Consideration for using net vs. gross and ceded data (DC-Re)

Hi, why under Data availability, if data is sparse, it may not be possible to directly estimate the present value of ceded cash flows? I am not quite getting what sparse mean here and why ceded CF only?

Thanks,
Wilson

Comments

  • Ceded = Gross - Net. If your net is really close to your gross amounts, then you could have 0 or close to 0 data for certain periods. This makes it hard to directly estimate the PV of ceded cash flows. You have much more data in each period with gross and net so it makes sense to do estimate gross and net then do ceded = gross - net

  • edited February 2023

    One more question, if reinsurance program has changed significantly wouldn't that make both net and ceded losses equally inappropriate?

    Note that the solution of 2015Fall-26 says net losses are actually preferred in this scenario, which contradicts the wiki?

    Similarly for discount rates I didn't understand why ceded is preferred if both ceded and net are different

  • If reinsurance program has changed substantially, you should use gross and net. You can retroactively apply the new reinsurance terms on the gross triangle. I think what the wiki is mentioning here is that you cannot use net triangles directly, which is true. For discount rates, technically the discount rate for net should be a blend of gross and ceded, which you do not have which means it is better to calculate gross and ceded separately

  • Hi @Staff-T1,

    I am also confused by this. If we have a change in reinsurance program, what method do you recommend and why?

    Clearly the solution and wiki contradict. I did not fully understand what your answer was above

    Thanks!

  • As I mentioned above, gross and net. It is easier to restate the net triangle by applying reinsurance terms to the gross triangle. Directly developing the ceded is difficult when you have changing limits as the ldfs will not be stable given there is only small amounts of data. That's basically the gist of it.

Sign In or Register to comment.