This preview shows pages 1–8. Sign up to view the full content.
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 DocumentThis 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: Econ 471: L6 Time Series I Sung Y. Park U of Illinois Time series I So far we have considered crosssectional data : data collected at given (flxed) time. For example, years of education, salary and years of work experience of individual i at flxed time, say, 2000. I Now we consider timeseries data : data collected over time. For example, salary t , t = 1973 ; 1974 ; ; 2009. I Thus, most importantly, we have to take care of the serial dependence of time series variables, y t , t = 1 ; 2 ; ; n . Some examples F A Static Model: y t = fl + fl 1 z t + u t ; t = 1 ; 2 ; ; n I Contemporaneous relationship between y and z variables. I A change in z at t has an immediate efiect on y ) y t = fl 1 z t when u t = 0. I Example: Static Phillips curve: inf t = fl + fl 1 unem t + u t ) Contemporaneous tradeofi efiect in ation and unemployment rates. Some examples F (Finite) Distributed Lag Models: gfr t = fi + 1 pe t + 2 pe t 1 3 pe t 2 + u t ; t = 1 ; 2 ; ; n where gfr t : the general fertility rate; pe t : real dollar value of the personal tax exemption. I Idea: Decision to have children is related with the tax value of a child. I Tax rate at t may not instantly afiect fertility rate at time t . Tax rates at t 1 or/and t 2 may afiect the fertility rate at time t . I Finite distributed model of order 2 : pt t , pe t 1 and pe t 2 . Some examples Consider y t = fi + z t + 1 z t 1 2 z t 2 + u t ; t = 1 ; 2 ; ; n What are 1 , 2 and 3 ? I Finite distributed model of order 2. I Suppose there is an unit increase at t and value of z at s 6 = t is a constant, say, c . ) ; z t 2 = c ; z t 1 = c ; z t = c + 1 ; z t +1 = c ; y t 1 = fi + c + 1 c + 2 c y t = fi + ( c + 1) + 1 c + 2 c y t +2 = fi + c + 1 c + 2 ( c + 1) y t +3 = fi + c + 1 c + 2 c Some examples I y t y t 1 = : immediate change in y due to the oneunit change in z at time t (impact propensity) I y t +1 y t 1 = 1 : the change in y one period after the temporary change I y t +2 y t 1 = 2 : the change in y two periods after the temporary change I y t +3 y t 1 = 0: after two periods y does not change. (why?) If we consider the change in y due to a permanent increase in z such that z s = c if s < t and z s = c + 1 otherwise. After one period : + 1 After two periods : + 1 + 2 After three periods : + 1 + 2 + 1 + 2 : longrun change (longrun propensity) Finite sample properties of OLS F Assumptions : A1 (Linear in Parameter) Time series f ( x t 1 ; x t 2 ; ; x tk ; y t ) : t = 1 ; 2 ; ; n g follows the linear model y t = fl + fl 1 x t 1 + fl 2 x t 2 + + fl k x tk + u t ....
View
Full
Document
This note was uploaded on 02/24/2010 for the course ECON 570 taught by Professor Staff during the Fall '08 term at UNC.
 Fall '08
 Staff
 Econometrics

Click to edit the document details