1 / 11

Econ 427 lecture 17 slides

Learn about the benefits of using multivariate regression models for forecasting, including bivariate regression models, conditional and unconditional forecasting models, distributed lag models, and transfer function models.

Download Presentation

Econ 427 lecture 17 slides

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Econ 427 lecture 17 slides Multivariate Modeling (intro)

  2. Multivariate Forecasting • So far we have used only univariate models for forecasting • In many settings, the evolution of one variable is related to developments in others; and can improve forecasts-multivariate regression modeling.

  3. Bivariate Regression model • where x helps determine (cause) y. • Terminology: dependent (endogenous) and independent (exogenous) variables.

  4. Conditional versus unconditional forecasting models • A conditional forecasting model takes x as given and forecasts y conditional on the value of x • Also called scenario analysis. Why?

  5. Conditional versus unconditional forecasting models • Unconditional forecasting models. we usually are interested in unconditional forecasts: • We need to have an optimal forecast of x in order to form an optimal unconditional forecast of y.

  6. Unconditional forecasts • separately model x (say, as an autoregressive model, and insert x-hat into our y forecasting equation. • usually better to estimate all params simultaneously, • We’ll come back to this next time when we talk about VARs

  7. Distributed lag models • Capture cross-variable dynamics using a sequence of lags of the other variable • The deltas are the “lag weights” and their pattern is called the lag distribution

  8. Estimation problems • Problem is there may be a large number (Nx) of parameters to estimate. • How can we overcome this problem? • polynomial distributed lags (see book) • rational distributed lags (we’ll see below)

  9. Adding own-variable dynamics • lag dependent variables. • We can write this more compactly as:

  10. Adding own-variable dynamics • Another way is to include ARMA disturbances

  11. Transfer function models • Or both: the transfer function model is the most general multivariate model and includes both types of influences

More Related