Glossary

GARCH (Generalised Autoregressive Conditional Heteroscedasticity)

A class of time-series models that capture the tendency of financial returns to exhibit volatility clustering. The basic GARCH(1,1) model specifies the conditional variance as a function of past squared residuals and past conditional variances. Extensions include EGARCH, GJR-G...

Definition

Why It Matters

Financial asset returns rarely have constant volatility; instead, they cluster into periods of high and low turbulence. Ignoring this conditional heteroscedasticity leads to underestimated risk, mispriced derivatives, and invalid inference. GARCH models provide a parsimonious yet flexible framework for capturing and forecasting time-varying volatility, which is central to Value-at-Risk calculations, portfolio optimisation, and monetary policy analysis in emerging markets like Turkey.

Example

Estimating a GARCH(1,1) model on daily BIST-100 returns reveals that the conditional variance spikes during the 2018 Turkish currency crisis and remains elevated for several weeks before gradually declining. The estimated persistence parameter (alpha + beta) is close to 0.95, indicating that volatility shocks are long-lasting but eventually mean-revert.

Related Terms

Software Notes

SPSS: Not natively supported; use R integration for GARCH estimation.
R: Use garch() from the tseries package for basic GARCH, or ugarchfit() from the rugarch package for a wide range of GARCH specifications including EGARCH and GJR-GARCH.
Stata: Use arch variable, arch(1) garch(1) for GARCH(1,1). For EGARCH, use arch variable, arch(1) egarch(1). Post-estimation commands include predict for conditional variances.