Skip to content
2003Sveriges Riksbank Prize · Behavioural, empirical, institutional

Robert Engle and Clive Granger

Citation: For methods of analyzing economic time series with time-varying volatility (ARCH) (Engle) and with common trends (cointegration) (Granger).

The key idea

Engle: ARCH/GARCH models — volatility clusters, time-varies, can be forecast. Granger: cointegration — non-stationary series can share a stationary long-run relationship; pairs trading and error-correction modelling follow.

The explanation

Engle's 1982 ARCH model captured the clustering of financial volatility — the empirical fact that big moves follow big moves. GARCH generalised it; the framework now underpins every VaR model. Granger's cointegration (with Engle, 1987) showed how to model variables that drift apart in the short run but are bound together long-term — exchange rate and inflation, money and prices, two stocks of similar firms.

Why Africa should care

Every NSE, JSE, NSEN volatility model is a GARCH variant. Pairs-trading between Safaricom and Vodacom, KCB and Equity, requires cointegration testing. South African Reserve Bank's monetary models use Granger's error-correction framework. Kenyan bond-yield decomposition into level/slope/curvature factors uses cointegrating regressions.

How to use it

Time Series Modules 6-7 of the Quant Math curriculum walk through ARCH/GARCH fitting and cointegration testing on real NSE data.

Canonical works

  • Robert F. Engle (1982) "Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation" Econometrica
  • Robert F. Engle and Clive W. J. Granger (1987) "Co-integration and Error Correction: Representation, Estimation, and Testing" Econometrica
Official Nobel Foundation page ↗