Old volatility - ARCH effects in 19th century consol data

Engle's autoregressive conditional heteroscedasticity (ARCH) model has been used successfully to model volatility in modern financial data. Here the returns on 3% Consols traded on the London market from 1821 to 1860 are examined for timevarying conditional heteroscedasticity. The series contains over 10,000 daily price changes. The analysis produces strong evidence for persistent ARCH effects in the data. Structural changes in the model and periods of increased volatility can be linked to important political and historical events.