Digital payment pressure and bank net fee margins
A documented research repository: bank-panel preparation, IMF FAS indicator extraction, main regressions and country-subsample robustness, plus a synthetic fixed-effects example that runs anywhere.
Yu-Chen Chao
Accounting and finance graduate with a public bank-panel research report
BSc (Hons) Accounting and Finance · Class II Division I · GPA 3.25
- 3.25 GPA
- 77/100 Accounting and Finance Research Project module
- 70/100 Business Analytics module
Class-relative figures are from Canvas cohort statistics.
Capability
What I bring
Four things I have already taken from raw input to a finished, checkable output — rather than tools I have read about.
Data work
Financial data from raw file to a finished table
I built the whole path myself: a restricted bank-level panel, an IMF Financial Access Survey indicator discovered and extracted, the country-year merge, and a complete-case sample with documented variable definitions. The dataset ends up auditable, not merely finished.
Measurement
Econometrics that survives being questioned
Bank and year fixed effects, bank-clustered standard errors, winsorisation, log and interaction specifications, country subsamples, and a between-within decomposition. I test the specification rather than reporting whichever result appeared first.
Communication
Analysis a non-specialist can act on
A five-dashboard Tableau story written for investors and FinTech strategists, with a stated insight and a commercial recommendation on every page — and the same work documented in Python so a technical reader can check it.
Reproducibility
Work that can be rerun, not just read
A locked Python environment, an automated test suite, a continuous-integration workflow, and a deterministic synthetic panel that lets anyone verify the estimation workflow in about a minute, because the restricted bank data itself cannot be shared.
Research
Digital payment pressure and bank net fee margins
A final-year panel study of 39 commercial banks in Indonesia, Malaysia, and Thailand, 2014–2022.
Research question
Is digital payment pressure associated with bank net fee margins in Indonesia, Malaysia, and Thailand?
What I did
- Constructed bank-level net fee margins and financial controls from a restricted bank panel.
- Extracted IMF Financial Access Survey indicator
IMF_FAS_FCMIBT: mobile and internet banking transactions as a percentage of GDP. - Merged the country-year measure into the bank-year panel.
- Estimated bank and year fixed-effects models with bank-clustered standard errors and 1st/99th percentile winsorisation.
- Tested log, interaction, nonlinear and country-subsample specifications, then checked the result with Driscoll–Kraay errors, a Mundlak decomposition, a one-year lag and a net-interest-margin placebo.
- Python
- pandas
- linearmodels
- PanelOLS
- IMF FAS
- WRDS-derived panel
What I found
- Linear specifications provide limited evidence of a direct association.
- The log specification shows a negative, statistically significant pooled association.
- Country subsamples are heterogeneous, and decomposition suggests cross-country structure may drive much of the pooled result.
- The design tests two competing hypotheses, substitution (Baumol, 1982) and complementarity (Thakor, 2020); the evidence is consistent with substitution in the pooled sample only.
These are conditional associations. The explanatory measure varies at the country-year level, only three countries are covered, and omitted factors cannot be ruled out.
Read the full research pageBecause the bank panel is restricted, the thesis regressions are documented and inspectable but cannot be rerun publicly. The repository ships a deterministic synthetic panel so that the estimation workflow still runs end to end.
Experience
Research, coursework, and practice
-
2025–2026 · Final-year research project
Digital payment pressure and bank net fee margins
University of Birmingham · Accounting and Finance
Research report component: 80/100 (65% of module; module total 77/100 on transcript) · above the component's upper quartile (73), i.e. top 25%; cohort mean 67, maximum 84
- Designed the research question, sample definition, and variable construction for a 39-bank ASEAN-3 panel (344 bank-year observations, 2014–2022).
- Built the Python pipeline: panel preparation, IMF FAS indicator discovery and extraction, country-year merge, and complete-case sample construction.
- Estimated two-way fixed-effects models with bank-clustered standard errors, then tested log, interaction, nonlinear, and country-subsample specifications.
- Published the design, results, limitations, and data-access boundaries as a public research report.
-
Business Analytics module
Mobile money adoption in the Global South — a five-dashboard Tableau story
Business Analytics module · individual Tableau assignment
Tableau assignment 70/100 (module mark 70/100 on transcript) · equal to the cohort's upper quartile (70); cohort mean 63.7, maximum 76
- Prepared the data first: an extract of 437 coded World Bank Findex fields across 174 economies and 12 population groups was mapped down to 24 business-readable measures, with decimals rescaled to percentages and missing binary usage flags imputed as zero.
- Dropped the 2011 and 2022 waves because their coverage is too thin, leaving 171 economies and four waves (2014, 2017, 2021, 2024).
- Built five interactive dashboards framed by a cover and a conclusion — Mobile Money Boom, Leapfrogging the Plastic Era, The Catch-Up, The Usage Paradox, and Platform Transformation — each with a written insight and a commercial recommendation.
- Findings: mobile money adoption expanded steadily from 2014 and leads in Sub-Saharan Africa; digital payment use runs ahead of debit-card ownership; savings lag far behind payment usage, most sharply in lower-income economies; and mobile money drives women's financial participation more than traditional banking does.
- Wrote a critical assessment of every dashboard, naming the limits of the data — no transaction-level detail, no qualitative evidence — and the chart types that would have carried the message better.
- Submitted December 2025. Its finding that digital payment use runs ahead of debit-card ownership raised the question my final-year project then tested: what happens to banks' fee income as payments move away from cards.
- Tableau
- Tableau Story and filters
- Tableau calculated fields
- Tableau Public
- Excel
- Python data preparation
- World Bank Global Findex
- GSMA industry data
-
July – September 2025
Intern, tax advisory team
Evershine CPAs Firm (永輝啟佳聯合會計師事務所)
- Context: applications for Philippines tax incentives, where eligibility turns on documented business activity, expenditure, and timing.
- What I did: prepared and checked the supporting documentation, including BOI Form 501, and reviewed business activity timelines and expenditure details against the stated eligibility conditions.
- Output: standard operating procedures mapped into flowcharts that the team used for the application workflow and for handover.
Internship certificate. Personal information has been redacted. Open certificate PDF
Education
Degree
BSc (Hons) Accounting and Finance — University of Birmingham
Qualification recommended: Class II (Division I) · GPA: 3.25 · Degree to be conferred 12 November 2026
SIM Diploma in International Business · April 2023 – June 2024
Skills
What I work with
Tools and methods I have used on real coursework and research data, not a list of everything I have read about.
Data and code
Python for data work
- Python 3.14
- pandas
- NumPy
- openpyxl
- requests
- BeautifulSoup
- matplotlib
- pytest
- uv
- Git and GitHub
Econometrics
Panel data and inference
- linearmodels PanelOLS
- statsmodels
- Bank and year fixed effects
- Bank-clustered standard errors
- 1st/99th percentile winsorisation
- Log and interaction specifications
- Between-within decomposition
- Country subsample robustness
Analytics and communication
Turning analysis into a story
- Tableau
- Tableau Story and dashboards
- Tableau calculated fields and filters
- Tableau Public story
- Excel and pivot analysis
- Field mapping and data dictionaries
- Data visualisation
- Research writing
- Markdown documentation
- Reproducible workflows
Domain
Finance and accounting
- Financial statement analysis
- Bank profitability and margins
- Non-interest income
- Digital payments and FinTech
- Financial Access Survey indicators
- Cross-border tax
Selected work
Projects you can open and check
Mobile money adoption in the Global South
A Tableau story in five dashboards, framed by a cover and a conclusion: a global bubble map and trend line, a leapfrogging comparison of bank accounts against mobile wallets, a catch-up scatterplot, a bow-tie and treemap of the savings gap, and a platform view of digital payments against bank accounts.
- Mobile money adoption expanded steadily from 2014 and leads in Sub-Saharan Africa.
- Digital payment use runs ahead of debit-card ownership, and savings trail both.
- Mobile money drives women's financial participation more than traditional banking does.
- Tableau Story
- Tableau Public
- Excel
- Python
Runnable synthetic fixed-effects example
Because the underlying bank data is restricted, the repository ships a deterministic synthetic panel and an automated test. Anyone can clone the project and verify the estimation workflow within a minute.
Contact
Get in touch
I am happy to talk about the research, the code, or opportunities in finance, banking, and data analysis.
Direct
- chaoglay1101@gmail.com Email address.
- LinkedIn Professional profile and updates.
CV as a PDF
- One-page CV (PDF)
- Full CV (PDF) The complete record, including methods, coursework, and notes.
The CV exists in English and Traditional Chinese; each link downloads the version that matches the language of this page.
Public work
- GitHub Research code and analytics projects.
- Full CV with a print-ready layout Use the print button on that page to save it as a PDF.