00:02
Part 1.
00:05
For 2 variables, approve and log a real gas price.
00:10
The first order autocorrelation coefficients are about 0 .93, very close to 1.
00:26
So we should worry about these 2 series being unit routes.
00:39
Given what we know in part 1, we might hesitate to estimate this model by ols because we will are going to violate one important ols assumption.
01:07
The series in the regression need to be weekly dependent, so central limit theorem can apply and we could yield unbiased estimates.
01:51
Unit root series are not weekly dependent.
02:19
You could try estimating the model, and you will find that the model has a high r -squared, and almost all explanatory variables are highly significant.
02:33
That is the expected result of a spurious regression.
02:41
In part 3, we will estimate the equation in part 2 by differencing all variables, including the dummy variables.
02:52
This is the result...