Time Series

Question 1

Conditions for Stationary?

Answer 1

1. The joint distribution of variables is the same for any number of successive periods in a row (PAST + FUTURE). If the mean is different but variance is still the same, it is an example of non-stationary

Question 2

What is the assumption of OLS which is the most similar to stationary?

Answer 2

identically distributed
Because stationary has constant jointly distribution function, which means data is identically distributed

Question 3

What are assumptions of OLS?

Answer 3

1. No outliners
2. Linear relationship
3. No endogeneity
4. No multicoliarlity
5. No serial correlation
6. Normality and homoschedasticity

Question 4

What is the function to test for stationary?
What test? and hypothesis?

Answer 4

Augmented Dicky Fuller Test
delta yt = beta0 + alpha t + delta yt-1 + gamma1delta yt-1 + gamma
H0: Unit root
H1: Stationary
check for t-test, value should be negative and smaller than critical value

Question 5

Assumption of time series

Answer 5

1. Serially uncorrelated error (no relationship between the error term and the regressors)
2. a. Stationary ~ iid of OLS
   b. (Yt, X1t, …, Xkt) and (Yt-1, X1t-1, …., Xkt-1) are independent from each other like NO MULTICOLLINEARITY
3. No outliers