You are given two variables y and x, which may or
may not contain a time trend t.
The data are provided in a csv le called \TimeTrendRemoval.csv".
If you are asked to estimate the model
y = 0 + 1t + 2x + u; (2)
where u is a random error.
Suppose that you believe that t a ects both x and y and want to remove
the information of t from y to obtain ~y and from x to obtain ~x. Then
you claim that the estimate of 1 in
~y = 0 + 1~x + v; (3)
where v is a random error, would be identical to the estimate of 2 in
equation (2).
Please verify this claim empirically and explain why you observe what
you have observed.
