Columbia-NYU Financial Engineering Colloquium
Registration is required.
4:00 pm-4:45 pm | Nicolas Barberis
4:45 pm-5:30 pm | Xuedong He
5:40 pm-6:25 pm | Yanwei Jia
6:25 pm-7:10 pm | Steve Kou
Nicholas Barberis
Stephen and Camille Schramm Professor of Finance at the Yale School of Management
Title
Model-free and Model-based Learning as Joint Drivers of Investor Behavior”
Abstract
Motivated by neural evidence on the brain's computations, cognitive scientists are increasingly adopting a framework that combines two systems, namely “model-free” and “model-based” learning. We import this framework into a financial setting, study its properties, and use it to account for a range of facts about investor behavior. These include extrapolative demand, experience effects, the disconnect between investor allocations and beliefs in the frequency domain, the insensitivity of allocations to beliefs, the inertia in household allocations, and the dispersion in these allocations. Our analysis suggests that model-free learning may play a significant role in the behavior of some investors.
Bio
Nicholas Barberis is the Stephen and Camille Schramm Professor of Finance at the Yale School of Management. He is one of the pioneers of behavioral finance, the scientific study of irrational thinking in financial markets. His research, which has focused on building psychologically realistic models of investor behavior, has been awarded the Paul A. Samuelson Prize, the FAME Prize, and the Jack Treynor Prize. He has also received multiple awards for his teaching, and has mentored dozens of doctoral students and junior faculty. Since 2015, he has served as the Director of the NBER Working Group on Behavioral Finance, and he is also the founder and lead instructor of the Yale Summer School in Behavioral Finance, which has introduced hundreds of graduate students to the latest ideas in the field. He holds a B.A. in Mathematics from Cambridge University and a Ph.D. in Business Economics from Harvard.
Xuedong He
Editor of the Financial Engineering Area of Operations Research and Associate Editor for Mathematics and Financial Economics, Operations Research Letters, and Digital Finance
Title
Reference-dependent asset pricing with a stochastic dividend-consumption ratio
Abstract
We consider a discrete-time endowment economy populated by a representative agent with reference-dependent preferences, following Koszegi and Rabin (2007). In this framework, the agent derives utility not only from absolute consumption, but also gain-loss utility from comparing realized current consumption and updated expectations of future consumption to prior expectations. The agent trades a risky stock and a risk-free asset. The aggregate endowment comprises stock dividends and labor income, with a stochastic dividend-to-endowment ratio. A rational expectations equilibrium exists if the agent's expectations regarding future consumption are rational when forming reference points, and her optimal consumption and asset holdings clear the market. We prove the existence and uniqueness of this equilibrium and examine how reference-dependent utility influences asset prices.
This is a joint work with Luca De Gennaro Aquino, Moris S. Strub, and Yuting Yang.
Bio
Xuedong He received the B.Sc. degree in Mathematics and Applied Mathematics from Peking University in 2005 and the Ph.D. degree in Mathematical Finance from the University of Oxford in 2009. He is a professor at the Chinese University of Hong Kong and was an assistant professor at Columbia University from 2009 to 2015.
Xuedong He’s research interests include behavioral finance and economics, risk management, stochastic control, and financial technology. He has published papers in leading journals such as Management Science, Operations Research, Mathematical Finance, and Mathematics of Operations Research. He is the Editor of the Financial Engineering Area of Operations Research and serves as associate editor for Mathematics and Financial Economics, Operations Research Letters, and Digital Finance.
Yanwei Jia
Assistant Professor, Department of Systems Engineering and Engineering Management at The Chinese University of Hong Kong
Title
When to Quit a Venture: Normative Theory and Structural Identification of Decoupled Belief and Decision
Abstract
In vast scenarios, people engage in a risky venture with unknown payoffs and decide whether to continue with it, like experimentation, R&D and entrepreneurship. The conventional approach formulates this process as an optimal stopping problem within the Bayes framework, where agents continuously update their beliefs based on statistical principles, thereby rigidly locking their decisions to these updated beliefs and predicting a strict, deterministic threshold-based decision. This paper develops a novel reinforcement learning framework to describe an agent's behavior in sequential experimentation that formally separates beliefs from decisions. By decoupling the evaluation and policy processes, we provide a unifying framework that yields both normative benchmarks and flexible positive dynamics. Normatively, using the workhorse Gaussian bandit model, we prove that by properly tuning learning rates, the framework achieves a logarithmic regret bound, matching the efficiency of Bayesian rationality. Positively, the decoupled policy generates distinct and testable predictions, such as experience-driven, path-dependent quitting dynamics, even when the belief is consistent with its Bayesian counterpart. Our framework can also be extended to general jump-diffusion bandits. Crucially, we prove the structural identifiability of these hidden learning dynamics under the workhorse model. By utilizing the method of simulated moments, we demonstrate how this framework can be structurally estimated from censored data.
Bio
Yanwei Jia is an assistant professor in the Department of Systems Engineering and Engineering Management at The Chinese University of Hong Kong. He obtained his Ph.D. degree from the National University of Singapore in 2020, and B.Sc. from Tsinghua University in 2016. Prior to joining CUHK in 2023, he was an associate research scientist and adjunct assistant professor in the Department of Industrial Engineering and Operations Research at Columbia University. His research interest falls broadly into financial decision-making problems and uses the structural approach to study the decision making.
Steve Kou
Allen and Kelli Questrom Professor in Finance at Questrom School of Business, Boston University
Title
Mitigating Endogenous Turnout Bias in Election Polls
Abstract
Most econometrics apply statistics to economics. This talk attempts to conduct a study that goes opposite. We introduce a polling method that addresses the issue that the poll itself may influence turnout. The proposed poll asks respondents to rate their satisfaction with four scenarios, defined by which of two candidates wins, and whether the respondent votes. These ratings determine a critical probability of winning beyond which the respondent will abstain from voting. The poll result is defined by an equilibrium condition in which individual voting decisions and the poll’s vote share are mutually consistent. We provide simple conditions under which an equilibrium outcome exists and is computationally tractable. This is a joint work with Paul Glasserman.
Bio
Steven Kou is the Allen and Kelli Questrom Professor in Finance at Questrom School of Business, Boston University. He teaches courses on FinTech, quantitative finance, and risk management. He was a co-area editor for Operations Research. Currently, he is a co-editor for Digital Finance and has served on editorial boards of many journals, such as Operations Research, Management Science, Mathematics of Operations Research, and Mathematical Finance. He is a fellow of the Institute of Mathematical Statistics and won the Erlang Prize from INFORMS in 2002. Some of his research results have been incorporated into standard MBA textbooks.