Putting It All Together Forecasting with EconometricRegression Model

1
Putting It All Together /Forecasting with
Econometric/Regression Model

• Diebold, Chapter 9

2
Review

• Trend, Chapter 4
• Seasonality, Chapter 5
• Cycles, ARMIA, Newbold/Bos

3
Consider a model of

• Time Trend
• Monthly Seasonal Dummies
• ARMA(p,q)

4
Model of Liquor Sales

• Regress Lsales on Time, Time2
• Now add Seasonal dummies
• Look at Residuals / Correlogram
• Now add AR(3)

5
Forecasting with Explanatory Models

• Three Sources of Uncertainty
• Specification Uncertainty
• Innovation Uncertainty
• Parameter Uncertainty
• Error of forecasts of RHS variables

6
Lagged variables

• Lags of independents variables
• PDL - Polynomial distributed lag
• Smooth, orderly lags of RHS variables
• Lags of dependent variables
• Transfer Function general case

7
Vector Autoregressions

• N-variable VAR(p)
• Estimate N different equations
• Each equation regress LHS variable on p lags of
  itself and p lags of every other variable
• Allows for cross-variable dynamics

8
Eviews

• VAR(4) Housing Starts and Completions