Lecture+6+Regression+with+panel+data

44beertax 1988data drunkdriverate 201015beertax 1513

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: t change between 1982 and 1988. The math: consider drunk driving rates in 1988 and 1982: DrunkDriveRatei1988 = β0 + β1BeerTaxi1988 + β2wi + ui1988 DrunkDriveRatei1982 = β0 + β1BeerTaxi1982 + β2wi + ui1982 Suppose the OLS conditional mean zero assumption E(uit|BeerTaxit, wi) = 0 holds, which means the errors is not correlated with BeerTaxit and wi. Subtracting 1988 from 1982 (that is, calculating the change), eliminates the effect of wi, i.e., DrunkDriveRatei1988 – DrunkdriveRatei1982 = β1(BeerTaxi1988 – BeerTaxi1982) + (ui1988 – ui1982) • The new error term, (ui1988 – ui1982), is uncorrelated with either BeerTaxi1988 or BeerTaxi1982. • This “difference” equation can be estimated by OLS. • The omitted variable wi doesn’t change, so it cannot be a determinant of the change in Y 4 Example: Drunk driving and beer taxes 1982 data: (n = 48) DrunkDriveRate = 1.86 + 0.44BeerTax 1988 data: DrunkDriveRate = 2.01 + 0.15BeerTax (.15) (.13) (n = 48) (.11) (.13) Difference estimation (n = 48) DDR − DDR = –.072 – 1.04(BeerTax1988–BeerTax1982) 1988 1982 (.065) (.36) Figure: Change in drunk driving rate and Beer taxes, 1988‐1992 Fixed Effects Regression What if you have more than 2 time periods (T > 2)? For state i Yit = β0 + β1Xit + β2wi + uit, i =1,…,n, t= 1,…,T Let’s assume that T=3, Yi1...
View Full Document

This document was uploaded on 03/11/2014.

Ask a homework question - tutors are online