Finance and Economics

The following resources are from the public. I would like to thank all distinguished scholars and contributors who made their teaching notes, materials, and other information online and accessible, your contributions to the future of finance and economics are invaluable. If you do not want to see your materials on this website, have any questions, or something you suggest adding, please feel free to get in touch below. (lasted update: April 2024)

Conference Information

1. Conferences led by Bristol Financial Market Research Group: here (paper submissions around Feb or March of the year and normally happen in the same summer)

2. Conferences led by Bristol Banking Group: here 

3. Conference information on AFA: here

4. Conference information on Conference Maker: here

5. Conference information on SSRN: here

6. Conference information on NBER: here

7. Conference information on CEPR: here

TBC

AI-Powered Research in Finance and Economics

Over the past few months, I have been learning a great deal from scholars and researchers who are thinking seriously about how AI is changing academic work. In economics and adjacent fields, I have found Anton Korinek, Paul Goldsmith-Pinkham, Benjamin Golub, Pedro H. C. Sant’Anna, and Chris Blattman especially useful. What I appreciate about their work is that none of them treats AI as magic. They treat it as something much more interesting: a tool that becomes powerful only when it is embedded in a good workflow.

A few examples make this point better than any abstract principle.

Paul Goldsmith-Pinkham has shown how an AI coding agent can take a vague empirical question and turn it into a working first result very quickly. In one example, starting from an empty folder, Claude Code found Census data, handled messy government spreadsheets, wrote parsers, produced clean CSVs, and recovered from a FRED download problem caused by a 403 error. In another example, it scraped SEC EDGAR filings, parsed 10-K risk-factor sections, and built a structured dataset to study changes in tariff-related disclosure. His broader point is not that AI can replace research judgment; it is that the distance between a research idea and an initial empirical prototype is becoming much smaller.

Pedro H. C. Sant’Anna offers a different lesson. His documented Claude Code workflow is not just about writing a script or drafting a paragraph. It is a full academic system for papers, slides, data analysis, and replication packages. The setup uses a plan-first workflow, specialized reviewing agents, adversarial critic-fixer loops, mandatory verification, and quality gates before outputs are treated as finished. That is important because it shows that serious academic use of AI is not just about getting a fast draft; it is about building review and correction into the process itself.

Chris Blattman’s open-source workflow system adds another useful perspective. His materials are built for academics and other knowledge workers who want repeatable systems rather than one-off prompts. He emphasizes a “Prompt, Plan, Review, Revise” loop, project dashboards that are maintained from emails, documents, and meetings, and an executive-assistant workflow that he describes as going from 5,000 unread emails to inbox zero. Even when the examples are not all econometrics-specific, the larger lesson is highly relevant for research: AI becomes much more useful when it is embedded in routines for planning, review, iteration, and knowledge capture over time.

Benjamin Golub’s work highlights yet another margin: AI as a serious review layer. In his Markus’ Academy talk, he discusses Cursor as an AI-native coding environment that can read a repository, edit multiple files, run agent tasks, and explain code in context. He also presents Refine.ink as a tool for generating referee-style feedback on academic papers, including problems in math, empirical strategy, clarity, and consistency. That is a powerful reminder that AI is not only useful for drafting or coding faster; it is also useful for auditing arguments, logic, and implementation.

Anton Korinek provides the broad conceptual frame that ties all of these examples together. In his NBER paper, he describes AI agents as systems that can plan, use tools, and execute multi-step research tasks. He explicitly discusses applications such as literature review, writing and debugging econometric code, fetching and analyzing data, and coordinating complex workflows. In other words, the frontier is not just “better chatbots.” It is systems that can actually do research work — but only under meaningful human supervision.

The main lesson I take from all of this is simple:

Using AI well in research is much less about clever prompting and much more about workflow design.

For PhD students in economics and finance, I do not think the key question is simply, “How do I use AI?” A better question is:

How do I build a research process that AI can improve safely, productively, and transparently?

My answer is that AI is an accelerator, not the foundation.

The foundation is still a workflow that is reproducible, auditable, reviewable, and easy to verify. That is why project documentation and replication standards matter so much. Project TIER emphasizes the README as the entry point for understanding software, project structure, and reproduction steps. The AEA similarly requires that data and code be clearly and precisely documented and deposited for replication. GitHub’s pull request and review system is built around discussing and reviewing changes before merging, and Claude Code’s own best-practices documentation emphasizes planning first and giving the model ways to verify its work.

So what does this mean in practice?

Start with the repository, not the prompt.
Before asking AI to clean data, rewrite code, or generate robustness checks, define the project structure clearly: where raw data live, where cleaned data go, which scripts generate which outputs, and what the main replication path looks like. A minimal README is often more valuable than an elaborate prompt, because it tells both humans and AI what the project is, what the files do, and how the results are supposed to be reproduced.

Use Git and GitHub as the control system around AI.
For research, the value of Git is not just abstract version control. It is that every important change becomes visible and recoverable. Branches create a safe place for experimentation. Commits create a history of reasoning. Pull requests create a review layer between “the AI changed something” and “this is now part of the project.” GitHub’s own documentation is very clear that pull requests are designed to let teams discuss and review changes before merging them, which is exactly what AI-assisted research needs.

Never confuse “it runs” with “it is correct.”
This is probably the biggest practical mistake in AI-assisted empirical work. Code can run successfully and still be wrong. In economics and finance, the dangerous errors are often silent ones.

For example, an AI-generated data pipeline may look perfectly clean but still be incorrect if:

• it performs a one-to-many merge when the design required one-to-one, silently duplicating observations;

• it joins on filing date when the research design required fiscal-year-end timing;

• it drops a fixed effect or changes the clustering level without making that obvious;

• it winsorizes the full raw dataset instead of the estimation sample;

• it uses future information when constructing lagged controls;

• it overwrites the baseline script directly on the main branch, making it hard to recover the last trusted version.

None of these mistakes is dramatic from a coding perspective. All of them are potentially fatal from a research perspective.

The same is true for external data collection.
Goldsmith-Pinkham’s EDGAR example is useful precisely because it shows that AI can help build a structured dataset from messy filings — but also that the user has to supply institutional knowledge, such as the fact that SEC EDGAR requests are subject to fair-access rules, including a maximum of 10 requests per second and a declared User-Agent header. If you do not build those constraints into the workflow, “faster scraping” can quickly become “faster failure.”

This is why I increasingly think the right workflow for AI-assisted research looks something like this:

Before starting: define the project structure, write the README, clarify the task, and decide what success looks like.
During execution: use Git/GitHub to track changes, review diffs, validate outputs, and isolate experimental edits in branches.
Afterwards: confirm that the code, regressions, tables, and figures are actually reproducible, and that someone else could follow the workflow without needing your memory or your chat history.

Agentic AI is powerful. But it works best inside a well-designed research system, not as a substitute for one. That is what I take from Korinek’s framework, Goldsmith-Pinkham’s empirical examples, Sant’Anna’s orchestrated academic workflow, Blattman’s repeatable systems, and Golub’s review-focused tools.

So for research, the goal should not be “use more AI.”

It should be: build a better workflow first, then let AI speed it up.

Introductory Level of Finance and Economics

1. For finance students who want to build a robust foundation on econometrics, I personally would suggest the book written by my lifelong friend and mentor: Chris Brooks (Bristol and Reading), Introductory Econometrics for Finance (4th Edition). You can watch these lecture videos by Chris on YouTube.

Related: Chris Brooks (Bristol and Reading) and Robert Wichmann (Reading), R Guide to Accompany Introductory Econometrics for Finance

Related: Ran Tao (Bristol) and Chris Brooks (Bristol and Reading), Python Guide to Accompany Introductory Econometrics for Finance

2. Jeffrey M. Woodridge (Michigan), Introductory Econometrics: A Modern Approach

TBC

Information about Pre-Doc, RAs, and Graduate Schools

1. Why do a predoc: here (For those who are certain about their academic research career, my personal suggestion is that if you can go to PhD programme directly, go there; if not, predoc can be a good option for you to get some good connections and experiences. For those who are still considering their way ahead and are initially interested in research, it is necessary to get some flavor of research and then you can decide.)

2. Opening Positions: Econ-RA, PREDOC, scholars' web/social media, universities' recruiting websites, etc.

3. Suggestions and ideas about how to prepare for Predocs (and early stage of academic research) can be found: here

4. For all research assistants/associates/professionals, data processing and management skills are extremely important, and you can find some resources on this website as well: Data and Processing.

5. Susan Athey, Advice for Applying to Graduate School in Economics

6. Chris Blattman, Advice for Applying to Graduate School in Economics – 

7. Torsha Chakravorty, Aishwarya Kekre, Prerna Kundu, Lakshya Narula, and Vaibhav Ojha, Applying to PhD Programs in Economics: An Extensive Guide

8. Kasey Buckles, Interviews of Graduate Admissions Officers

9. CSWEP Newsletter, Considering Graduate Education in Economics? A Few Things to Ponder

10. The Chronicle of Higher Education, Your First Year in a PhD Program

11. TBC

PhD Lecture Notes

1. Ralph S.J. Koijen (Chicago Booth) and Motohiro Yogo (University of Princeton), PhD Lecture Notes: Financial Economics of Insurance

2. Ralph S.J. Koijen (Chicago Booth) and Stijn Van Nieuwerburgh (Columbia Business School), PhD Lecture Notes: Empirical Asset Pricing, 

3. Jiang, Zhengyang (Kellogg School of Management), PhD Lecture Notes: International Finance

4. Todd A. Gormley (Olin Business School), PhD Lecture Notes: Empirical Methods in Corporate Finance

5. Zhang, Johnew (Columbia Business School), PhD Lecture Notes: Finance and Economics

6. Bernt Arne Ødegaard (UiS Business School), PhD Lecture Notes: Finance

7. Paul Söderlind (University of St. Gallen), PhD Lecture Notes: Econometrics for Finance

8. Itay Goldstein (Wharton School), PhD Lecture Notes: Information in Financial Markets and Its Real Effects 

9. Itay Goldstein (Wharton School), PhD Lecture Notes: Moral Hazard

10. Itay Goldstein (Wharton School), PhD Lecture Notes: Theory of the Firm

11. Itay Goldstein (Wharton School), PhD Lecture Notes: Capital Structure

12. Itay Goldstein (Wharton School), PhD Lecture Notes: Financial Contracting

13. Itay Goldstein (Wharton School), PhD Lecture Notes: Corporate Control

14. Itay Goldstein (Wharton School), PhD Lecture Notes: Financial Intermediation and Crises

15. Itay Goldstein (Wharton School), PhD Lecture Notes: Financial Markets

16. Itay Goldstein (Wharton School), PhD Lecture Notes: Feedback Effects

17. Itay Goldstein (Wharton School), Lecture on "Short Term Debt and Incentives in Banks"

18. Michael Tehranchi (Cambridge), PhD Lecture Notes: Advanced Financial Models

19. Jeffrey Wooldridge (Michigan State), Works Mostly DiD & Others

20. Theresa Kuchler and Professor Johannes Stroebel (NYU), PhD Lecture: Empirical Household Finance

21. Prof. John H. Cochrane (Chicago), Time Series for Macroeconomics and Finance

Things Related

1. "Turning off my phone improved my science: how stepping back from 24/7 connectivity helped to restore Adam Weiss's focus in the lab" -- Adam Weiss

2. "How we landed job interviews for professorships straight out of our PhD programmes: follow these tips for an uber-organized and successful job hunt" -- Violeta Rodriguez and Qimin Liu

3. "How two PhD students overcame the odds to snag tenure-track jobs: Between us, we got several offers to lead labs before we had finished our PhDs." -- Violeta Rodriguez and Qimin Liu

4. "Three actions PhD-holders should take to land their next job: A hiring manager reveals the lessons he learnt when transitioning from a PhD programme to industry." -- Fawzi About-Chahine

5. How to Succeed in Academia or Have Fun Trying (updated 2024) -- Lasse Heje Pedersen

6. The Academic Job Market in Finance: An Updated Rookie's Guide -- Alexander W. Butler and Timothy Falcon Crack 

7. How to Write An Empirical Paper (and Referee One) -- Jay Ritter 

8. Writing Tips for PhD Students -- John Cochrane

9. Reviewing Less—Progressing More -- Matthew Spiegel

Advice for Aspiring Economists

by Professor Gregory Mankiw (Harvard)

Here is some advice for, say, an undergraduate considering a career as an economist.

1. Take as many math and statistics courses as you can stomach.

2. Choose your economics courses from professors who are passionate about the field and care about teaching. Ignore the particular topics covered when choosing courses. All parts of economics can be made interesting, or deadly dull, depending on the instructor.

3. Use your summers to experience economics from different perspectives. Spend one working as a research assistant for a professor, one working in a policy job in government, and one working in the private sector.

4. Read economics for fun in your spare time. To get you started, here is a list of recommended readings.

   1. 1. Milton Friedman, Capitalism and Freedom

      2. Robert Heilbroner, The Worldly Philosophers

      3. Paul Krugman, Peddling Prosperity

      4. Steven Landsburg, The Armchair Economist

      5. P.J. O'Rourke, Eat the Rich

      6. Burton Malkiel, A Random Walk Down Wall Street

      7. Avinash Dixit and Barry Nalebuff, Thinking Strategically

      8. Steven Levitt and Stephen Dubner, Freakonomics

      9. John McMillan, Reinventing the Bazaar

      10. William Breit and Barry T. Hirsch, Lives of the Laureate

5. Follow economics news. The best weekly is The Economist. The best daily is the Wall Street Journal.

6. If you are at a research university, attend the economic research seminars at your school about once a week. You may not understand the discussions at first, because they may seem too technical, but you will pick up more than you know, and eventually, you’ll be giving the seminar yourself.