The following is from Ch13 p10. Aren't they basically saying the same thing that claims from a underwriting year cohort span over 2 years or more? 1. The disadvantage is that it will take more than one year before all the claims under that cohort have occurred (for example, if all policies are of a maximum period of 12 months, then it may be up to two years from the beginning of the underwriting year before all claims have occurred). 2. Claims in a particular cohort will also have occurred over a wider risk period than a single accident year cohort. 3. The claims will generally relate to at least two accident years.
I am tending to agree with you here. The only thing that I can think of which is a distinction is that in the first point they're saying that it takes longer to find out what claims will have occurred and in the second point they're focussing on the idea that the risk environment (e.g. how much the claim could be) will vary due to different conditions in the two accident years. The first point means that you're going to have to wait longer to get stable results and the second means that when you try and rationalise movements in reserves it's going to be more complicated because you're going to get "underwriting year" effects which are harder to explain than "accident year" effects. They're making a poor attempt at explaining whatever point they're trying to make anyway. Note also that just because it's on an underwriting year doesn't mean you have to wait two years to find when all the claims have happened. If you take the example of a travel insurance company which is reserving on an underwriting year it's likely that most claims will occur by the end of January rather than having to wait a whole extra year. On the other end of the spectrum, an insurance company which sells only construction insurance could have underwriting year cohorts spanning 10+ years!
They are certainly related points. However the key thing to understand with an underwriting year cohort is that it can be up to 2 years from the beginning of the underwriting year until all claims from the cohort have happened. For example, consider an annual policy written on 31 December 2014, this will belong to the 2014 underwriting year. In theory claims under this policy can happen right up until the end of 2015, which is two full years from the start of 2014. Because of this claims from the 2014 underwriting year happen in both the 2014 and 2015 accident years where the risk environments might be quite different for whatever reason. With an accident year cohort all claims by definition come from the same accident year.
