April 2013 q9

[George88]

The solution applies severity inflation of 5% to the losses. Isn't this ground up claims severity? Don't we have to take account for the limits either gearing inflation at the bottom or capping it at the top?

I tried to use adjust the ilf curve to adjust the losses which was barking up the wrong tree. Do you ever need to use that inflation adjustment to an ilf?

 

[Katherine Young]

Isn't this ground up claims severity? Don't we have to take account for the limits either gearing inflation at the bottom or capping it at the top?

The question makes no mention of a stability clause (in fact I wouldn't expect it to, seeing as this isn't a reinsurance question). Therefore I don't think there's any need to inflate the XS or limit. We simply need to inflate the past claims to allow for the fact that the past data is smaller than it would be in today's terms, ie to allow for the fact that today's claims would be more likely to breach these limits.

The method followed by the examiners is to:

• Use the ILFs to get the hypothetical claims cost, assuming the 2013 XS and limit had been in place.
• Then inflate these to 2013 values.
• Then develop losses.

I tried to use adjust the ilf curve to adjust the losses which was barking up the wrong tree.

I'm not sure what you mean here I'm afraid. You do need to adjust the past claims using the ILF, to work out what the cost would be under today's XS and limit.

I wouldn't go adjusting the ILF for inflation though, because you'd have to do that for each year separately, which is a bit of a pain.

Do you ever need to use that inflation adjustment to an ilf?

Quite possibly. Have a look at April 2010 Q5 part (iv).

 

[George88]

Hi Katherine,

Thanks ill have a go at that question

i didnt mean that there would be a stability clause - i meant that the notified claims are net of the policy excess and per claim limit, so wasnt sure whether the severity inflation would be applicaple for some claims, for example those that have already hit the limit (and so couldnt get any large in the light of inflation)

Thanks,

George

 

[Katherine Young]

The question says the ILFs are applicable without the need to adjust for trends, so luckily we don't need to worry about that!

 

[tatos]

Hi Katherine

In the examiners' report, it mentions that only 50% credit was given for making the assumption that claims are incurred at the midpoint of each policy year. Why is that?

I thought that when we converted exposures to policy year exposures to get correspondence with the claims data we were already assuming (although it wasn't stated in the solution) that claims are incurred at the midpoint of the policy year. And that's why we needed exposures at the ends of calendar years.

Instead the full credit assumption is that claims in each policy year are paid / incurred on average at the same time relative to the start of that policy year. I don't see why or how this would actually change the calculation? Wouldn't the time period between any policy year and the year we are projecting to, just be the same as the time period under a midpoints assumption?

Also what is the relevance of telling us that claims notified are " as at 31 March"

Thanks

 

[Katherine Young]

Please don't direct your questions to individual tutors, it discourages other people from responding. This forum is a place for students to help each other, not a place to get personal tuition for free. See https://www.acted.co.uk/forums/index.php?threads/guidance-on-posting-queries.8779/#post-33197

I thought that when we converted exposures to policy year exposures to get correspondence with the claims data we were already assuming (although it wasn't stated in the solution) that claims are incurred at the midpoint of the policy year. And that's why we needed exposures at the ends of calendar years.

Converting exposures to correspond to the claims data has nothing to do with the inflation assumption. It's to makes sure we know how many solicitors are capable of giving rise to each claim figure.

Instead the full credit assumption is that claims in each policy year are paid / incurred on average at the same time relative to the start of that policy year. I don't see why or how this would actually change the calculation? Wouldn't the time period between any policy year and the year we are projecting to, just be the same as the time period under a midpoints assumption?

It doesn't change the calculation, it just changes the logic behind the calculation, that's why you do get some marks for it. The point is that, so long as the claim delays are the same in each year, the year-on-year inflation adjustment will then be one whole year each time. So, eg, if claims always occur at the end of a policy year, the year-on-year inflation increase will be for one whole year each time. But if, say, claims occur at the start of a policy year in 2011 and the end of a policy year in 2013, the inflation assumption for 2011 will be more than 2 years, rather than exactly 2 years. In short, changes in earnings patterns will affect the inflation calc.

Also what is the relevance of telling us that claims notified are " as at 31 March"

Not a lot. It tells you you don't have to adjust the development profile, since this is also as at 31 March.

 

[o.menary11]

Hi,

How has linear interpolation been used to find the adjusted exposure (number of solicitors) ? and what clues in the question point to such an adjustment is needed?

Thanks

 

[Katherine Young]

What clues in the question point to such an adjustment is needed?

The claims data is by policy year whereas the solicitors data is by calendar year. So you need to adjust the solicitors data to a calendar year basis.

It's a golden rule: Claims data and policy data always need to be on the same basis.

How has linear interpolation been used to find the adjusted exposure (number of solicitors) ?

Let's look at the 2008 policy: It runs from 1st July 2008 to 30th June 2009. In other words, it has exposure from half of 2008 and half of 2009. So the exposure for the 2008 policy year is 0.5*210 + 0.5*208.

Now apply the same method for the later years.