PDA

View Full Version : Marker/Mohl: Future PP Projection

Examinator
03-02-2005, 05:33 PM
This is the one section of this paper that isn't clear to me (beginning on page 22 of the text). I understand the general concept, which is determining the mature PP, then distributing back to less mature lags. The actual process of doing this at first seems straight-forward and logical, but I get lost in it once I try to understand it completely. They surely won't ask a candidate to perform any regression on this exam, but I'd like to understand the process for those pesky T/F questions. This is what I gather.

In the first step (given that the state's trend is not credible), you calculate CW PPs using regression. What exactly are you regressing to get these?

Secondly, you regress these CW PPs with the state PPs in order to get a slope (and intercept, if desired) which assumes that the state and CW trends will be similar to one another with the same percent change from year to year. This step seems reasonable. But where exactly is the mature PP calculated? Is it the slope (and intercept) applied to the state's PP for each year?

Distributing the trended mature PPs backward also seems a little laborious to ask on an exam, but asking about the proceses should be fair game. Should it be enough to be able to regurgitate the three approaches listed and why they should(n't) be used? I don't recall any past exam problems testing the mechanics of this section of the text. Thanks.