Hi Paul
I am a bit puzzled by this
At this stage,a nyway, we are not looking at th validity 9or otherwise0 of the calculations of the lifetime costs.
But we are trying to get a clear picture of the over all landscape of motorcycles claims
Now, you say that you used the matched claims to establish relativites.
But the matched claims , expecially those which have a CC rating , form only a small proportion of the total - less than one thrid. And they are a selected subset, not a random selection. Only those crashes serious enough to come to official police attention will be included (and not all of them)
So I do not see how that data can give an undistorted view.
I am also puzzled by your comment that lifetime estimates are not performed on all claims. Is this because trivial claims it is not sensible to go the effort? if that is the I have no issue with it, but surely then the lifetime cost would simply be the actual cost in the year of occurrence, a very simple matter to determine.
If it is because lifetime estimates are done only for selected serious claims then that opens the question of how those claims are selected.
I am also still unclear why the provision of a full list of motorcycle crash claims should be difficult . It is quite obvious that the matched claims were generated by taking the recordset of all claims and joining to the recordset of police records. Those that matched, are the matched claims. In which case the original primary recordset is, by definition, the list of all motorcycle claims.
It is not a big deal to us, we can work around it . But as a BI analyst myself, I am puzzled at the poverty of data available within such a large organisation.
In passing , also, I fail to see how it is possible to establish a risk rating for a vehicle. The vehicle itself is neither risky or riskfree. It all depends on how it is ridden or driven. Leaving the driver/rider out of the equation seems to me to make the result meaningless.
Best regards
Les
On Thu, Nov 5, 2009 at 10:21 AM, Paul Gimblett
wrote:
Hi again Les,
I wanted to say a few words about this statement "OK. So you are saying that the remaining claims are not taken into consideration at all. Is there a data source that shows ALL motorcycle (in particular, though other vehicles being included would be fine) claims, whether matched or not. We are looking at the entirety of the motorcycle cost landscape not just that relating to different capacities"
If you review the technical document provided previously you will see that all claims experience to date has been used to build the estimate of life time costs for new claims lodged in 2010/11. As far as building the relativities which determine how the required levy is distributed across the various vehicle classes this year we used the matched claims. Previous levies have been based on one-off pieces of analysis using (often) external data to assess the risk profiles of different vehicles. The relativities are to produce an equitable distribution of the required levies rather than reflecting the true risk profile of the vehicles involved.
I have had a chat to the actuaries and there is no data set that will provide what you are after for two reasons: first lifetime estimates are not performed on all claims, and second the data collected by ACC at the time of claim lodgement is primarily targeted at providing case and claims management information rather than for use in risk rating or levy setting which is why we have started to use the MOT and vehicle registration data to enhance our data and allow it to be used to calculate relativities.
| |
| | | |
| | Paul Gimblett, Product & Scheme Manager, ACC
Tel +4 918 7554 / Fax +4 918 7351
|
Reply With Quote
Originally Posted by Reckless
Bookmarks