You are on page 1of 14

Box Jenkins

or
Arima
Forecasting

H:\My Documents\classes\eco346\Lectures
\chapter 7\Autoregressive
Models.doc

All stationary time series can be modeled as


AR or MA or ARMA models
A stationary time series is one with constant
mean ( ) and constant variance.
Stationary time series are often called mean
reverting seriesthat in the long run the
mean does not change (cycles will always die
out).
If a time series is not stationary it is often
possible to make it stationary by using fairly
simple transformations

Nonstationary Time series

Linear trend
Nonlinear trend
Multiplicative seasonality
Heteroscedastic error terms (non constant
variance)

How to make them stationary


Linear trend
Take non-seasonal difference. What is left
over will be stationary AR, MA or ARMA

Nonlinear trend
Exponential growth
Take logs this makes the trend linear
Take non--seasonal difference

Non exponential growth ?

Multiplicative seasonality
Take logs
Multiplicative seasonality often occurs when
growth is exponential.
Take logs then a seasonal difference to
remove trend

Heteroscedsatic errors
Take logs
Note you cannot take logs of negative
numbers

Box Jenkins Methodology

Identification
Estimation
Forecasting
Examine residuals
Reestimate
Repeat until you only have noise in
residuals

Identification
What does it take to make the time series
stationary?
Is the stationary model AR, MA, ARMA
If AR(p) how big is p?
If MA(q) how big is q?
If ARMA(p,q) what are p and q?

Seasonality
Is the seasonality AR, MA, ARMA
What are p, q?

AR(p) models
The ACF will show exponential decay
The first p terms of the PACF will be
significantly different from zero (outside
the parallel lines)

MA(q) models
The first q terms of the ACF will be
significantly different from zero
The PACF will decay exponentially
towards zero

ARMA models
If you cant easily tell if the model is an AR
or a MA, assume it is an ARMA model.