financial-time-series

To model stock return volatility with GARCH in Python, you fit GARCH-family models to daily returns to capture volatility clustering and estimate conditional risk metrics. This approach quantifies dynamic volatility dynamics for financial time series.

Cointegration testing for pairs trading evaluates long-run relationships between two asset price series using Engle-Granger tests. By estimating a hedge ratio and computing spread Z-scores, it generates mean-reversion trading signals for entry and exit decisions.

You can run Fama-French factor regressions with rolling windows to estimate rolling factor betas and alpha. This performs OLS factor regression on time-indexed return series to explain portfolio return attribution across equities.

You need time-indexed return or price series formatted as pandas DataFrames to run financial time-series analysis. These inputs satisfy the statistical requirements for GARCH fitting, cointegration evaluation, and OLS factor regression.

The best way to calculate rolling risk metrics for VaR estimation is using rolling-window statistics on return series. This supports VaR and expected shortfall style risk monitoring alongside drawdown-aware analysis.

GARCH fitting may fail on high-frequency tick data because the models require evenly spaced time-indexed return series to satisfy statistical requirements. Aggregating tick data to daily or regular intervals ensures proper GARCH volatility estimation.