blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Oct 1, 2014 11:09:10 GMT
Because of the discussion on this board about the apparent quality of the partial business loans being offered on the platform, particularly at A+, I have done some anaylsis of the effect of whole loans. The results are concerning. Taking the last 100 business loans agreed in today’s loan book, by loan number up to 8060, we see that 72 were first offered as whole loans, with 15 rejected and added to the 28 partial loans to end in the loan book as 57 whole loans and 43 partial loans. This means that, for this sample, the great majority are being offered first as whole loans, and that of the partial loans appearing on the platform 35% are rejects from the partial loan offers. I can see no particular category that is being rejected as unacceptable to the whole loan lenders, except that the higher risk bands are much more likely to be rejected – and this makes sense considering that the average interest rates for the lower risk bands have improved and it is the average rate that is offered to the whole loan lenders, about 8.2% for A+. Assuming that the rejected whole loans have been rejected on a perceived risk/reward basis it does mean that 35% of business loans accepted as partial loans may be of a lesser than average quality for the band. This might explain why some on this forum have commented on perceived reduction in quality of loans offered. If so and maintained, then that suggests that over time whole loan lenders may experience better than average default performance while partial loan lenders, and in particular non-selective Autobidders, may experience, systematically, a worse performance. However this is only a 100 loan sample. A further analysis by band gives some surprising results, considering that FC has given assurances that whole loans are selected and offered by a random process.
| A+ | A | B | C | C- | Total | | All business loans | 27 | 23 | 24 | 13 | 13 | 100 | | Whole offered | 26 | 14 | 19 | 8 | 5 | 72 | | Whole accepted | 24 | 13 | 14 | 5 | 1 | 57 | | Whole rejected | 2 | 1 | 5 | 3 | 4 | 15 | | Partial offered | 1 | 9 | 5 | 5 | 8 | 28 | | Total partial | 3 | 10 | 10 | 8 | 12 | 43 |
Source: last 100 business loans accepted, 1 Oct 2014 loanbook, to loan 8060. Firstly the proportion of business loans which qualify as A+ nowadays seems higher than before, and I will leave others to speculate as to why. I am more concerned with the random process which selects 26 A+ out of 27 for whole loans, and leaves the partial board with one, plus two rejects. On average you would expect to select 19 and leave 8. When you consider that the taking of A+ business loans as whole loans makes it easier to fill the A+ property loans on the platform, then the short term excursions of this random selector towards the A+ must be quite fortuitous. Of course this is a small sample, but perhaps we will see how things average out going forward. |
|
oldgrumpy
Member of DD Central
Posts: 5,087 
Likes: 3,233
|
Post by oldgrumpy on Oct 1, 2014 12:43:57 GMT
That could be why I have been bidding on almost nothing for the last couple of months!! I wonder how many more people have given up devoting time searching through mediocrity to find the good prospective loans to sound looking companies. When the likelihood of returns on anything but large loans, after fees and losses (and the big impact of those for tax purposes) is (they say) barely above 6%, why spend time on FC when RS gives (currently) 6+% with no enhanced tax implications following losses, and almost no time consumption. If I want to spend time examining loans to assess their merits, I'm not prepared to spend that time on £20 bids!!! The same time spent on AC or SS leads to three figure (sometimes four) at 10%-12% (no fees), and I don't see the overall risk as any higher than FC's. edit... curses.... I've just destroyed the impression that I am a HNW primate  |
|
blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Oct 1, 2014 13:14:15 GMT
edit... curses.... I've just destroyed the impression that I am a HNW primate But surely you are. I don't know any other gorillas who can splash out almost a pony on an FC loan. No-one expects you to bid a monkey.
|
|
oldgrumpy
Member of DD Central
Posts: 5,087
Likes: 3,233
|
Post by oldgrumpy on Oct 1, 2014 13:20:41 GMT
Pain in the a**e, monkeys, though I often wish I had a few more monkeys to play around with .... I'd have a grand time .... come to think of it, my last C&G bond comes up at the end of the year - but that will probably fund something via Hargreaves Lansdowne .... must look into the current state of play with Income Investment Trusts ....  |
|
|
|
Post by jackpease on Oct 1, 2014 13:27:46 GMT
If wonder if the 'favoured lenders' pay a lower rate than would have been the case had it been left to an auction?
I wonder if they bid against each other? Presumably if borrowers suspect the rate is higher than an auction would have been - they can reject it? Would they get fees returned?
Are borrowers allowed to insist it goes to auction to guarantee best rates?
Jack P
|
|
blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Oct 1, 2014 14:40:01 GMT
For the whole loans the interest rate is fixed for the band, according to the moving average on the platform for the band. You can see the rate paid for each loan in the loan book. Because the interest rate is fixed for the band I assume that the riskier looking ones do not get taken and become partial loans. Each whole loan goes to the first to accept it (with money presumably). You might ask whether they pay the 1% fee - presumably they do. Borrowers do not get the choice as far as I know - they know the offered rate. Whole loans cannot be traded at present.
|
|
adrianc
Member of DD Central
Posts: 10,000
Likes: 5,139
|
Post by adrianc on Oct 1, 2014 17:36:37 GMT
For the whole loans the interest rate is fixed for the band, according to the moving average on the platform for the band. So the whole loans - which are the better-than-average loans for each band - are offered at the average rate for the band, which will now be set solely by the worse-than-average loans in the band...? I can see this being a not-terribly-successful long-term plan.
|
|
blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Oct 1, 2014 21:32:57 GMT
For the whole loans the interest rate is fixed for the band, according to the moving average on the platform for the band. So the whole loans - which are the better-than-average loans for each band - are offered at the average rate for the band, which will now be set solely by the worse-than-average loans in the band...? I can see this being a not-terribly-successful long-term plan. That's about right AdrianC, though the average may include the whole loans- I think. The original assumption was that the partial loans would be the clear majority and so the average rate would not be increased by the mechanism you point out. But for the partial loans the size of the loan continues to be a greater determinant of interest rate within a band than the quality/risk perceived by the manual bidders - because the autobidders bid blind. A lot of the problem for FC comes from using the partial loan platform for the property loans. It was an obvious way of kicking off the property loans, but of course the recommended diversity rules apply even though they are inappropriate for A+ loans secured on real property. If you take the autobidders money blind on property loans you only get £20 for each £2M project, whereas if the property was a separate sub-platform with different rules you could justify much larger holdings. Now the shared diversity rules are causing FC real problems in filling the property loans - like having to use their own money. In turn to some extent the property loan tail is wagging the business loan dog, especially at A+ where the interest rate achieved by the few and sometimes dubious A+ partial loans will move the average rate used for the majority whole loans - as you say, AdrianC. I do not see how FC will move forward with property loans - but that's not my job.
|
|
|
|
Post by gingergent on Oct 2, 2014 17:25:21 GMT
if the property was a separate sub-platform with different rules you could justify much larger holdings That depends on how much you trust FC's recovery capability. You also need to offset the theoretical improvement in expected recovery against the known increase in amount to recover, since they're interest-only. Personally, I'm not yet willing to bet that they have significantly better odds on FC. edit: fixed the quote
|
|
blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Oct 2, 2014 23:08:30 GMT
if the property was a separate sub-platform with different rules you could justify much larger holdings That depends on how much you trust FC's recovery capability. You also need to offset the theoretical improvement in expected recovery against the known increase in amount to recover, since they're interest-only. Personally, I'm not yet willing to bet that they have significantly better odds on FC. edit: fixed the quote I was assuming that the average 0.6% loss of A+ property loans was real - but I have no crystal ball. On a separate property sub-platform with the current model you could not have sufficient diversity to spread the losses, and so would have to fix the interest rate to the lenders (as SS do) or have a provisions fund - but having done that you could at least take a much larger proportion of the consumer lender's (Autobidder's) funds into property loans more quickly - assuming that you could find an acceptable way of making that division or obtaining the Autobidder's instructions.
|
|
blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Oct 6, 2014 11:13:31 GMT
Thanks to GSV for raising this issue on the FC forum. It is true that the analysis in the OP does not account for loan offers rejected by the Borrower - though there is no way of doing that from the loan book, or at all for partial lenders. To be clear, there is no suggestion whatsoever that the quality of loans offered to the whole loan lenders is any different from the quality offered directly to partial loan lenders, though the random distribution by band is questioned by the figures. The point is that if the total offered on the partial board includes a large proportion of loans which have been rejected by the whole loan lenders, then the overall quality of the partial loan board will be diminished, judged by the standards of each band and predicted loss performance.
The intention not to proceed with the new front end for whole and partial loans is probably sensible - though why not say so earlier rather than just fail to meet commitments made?
Edit : last 200 loans in current loan book 20:00hrs 6 Oct, WL offered 124, WL accepted 92, WL rejected by lenders 32, + 76 direct partial loans = 108 loans accepted as partial, 30% of which are WL rejects.
(This is a sufficient number of loans to set the 'current estimated return' in FC's front-page advertising).
|
|
blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Nov 10, 2014 11:23:29 GMT
Because of the discussion on this board about the apparent quality of the partial business loans being offered on the platform, particularly at A+, I have done some anaylsis of the effect of whole loans. The results are concerning. Taking the last 100 business loans agreed in today’s loan book, by loan number up to 8060, we see that 72 were first offered as whole loans, with 15 rejected and added to the 28 partial loans to end in the loan book as 57 whole loans and 43 partial loans. This means that, for this sample, the great majority are being offered first as whole loans, and that of the partial loans appearing on the platform 35% are rejects from the partial loan offers. I can see no particular category that is being rejected as unacceptable to the whole loan lenders, except that the higher risk bands are much more likely to be rejected – and this makes sense considering that the average interest rates for the lower risk bands have improved and it is the average rate that is offered to the whole loan lenders, about 8.2% for A+. Assuming that the rejected whole loans have been rejected on a perceived risk/reward basis it does mean that 35% of business loans accepted as partial loans may be of a lesser than average quality for the band. This might explain why some on this forum have commented on perceived reduction in quality of loans offered. If so and maintained, then that suggests that over time whole loan lenders may experience better than average default performance while partial loan lenders, and in particular non-selective Autobidders, may experience, systematically, a worse performance. However this is only a 100 loan sample. A further analysis by band gives some surprising results, considering that FC has given assurances that whole loans are selected and offered by a random process.
| A+ | A | B | C | C- | Total | | All business loans | 27 | 23 | 24 | 13 | 13 | 100 | | Whole offered | 26 | 14 | 19 | 8 | 5 | 72 | | Whole accepted | 24 | 13 | 14 | 5 | 1 | 57 | | Whole rejected | 2 | 1 | 5 | 3 | 4 | 15 | | Partial offered | 1 | 9 | 5 | 5 | 8 | 28 | | Total partial | 3 | 10 | 10 | 8 | 12 | 43 |
Source: last 100 business loans accepted, 1 Oct 2014 loanbook, to loan 8060. Firstly the proportion of business loans which qualify as A+ nowadays seems higher than before, and I will leave others to speculate as to why. I am more concerned with the random process which selects 26 A+ out of 27 for whole loans, and leaves the partial board with one, plus two rejects. On average you would expect to select 19 and leave 8. When you consider that the taking of A+ business loans as whole loans makes it easier to fill the A+ property loans on the platform, then the short term excursions of this random selector towards the A+ must be quite fortuitous. Of course this is a small sample, but perhaps we will see how things average out going forward. It is time to give an update on the statistics in the OP. This time the last 500 loans in the current loan book have been analysed. The equivalent table as at 10am on 10 November 2014 (to loan 8766) is as follows:
| A+ | A | B | C | C- | Total | | All loans accepted | 119 | 125 | 107 | 89 | 60 | 500 | | Whole loans offered | 73 | 91 | 73 | 58 | 43 | 338 | | Whole loans taken | 46 | 53 | 37 | 19 | 16 | 171 | | Whole loans rejected (by lenders) | 27 | 38 | 36 | 39 | 27 | 167 | | Partial offered | 46 | 34 | 34 | 31 | 17 | 162 | | Total partial | 73 | 72 | 70 | 70 | 44 | 329 |
We can now agree that the whole loans are being randomly selected, whereas before too few A+ loans were becoming partial, which seemed to favour filling the property loans. But clearly about two thirds of the loans are first being offered as whole loans and only a third as partial loans. Half of those whole loans are being rejected and so the make up of partial loans is now about 50% fresh loans and 50% whole loan rejects. The 500 sample is plenty large enough. In these circumstances one cannot assume that the quality of the partial loan book and the whole loan book are the same. It is reasonable to suspect that the rejected whole loans were in part rejected on grounds of perceived quality within the band, and that partial loan lenders might experience a worse performance, if this continues.
|
|
oldgrumpy
Member of DD Central
Posts: 5,087
Likes: 3,233
|
Post by oldgrumpy on Nov 14, 2014 10:26:48 GMT
I suppose FC think this is a wonderful idea, whole loans, so much less hassle, and probably developed to feed those new so called P2P "funds" or "trusts" which need to invest in lots of loan value with minimum "paper" work. They won't want to be dealing in £20 loan units!!! I think FC will increase whole loan business as those funds catch on in ISA land. Those take huge "management" fees and commissions....what a con!!!  My involvement in FC is very much reduced these days, just careful recycling into larger A A+ and a selected few BC on short term, and selling of older loan parts. No more crappy Dai Bach scrappy or Top Tier Annoyer or Death of fresh Bare ............. I hope. Still like PDQ Bach though |
|
|
|
Post by davee39 on Nov 14, 2014 11:21:51 GMT
I think FC will increase whole loan business as those funds catch on in ISA land. Those take huge "management" fees and commissions....what a con!!! Not entirely true. Investing through a fund platform I pay 0.5% to the platform and fees of 0.5 to 0.75% to the funds. An Investment trust could be set up to cost a total of 1% including platform fees. A fund would be diversified to ensure 'average' losses and units would be marketable, so an FC only fund might easily offer 8% after fees. The first trust invested in several platforms and has a lower return. FC is hitting the limits of the current model. The secondary market is clogged with property and now with flipped 1% cashback parts. Change is always annoying, but inevitable as P2P grows. The upside might be to a create a space for the smaller players to fill the gap.
|
|
oldgrumpy
Member of DD Central
Posts: 5,087
Likes: 3,233
|
Post by oldgrumpy on Nov 14, 2014 13:01:12 GMT
davee39Yes, I tend to accept the 1% management fees on such things (sometimes a little higher). I'm thinking more of that fund recently which, in addition to those fees, takes 15% of any returns on top!!
|
|
