Fall 2018 #5
Hello,
Quick question, why is the midpoint of the 2 step trending July 1 2017 and not November 15 2017? the 2 step trending for losses as I understand follows the same logic as the 2 step trending for premium:
- Mid point of historical period = July 1 2015
- Midpoint of latest data available (we were given both annual and quarterly data but the quarterly data was used to find the trend so I assumed since the trend comes from quarterly data that's our reference) = Nov 15 2017
- Mid point of effective period + 0.5 x term = April 1 2020
If you are given both yearly and quarterly data, which are we suppose to use to then know the midpoint of the latest data available?
Comments
Actually, the data is sort of a mixture between annual and quarterly. According to the column heading "Year Ending Quater", each row in the table represents 12 months of data, but it's "rolling years". That means that at the end of each quarter, 3 new months of data are added at the end and 3 months of data are removed at the beginning. The last row in the table still represents 12 months of data, so it's probably more reasonable to use the mid-point of the year, which is July 1, rather than the mid-point of the last quarter.