2012 Exam 8 Q4

downthesun01downthesun01
edited May 2024 in GLM.ModelSelection

Hi. Is what's being described in part b. not price optimization? It sounds like they're going to take the model result and use that to figure out a rate change that will result in at least 78% retention. I thought using models in such a way isn't allowed and put that as one of my answers against against the plan in part b.

Thanks.

Comments

  • adminadmin

    I could be wrong but in my opinion price optimization isn't occurring here. Price optimization would be if we had two insureds with identical rating characteristics that receive different rate changes based on our expectation of the likelihood of each shopping around for a better deal.

    Here, we're incorporating a new rating variable into the classification plan so now two otherwise identical insureds will receive different rates based on the number of calls they make.

    I think the bigger issues are:

    1. Relation to loss - is there a demonstrable link between the number of calls made and the loss? Probably not - it sounds like the number of calls affects the expenses rather than the loss. Expenses (other than maybe claims adjustment expenses) aren't generally included in the loss. So it's highly unlikely this variable will be predictive.
    2. Avoiding ambiguity - what counts as a call? Company initiated? Does a dropped call that calls back within 5 min count as 1 or 2 calls? Claims only calls or customer service calls as well, etc.

    I think you're perhaps going a level beyond what the question is asking by answering based on what the company might use the model for rather than addressing issues about including the number of calls in a rating plan irrespective of what that model would be used for. Keep it simple!

    p.s. Even if the company used it to keep retention above 78% by limiting the rate change, it doesn't mean they can't take multiple rate changes across several renewals to get to rate adequacy across all segments, thus avoiding price optimization concerns.

Sign In or Register to comment.