The last sentence of paragraph 3 claims that paid development ratios will be higher for early development periods for 2010/11 due to faster claims handling. I thought that faster claims handling would lead to more paid claims in earlier development periods, leading to lower development factors? Simple example using cumulative paids: 10 15 18 19 20 (Original claims handling) eg. Df 1->2 =1.5 13 18 20 20 20 (New claims handling (getting to ultimate faster)) eg. Df. 1->2 = 1.38 Am I being silly?
I wouldn't say that's silly. I just think you're confused by the profiles you've chosen, because so much of your development occurs in your first year. Take a look at the attached. The development factors for the short-tail class will be higher until about time t (because the gradient is steeper), but lower after time t (where the gradient is more shallow). It's always difficult to "fabricate" a profile to fit a theory. Better to compare some real classes from work.
If the gradient is steeper then the development factor would be lower no? As the graph you showed is the % developed and the % dev = 1/cdf. So would you not expect lower CDFs because we are closer to ultimate quicker with a new claims handling process? The only reason I can think why you expect higher development factors is if you look at it monthly and see the individual development factors you might see an uptick, but comparing the 12-24 cumulative development factor with and without the new claims handling process surely it should be lower after the claims handling process?
I wouldn't get too hung up on this Phos2, it's easy to 'fabricate' a story depending on whether you're looking a factors to ultimate, or % developed, or link ratios. Broadly speaking though, if the graph of % developed is steep, then claims are developing quickly at that 'age'. Similarly, if individual link ratios are high, then claims are developing quickly.
