beta-research

Research using different methods to estimate betas between assets and a dive into computing returns in execess of market beta.

Running

• Clone repository
• Run notebook

Overview

This research project uses financial data to estimate the beta of various assets relative to the S&P 500 index. Two methods are employed for beta estimation:

• Kalman Filter
• Rolling Ordinary Least Squares (OLS)

Data

The data is fetched from Yahoo Finance and includes the following tickers:

• JPM
• ^GSPC (S&P 500)
• GLD
• WMT
• AAPL
• BRK-B
• BIL

Methods

Kalman Filter Beta

The Kalman Filter is used to estimate time-varying betas.

Rolling OLS Beta

Rolling OLS is used to estimate betas over a specified window period.

Visualization

The project includes visualizations of the estimated betas and portfolio returns.

Usage

To run the analysis, execute the provided Python Jupyter Notebook. Ensure you have the necessary dependencies installed:

• pandas
• numpy
• matplotlib
• yfinance
• cycler
• jinja2

Results

The results include:

• Estimated betas for each asset
• Portfolio returns and betas
• Hedged portfolio returns
• Performance metrics

Reference

For more information on Kalman Filters, refer to the following resource:

• Kalman and Bayesian Filters in Python

License

This project is licensed under the MIT License.