2
answers
0
watching
231
views
28 Sep 2019
we have a true model
Yt = b0 + b1Xt + b2Yt + Et
where the error term follows a weird serial correlation: Et = Et-1 + Ut. Ut is perfectly well-behaved random variable with 0 mean and constant variance. cov(Ut,Us) = 0 when t is different from s.
a. state the consequences of estimating this model using OLS? what kind of Gauss-Markow assumptions that it happens to violate.
b. find a transformation of the data to be able to use the same data to estimate a model that satisfies the Gauss-Markov assumptions. be clear and explicit about the process. clearly explain why the transformed model meets the Gauss-markov assumption.
* this kind of model is random walk serial correlation.
we have a true model
Yt = b0 + b1Xt + b2Yt + Et
where the error term follows a weird serial correlation: Et = Et-1 + Ut. Ut is perfectly well-behaved random variable with 0 mean and constant variance. cov(Ut,Us) = 0 when t is different from s.
a. state the consequences of estimating this model using OLS? what kind of Gauss-Markow assumptions that it happens to violate.
b. find a transformation of the data to be able to use the same data to estimate a model that satisfies the Gauss-Markov assumptions. be clear and explicit about the process. clearly explain why the transformed model meets the Gauss-markov assumption.
* this kind of model is random walk serial correlation.
abhisheksinghLv10
26 Sep 2023
Mahe AlamLv10
29 Sep 2019
Already have an account? Log in
