Volatility Clustering and Weak-Form Efficiency in Emerging Markets

This repository contains the Python implementation of my MA Economics dissertation project at IIFT Delhi, investigating volatility clustering and weak-form market efficiency in eight emerging stock markets: India, Brazil, China, Indonesia, Mexico, Saudi Arabia, South Africa, and Turkey.

The analysis is performed in country-specific Jupyter notebooks, each using daily stock index data (2000–2024) and covering both full-sample and crisis-period studies (2008 Global Financial Crisis and COVID-19 pandemic). Each notebook follows the same structure - from data preprocessing, to volatility and efficiency modeling, and visualization - for a fully reproducible workflow, however using country-specific stock index data.

Link to dissertation report: https://drive.google.com/file/d/1dDeyyBxuX34247QOcj_HONgxgbbUOihP/view?usp=drive_link

Project Overview

The objective of the project was to investigate:

• Whether volatility clustering exists in emerging markets
• How different GARCH-family models capture volatility dynamics
• Whether weak-form efficiency holds in normal and crisis periods

The analysis was conducted for three scenarios:

• Full sample (2000–2024)
• Global Financial Crisis (2008)
• COVID-19 pandemic period (2020)

Workflow

1. Data collection and preprocessing of daily stock indices
2. Model estimation using univariate GARCH-family models:
   • ARCH(1)
   • GARCH(1,1) with normal and Student’s t-distributions
   • EGARCH
   • GJR-GARCH
3. Model selection using residual diagnostics and information criteria (AIC/BIC)
4. Testing weak-form efficiency using AR(1) regressions and AR(1)-GJR-GARCH models
5. Comparative analysis across countries and crisis periods

Libraries Used

• pandas
• numpy
• matplotlib
• seaborn
• statsmodels
• arch
• scipy

Project Structure

Full Sample Analysis/
- Jupyter notebooks for full-sample analysis
- Data/ (CSV files for full period)

GFC-2008/
- Jupyter notebooks for 2008 Global Financial Crisis analysis
- Data/ (CSV files for GFC period)

COVID-19/
- Jupyter notebooks for COVID-19 period analysis
- Data/ (CSV files for COVID19 period)

Notes

• The focus of this repository is on presenting the workflow, methodology, and findings of the project.
• The notebooks demonstrate the complete code and model outputs for each country and period.