Transcript Chapter 6
x x ECONOMETRICS x x x CHAPTER 6 Dummy Variable Regression Models Dummy, or indicator, variables take on values of 0 or 1 to indicate the presence or absence of a quality. They can be included in regressions just like quantitative variables. Yi = B1 + B2 Di + B3Xi + ui Di = 1 for female 0 for male OR Di = 1 for prior to 2001 0 for 2001 and later A Dummy Variable Indicates Change in Intercept Y Yi = 60 + 30 Di - 0.4 Xi + ui 120 If D = 0 100 Y = 60 + 30 D - 0.4 X 80 D=1 60 40 D=0 20 Y = 60 – 0.4 X If D = 1 Y = 60 + 30 D - 0.4 X 20 40 60 80 100 120 Y = 90 - 0.4 X X B2=30 is the differential intercept coefficient. A Dummy Variable Can Allow for Change in Slope, Too Y Yi = 60 + 30 Di – 0.4 Xi + 0.3(DiXi) + ui 120 100 If D = 0 D=1 80 Y = 60 + 30 D – 0.4 X + 0.3 DX Y = 60 – 0.4 X 60 40 D=0 20 20 40 60 80 100 120 If D = 1 Y = 60 + 30 D – 0.4 X + 0.3 DX Y = 60 – 0.1 X X B4=0.3 is the differential slope coefficient. Variables with More than Two Classes Don’t allow the dummy variable to take on multiple values: Yi = B1 + B2 Di + B3Xi + ui Di = 0 for first quarter 1 for second quarter 2 for third quarter 3 for fourth quarter This assumes a pattern in the effect that might not exist. Instead, use several indicator variables. Yi = B1 + B2 D2i + B3D3i + B4D4i + B5Xi + ui D2i = 1 if 2nd qrt 0 otherwise D3i = 1 if 3rd qrt 0 otherwise D4i = 1 for 4th qrt 0 otherwise