Hi Ben, Your videos are great. I was wondering if you can expand on whether first order differences and/or adding the time dummies corrects for non stationary data. Thank you
How do we know the intercept B0 doesn't vary through time? Why can't it be different for the different time periods? If I'm not mistaken it just stands for the general house price in a certain time period if there was no crime. On the contrary, how can we have a difference in "overall trends" between time periods? If the sigmas are dummies for the effect of a certain time period (which in itself doesn't make sense to me as we are estimating the house price at time t, in the most upper model); how can we interpret a "change" in them from one period to the next?
We have an intercept which is added to b0 in each time period, namely our deltas. Your second question: we are estimating the effect of a change on a change. That will result in a type of elasticity. If I remember correctly there are ways to get back to the original variables after estimating the model, though we are mainly interested in the "effect", where the measure becomes less relevant.
@@lastua8562 Hey la stua, thanks for the answer. I asked these some time ago so by now I've figured out most, but I still appreciate your replies very much. Re the dummies, I dont think it is elasticities - the dummy variable will have almost the same values as in an OLS. I.e. in the first difference model between t=2 and t=1, the dummy variable for t=2 will have the value 1-0 = 1, and for t=0 it will be 0-0 = 0. For t=1 it will be 0-1 = -1. This is because otherwise the dummy variable for t=2 only explains the change from t=0 to t=2, but by subtracting the change from t=0 to t=1 we can get the change from t=1 to t=2.
@@ahmadghaemi2192 I think your response has been cut off. You may be right that they might not even be a type of elasticity though. I know you may have figured out most. I am answering for future students taking this unique course, as I like the questions you ask very much.
Really like your videos and you deserve all the praise you get, but what annoys me is how your notations are almost always different from the book (Stock, Watson). This make it really confusing and frustrating sometimes, which would be eliminated if you just were using the same notations as in the book.
Hello Ben ..do you have video explaining difference-in-difference (DD) estimator
I love you Ben. I have my exam tomorrow and I was in panic I wouldn't get this in time... you're a savior
Hi Ben, Your videos are great. I was wondering if you can expand on whether first order differences and/or adding the time dummies corrects for non stationary data. Thank you
This video is super helpful thanks!
Would the high standard error also apply to differencing a basic time series?
Awesome video!
you are doing my degree
thank u
I wonder what is the first difference of the time dependent dummy variable i.e delta dt in the above equation?
was this the full video? it seems it was cut off at the very end?
If the regressor is lagged dependent variable, can the first difference be used?
Thank youuu
By using change in HP, would you first adjust housing to real prices?
What would you say now?
How do we know the intercept B0 doesn't vary through time? Why can't it be different for the different time periods? If I'm not mistaken it just stands for the general house price in a certain time period if there was no crime.
On the contrary, how can we have a difference in "overall trends" between time periods? If the sigmas are dummies for the effect of a certain time period (which in itself doesn't make sense to me as we are estimating the house price at time t, in the most upper model); how can we interpret a "change" in them from one period to the next?
We have an intercept which is added to b0 in each time period, namely our deltas.
Your second question: we are estimating the effect of a change on a change. That will result in a type of elasticity. If I remember correctly there are ways to get back to the original variables after estimating the model, though we are mainly interested in the "effect", where the measure becomes less relevant.
@@lastua8562 Hey la stua, thanks for the answer. I asked these some time ago so by now I've figured out most, but I still appreciate your replies very much. Re the dummies, I dont think it is elasticities - the dummy variable will have almost the same values as in an OLS.
I.e. in the first difference model between t=2 and t=1, the dummy variable for t=2 will have the value 1-0 = 1, and for t=0 it will be 0-0 = 0. For t=1 it will be 0-1 = -1. This is because otherwise the dummy variable for t=2 only explains the change from t=0 to t=2, but by subtracting the change from t=0 to t=1 we can get the change from t=1 to t=2.
@@ahmadghaemi2192 I think your response has been cut off. You may be right that they might not even be a type of elasticity though.
I know you may have figured out most. I am answering for future students taking this unique course, as I like the questions you ask very much.
@@lastua8562 My answer should end with t=2, in that case it's all:-)
Really like your videos and you deserve all the praise you get, but what annoys me is how your notations are almost always different from the book (Stock, Watson). This make it really confusing and frustrating sometimes, which would be eliminated if you just were using the same notations as in the book.