Post by GSV3MIaC on Oct 19, 2015 10:01:54 GMT
I just ran the calculator over the loan book, and assuming an exit cost of 0.25% and half a month's interest (which is now a fixed-ish value - I used the 60 month rates) the month at which the average loss %age is more than the exit cost %age, came out as follows (use at your own risk, E&OE!)..
1) A/A+ property loans, interest only ... never (but might be smart to exit before full term!). Nothing longer than 15 months so far anyway.
2) A+ overall 18 Months, A+ WL Rejects 7 Months
3) A overall 18 Months, A rejects 4 Months
4) B overall 9 months, B rejects 7 Months
5) C or D overall 8 months, C or D rejects 6 Months
6) E overall >3 months, E rejects >2 months, but no statistically significant data (yet).
This assumes failure rates for B are always at least as bad as A (although loss of interest from bailing out is more, so you can stand slightly more expected loss .. hence the 7 months).
A few categories have very little data, and I still don't have a good set of early repayment dates, so loss %ages are probably worse than stated (because money I have assumed to be 'at risk' in month N had actually been repaid by then, but you can't figure that out from the loan book).
So yeah, OK, 6 months looks fairly safe, apart from some WL rejects which go pear shaped rather faster.For a quiet life, interest only property loans, especially with a cashback, look fine (and will continue to do so until the great bubble bursts?). I suppose I should look at 'excluding known WL rejects', which might give longer safe hold times (on ever thinner amounts of data .. most of this is not statistically significant yet anyway, despite aggregating all loan terms, etc. etc). Because the data is so sparse, you can decide that, having missed the 18 month breakdown on As, if you get back in on month 19, it is clear sailing to the end of the term .. I wouldn't, but you could.
This is based on losses, so future recoveries (oo, look, flying porkers!) could change the picture, although I'd much rather not have a default, than have a default which is recovered, eventually. It is also based on the whole of history .. so the 'overall' numbers go back further than the 'WL reject' numbers, which doubtless affects the results too (FC's risk banding was notoriously flaky for the first year or so).
Oh, and if you have a 15% A rated part, the answer will be more than the 18 months quoted, since (apart from losing 2 weeks interest while you recycle) you won't be able to get it back into the market as a 15% A. It is left as an exercise for the reader to decide if flipping a 60 month A every 18 months is more or less rewarding than flipping a 60 month D every 8 months. 8>.
1) A/A+ property loans, interest only ... never (but might be smart to exit before full term!). Nothing longer than 15 months so far anyway.
2) A+ overall 18 Months, A+ WL Rejects 7 Months
3) A overall 18 Months, A rejects 4 Months
4) B overall 9 months, B rejects 7 Months
5) C or D overall 8 months, C or D rejects 6 Months
6) E overall >3 months, E rejects >2 months, but no statistically significant data (yet).
This assumes failure rates for B are always at least as bad as A (although loss of interest from bailing out is more, so you can stand slightly more expected loss .. hence the 7 months).
A few categories have very little data, and I still don't have a good set of early repayment dates, so loss %ages are probably worse than stated (because money I have assumed to be 'at risk' in month N had actually been repaid by then, but you can't figure that out from the loan book).
So yeah, OK, 6 months looks fairly safe, apart from some WL rejects which go pear shaped rather faster.For a quiet life, interest only property loans, especially with a cashback, look fine (and will continue to do so until the great bubble bursts?). I suppose I should look at 'excluding known WL rejects', which might give longer safe hold times (on ever thinner amounts of data .. most of this is not statistically significant yet anyway, despite aggregating all loan terms, etc. etc). Because the data is so sparse, you can decide that, having missed the 18 month breakdown on As, if you get back in on month 19, it is clear sailing to the end of the term .. I wouldn't, but you could.
This is based on losses, so future recoveries (oo, look, flying porkers!) could change the picture, although I'd much rather not have a default, than have a default which is recovered, eventually. It is also based on the whole of history .. so the 'overall' numbers go back further than the 'WL reject' numbers, which doubtless affects the results too (FC's risk banding was notoriously flaky for the first year or so).
Oh, and if you have a 15% A rated part, the answer will be more than the 18 months quoted, since (apart from losing 2 weeks interest while you recycle) you won't be able to get it back into the market as a 15% A. It is left as an exercise for the reader to decide if flipping a 60 month A every 18 months is more or less rewarding than flipping a 60 month D every 8 months. 8>.
