A case where IBNR cannot be negative

[Pacted]

Hi all,
I am quite new to general insurance, and encountered negative IBNR at work. Upon internet research, I learned that IBNR can be negative due to things like salvage and subrogation (non-reinsurance recoveries).

My colleague mentioned to me that when IBNR is calculated from claims incurred to reported (as opposed to incurred to paid-- both of which are incurred claims, apparently), without taking into account salvage or subrogation, it can never be negative. I am failing to wrap my head around this.
1. I understand the difference between paid and incurred claims. This incurred to reported vs. incurred to paid are both apparently incurred claims, with the difference to do with their dates (columns of the triangle).
2. How can IBNR be negative (I understood it as when the case estimates are too high the IBNR is used as a correction term etc. Is this true?) and how can it NOT be negative?

Any help would be appreciated. Thanks in advance!

 

[Katherine Young]

Hi Pacted,

With regards to point number 2:

IBNR can be negative for any number of reasons, the most significant probably being when claims settle for less than their case estimates. Other reasons could include salvage, subrogation, recoveries from other third parties (such as other insurers for example), etc.

With regards to point number 1:

When you say reserves are calculated "from claims incurred to reported" or from "incurred to paid", these are not standard terminology, and really don't make much sense. Do you mean you have calculated a triangle of the ratios of incurred claims to paid claims?

Also, generally, "incurred" and "reported" claims are often taken to mean the same thing, ie the sum of paid claims + "reported but not yet paid" claims (although in practice the term "incurred" is used much more often than "reported"). Therefore, your notion of "incurred to reported" really doesn't sound normal to me at all. (Conceivably your colleague is including IBNR in your definition of incurred, but this is quite unusual.) Are you sure you have understood what your colleague said to you?

In any case, I think we're getting dragged into unnecessary detail here. In short:

• It is quite common for a triangle of incurred claims (ie reported claims, or "paid + outstanding reported claims") to produce negative IBNR, particularly on the earlier (more developed years), if the case estimates are prudent. For the later (less developed) years, the effect of prudence in the case estimates will be more than offset by claims that haven't yet been reported to the insurer, so you would be much less likely to see negative IBNR here.
• It is much less common to see negative IBNR (or at least, large negative IBNR) when using a triangle of paid claims, although it is still possible, as you suggest, due to salvage / subrogation.
• How can IBNR not be negative? I'd say there's always a chance it can be negative. Don't be surprised by it.

Pacted, this is way way beyond CT6 level. If you have any further questions, I'd recommend you post your discussions on the ST7 forum.

Best wishes,

Katherine.

 

[Pacted]

Katherine,
Thank you so much for your detailed response, and apologies for my late reply. I did suspect that the terms 'incurred to reported' or 'incurred to paid' were unconventional (as I have only read about incurred (aka. reported) vs. paid).