Slightly suffering from information overload so maybe I missed it but is there an analysis of defaults for those who borrow the maximum amount vs others in the same grade? I mostly avoid loans over $40k but admittedly it's pure guesswork.
Printable View
Slightly suffering from information overload so maybe I missed it but is there an analysis of defaults for those who borrow the maximum amount vs others in the same grade? I mostly avoid loans over $40k but admittedly it's pure guesswork.
Yep definately, doesn't matter where the improvement comes from. However I'll have some reports on the default minimiser set, just wont be RAR.
The "stable set" (eg less likely to repay early), was formed by selecting variables that maximised active loan days. iirc - the criteria included repay/loan < 8%, 1 enquiry max, in job 3+ years, married, exclude F loans, and excluded wedding exp, buy new car, buy new boat, tax bill. Pretty sure there was something else but can't recall. That set according to my scrawled notes gave 2.8% default rate (compared against the average 7% for the whole set through to Dec16), and extended the loan period an extra 2 months on average - to about 16.5 months. The return wasn't too different from the overall average though - mainly due to the set containing over 60% A and B grade loans. In contrast the default minimiser set contained 60% at CDE grade and turned in a lower default rate and higher return.
I'd say its just lazy programming. Easy enough to do without thinking. If somebody did mention to HM, am sure they'd add it to their list of fixes!
Will help Myles out on this one.
Attachment 10088
have to say didn't expect that result. So had a look at the data, and it appears that loans above the max have been invested on average at a later date. Cannot really draw a conclusion based on the whole dataset.
ave date below max ave date at max A 16/02/2017 14/11/2017 B 09/01/2017 20/11/2017 C 19/01/2017 21/11/2017 D 22/12/2016 17/02/2018 E 22/09/2016 18/02/2018 F 26/05/2016 10/03/2018
The loans above max would be a really interesting dataset - with the rewrite option not available early repayments should be much less.
Prior to August 2017 no max loans in grade DEF had been written. Looks like only one investor in the overall dataset has taken on the strategy - will be interesting to see how they perform
That will probably be because on august 3rd 2017 the maximum borrowing limits on many grades increased - prior to then borrowing todays maximum amount would not be possible
below is from harmoney email
Hi ...........,
On Thursday, 3 August 2017, we are implementing our next generation borrower scorecard. This will not affect your current Harmoney portfolio, but may affect your lending strategy moving forward.
What is changing
- From Thursday, 03 August 2017, all new loan applications will be assessed with the new generation scorecard.
- The scorecard will increase the accuracy of pricing risk significantly from the first generation scorecard.
- The probability of a default for each grade will change with the new scorecard.
- Borrower interest rates will reduce across risk grades A1-F5. The minimum rate will be 6.99% p.a. and maximum rate will be 29.99% p.a. This reflects the underlying risk and improved ability to price risk.
- The maximum loan limits are being increased in some grades. However, it is important to note that the affordability testing that is done will not change.
Why we are changing it
After 3 years and over 300,000 applications to date we continue to demonstrate our commitment to innovation, and creating value for both Lenders and Borrowers, with the launch of our next generation scorecard.
Here are the pre August 2017 Limits (as at August 2016 to be exact)
A1 - A5 $35,000
B1 - B5 $30,000
C1 - C5 $25,000
D1 - D5 $25,000
E1 - E5 $10,000
F1 - F5 $5,000
Leesal, if you could kindly redo your graph with only loans since August 2017.
And please don't try to confuse us with swapping the colours between the two graphs :p.
Thanks.
Or continuing on Humvee above post, do it from August 2016 but with the max set at different levels for the different dates - more complicated.
Good to have that gem of information :)
Here goes. Shows a not insignificant difference in default between frequency between max and non max loans. There is a slight mitigant, that max loans (particularly at the mid grade) stay current longer. Potential for more interest payments needs to be weighed up against default frequency.
Attachment 10089
Further - above is drawn from pre Aug 2017 data only