Publication & Under Revision:
Investing in Lending Technology: IT Spending in Banking (with Zhiguo He, Sheila Jiang and Xiao Yin), Management Science, 2025
Bank technology spending is not monolithic: this paper separates it into communication capital (soft-information gathering) and software capital (hard-information processing), by constructing a novel dataset spanning the U.S. banking system. Regionally differentiated policy shocks reveal the two are demand-driven substitutes — communication IT tracks local small-business credit demand, while software IT responds to refinancing demand and fintech competition. The result reframes bank IT investment as a technology-choice problem between relationship-based and algorithmic screening, a tradeoff central to how any intermediary allocates capital between human judgment and automation.
Rise of Factor Investing: Asset Prices, Information Efficiency, and Security Design (with Lin Will Cong and Shiyang Huang), Conditionally accepted, Journal of Finance, 2026
This paper develops a general-equilibrium framework of composite security design, in which competitive intermediaries assemble liquid baskets of assets spanning the economy's priced risk factors. As designers create better factor-mimicking securities, factor investors are pulled out of individual-asset markets and into these baskets — an endogeneous market structure shift with sharp, testable price implications. Security-level data confirm the model's predictions: assets bundled into more baskets show higher informational efficiency, greater price variability, and stronger factor co-movement.
Monitoring with Small Stakes (with Sheila Jiang and Shohini Kundu), Revise and resubmit, Management Science, 2025
Bank lenders' share in leveraged loans (skin in the game) has shrunk for over a decade as institutional investors take on a growing share of leveraged loan syndications — yet banks keep monitoring diligently, a puzzle this paper resolves by identifying rent extraction through renegotiation as a second incentive that substitutes for direct retention. The mechanism is formalized through split-control loan structures, where banks retain covenant-heavy debt and renegotiation rights while institutions hold covenant-lite term debt, trading loan share against renegotiation power in the optimal contract. Exploiting a tax policy that lowered creditors' renegotiation costs, the paper finds easier renegotiation causally sharpens monitoring and improves borrower performance, concentrated in exactly these split-control deals.
Working Papers:
Tech-Driven Intermediation in the Originate-to-Distribute Model (with Zhiguo He and Sheila Jiang), 2026
This paper develops a general-equilibrium model of technology-driven intermediation in originate-to-distribute markets. When intermediaries possess sufficiently accurate information, illiquidity-induced retention incentivizes them selectively purchase high-signal assets, endogenously shutting down the “lemon” market and solving the classic “monitoring the monitor” problem. By contrast, poorly informed intermediaries merely add balance-sheet capacity while weakening market discipline, potentially reducing welfare when adverse selection is severe. Thus, the social value of intermediation depends critically on information technology, not simply on intermediaries’ capacity to absorb assets.
Bank Technology Adoption and Loan Production in the U.S. Mortgage Market (with Sheila Jiang and Adam Jorring), 2025
This paper models mortgage loan market as a screening problem under borrower-side adverse selection, under which lenders' investment in screening technology, optimal credit approval policy and risk-based pricing menu is derived. We assemble a novel loan-level dataset that covers the trajectory of mortgages from application to repayment and detailed information about IT investment by lenders. Technology investment is instrumented using lenders' house-price exposure and physical distance to key suppliers, isolating adoption's causal effect from lenders' own selection into investment. Better screening technology widens credit approval and re-prices the risk menu more finely, lowering two-year delinquency — with the largest gains for marginal borrowers hardest to screen traditionally.
Optimal Banking System for Private Money Creation, 2021
This paper studies how a private banking system should be designed to create liquid, safe deposits efficiently when markets and contracts are incomplete. A Salop-style spatial-competition model uncovers three distinct market failures: an incentive problem that under-produces a shared liquidity pool ex ante, a commitment problem that over-uses it ex post, and a coordination failure that can halt safe-asset creation once markets grow sufficiently competitive — with competition itself cutting both ways, improving incentives while worsening commitment.
Covenant Amendment Fee and Value of Creditor Intervention after Covenant Violations (with Sheila Jiang), 2020
This paper studies covenant amendment fees — discretionary charges levied by whichever party holds bargaining power when a covenant is violated and the loan is renegotiated. Using a LLM-extracted dataset of renegotiation outcomes built directly from the text of credit agreements, the paper studies the economics of this fee and shows how it reflects the value of creditor intervention. Exploiting exogenous variation in the fee charged identifies the causal value creditors add by exercising intervention rights rather than renegotiating passively.
Rise of Domestic Banks in EME Cross-border Credit Intermediation (with Sheila Jiang), 2021
Before the 1990s, foreign banks channeled over 90% of cross-border credit to emerging markets directly; by 2010, domestic banks handled more than half of it instead — a structural shift this paper traces to the transformation of the U.S. money market. Detailed cross-border loan data reveal why the shift matters: foreign lenders favor transparency-based, covenant-driven credit, while domestic lenders rely on tangible collateral, so the identity of the intermediary now shapes who gets capital and on what terms. A cross-country extension shows this micro-level difference scales up, reshaping recipient economies' industry structure and heightening their exposure to global financing cycles.
Domestic Bank Channeled Foreign Credit– A Blessing or a Curse: Evidence from China (with Sheila Jiang), 2021
Is domestic-bank-intermediated foreign credit actually good for the real economy? Exploiting cross-region heterogeneity in domestic global banks' presence across Chinese cities during the 2003–2009 global financing cycle, this paper finds regions receiving more of it experienced significantly more volatile real outcomes. Firm-level data expose the mechanism: easier credit access led firms to over-accumulate tangible, collateralizable assets during the easing phase, setting up a sharper collateral-value collapse — and a deeper downturn — once global conditions tightened. The verdict is a genuine mixed blessing: more efficient credit allocation in good times, bought at the price of amplified fragility in bad ones.