Research
Research
"Distinguishing Permanent from Transitory Changes: Experimental Evidence on Human Forecasting Ability" with Elise Payzan-Le Nestour and Qihe Tang.
Paper available here. (Job Market Paper)
Abstract: Distinguishing permanent from transitory shocks is a ubiquitous task, particularly critical in financial decision-making. In a forecasting experiment, success required inferring the task's hidden Data-Generating Process (DGP)---a step where participants showed surprising but inconsistent ability. Moreover, their forecasting was flawed by a systematic bias. To remedy these imperfections, we tested two solutions. A ``meta-nudge'' that channels attention reliably improved DGP discovery. For execution, a model-based AI capable of structural inference avoided this bias in simulations, suggesting a path to optimal performance. These findings provide a blueprint for improving financial decisions via both enhanced information design and human-AI collaboration.
Presented at:
University of Melbourne, 2025 (scheduled)
RMIT, 2025 (scheduled)
University of Sydney, 2025
Summer school on Cognitive Foundations of Decision-Making, 2025
School of Economics seminar, UNSW Sydney, 2024
School of Banking & Finance seminar, UNSW Sydney, 2024
Seminar at the University of Chicago, 2024
Geneva Decision-Making workshop, 2024
Vienna University of Economics and Business, 2024
The 26th International Congress on Insurance: Mathematics and Economics, Heriot-Watt University, 2023
"Stubborn by Design: The Behavioral Trap of Clinging to Soured Reward Sources" with Elise Payzan-Le Nestour, Bernard Balleine, and Samuel Thelaus.
Paper available here.
Abstract: We document a spillover of past rewards into subsequent gambling decisions in a controlled laboratory experiment: after receiving rewards from a profitable gambling strategy, participants persist with it despite being explicitly informed it has become unprofitable. Through experimental manipulations and computational modelling, we isolate a Pavlovian mechanism driving this maladaptive behavior: past rewards create an automatic attraction to gambling that can derail rational choice. This Pavlovian influence intensifies under uncertainty, suggesting it extends beyond gambling for monetary rewards to include "sign-tracking"-an attraction to gambling cues independent of monetary outcomes. These findings may help explain behavioral persistence in financial markets and suggest dual-focused interventions for problem gambling: both recalibrating outcome expectations and redirecting attention away from gambling cues.
Presented at:
Paris School of Economics, 2025
Paris Brain Institute (ICM), 2025
École normale supérieure de Paris (ENS), 2025
University of Chicago, 2023
GATE-Lab at the University of Lyon, 2023
École normale supérieure de Paris (ENS), 2023
School of Psychology, UNSW Sydney, 2023
The 2023 Economic Science Association World Meeting
The 2022 Asian Meetings of the Econometric Society
"High-quality Credit Portfolios under Multilevel Risks" with Qihe Tang and Yang Yang.
Paper available here.
Abstract: Consider a large investment portfolio that is crucial for social and economic stability, and therefore requires a prudent examination of the portfolio loss due to defaults. Suppose the portfolio is exposed to multilevel risks, categorized as an idiosyncratic risk factor, a systematic risk factor, and a common shock factor. To quantify the portfolio loss, we employ a one-period structural model in which latent variables governing individual credit rating migrations and defaults follow a mixture structure integrating the multilevel risks. At the heart of this work lies our argument that the common shock factor and the systematic risk factor may interplay to create a jointly extreme scenario. An asymptotic study of portfolio losses is carried out under a joint regular variation structure for the common shock factor and the systematic risk factor. Our main finding is that their tail dependence acts as an additional driving force behind portfolio losses. We show, both analytically and numerically, that this tail dependence, if underestimated or even ignored, can lead to catastrophic consequences.
Presented at:
The 2022 Australasian Actuarial Education and Research Symposium, Australian National University
School of Risk & Actuarial Studies, UNSW Sydney, 2022
The 25th International Congress on Insurance: Mathematics and Economics, Sun Yat-Sen University and Macquarie University, July 13--15, 2022
"Adapt and Endure: Individual Responses to Permanent shifts and Transitory Shocks" with Elise Payzan-Le Nestour.
Work in progress.
"Profit allocation in investment-based crowdfunding with investors of dynamic entry times" with Lindong Liu and Gongbing Bi.
European Journal of Operational Research, volume 280, issue 1, 2020.
Abstract: Even distribution is a normal profit allocation mechanism for investment-based crowdfunding projects on many platforms. In other words, the investors with the same pledging funds will be paid evenly when the investment ends. The even allocation mechanism works well under the assumption that the investors arrive at the platform simultaneously. However, in practice, the investors are sequential, therefore, the stories are different when considering the dynamic entry times of the investors. In this paper, we study ways to design appropriate profit allocation mechanisms to enhance the success rate of an investment-based crowdfunding project. The basic model focuses on the two-investor case, where only two investors with dynamic entry times are considered. The profit allocation mechanism is shown to have great impacts on the pledging probabilities of investors, as well as the success rate of a project. After that, we shift our focus to the two-cohort case, where dynamic investors are assumed to arrive at the platform as two sequential cohorts. By taking the sizes of each cohort into consideration, we are able to analyze the success rate of a project under various practical situations. Finally, we implement some numerical experiments to generalize our studies to the situations where (i) there are more than two pledging periods for the investors, (ii) the herding effect of the investors is considered, and (iii) the valuations of the investors are assumed to be normally distributed. Our main results still hold under these general situations.
"Worst-case moments under partial ambiguity" with Qihe Tang.
ASTIN Bulletin: The Journal of the IAA, volume 50, issue 2, 2023.
Abstract: The model uncertainty issue is pervasive in virtually all applied fields but is especially critical in insurance and finance. To hedge against the uncertainty of the underlying probability distribution, which we refer to as ambiguity, the worst case is often considered in quantifying the underlying risk. However, this worst-case treatment often yields results that are overly conservative. We argue that, in most practical situations, a generic risk is realized from multiple scenarios and the risk in some ordinary scenarios may be subject to negligible ambiguity so that it is safe to trust the reference distributions. Hence, we only need to consider the worst case in the other scenarios where ambiguity is significant. We implement this idea in the study of the worst-case moments of risk in the hope of alleviating the over-conservativeness issue. Note that the ambiguity in our consideration exists in both the scenario indicator and the risk in the corresponding scenario, leading to a twofold ambiguity issue. We employ the Wasserstein distance to construct an ambiguity ball. Then we disentangle the ambiguity along the scenario indicator and the risk in the corresponding scenario so that we convert the two-fold optimization problem into two one-fold problems. Our main result is a closed-form worst-case moment estimate. Our numerical studies illustrate that the consideration of partial ambiguity indeed greatly alleviates the over-conservativeness issue
Presented at:
The 11th Conference in Actuarial Science & Finance on Samos, University of the Aegean, 2022
The 24th International Congress on Insurance: Mathematics and Economics, co-hosted virtually by the University of Illinois at Urbana-Champaign, the Pennsylvania State University, Ulm University, and UNSW Sydney, 2021
Reviewing service: Journal of Behavioral and Experimental Finance
Reviewing service: PLOS Computational Biology
Reviewing service: Insurance: Mathematics and Economics
Member of the Organizing Group, the 24th International Congress on Insurance: Mathematics and Economics, co-hosted virtually by the University of Illinois at Urbana-Champaign, the Ulm University, the Pennsylvania State University, and UNSW Sydney, July 5--9, 2021
Member of the Research Committee and Teaching Committee, School of Risk & Actuarial Studies, UNSW Sydney, 2020-2021