PolicyPulse: RBI Repo Shocks and Nifty Volatility Regimes

A complete end-to-end macro-finance and data-science project studying how RBI repo-rate changes relate to Nifty 50 volatility from 2012 to 2025.

Project Summary

This repository brings together policy data, market data, and quantitative analysis to answer a focused question:

Are RBI repo-rate changes informative about future volatility regimes in the Indian equity market, and can they improve short-horizon prediction beyond simple persistence-based baselines?

The work combines data preparation, exploratory analysis, event studies, model benchmarking, and a GJR-GARCH volatility framework. The final conclusion is that repo-rate moves are informative about regime context and market conditions, but they are not strong enough on their own to produce precise short-term forecasting power. Volatility persistence and market structure dominate the dynamics.

Final Repository Status

This project is complete and includes:

• a reproducible data preparation pipeline
• an exploratory analysis notebook
• an interactive Streamlit dashboard
• a formal written report with methodology and results
• prepared data files for downstream analysis

Repository Structure

• 00_data_preparation.py — prepares the master dataset by merging RBI repo-rate data with Nifty realized volatility.
• 01_analysis.ipynb — contains the main exploratory work, event-study analysis, predictive benchmarking, and volatility modeling.
• app.py — Streamlit application for visual exploration of repo-rate and volatility dynamics.
• REPORT.md — final narrative report summarizing the research and conclusions.
• data/ — cleaned and processed datasets used across the project.
• requirements.txt — Python dependency list.

Data and Methodology

Data Sources

• RBI repo-rate records from the official RBI data portal
• Nifty 50 daily market data from Yahoo Finance using yfinance

Core Workflow

1. Load RBI repo-rate records and standardize dates.
2. Resample the policy series to month-end frequency using forward fill.
3. Download Nifty 50 daily data and compute daily log returns.
4. Convert daily returns to monthly realized volatility using a 21-trading-day scaling factor.
5. Merge policy and market data into a monthly benchmark dataset.
6. Run exploratory statistics, event studies, predictive benchmarks, and GJR-GARCH analysis.

Analyses Included

1. Exploratory Relationship Analysis

• measurement of correlation between repo-rate changes and realized volatility
• interpretation of policy versus market regime dynamics

2. Event-Study Analysis

• comparison of volatility before and after hikes and cuts
• abnormal volatility analysis around policy decisions

3. Prediction Benchmarks

• naive persistence benchmark
• Logistic Regression and Random Forest comparisons
• walk-forward validation to preserve temporal ordering

4. Volatility Modeling

• GJR-GARCH applied to daily Nifty returns
• analysis of persistence, asymmetry, and heavy-tailed behavior

Main Findings

The final evidence from the project is consistent across methods:

• Repo-rate changes display a moderate inverse relationship with subsequent realized volatility.
• Cuts are associated with elevated pre-event volatility and normalization after the policy move.
• Hikes often occur in calmer volatility environments.
• Volatility regimes are highly persistent and typically outperform policy-based models in short-horizon prediction.
• Standard machine learning models do not reliably beat naive persistence baselines.
• GJR-GARCH confirms strong clustering, asymmetry, and fat-tailed behavior in Nifty returns.

Overall Interpretation

RBI policy is most informative as a regime and context indicator rather than as a standalone short-horizon forecasting signal. The project shows that Nifty volatility is driven substantially by its own persistence and market structure, while repo-rate actions provide useful but not dominant explanatory power.

How to Run the Project

1. Install dependencies

pip install -r requirements.txt

2. Prepare the master dataset

python 00_data_preparation.py

3. Explore the analysis

Open and run the notebook in 01_analysis.ipynb.

4. Launch the dashboard

streamlit run app.py

Project Outputs

• data/master_data.csv — combined monthly RBI and Nifty dataset
• data/rbi_rates.csv and data/rbi_rates_raw.csv — source policy data
• REPORT.md — detailed project report
• app.py — interactive policy and volatility dashboard

Final Assessment

This project is a complete, evidence-based macro-finance study. It demonstrates a realistic workflow for turning policy data and market prices into a coherent analytical narrative, and it reaches careful conclusions instead of overpromising predictive power. The repository is intentionally framed as a benchmark study of policy-volatility interaction, not as a claim of a highly predictive trading model.