This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 1 / 30 Introduction to Econometrics Econ 322 Fall, 2010 Lecture 27: Introduction to Time Series Methods – II December 8, 2010 Topics Covered triangleright Topics Covered Causal Models using Time series When does OLS make sense for estimating a linear regression with time series included? Models where epsilon1 t exhibits serial correlation Testing for Serial Correlation Time series with trends Dealing with Deterministic Trends Checking for trends Testing for unit roots DickeyFuller Test Summary of Trends Dynamic Effects Impulse Response of Y for a Temporary Change in X Impulse Response of Y for a Permanent Change in X Autoregressive distributed lag models (ADL(p,q)) An Example 2 / 30 1. serial correlation with covariates 2. dynamic effects 3. trends 4. unit roots Causal Models using Time series Topics Covered triangleright Causal Models using Time series When does OLS make sense for estimating a linear regression with time series included? Models where epsilon1 t exhibits serial correlation Testing for Serial Correlation Time series with trends Dealing with Deterministic Trends Checking for trends Testing for unit roots DickeyFuller Test Summary of Trends Dynamic Effects Impulse Response of Y for a Temporary Change in X Impulse Response of Y for a Permanent Change in X Autoregressive distributed lag models (ADL(p,q)) An Example 3 / 30 square sometimes we are interested in the causality between two time series square e.g. is there a (contemporaneous) relationship between inflation and the “state” of the economy square modern economists talk about the relationship between inflation and the output “gap”. square in recent history we had the Phillip’s Curve which noted that inflation and unemployment were negatively related. square such a model would be in the form of y t = β + β 1 x t + epsilon1 t When does OLS make sense for estimating a linear regression with time series included? Topics Covered Causal Models using Time series triangleright When does OLS make sense for estimating a linear regression with time series included? Models where epsilon1 t exhibits serial correlation Testing for Serial Correlation Time series with trends Dealing with Deterministic Trends Checking for trends Testing for unit roots DickeyFuller Test Summary of Trends Dynamic Effects Impulse Response of Y for a Temporary Change in X Impulse Response of Y for a Permanent Change in X Autoregressive distributed lag models (ADL(p,q)) An Example 4 / 30 square Estimating a GLRM using OLS is only valid when the error term epsilon1 t is stationary. square OLS is only efficient if the error term is stationary and is not serially correlated. square How can epsilon1 t be not stationary?...
View
Full Document
 Fall '11
 LANDONLANE
 Econometrics, Regression Analysis, Time series analysis, Dickey–Fuller test, Augmented Dickey–Fuller test

Click to edit the document details