Time Series Basics Fin250f: Lecture 3.1 Fall 2005 Reading: Taylor, chapter
Outline Random variables Distributions Central limit theorem Two variables Independence Time series definitions
Random Variables: Discrete
Random Variables: Continuous
Important Distributions Uniform Normal Log normal Student-t Stable
Normal/Gaussian
Normal Picture: Sample = 2000
Normal Exponential Expectations
Why Important in Finance? Central limit theorem Many returns almost normal
Log Normal
Not symmetric Long right tail
Log Normal Histogram (Sample = 5000)
Chi-square
Student-t
Student-t Moments All moments > r do not exist
Stable Distribution Similar shape to normal Infinite variance Sums of stable RV’s are stable
Central Limit Theorem (casual)
Consequence of CLT and continuous compounding
Two Variables
More on Two Variables
More Two Variables
Independent Random Variables
More than Two RV’s
Multivariate Normal
Independence
Independent Identically Distributed All random variables drawn from same distribution All are independent of each other Common assumption IID IID Gaussian
Stochastic Processes
Time Series Definitions Strictly stationary Covariance stationary Uncorrelated White noise Random walk Martingale
Strictly Stationary All distributional features are independent of time
Covariance Stationary Variances and covariances independent of time
Uncorrelated
White Noise Covariance stationary Uncorrelated Mean zero
Random Walk
Geometric Random Walk
Martingale
Autocovariances/correlations
Outline Random variables Distributions Central limit theorem Two variables Independence Time series definitions