Schedule P anomaly

[fargo]

I am stumped and need to go to the experts on this one. How can you get a negative ceded loss ratio? I understand negative IBNR is not uncommon, but there are no ceded paid losses. The ceded DCC looks bizarre in isolation.
Here is a table (with rounding to protect the innocent) from the 2005 year of Schedule P. Any ideas??? Please see the attached truncated summary if you care to comment.

Other notes:
Mostly Claims Made - Other Liability
No commutations mentioned in Notes - or anything else unusual
No LPT or financial reinsurance mentioned in latest actuarial opinion
No ceded A&O

[E. Blackadder]

You could check the terms of the reinsurance treaty. :popcorn:

Just to emphasize how stupid the annual statement is, I once worked with a firm that broke out LOBs in a way that we had negative direct for a particular line. The Insurance Department called us, and asked about a different number on the same page. :lol: Maybe they were being subtle.

[Westley]

Just a guess:

A three-year policy that has negative DCC IBNR based on DCC case in two of the years. The person preparing didn't know how to spread it, and just said "screw it, I'll split it evenly".

[Arlie_Proctor]

There are several data anomalies that might cause this, but I think the likely cause is the allocation of ceded IBNR for Schedule P reporting purposes. Most companies don't conduct their reserve reviews using strict Annual Statement Line definitions; frequently smaller lines with similar development are combined together. IBNR is then allocated to each line by year to fill out Schedule P. In that process, you sometimes get strange results in a few cells, especially with ceded data.

[Morrison]

Maybe an intercompany calendar year (not accident year) based re-insurance contract?

[fargo]

Some additional info:
No intercompany pooling per the notes, although they own some smaller companies.
Only one LOB - Claims Made professional liability with a small tail component.
2005 reinsurance treaty is flat-rated excess of loss.
Policies are assumed to be one-year.

Thanks to everyone for your input.

On first pass, it looks like the actuary may have calculated D&A ultimates, then Net ultimates, then backed into a ceded value. If development doesn't line up consistently then one can get negative ceded amounts.

[joeorez]

Just a thought: Note that losses are accident year and premium is calendar year. Suppose after year-end, the premium for a July 1 treaty turned out much different than originally expected. If you could adjust the prior calendar year premium you would, but it's calendar and you can't. Couldn't there be a negative adjustment on the losses corresponding to the negative adjustment you would have liked to make on the premium?