CH 17-Friedland

[Kongo]

Taken from Friedland:
"
Challenges of the Classical Technique
Recall that the key assumption underlying the application of 50% of the ULAE ratio to case
outstanding and 100% of the ULAE ratio to IBNR is that 50% of the expenses are sustained when
opening the claim and the remaining 50% when closing the claim. One challenge with the
classical technique is that “closing” a claim and “paying” a claim do not necessarily mean the
same thing. For some lines of business, a single payment may be the norm, and thus, such
payment may in fact represent settlement (i.e., closure) of the claim, and therefore the end of the
claims handling activity. An example is glass coverage to replace a shattered windshield under
automobile physical damage insurance. (Note, not all automobile physical damage insurance can
necessarily be categorized as single payment where payment equates to closure of the claim.) An
example of insurance where a claim payment and closing of the claim often differ is U.S. workers
compensation; for this coverage, regular payments can replace lost wages for an extended period
of time.
Some actuaries address this challenge by adjusting the percentages applied to the case
outstanding and the IBNR to reflect their expectations for the particular company. For example,
an actuary of an insurer with a portfolio of long-tail professional liability coverage, which is
characterized by very long-tailed liabilities and substantial claims-handling work during the life
of the claim, estimates unpaid ULAE assuming ratios of 25% applied to case outstanding and
75% to IBNR, which includes development on case outstanding. Thus, they assume a greater
proportion of the expenses are related to closing the claims rather than opening claims."

Can someone explain this to me. If we are assuming that there are more costs involved in closing the claim, shouldn't the % applied to Case Outstanding be greater than 50%, why does it go down to 25%. Same thing with IBNR, shouldn't it be 125%, why does it go down to 75%

Also-Did anyone hear of an ETA when TIA will post the video on ch 17? Seems that there is only 1 new video to do, you would think it would be done.

Thanks.

[Victor60]

Vorian,

Do you want to take a stab at it?
It seems to be a strange statement...

[Vorian Atreides]

I think the biggest key here is to keep in mind that we're estimating the total unpaid ULAE. Given that r = ULAE ratio to ultimate claims, applying r to a "basis" will provide the estimated costs.

The "basis" is what "unpaid claim" amount is used to measure the unpaid ULAE. (Note that this is also the key to looking at Kittel's refinement, IIRC.) One assumption (that's not made very explicit) of the classical method is that case outstanding (OS) is associated with claims that already have a payment associated with them. This assumption isn't very important to understanding the development of the classical method, but it's likely to be pivotal in understanding the second point.

In the paragraph in question, I will posit the following scenario that does result in what's stated there. (This is the result of my understanding of the rest of the chapter.)

Assume that 25% of ULAE is for opening claims, 25% is for maintaining claims (opened claims that will have activity, but will not close in the time frame of the analysis), and 50% is for closing the claims.

Keep in mind the comment about IBNR--namely that it's not often "well defined" on an industry level. If I'm dealing with a claims-made book, my "IBNR" will be purely development on OS and won't include "new claims" (that would incur opening costs).

So, my basis for estimating the allocation of ULAE expenses between "opening", "maintaining", and "closing" are going to be Paid Claims, total unpaid claims (OS + IBNR), and IBNR, respectively. (Note that one could choose different "bases" than these--just that someone provided a reasonable argument as to why this would work out this way. In any event, I wouldn't get hung up on the validity of this selection--accept it and pass the damn test so you don't have to worry about any more!! :-P)

So, my estimate for unpaid ULAE will be:
0.25 * r * (OS + IBNR) + 0.50 * r * (IBNR) = 0.25 * r * OS + 0.75 * r * IBNR.

Hope this helps.

[Rice]

Here's how I see it:

Case outstanding refers to the part of the reserve that has already been determined. That is, much of the adjusting and reserve analysis of the claim has already been done and thus much of the ULAE has already been paid, the work being in the past, leaving a smaller amount still to be distributed.

IBNR refers to the part of the reserve that has yet to emerge. It relates to future claims work that has not yet been performed. Given that the claim in question is in a long-tail line, this represents a lot of work still to come, and thus a larger percentage of ULAE is "allocated" to it.

So we get something like 25% for case and 75% for IBNR.

You can't have 50%+ for case outstanding and 125% for IBNR. That makes a total of 175%. You can't allocate more than 100% of something. The question should be something like "How do we divvy up the ULAE dollars across these certain claim functions?" We are splitting a pie. Only the government can redistribute more than it has. :)

[Vorian Atreides]

You can't have 50%+ for case outstanding and 125% for IBNR. That makes a total of 175%. You can't allocate more than 100% of something. The question should be something like "How do we divvy up the ULAE dollars across these certain claim functions?" We are splitting a pie. Only the government can redistribute more than it has. :)
But, under the classical method, 50% is applied with OS and 100% is applied with IBNR. Based on the bolded argument, this would be 150%. Care to elaborate on that?

Note that the "25%" of OS and "75%" of IBNR is not an allocation of ULAE, it's how one determines unpaid ULAE. It's an unfortunate coincidence that 25% + 75% were used here.

Also, the government can't redistribute more than it has . . . but it has the power to collect "advance" payments from future taxes to pay for their pork barrels and free-loaders. :-P

[Kongo]

Rice- I think, as VA comments, your missing the point.

There is 50% applied to Case OS and for IBNR (Pure IBNR) is 100%, meaning a different case, adding them together means you should reread the section because that is not what is being discussed. I still don't understand why in the world it goes down to 25%.

Anyone have any experience shooting a letter to Friedland? This is a new chapter and hasn't been tested b4 so I wonder if anyone would have picked up on this errata.

[Vorian Atreides]

Rice- I think, as VA comments, your missing the point.

There is 50% applied to Case OS and for IBNR (Pure IBNR) is 100%, meaning a different case, adding them together means you should reread the section because that is not what is being discussed. I still don't understand why in the world it goes down to 25%.

Anyone have any experience shooting a letter to Friedland? This is a new chapter and hasn't been tested b4 so I wonder if anyone would have picked up on this errata.
While it's a new chapter, it's based pretty much on the previous material that covered this topic. It was available last sitting (albeit not formally part of the syllabus); I spent some time reconciling the info between the previous material and didn't find much different (apart from terminology--"claims" vs. "losses"--and specific notation).

And I think the piece that you're missing is that the use of "25%/75%" should not be interpreted in light of the classical method; but rather one should reframe the classical method in view of Kittel's refinement (which was the attempt I tried to make in the previous post).

[Gyudon]

VA, correct me if I'm wrong but I think you mean 25% to open, 50% to maintain, 25% to close, or 50% would be applied to O/S to close it. They're talking about long-tail lines (like PL in the second paragraph) where there are substantial claim maintenance costs so this would make more sense. I also agree that IBNR needs to exclude pure IBNR like in claims-made PL so that the cost of opening is excluded in IBNR (hence the 75%).

At the end of the day though this is arguably wrong too since how can one assume O/S has no maintenance costs only closing costs, i.e. assuming case is only put up for claims about to close soon without significant further work to be done. All this is the unfortunate side effect of trying to approach actuarial science as a real science, which leads back to your original comment of just past the damn test :).

[Vorian Atreides]

No, I meant what I wrote earlier; however, the situation you present below could also work as well.

In the situation I described, the IBNR also included development of the OS; so the view there is that the OS losses aren't associated with closing those claims, just maintenance. ULAE associated with closing the claims is associated only with the IBNR. (Note that in WC, IBNR can be substantially greater than OS early on in the development.)

(Also note that IBNR in that situation was also part of the basis for the maintenance.)

In your situtation, the view would be that the OS needs only to be closed (not maintained) while the IBNR would need to be maintained plus closed. Nothing wrong with this situation (and it would generate the same 25/75 application mentioned in the reading); however, it wouldn't satisfy the aspect that "payment on a claim doesn't necessarily close that claim" (as you recognize in your post). The situation I proposed above was attempting to account for all aspects of what was presented in the paragraph(s) quoted by the OP.

I'll reiterate what I think is critical for this chapter (especially wrt the generalized Kittel method): Identifying what basis is being used to determine the unpaid ULAE estimates.

VA, correct me if I'm wrong but I think you mean 25% to open, 50% to maintain, 25% to close, or 50% would be applied to O/S to close it. They're talking about long-tail lines (like PL in the second paragraph) where there are substantial claim maintenance costs so this would make more sense. I also agree that IBNR needs to exclude pure IBNR like in claims-made PL so that the cost of opening is excluded in IBNR (hence the 75%).

At the end of the day though this is arguably wrong too since how can one assume O/S has no maintenance costs only closing costs, i.e. assuming case is only put up for claims about to close soon without significant further work to be done. All this is the unfortunate side effect of trying to approach actuarial science as a real science, which leads back to your original comment of just past the damn test :).

[peem]

Using VA's terminology, usually, under the the Classical Method, the basis is calculated as
= 0.50(Case O/S + IBNR) + 0.50(IBNR)
= 0.50(Case O/S) + 1.00(IBNR)

Theoretically (pg. 389 Friedland), Classical Method basis should be
= 0.50(Case O/S + IBNER + IBNYR) + 0.50(IBNYR)
= 0.50(Case O/S) + 0.50(IBNER) + 1.00(IBNYR)

Based on older past exams (See 2002 #3 and 2003 #1), if you are given a “opening” or “initial cost” % and asked to use the Classical Method, it’s calculated as
= (ratio)[(1 – opening %)( Case O/S + IBNR) + (opening %)(IBNR)]
= (ratio)[(1 – opening %)(Case O/S) + 1.00(IBNR)]

I can see why Konger is confused or wondering why 75% isn’t applied to Case O/S (because applying 25% to Case O/S is equivalent to saying that 75% is the “opening %”).

Assuming Friedland really did mean to apply 25% to Case O/S and 75% to IBNR, my thinking is that it is an adjustment to the Theoretically-based formula where the adjusted basis is calculated as
= 0.25(Case O/S) + 0.75(IBNER) + 1.00(IBNYR)
= 0.25(Case O/S) + 0.75(IBNR) + 0.25(IBNYR)

The last term accounts for the opening ULAE costs on IBNYR.

[Vorian Atreides]

Using VA's terminology, usually, under the the Classical Method, the basis is calculated as
= 0.50(Case O/S + IBNR) + 0.50(IBNR)
= 0.50(Case O/S) + 1.00(IBNR)

Theoretically (pg. 389 Friedland), Classical Method basis should be
= 0.50(Case O/S + IBNER + IBNYR) + 0.50(IBNYR)
= 0.50(Case O/S) + 0.50(IBNER) + 1.00(IBNYR)

Based on older past exams (See 2002 #3 and 2003 #1), if you are given a “opening” or “initial cost” % and asked to use the Classical Method, it’s calculated as
= (ratio)[(1 – opening %)( Case O/S + IBNR) + (opening %)(IBNR)]
= (ratio)[(1 – opening %)(Case O/S) + 1.00(IBNR)]

I can see why Konger is confused or wondering why 75% isn’t applied to Case O/S (because applying 25% to Case O/S is equivalent to saying that 75% is the “opening %”).

Assuming Friedland really did mean to apply 25% to Case O/S and 75% to IBNR, my thinking is that it is an adjustment to the Theoretically-based formula where the adjusted basis is calculated as
= 0.25(Case O/S) + 0.75(IBNER) + 1.00(IBNYR)
= 0.25(Case O/S) + 0.75(IBNR) + 0.25(IBNYR)

The last term accounts for the opening ULAE costs on IBNYR.
Agree with this. Plus, if IBNYR = 0 (which would be the case if IBNR is only development on OS) . . . well . . .

[Gyudon]

Fair enough, though I've never heard of ULAE only being expended to close claims with insufficient case, I suppose it's possible. Consider an example where the vast majority of claims have sufficient case reserves (I realize that's unlikely). That means it includes future payments including loss payments to close. We are assuming no ULAE is expended to close these claims. Only those claims with insufficient case (IBNER) will have ULAE associated with closing.

[Kongo]

:bump:

Anyone listen to the recently posted TIA lesson, on this chapter.

I could've sworn Ken said the opposite then said in Friedland?