A4. Mahler

Question 1

How to test risk parameter shifted over time?

Answer 1

Chi-Square Test:
* Test Statistic = sum[(Actual - Expected)^2 / (Expected)]
* if TS > value –> risk parameter has shifted over time

Compare Correlations
* For each separation in time (length t=1,2,3…), calculate the correlation between risk parameter and take average for each t
* Corr(X,Y) =… just use excel function
* If years closer together (length t=1,2,3…) are more correlated than years further apart (length t = 50,51…) –> risk parameter has shifted over time