take X-Y plane with observations (x,y) such that the correlatiob coefficient is 0 between them...
if u take linear regression Y=f(X) then the slope coefficient will come out to be 0 because of 0 correlation, there will be only intercept term Y= some constant (this will be parallel to X axis)
now if u take linear regression X=f(Y) then the slope coefficient will come out to be 0 because of 0 correlation, there will be only intercept term X= some constant (this will be parallel to Y axis)
Two lines, one is parallel to X axis and other is parallel to y axis are perpendicular to each other
Akshay Jain
Masters in Economics
Delhi School of Economics
2013-15