Exam 5 2008 # 26

[Buche]

On chapter 5 of Werner and Modlin,

They explain that the trend period ( One step trending) is "the lenght of time from the average written date of policies with premium earned during the historical period to the average written date for policies that will be in effect during the time the rates will be in effect.

On exam 5 2008 # 26, we have 2 CY : 2006 and 2007. We have annual policies.
So for 2006, the average written date of policies with premium earned during the historical period (2006) is supposed to be 01/01/2006. Is someone understand why the correct answer is 07/01/2006?

[Kongo]

I got stuck on this too.

I figured out that when you are 2 step trending, the first step is dividing AVG Written over AVG earned, Therefore you have brought all past experience to the AVG Written of the latest CY, and therefore answer will be 7/1.

In One Step Trend, you are trending AVG Earned, and therefore will be 1/1.

(Note, I have not looked up your specific question)

[Vorian Atreides]

Originally Posted by Buche

On chapter 5 of Werner and Modlin,

They explain that the trend period ( One step trending) is "the lenght of time from the average written date of policies with premium earned during the historical period to the average written date for policies that will be in effect during the time the rates will be in effect.

On exam 5 2008 # 26, we have 2 CY : 2006 and 2007. We have annual policies.
So for 2006, the average written date of policies with premium earned during the historical period (2006) is supposed to be 01/01/2006. Is someone understand why the correct answer is 07/01/2006?

Total time period of the historical period is from 1/1/2006 to 12/31/2007 . . . two full years.

The mid-point date of CY is the average earned date . . . which would be 1/1/2007.

The average written date will be 1/2 policy term before the average earned date . . . making it in this case to be 7/1/2006.


Alternatively, the policies providing the earned premium data for the two CY in question are written over a span of time ranging from 1/1/2005* through 12/31/2007--three years. The mid-point date of this would be the average written date; namely middle of 2006 or 7/1/2006.

*Technically, a policy written on 1/1/2005 would be completely earned by the end of the year 2005; but the difference of a day in this context is immaterial for the purposes at hand.

[Buche]

Your answer seems not to be consistent with the answer.

For CY/AY 2006, the trend period is 7/1/06 to 1/1/2010 (3.5 years)
For CY/AY 2007, the trend period is 7/1/07 to 1/1/2010 (2.5 years)

So the assumption that they use a 2 years experience seems not to be good.

[Vorian Atreides]

While the technicality of using average written date to calculate the trend period length . . . namely from avg written date in experience period to avg written date in the projected period . . . calculating that length using the avg earned dates (avg earned date to avg earned date as shown in the sample solution) results in the same length for the trending.

Using avg written dates:

1/1/06 to 7/1/09 ==> 3.5 years
1/1/07 to 7/1/09 ==> 2.5 years

Since it's the 3.5 and 2.5 that are the critical parts for the problem, making that determination using an equivalent method (dates that shifted by the exact same amount) is acceptable.

[Gyudon]

Originally Posted by Buche

Your answer seems not to be consistent with the answer.

For CY/AY 2006, the trend period is 7/1/06 to 1/1/2010 (3.5 years)
For CY/AY 2007, the trend period is 7/1/07 to 1/1/2010 (2.5 years)

So the assumption that they use a 2 years experience seems not to be good.

To put simply, I think your confusion is that the sample solution trends from the avg earned date to the avg projected earned date. As VA stated, trending from written to written instead will give you the same length.