Extreme Events – Specimen Question A.4.1

[this page | pdf | references | back links | custom searches]

Return to Question and Answer Summary


You are an investor trying to understand better the behaviour of Index B in A.2.1. You think that it is likely to be best modelled by an AR(1) autoregressive model along the lines of  with random independent identically distributed normal error terms .


(a)    Estimate the value of c 16 times, the first time assuming that you only have access to the first 5 observations, the next time you only have access to the first 6 observations, etc.




(b)   Do these evolving estimates of   appear to be stable? How would you test such an assertion statistically?




Contents | Prev | Next | Chapter Questions

Desktop view | Switch to Mobile