September 2010 question 4ii

[zuglubuglu]

In discussing the pattern I said that

• if written on a claims-occurring bases
  • 
    [a] We would expect half of all claims to occur in first year of development and half to occur on second.
    We note that only 18.6% were reported by end of first year, so 31.4% have not been notified yet.
    [c] Even if we assume that all of these are reported by second year of development, shouldn't claims occurring in second year also have the same delay. However the second year delay is 100-87% = 13%.
    [d] this anomaly needs to be investigated.
    
  
  
  [*] The above might indicate that these are written on claims-reported basis.
  
  • 
    [a] possibly we are underreserving in first year.
    maybe it is a case such that policyholders need to submit all their PL claims by end of their policy and tend to do it at that point. All of the policies would end in the second year (unless written on 1 January).
    
  
  

Am I completely off track here? What do you think guys?

 

[LastHurdles]

I would assume it is written on a LOD basis since your given UWYR triangles.

A) would be correct if the business is written evenly over the year. However since experience tells us otherwise its probably not written evenly over the year.

B and C) the delays are due to reporting and settlement and in the first development period it is likely to be more to do with reporting delays and in the second development year it is likely to be 'more' due to settlement so the 13% you talk about is likely to be a moisture of reporting and settlement

That's my thought! Not saying its correct though

 

[zuglubuglu]

Yes I guessed I assumed business being written evenly which is probably not the case here!

Re B& C; If settlement delay is an issue, then wouldn't we be under-reserving? This would be a bit strange since there is over-reserving in the following two years.

 

[Darren Michaels]

When discussing the development pattern it is more important to think about the different types of delays rather than speculate on the basis of coverage.

It is important to remember with incurred claims triangles were are dealing with when the claims are reported not when they happen. So we need to think about reporting delays as well as event delays here too. Settlement delays also affect the incurred claims development because they effect when the case reserves become paid claims.

Also remember with an underwriting year triangle as in this case, new claim events may still occur up to the end of the second year (based on a policy written on 31 December in the underwriting year).

Therefore it would be very unlikely for the incurred claims triangle to be 50% developed by the end of the first year as you suggest.

 

[zuglubuglu]

I am finding this slightly confusing. Claim events for claims written in year X on underwriting year basis can occur up to end of year X+1.

Assuming no reporting or event delays, perfect reserving, and uniform occurrence of claims and sales of policies; wouldn't we expect half of the incurred claims for each of the first two years of development?

If so isn't it an anomaly that in this case only 18.6% are reported by end of first year while a staggering 87% by end of second year of development?

 

[Darren Michaels]

It is very unrealistic to assume no event, reporting or settlement delays. The question does not ask you to assume this, so you should not assume it when discussing the reasonableness of the development pattern.

Therefore I would not regard the development pattern as anomalous. In fact it looks perfectly plausible for public liability business.

Remember that the incurred claims pattern is about when claims are reported, not about when the claim events themselves happen, although of course they have to happen before they can be reported!