Last Draft: December 4, 2024

Dealers intermediate trades in OTC markets through trading networks. In the municipal bond market, we document greater complexity than the typical core-periphery structure. Analyzing dealer reciprocity-the tendency to repay favors-we find reciprocity generally reduces markups. Dealers trade lower markups today for future liquidity. However, in small trading communities, reciprocity can foster collusion via quid-pro-quo agreements, inflating transaction chain markups. Among high-centrality dealers in large communities, high reciprocity lowers average markups by 80 basis points, while among low-centrality dealers in small communities, it raises markups by 72 basis points. Although only around 2% of transaction chains suggest collusive behavior, these significantly affect regression results, highlighting the importance of controlling for such outliers to accurately estimate centrality premiums or reciprocity discounts.

Presentations: 

2025: Virtual Municipal Finance Workshop, 14th Annual Municipal Finance Conference at Brookings, Financial Management Association Annual Meeting, German Finance Association Annual Meeting, Southern Finance Association Annual Meeting 

2026: Midwest Finance Association Annual Meeting, Eastern Finance Association Annual Meeting, Financial Management Association European Conference

Media Footprint: 

https://www.bondbuyer.com/news/dealer-quid-pro-quo-shapes-bond-price-markups-paper-says 

Awards and Nominations: 

Best Microstructure Paper (Semifinalist) - 2025 Financial Management Association Annual Meeting 

Table below: 

Average transaction markups, transaction chain length, and network size based on the chain-initiating dealer's degree of centrality and reciprocity

Last Draft: December 10, 2025

This paper examines how shifts in banks’ exposure to repo markets and the expansion of deposit franchises have influenced their financial performance and balance sheet strength before, during, and after the Global Financial Crisis. Using linear models, vector autoregressions, and a theoretical framework, I show how the monetary policy transformations shifted banks' equilibrium behavior based on risk preferences, collateral constraints, and the cost of funding. The lending and financing strategies diverged before the crisis, with risk-seeking banks prioritizing the maximization of their interest income and risk-averse banks benefiting from favorable repo terms by maintaining strong balance sheets. After the crisis, strategies converge as risk-seeking banks shift toward safer short-term funding, while risk-averse banks extend more credit amid diminishing repo market advantages, both of which are connected to the post-crisis policy environment.

Presentations: 

2025: University of Georgia Brownbag Seminar

2026: Southwestern Finance Association Annual Meeting, Case Western Reserve University Brownbag Seminar

Figure below: 

Effects of changes in repo dependency on the long-term net interest margins for risk-seeking banks

Last Draft: April 1, 2026

This paper exploits the tonal and topical structure of press conferences following Federal Open Market Committee (FOMC) meetings to train large language models (LLMs) that generate realistic responses to journalists’ questions by Federal Reserve Chairs. Using both actual and LLM-generated responses, I study the extent to which language and information-based features predict short-horizon movements in Treasury ETF trading volume, returns, volatility, and quoted and effective spreads during the press conference window. Forecasts based on LLM-generated responses exhibit strong average predictive performance across a range of econometric and machine-learning models, often delivering higher explanatory power and lower forecast error for liquidity and trading-activity measures. The results highlight a clear distinction between mechanical market responses to linguistic structure and informational reactions to policy communication, demonstrating both the usefulness and the limitations of LLMs in analyzing central bank communication and short-horizon market dynamics.

Figure below: 

Feature importance (based on the percentile of contribution to the forecast) of the questions and answers' tone-topical content in forecasting microstructure variables 

Presentations: 

2025: University of Georgia Fall Finance Conference Early Idea Rapid Session

Figure below:

Left: Number of dollar stores per 1,000 residents, by county

Right: Net number of business closures in the counties that experience the net outflow of the local businesses following the opening of a dollar store within a 10,000-foot radius, per 1,000 residents