Research

Working Papers

1. “The Demand for Robo-Advising: Adoption and Delegation Intensity”, with Richard Harris (Bristol) and Manuela Pedio (Bristol&Bocconi)

SSRN Working Paper Here

Abstract: Robo-advisors make delegated portfolio management scalable for retail investors, yet evidence on who uses them remains conflicting. We argue that this conflict reflects a missing distinction between adoption and conditional delegation intensity. Using proprietary monthly Alipay data on 100,000 Chinese mutual fund investors from September 2020 through September 2024, we estimate Heckman selection models that separate the two margins. We find systematic opposite signs: the same investor characteristic can predict entry in one direction and delegated share in the other. Underdiversification and turnover reduce adoption but raise delegated share among adopters; confirmation bias raises adoption but reduces delegated share. Fund experience raises participation in general but lowers conditional delegation, while wealth and age operate mainly on entry after behavioral controls are included. We propose and test three mechanisms. Selective awareness explains why investors with correctable mistakes may be less likely to seek correction ex ante, but delegate more and reduce underdiversification and turnover after exposure. Decision authority explains why robo-advising is not equivalent to generic automation: within-investor-month comparisons across self-directed investing, automatic investment plans, and robo-advising show that automating execution and delegating fund-selection authority are distinct choices. Reclaiming control explains post-adoption dynamics. More experienced investors delegate less over tenure, consistent with self-directed capability raising the cost of ceding authority. Return tests do not explain away the main cross-margin patterns. The paper contributes by reconciling conflicting robo-advisor adoption evidence, showing that behavioral factors systematically shape selection into corrective financial technology, and isolating delegation of decision authority from generic financial automation.

Note: This work is the first chapter of my PhD dissertation.

Presentation: USI Lugano Brownbag (2026)

2. “What's in the Retail Investors' Information Set? Understanding Retail Investor Behaviour with CNNs”, with Manuela Pedio (Bristol&Bocconi)

Draft Coming Soon

Abstract: In this paper, we use convolutional neural network (CNN) models to predict retail investors’ trading behaviour on Robinhood. Understanding retail investor trading behaviour is interesting and important, but it is difficult as it can be driven by sentiment and other behavioural factors that are hard to quantify. Among other things, retail investors’ activity has been shown to be driven by the visual features of price charts. CNN models enable us to build a time-series prediction system that exploits both graphical features and a host of stock characteristics to forecast retail trading activity. Specifically, this study seeks to predict investors’ holdings in the stocks that comprise the S&P 500 index over a sample spanning from 2018 to 2020. The CNN models that combine graphical and stock characteristics perform very strongly in predicting aggregate retail investor trading behaviour, achieving low test errors and economically meaningful R^2 at medium horizons, confirming the importance of price-chart information to explain retail investor trading activity. We then apply SHAP (Shapley Additive exPlanations) to help us understand which among the stock characteristics are the most relevant to explain aggregate investor behaviour. We find that technical indicators, such as the volumes and past open and close prices, are the most relevant features to drive retail investors' behaviour.

Note: This work was based on my Master's dissertation under the supervision of Prof. Vincent Han, with a previous title as "Trading by Charts: A Multivariate CNN System to Predict Retail Investor Trading".  Currently, I am working with Manuela to revise the paper significantly and explore the current direction. The work is now at a very early stage. 

Presentations: 45th International Symposium on Forecasting, the Ninth PKU-NUS Annual International Conference on Quantitative Finance and Economics, 2nd CUFE International Doctoral Forum, etc.

Selected Work in Progress

"AI and Robo-Advising"

"Whose Loss Is It?" with Tai-lo Yeung (USI Lugano)

Other Research (for self-interests)

1. “Understanding Financial Markets by Ancient Chinese Wisdom” (Sole Author) 

Highlights: This paper shows that ancient Chinese wisdom can help investors, academics, professionals, and regulators better understand the financial markets and make more feasible decisions accordingly. This is a pioneering study that systematically understands how Chinese wisdom can be used to understand financial markets in different situations. As there are countless experiences and histories that may be used to learn and reflect on the financial market and investor behaviours, this paper provides an example of how to learn these lessons and shows some insights about the future of finance using ancient Chinese wisdom. 

Note: This work has been presented at the 2024 ESRC SWDTP PhD Student Annual Conference. 

Major revisions are expected. 

Here is Jay's story of his Academic Research Journey.

I am currently a PhD student at the University of Bristol. My research interests include FinTech, Retail Investments, Machine Learning, Artificial Intelligence, ESG, Social Welfare, Investor Behaviour, Applied Financial Econometrics, Chinese Financial Market, Financial Theory, etc.

On my academic path, I am grateful to receive funding support from the following supporters.

1. My family (love you Mom and Dad ❤️)

2. The University of Bristol

3. The University of Bristol Business School

4. The University of Bristol Faculty of Social Science and Law

5. The University of Bristol Doctoral College

6. The University of Bristol SEC Alumni Grants Group

7. Jean Golding Institute at the University of Bristol

8. The Alan Turing Institute

9. Bocconi University

10. International Associate of Applied Econometrics

None of my achievements or progress would have happened without these generous supporters. 

I am the first generation to go to university in my family. I am not talented, I am not smart, and I am not even working very hard. I am a normal person who is resilient and determined in his dream, which is the most important thing for the academic journey, and maybe for all journeys. Besides, it is important to put every effort into chasing dreams.

My academic journey begins with my bachelor's degree in Finance at the University of Bristol. When I chose the program, I knew nothing about finance and I thought, similar to many, finance is a major that allows us to make a lot of money in the industry (which is obviously not the case). Luckily, later on, I found my interest in finding the truth behind the cover (it is important to keep curiosity) and in contributing to the future of human well-being. If this is the case for you, research is a good option. I then tried both industrial research and academic research (my first research assistantship with Dr Manuela Pedio and Dr Silvina Rubio) and subjectively concluded that academic research is actually the thing I want. After careful consideration, I decided to start an academic career in finance and applied economics.

The journey is never smooth, due to non-outstanding academic performances in UG, I faced a bunch of difficulties in the application of MSc, Pre-Doc, and PhD programs: Applied, Rejected; Applied, Rejected; etc (Sometimes I can get interviewed or future tasks).  But I am grateful for these experiences, which made me different and, more importantly, resilient and determined. I am even thinking of making a CV of Failures on this webpage, I will do so if I find some time later.

TBC...

What It Takes to Be A Good Scholar?

In this part, I am going to share my naive thoughts about my understanding of 1. what is a good scholar; 2. how to move toward to that goal.

1.  What is a good scholar? 

Before answering this question, it is important to define what is a scholar? I prefer the word "Scholar" to "Researcher" because the previous word contains the meaning of a person who focus on research but goes beyond that. Scholar is a status of lifelong learning and contributing to the field. 

It is important to differentiate Good Scholar from Impactful Scholar.

2. The Economist’s Craft: A Scholar’s Introduction to Research, Publishing, and Professional Development by Michael S. Weisbach

3. Writing Tips for PhD Students by John Cochrane

4. How to Succeed in Academia or Die Trying Have Fun Trying by Lasse Heje Pedersen*

5. Investments Notes by John Cochrane

What It Takes to Be An Impactful Scholar?

An impactful scholar is someone who caused significant effects to participants in the field, sector, and the world. These effects can be academic outputs, personal characteristics, and brilliant thinking.

Influential academic outputs normally are

1. New theories and methodologies that can be widely applied in the field and in other related fields.

2. Important empirical findings in a hot but long-lasting topic. 

3. Widely used textbooks in the field around the world.

TBC.

If you have any questions or suggestions for me, please feel free to get in touch below.