Celebrating the 2026 Computer Science and Economics (CSEC) Senior Prize Winners
The Department of Economics is proud to recognize this year’s winners of the Yale Computer Science and Economics (CSEC) prizes, honoring seniors for outstanding coursework and senior essays.
Launched in 2019, the CSEC interdepartmental major combines training in economics, computer science, data analysis, and empirical methods. The annual prizes celebrate students whose academic work reflects the rigor and range of the major, including original research at the intersection of economics and computer science.
The prizes include:
Donald Brown Prize: Arnav Bhakta and Neil Mathew
This award is granted to seniors in the joint CSEC major who have demonstrated excellence and achieved outstanding records during their tenure at Yale.
Herbert Scarf Prize: Dylan Bober and Haroon Mohamedali
This award is granted to seniors in the joint CSEC major who have presented the best senior thesis focusing on Computer Science.
Martin Shubik Prize: Henry Chen and Way Lee
This award is granted to seniors in the joint CSEC major who have demonstrated excellence in their senior thesis within the field of Economics.
Congratulations to all of these seniors for their exceptional accomplishments! Learn more about the prize winners below.
Arnav Bhakta: Donald Brown Prize
Senior Essay: On the Application of Machine Learning to Bankruptcy Prediction
Bankruptcy is a recurring and economically consequential event, relevant to nearly 10% of all corporations annually. Yet, to date, there has been a remarkable lack of quantitative tools developed to predict it and its resulting effect on companies’ capital structures. In this paper, Random Forest, eXtreme Gradient Boosting, Long Short-Term Memory Networks, and Temporal Convolutional Networks are applied to assess the tractability of using machine learning to solve these two problems. That is, each method is first used to classify whether a firm is on a trajectory toward bankruptcy, and second, to forecast its post-emergence leverage structure as a quantitative proxy for creditor recovery prospects. Based on a comprehensive experimental analysis that evaluates each of these problems across varying corporate population distributions, input window lengths, and forecasting horizons, our results reveal that machine learning models consistently outperform the Altman Z-Score, the current qualitative gold standard for default risk classification.
“I landed on this essay topic after taking Serkan Savasoglu's "Financial Risk Management," which prompted me to consider how recent technological advancements might allow us to better manage and predict the single greatest risk companies face: bankruptcy. Having seen how damaging an overextended capital structure can be to an otherwise fundamentally sound business, I wanted to explore whether machine learning methods could predict bankruptcy well in advance of a filing – long before conventional warning signs emerge.”
— Arnav Bhakta, Donald Brown Prize Winner
Neil Mathew: Donald Brown Prize
Senior Essay: AI Web Agent for Unstructured Golf Course Data Collection and an Econometric Analysis of Green-Fee Pricing
This project addresses two linked problems: the difficulty of collecting structured data from the heterogeneous web, and the absence of a systematic econometric model explaining green-fee price differences. Part I introduces an AI-powered web agent that intelligently navigates golf course websites to extract course information into a JSON scaffold, while Part II uses the resulting database to estimate a hedonic pricing model. Together, the two parts establish a modern LLM-powered data pipeline and a transparent econometric framework for studying pricing in the golf industry. The AI agent can be generalized as well to transform any kind of unstructured web into a structured dataset, useful for all kinds of applications from financial due diligence to academic data collection.
“I was drawn to this problem because econometrics fundamentally relies on quality structured data, which is limited compared to the vast world of unstructured data on the web. The advent of LLMs offered a basis for a solution I could build upon.”
— Neil Mathew, Donald Brown Prize Winner
Dylan Bober: Herbert Scarf Prize
Senior Essay: Mitigating and Simulating Algorithmic Price Collusion
Algorithmic pricing agents trained via Q-learning reliably converge to tacit collusion, yet regulators currently lack effective tools to address it in real-world marketplaces. This paper evaluates whether a platform-imposed price threshold can meaningfully reduce collusion among competing Q-learning sellers. In multi-seller simulations, a price threshold mechanism reduces collusion by up to 62%, with consumer surplus rising by over 23%. These results suggest that price thresholds represent an effective, market-compatible policy instrument—one that regulators could mandate without access to a firm’s proprietary algorithms.
“I chose this project because I was interested in the impacts that pricing algorithms can have on consumers. They have rapidly advanced in their ability to raise prices through sophisticated price discrimination, and in some cases through tacit collusion with competing algorithms.”
— Dylan Bober, Herbert Scarf Prize Winner
Haroon Mohamedali: Herbert Scarf Prize
Senior Essay: Evaluating ReBeL-Style Public Belief Modeling for Two-Player Cribbage
Cribbage has one of the most interesting imperfect-information gameplay dynamics. More specifically, the pegging phase requires constant reasoning about the opponent’s hidden hand, their potential future replies, and the value of the current board position. This thesis evaluates the strength of applying a ReBeL-style public-belief approach to two-player Cribbage.
“After taking CPSC 4740 with Professor Glenn, I developed a deep love for finding different ways to approach games. This thesis represents my attempt to rethink how Cribbage can be played through belief-aware AI and imperfect-information reasoning.”
— Haroon Mohamedali , Herbert Scarf Prize Winner
Henry Chen: Martin Shubik Prize
Senior Essay: Counting from the Clouds: Population Density Estimation in Mexico using Geospatial Foundation Models, 2010–2020
Accurate population data underpins critical decisions in public life like infrastructure investment and labor market planning, yet traditional decennial censuses are too expensive and infrequent to track rapid demographic shifts. This project evaluates a more efficient alternative by leveraging state-of-the-art geospatial foundation models to predict population density across Mexico directly from satellite imagery. The results show strong promise, with the models explaining up to 88% of density variation at the municipality level and reliably identifying which local areas are growing versus shrinking over time.
“I was first introduced to geospatial data as a Herb Scarf research assistant for Professor Luke Sanford during my first-year summer, and quickly became fascinated by the power of satellite imagery to solve real problems. When it came time to choose a thesis topic three years later, returning to that toolkit to tackle a question in economics felt like the natural choice—and having Professor Sanford advise the thesis made it a full-circle moment.”
— Henry Chen, Martin Shubik Prize Winner
Way Lee: Martin Shubik Prize
Senior Essay: Pricing the End of Renewal Expectations: Lease Maturity, Redevelopment Beliefs, and HDB Resale Prices in Singapore.
My senior essay studies how Singapore’s public housing resale market prices long-horizon policy expectations. Using a 2025 policy announcement as a shock to redevelopment expectations, I examine whether older HDB flats and flats with shorter remaining leases experienced stronger price responses, suggesting that resale prices reflect not only current housing services and lease duration, but also beliefs about future government renewal.
“I chose this topic because I have always been interested in Singapore’s housing policy, especially the question of what happens as public housing leases begin to expire. I am very grateful to my advisor, Professor Cody Cook, whose advice and expertise in housing economics helped me refine my research question and guided me immensely throughout the process.”
— Way Lee, Martin Shubik Prize Winner
The three prizes, awarded annually, are presented in honor of Professors Donald J. Brown, Herbert Scarf, and Martin Shubik.
Donald Brown Prize
Professor Donald J. Brown is the Phillip R. Allen Professor of Economics and Professor of Mathematics and is a leading mathematical economist. Applying his superb talents as a mathematician to the theory of economics, he has contributed significantly to the discovery and understanding of principles underlying economic models and has developed analytic methods that invigorate the mathematical theory of economics.
Herbert Scarf Prize
Professor Herbert Scarf was the Sterling Professor Emeritus of Economics. Professor Scarf made foundational contributions to game theory and to general equilibrium theory, but much of his best-known work concerned the application of theory. His work on inventory management proved the optimality of (S,s) rules, and he made pioneering advances on economies with increasing returns or with indivisibilities. Professor Scarf’s long-lasting influence on economics exemplifies the value of work that interfaces theory and computation.
Martin Shubik Prize
Professor Martin Shubik was the Seymour H. Knox Professor Emeritus of Mathematical Institutional Economics and was part of the Yale faculty since 1963. Throughout his career, he used the tools of game theory to better understand numerous phenomena of economic and political life. Professor Shubik was among the early scholars who recognized that mathematical and statistical tools could be applied to the social sciences, including economics.