Some Important Time Series Processes Fin250f: Lecture 8.3 Spring 2010
Outline Basic finance price dynamics Random walk Geometric random walk Martingale Random walk + noise Trend + noise
Random Walk
Geometric Random Walk
Martingale
Random Walk + Noise
Why interesting? Exponential smoother is the optimal forecast History: Muth, and rational expectations
Trend + Noise Model Model returns as Small persistent trend Plus noise Can generate significant predictability which is invisible to most tests
Covariance Reminder
Trend + Noise
Trend Plus Noise
Parameter Example A small big A = 0.02,
Trend Plus Noise ACF
AR(1) ACF’s
Trend Plus Noise Can show Returns are ARMA(1,1) ARMA’s can generate persistent, but small autocorrelations for certain parameters This is useful to model returns Low correlations Trends for momentum/moving average trading rules