existing research on the monetary policy, deposits channel, and reaching-for-yield literature.