Quote Originally Posted by beacon View Post
And I'd do the same for the separate PP dataset beginning from the date of the first PP loan.
Have done the PP one - see note on side of graph.

I'm tempted to just go with all loans from start to current, the watering down effect is likely to be minimal and one would expect it would effect everything by the same proportion (assuming similar loan characteristics throughout the year)?

Happy to go with the flow with this, just don't want to over complicate things too early