Inflation adjusted chain ladder Assumptions

Discussion in 'CT6' started by kaysee, Nov 27, 2014.

  1. kaysee

    kaysee Member

    Hi,

    the assumption underlying the method are:
    1) for each origin year, the amount of claims paid, in real terms, in each development year is a constant proportion of the total claims, in real terms, from that origin year.

    Can someone explain this assumption to me? I don't understand what it's talking about at all :/
     
    Last edited by a moderator: Nov 27, 2014
    kaysee, Nov 27, 2014
    #1
  2. Hemant Rupani

    Hemant Rupani Senior Member

    After solving problems so, I got the statement as
    "for each origin year, the amount of cumulative claims paid, in real terms, in each nth development year is a constant proportion of the total cumulative claims (n-1)th development year, in real terms, from that origin year."
     
    Hemant Rupani, Nov 28, 2014
    #2
  3. Katherine Young

    Katherine Young ActEd Tutor Staff Member

    I've knocked up an example to talk you through this. Have a look at the attached document.
     

    Attached Files:

    Last edited: Nov 29, 2014
    Katherine Young, Nov 29, 2014
    #3
    Deepesh likes this.
  4. Hemant Rupani

    Hemant Rupani Senior Member

    Hi,
    I don't know if I correctly get the meaning of the assumption, do I?
    Let's consider origin year 2009, now for getting cumulative claims for 4th development year.
    As per the assumption, it will be 1088*1.73*1.53*1.09*1.027=3223
    But it is 3119*1.027=3203
     
    Hemant Rupani, Nov 29, 2014
    #4
  5. Katherine Young

    Katherine Young ActEd Tutor Staff Member

    In a perfect world, if claims experience emerges exactly as expected, you'd be right.

    Inevitably though, the real world isn't that neat and tidy. There will always be randomness present, which means that the model isn't a perfect fit.

    You can see this by noticing the variation in each column of ratios. If the model were a perfect fit then every ratio in a given column would be exactly the same.

    So instead of projecting the triangle using the first column of figures, we take the most recent data available (which as you point out is 3119). There are two good reasons for doing this.
    • It is fact after all, it's actually happened, so there's no point ignoring it.
    • It includes more information (ie 4 development periods of data, rather than just 1), so it will be less volatile and more credible.
     
    Katherine Young, Nov 29, 2014
    #5
  6. Hemant Rupani

    Hemant Rupani Senior Member

    Thanks!:)
     
    Hemant Rupani, Nov 29, 2014
    #6

Share This Page