Atlanta Finance Workshop
Program
Afternoon research workshops held at the Federal Reserve Bank of Atlanta, 1:00 to 6:00 PM.
Workshop 2
Friday, May 22, 20261:00 – 1:30 pm
Digitally Inclined Small Businesses: Effects of E-Commerce on Small Business Credit
Tetyana Balyuk (Emory)
Abstract +We study how e-commerce affects credit provided to small businesses in the U.S. We find that small businesses have higher sales, larger orders, and more stable revenue after e-commerce adoption. Despite lower cash flow volatility (which should reduce credit demand), businesses obtain more credit after e-commerce adoption. Results are confirmed using staggered entry of Uber Eats in the food industry. Consistent with e-commerce relieving credit constraints both directly (higher expected payoffs to lenders) and indirectly (lower lending costs from hard data generated in e-commerce). We also find redistributional effects of e-commerce on credit across industries in local credit markets.
1:30 – 2:00 pm
Corporate Debt Refinancing and Path-Dependent Effects of Monetary Policy
Indrajit Mitra (Atlanta Fed)
Abstract +Corporate debt refinancing generates strong path dependence in monetary policy transmission. A prolonged period of low interest rates allows firms to lock in cheap, long-term financing, muting the impact of subsequent rate hikes. We formalize this in a model highlighting a firm-level rate gap, the difference between the interest rate on existing debt and the current market rate. The gap governs refinancing and investment decisions. Monetary policy operates by shifting the cross-sectional distribution of gaps, implying weaker transmission following prolonged periods of low rates. We use detailed firm-level bond data to construct gap measures with empirical support.
2:00 – 2:30 pm
Break
2:30 – 3:00 pm
When AI Takes the Call: Automation, Handoffs, and Residual Human Work
Rik Sen (UGA)
Abstract +How does AI change work when firms deploy it both to automate tasks and to assist workers on the tasks that remain? We study 86,589 customer-service tickets from a fintech firm that introduced an AI copilot for agents and, shortly after, an AI bot for selected tickets. Using difference-in-differences based on predicted automation exposure, we show that more exposed tickets are resolved faster and more durably, as task-based models would predict. Yet human performance metrics worsen because automation removes easier tickets and leaves a harder residual workload. A realized-path decomposition shows that gains come from clean bot resolution, while bot-to-human handoffs perform much worse on eventual quality, creating a task margin absent from standard models. Extending a task-based model to incorporate handoffs and assistance, we predict and find that copilot-related quality gains are larger after bot rollout.
3:00 – 3:30 pm
Nocturnal Trading
Andriy Shkilko (UGA)
Abstract +Although still relatively new, nocturnal trading in U.S. equities, defined as trading between 8:00 p.m. and 4:00 a.m., has grown rapidly. It is largely retail-driven, concentrated in a small set of securities, and marked by substantial order imbalances. Using unique transaction-level data, we show that nocturnal execution costs exceed regular-hours benchmarks but are broadly consistent with elevated adverse selection faced by liquidity suppliers. Nocturnal returns generally do not reverse during the subsequent regular-hours session, except in a small subset of high-sentiment stocks. Overall, the nocturnal session appears to be an important source of price discovery and may create profit opportunities for retail liquidity demanders despite higher transaction costs.
3:30 – 4:00 pm
Break
4:00 – 4:30 pm
The Limits of Unilateral Financial Enforcement
Hala Moussawi (Georgia Tech)
Abstract +The centrality of the U.S. dollar allows the U.S. to impose sanctions and project influence abroad, but unilateral enforcement may shift financial activity outside U.S. jurisdiction. Using 7,085 enforcement actions against global banks during 2011-2022, I document that the U.S. levies 89% of all monetary penalties across 45 jurisdictions. U.S. settlements raise funding costs, contract cross-border intermediation, and partially reallocate credit toward lenders outside U.S. oversight, even as borrowers remain anchored to dollar-denominated credit. A model disciplined by observed switching patterns implies a roughly 15-year half-life for U.S. jurisdictional reach at the current enforcement intensity.
4:30 – 5:00 pm
Monopolizing Minds: How M&As Stifle Innovation Through Labor Market Power
Sophia Jing Xue (GSU)
Abstract +We argue that M&As reduce inventors' innovation incentives and outputs by increasing firms' labor market power and limiting the rents inventors can capture. Using individual-level longitudinal data from the U.S. Census Bureau, we find that at both target and acquiring firms, inventors exposed to greater increases in labor market concentration in already concentrated labor markets produce fewer patents, earn lower wages, and exhibit reduced job mobility following mergers. The negative impact of increased labor market power on inventor productivity outweighs the potential benefits from innovation synergies.
5:00 – 6:00 pm
Reception
Previous Workshops
Workshop 1
Friday, November 21, 20251:00 – 1:25 pm
Extensive and Intensive Demand: Origins and Impacts
Xindi He (Georgia Tech)
1:25 – 1:50 pm
Revealed Preference Discount Rates
Joseph Hall (Georgia Tech)
1:50 – 2:00 pm
Shock Propagation via Bank–Non-Bank Linkages: Evidence from U.S. Credit Registries
Manasa Gopal (Georgia Tech)
2:00 – 2:30 pm
Break
2:30 – 3:00 pm
Charitable Capital Allocation
Gene Lu (UGA)
3:00 – 3:30 pm
Buying In and Selling Out: The Dynamic Returns to Investing in Expertise
Felix Feng (UGA)
3:30 – 4:00 pm
Break
4:00 – 4:30 pm
Move to Creativity: Inventor Household Mobility and Innovation Productivity
Sophia Jing Xue (GSU)
4:30 – 5:00 pm
When Banks Fail: Depositor Attention and the Cost of Funding for Survivors
Brian Jonghwan Lee (Emory)
5:00 – 6:00 pm
Reception
