BF and LDFs

[Sox34]

When you are using the BF approach, will we always have to calculate the LDFs as the article does in Exhibit B - i.e. using three year data and looking at the average beginning amount and ending amount, and the LDF = quotient of the two? This approach seems counterintuitive to me especially when a problem also gives us the Age-to-Age LDFs.

[The Sad Man]

I think if the LDFs are already given you won't have to calculate them from the data again? I'm not sure I understand your question. Can you provide an example?

[Sox34]

I'm not sure if you have the All 10 manual or not... but the sample question on page 9 of volume 1a is a good example. We are given triangular incurred data and Age-to-Age LDFs, and instead of using some average of these age-to-age LDFs to get a selected LDF, we are supposed to use the approach that the BF article uses in Exhibit B of three year average beginning and ending amounts with the LDF equal to the average ending amount/average beginning amount.

[Sox34]

maybe I shouldn't think too much about it, but I guess my main question is if we are given triangular incurred data in a problem and told to find the BF IBNR estimate, should we 1. calculate the age-to-age LDFS and use an average of those LDFS or 2. follow the method of Exhibit B using the three year beginning and ending amounts and calculate LDFs that way. Personally, my natural approach would be to do 1. but the article explains method 2. so I'm just wondering if it's a black and white situation. As in - if the question says "BF" method then you use 2. to find LDFs regardless of if the question also shows the age-to-age LDFs.

I'm sorry if i'm not articulating my question very well - please let me know if this is still confusing...

[Sox34]

Ok well I solved part of my problem.... I was getting thrown off partly by the BF article's notation in Exhibit B. Essentially it's just a three year weighted average approach. Again though, shall we always use a three year weighted average in selecting LDFs when we do BF questions?

[SouthOfSanity]

I think there is somewhere in the B-F's paper that says that the 3-year average was selected purely for illustrative purposes, and I think the authors did mention in the paper that the LDF selections in the real world are not going to be as straightforward as used in the paper. I do not think it makes sense to restrict our LDF selections to one method.

[2M]

Unless a question specifically says use a specific averaging method or references a paper that uses a specific averaging method, it appears to me, that the averaging method used is one of judgement

[The Sad Man]

It's definitely a judgement call unless stated otherwise. If they don't say anything during the exam, I'd just use what the author of that particular paper decided on.

[SouthOfSanity]

It's definitely a judgement call unless stated otherwise. If they don't say anything during the exam, I'd just use what the author of that particular paper decided on.

I would do the same too.