EVAN / SUN Résumé
Research & project showcases

02 · Empirical finance research

HKEX IPO Performance Study

A 441-firm regression study of how industry, assets, and revenue relate to ROE among Hong Kong IPOs from 2020 to 2025.

HKEX IPO sample2020-2025
441

OLS · log transforms · clustered inference

Role

Academic research - data preparation, model design, robustness checks, visual analysis, and written report.

Context

The paper examines HKEX IPO data as a transparent proxy environment for thinking about financial fundamentals and IPO selection dynamics.

Tools

R · OLS regression · Clustered inference · Data visualization · Wind Financial Database

Project overview

This study uses ROE as a defined proxy for post-IPO financial performance and evaluates industry, firm size, and revenue with explicit caveats about inference.

Research question

Do industry classification, total assets, and total revenue explain variation in ROE around the IPO period?

Workflow

  1. 01Compile a 441-firm HKEX IPO sample from the Wind Financial Database.
  2. 02Use ROE as the outcome and log-transform heavily skewed assets and revenue variables.
  3. 03Estimate cross-sectional OLS with industry dummy variables and Consumer Discretionary as the reference group.
  4. 04Compare conventional and industry-clustered standard errors, then interpret results alongside omitted-variable and proxy limitations.
02

Data & methodology

Data sources

  • Wind Financial Database
  • 441 HKEX IPO firms
  • IPO cohort spanning 2020-2025

Methods

  • Cross-sectional OLS regression
  • Log transformations
  • Industry one-hot encoding
  • Industry-clustered standard errors
Evidence boundary
Input→Preparation→Analysis→Output

Content is based on the local paper and its HTML output; the underlying Wind dataset is not published.

Results & conclusion

Key outputs

  • 17-page research paper
  • Regression specification and coefficient output
  • Confidence-interval visualizations
  • HTML research deliverable

Documented findings

  • The report documents industry-level differences in the analysis narrative.
  • The reported adjusted R-squared was -0.0045, limiting the strength of broad explanatory claims.

Conclusion

Industry and accounting variables alone did not provide a sufficiently precise basis for strong general claims about IPO performance in the documented specification.

Decision relevance

  • The result favors a cautious due-diligence approach that supplements sector labels with firm-level drivers.
  • ROE can inform a fundamental view but does not capture market reception or first-day performance.

Limitations

  • ROE is only a proxy for IPO success.
  • The cross-sectional design does not establish causality.
  • Firm age, leverage, ownership, and market conditions were not fully modeled.
Next project showcase
03 · Market data engineering

Stock Market & Cryptocurrency Prediction

View all research & project showcases