|
Post by propman on Oct 17, 2019 16:40:11 GMT
Looking increasingly like the "income" funds of yesteryear which agreed to pay outs higher than realistic returns (ie out of capital). Ever harder to see how they can hope to meet defaults from the Shield. Matthew, it would be good to understand why you believe that future defaults will be at significantly lower levels than those experienced to date on the same cohorts or why you expect future recoveries to be so high a proportion of expected defaults?
In light of the fact that cash in the shield declined by nearly £500k in July & August, this suggests that if performance has not improved then the Shield will have little cash remaining. Please would you report on progress in improving this position or the likely implementation of a haircut.
Many thanks
|
|
alanh
Posts: 556
Likes: 560
|
Post by alanh on Oct 17, 2019 16:52:37 GMT
Thanks for pointing out the new stats r00lish67. I also monitor these and agree with the £1m drop in shield cash since Feb. If we annualised this we get 12/8*£1m = £1.5m per year.
On a loan book of £90.8m that is 1.65% which means that if the shield cash keeps falling at the same rate then it could be stabilised by 1.65% of the interest on the loans.
I only invest in the 5 year or "growth" category so this implies the rate would need to reduce from 6.5% to 4.85% to achieve this.......funnily enough under the "poor" portfolio projection they are using an interest rate of 4.88%
I don't think 4.88%, if it were to happen, is disastrous as it just seems to bring LW more into line with the likes of RS who are now at 5% in their "improved rate scenario"
A comment from LW would be useful.
|
|
r00lish67
Member of DD Central
Posts: 2,691
Likes: 4,048
|
Post by r00lish67 on Oct 17, 2019 17:17:06 GMT
Hopefully that's not why the new portfolio projection tool was added recently with the performance based on Good-Moderate & Poor options? I hadn't clocked that there was a new tool, that is interesting. FWIW, I calculate the current deficit of the currently managed loans to be £560k (this is the sum of PF cash + expected inflows minus expected outflows). I mention that as LW state in their FAQs : "If a Pooling Event occurs <due to the shield failing>, the value of the outstanding deficit will be split or 'pooled' among all existing lenders on a pro-rata basis". <speculation> - If the rates for new investors were to be kept the same, then no further stick would be required to stop existing lenders withdrawing as to try and withdraw would incur significant penalties given the new discrepancy between the new diminished rate for existing investors and the old rate that new investors would still benefit from. I doubt existing lenders would feel very happy about that though! I'm not quite sure how the detail of this would work though, as whilst existing investors would be paying the price to aim to restore the PF to a net zero position, why would new investors feel comfortable investing? Where's their provision fund? The final thing which bears repeating is that the not particularly great implication of the stats at present still includes what I view as a highly optimistic view of the 2019 cohort - namely that bad debt will be 30% less than in 2017 and 2018 despite the blended borrower rate being significantly higher (2019 = 14.3%, 2018 = 12.2%, 2017 = 9.7%). i.e. How are less creditworthy borrowers going to deliver significantly better results? One would ordinarily expect a higher bad debt rate in this case, not lower.
|
|
zlb
Member of DD Central
Posts: 1,412
Likes: 331
|
Post by zlb on Oct 17, 2019 21:22:36 GMT
This sentence:
Shield utilisation : Total payments covered by the Shield, minus recoveries, as a proportion of total loan contributions to the Shield in respect of loans issued in each origination year. (actual 2014-2019) 67.5% 214.4% 285.6% 155.5% 64.5% (2019)6.7% (total)70.1%
means? = ((shield depletion) - (recoveries -?in year??)) / (portion of all loans which go toward the shield - ?within that year?)
edit: That doesn't look so worrying if it's moved from extreme multiple of 285% to 64.5% in 2018 - if I understand it correctly, and if it's a valid statistic does comparing in-year actually mean anything?
|
|
|
Post by propman on Oct 18, 2019 9:27:06 GMT
This sentence: Shield utilisation : Total payments covered by the Shield, minus recoveries, as a proportion of total loan contributions to the Shield in respect of loans issued in each origination year. (actual 2014-2019) 67.5% 214.4% 285.6% 155.5% 64.5% (2019)6.7% (total)70.1% means? = ((shield depletion) - (recoveries -? in year??)) / (portion of all loans which go toward the shield - ? within that year?)edit: That doesn't look so worrying if it's moved from extreme multiple of 285% to 64.5% in 2018 - if I understand it correctly, and if it's a valid statistic does comparing in-year actually mean anything? I believe this statistic is only looking at the loans in each cohort that have defaulted. This means that the 2018 loan defaults have already burned through 2/3 of the amount that they are expected to contribute to the Shield. That with 71% of the amount lent still O/S. That sounds worrying to me. 2019 is, of course, too early to tell.
Essentially the Shield is dependent on new loans yielding more than the old loans default. Unfortunately the growth this year is low so new kloans are almost required to pay for the defaults on approx. twice the amount borrowed.
|
|
zlb
Member of DD Central
Posts: 1,412
Likes: 331
|
Post by zlb on Oct 18, 2019 10:01:07 GMT
This sentence: Shield utilisation : Total payments covered by the Shield, minus recoveries, as a proportion of total loan contributions to the Shield in respect of loans issued in each origination year. (actual 2014-2019) 67.5% 214.4% 285.6% 155.5% 64.5% (2019)6.7% (total)70.1% means? = ((shield depletion) - (recoveries -? in year??)) / (portion of all loans which go toward the shield - ? within that year?)edit: That doesn't look so worrying if it's moved from extreme multiple of 285% to 64.5% in 2018 - if I understand it correctly, and if it's a valid statistic does comparing in-year actually mean anything? I believe this statistic is only looking at the loans in each cohort that have defaulted. This means that the 2018 loan defaults have already burned through 2/3 of the amount that they are expected to contribute to the Shield. That with 71% of the amount lent still O/S. That sounds worrying to me. 2019 is, of course, too early to tell.
Essentially the Shield is dependent on new loans yielding more than the old loans default. Unfortunately the growth this year is low so new kloans are almost required to pay for the defaults on approx. twice the amount borrowed.
They are not comparing like with like? They are looking at a count of loans on the one hand, and '£value as a %', on the other, all in the same equation? Thanks. So to check what you mean - there isn't 'twice the amount borrowed', therefore new loans are only matching approx 50% of what is required by the shield? So in example, there are defaults yet to occur from borrowers who started borrowing in 2018; and as all new money is going into covering the 2019 shield requirements, there is a (high/low?) risk that there is more required by 2018 defaults than the 33%?
|
|
|
Post by propman on Oct 18, 2019 10:56:33 GMT
I believe this statistic is only looking at the loans in each cohort that have defaulted. This means that the 2018 loan defaults have already burned through 2/3 of the amount that they are expected to contribute to the Shield. That with 71% of the amount lent still O/S. That sounds worrying to me. 2019 is, of course, too early to tell.
Essentially the Shield is dependent on new loans yielding more than the old loans default. Unfortunately the growth this year is low so new kloans are almost required to pay for the defaults on approx. twice the amount borrowed.
They are not comparing like with like? They are looking at a count of loans on the one hand, and '£value as a %', on the other, all in the same equation? Thanks. So to check what you mean - there isn't 'twice the amount borrowed', therefore new loans are only matching approx 50% of what is required by the shield? So in example, there are defaults yet to occur from borrowers who started borrowing in 2018; and as all new money is going into covering the 2019 shield requirements, there is a (high/low?) risk that there is more required by 2018 defaults than the 33%? I think these statistics are fine. What you should be comparing is the projected defaults as a proportion of actual defaults to date anmd then see how this compares with the shield usage in each year. This gives the expected shield shortfall. I make this 46%, -83%, -138%, -65%, -1% and 44% for 2014-19. ie they are expecting the 2018 shield to pay out in full and the accumulated shortfall from earlier years to be met by the 2019 shield.
Also, as the projection is just that, you also need to compare whether the projected default is consistent with recent years performance (r00lish's comment above).So the adequacy of the Shield rests on the 2018 & 2019 default projections being correct when 2018 is only 29% repaid and 2019 5% repaid. ie where we have data the Shield has been inadequate since 2014. Even with the current projections, they expect to need a further £400-500k surplus from the future contributions. Further, 2017 assumes that recoveries on the defaults to date are only 1% of the outstanding loans less than further defaults (against the 6.7% to date) nearly 2/3 of the net defaults on 2018 have occurred while only 30% has been repaid and that in 2019 more expensive loans will have only 70% of the defaults now assumed for 2018. I am sure that they will have improved their credit assessment as a result of the recent experience, but are these levels of improvements realistic?
|
|
zlb
Member of DD Central
Posts: 1,412
Likes: 331
|
Post by zlb on Oct 18, 2019 20:34:37 GMT
They are not comparing like with like? They are looking at a count of loans on the one hand, and '£value as a %', on the other, all in the same equation? Thanks. So to check what you mean - there isn't 'twice the amount borrowed', therefore new loans are only matching approx 50% of what is required by the shield? So in example, there are defaults yet to occur from borrowers who started borrowing in 2018; and as all new money is going into covering the 2019 shield requirements, there is a (high/low?) risk that there is more required by 2018 defaults than the 33%? I think these statistics are fine. What you should be comparing is the projected defaults as a proportion of actual defaults to date anmd then see how this compares with the shield usage in each year. This gives the expected shield shortfall. I make this 46%, -83%, -138%, -65%, -1% and 44% for 2014-19. ie they are expecting the 2018 shield to pay out in full and the accumulated shortfall from earlier years to be met by the 2019 shield.
Also, as the projection is just that, you also need to compare whether the projected default is consistent with recent years performance (r00lish's comment above).So the adequacy of the Shield rests on the 2018 & 2019 default projections being correct when 2018 is only 29% repaid and 2019 5% repaid. ie where we have data the Shield has been inadequate since 2014. Even with the current projections, they expect to need a further £400-500k surplus from the future contributions. Further, 2017 assumes that recoveries on the defaults to date are only 1% of the outstanding loans less than further defaults (against the 6.7% to date) nearly 2/3 of the net defaults on 2018 have occurred while only 30% has been repaid and that in 2019 more expensive loans will have only 70% of the defaults now assumed for 2018. I am sure that they will have improved their credit assessment as a result of the recent experience, but are these levels of improvements realistic?
Should people really have to pick it apart this much - it makes me question their transparency and new FCA demands. I'm still struggling with your last bit after 'further' but that's not your fault. Thanks for trying. So borrower fees are to be used to bolster LW as they have been shoring up defaults <2017. So the shield doesn't really exist in-year - it's only based on an endless supply of future borrowers - but LW are asserting that their credit checks are so good that default rates will drop... I wonder where all of those borrowers will go. I've seen other posts elsewhere, it seems quite damning.
|
|
|
Post by df on Oct 18, 2019 21:15:25 GMT
I think these statistics are fine. What you should be comparing is the projected defaults as a proportion of actual defaults to date anmd then see how this compares with the shield usage in each year. This gives the expected shield shortfall. I make this 46%, -83%, -138%, -65%, -1% and 44% for 2014-19. ie they are expecting the 2018 shield to pay out in full and the accumulated shortfall from earlier years to be met by the 2019 shield.
Also, as the projection is just that, you also need to compare whether the projected default is consistent with recent years performance (r00lish's comment above).So the adequacy of the Shield rests on the 2018 & 2019 default projections being correct when 2018 is only 29% repaid and 2019 5% repaid. ie where we have data the Shield has been inadequate since 2014. Even with the current projections, they expect to need a further £400-500k surplus from the future contributions. Further, 2017 assumes that recoveries on the defaults to date are only 1% of the outstanding loans less than further defaults (against the 6.7% to date) nearly 2/3 of the net defaults on 2018 have occurred while only 30% has been repaid and that in 2019 more expensive loans will have only 70% of the defaults now assumed for 2018. I am sure that they will have improved their credit assessment as a result of the recent experience, but are these levels of improvements realistic?
Should people really have to pick it apart this much - it makes me question their transparency and new FCA demands. I'm still struggling with your last bit after 'further' but that's not your fault. Thanks for trying. So borrower fees are to be used to bolster LW as they have been shoring up defaults <2017. So the shield doesn't really exist in-year - it's only based on an endless supply of future borrowers - but LW are asserting that their credit checks are so good that default rates will drop... I wonder where all of those borrowers will go. I've seen other posts elsewhere, it seems quite damning. I think all of consumer lending p2p platforms claim that the loans are to creditworthy borrowers , but if they tighten their conditions the loan flow will drop. There will always be somewhere else for "unqualified" borrowers to go.
|
|
zlb
Member of DD Central
Posts: 1,412
Likes: 331
|
Post by zlb on Oct 18, 2019 21:51:04 GMT
So, I think their stats descriptions are pretty poor. But I think I've interpreted these two figures correctly,
"actual arrears rate: Outstanding balance of loans with one or more payments more than 45 days overdue, as a proportion of total amount lent in that year "
Percentage 0%(2014) 0.2% 0.1% 0.7% 0.8% 0.5%(2019)
Lent in each year: £4,643,783(2014) / £11,252,059 / £17,992,943 / £43,393,227 / £58,010,300 / £49,574,008 (2019)
My calculation of the values >45 days in arrears: £0(2014) / £22,504 / £17,992 / £303,752 / £464,082 / £247,870
Total >45 days currently in arrears (if my interpretation of their descriptions is correct) £1,056,202 - so this value is yet to be added to official defaults??
When they say "Total blended annual return to date of loans in origination year (after fees and bad debts)" do they mean that is profit? Lender returns, or ? It's less than the % they are offering? or is that an incorrect understanding?
|
|
ashtondav
Member of DD Central
Posts: 1,805
Likes: 1,087
|
Post by ashtondav on Oct 19, 2019 8:07:54 GMT
Surely some of the defaults are covered by “insurance” and have no claim on the PF?
|
|
r00lish67
Member of DD Central
Posts: 2,691
Likes: 4,048
|
Post by r00lish67 on Oct 19, 2019 9:22:32 GMT
Surely some of the defaults are covered by “insurance” and have no claim on the PF? They are, but not as much as one might think, and when I've stated my assumption previously that any reclaims via insurance are already baked into the forecasts, LW's Matthew hasn't corrected it.
|
|
macq
Member of DD Central
Posts: 1,924
Likes: 1,192
|
Post by macq on Oct 19, 2019 9:56:45 GMT
Surely some of the defaults are covered by “insurance” and have no claim on the PF? They are, but not as much as one might think, and when I've stated my assumption previously that any reclaims via insurance are already baked into the forecasts, LW's Matthew hasn't corrected it. is it not the case that institutional lenders money is not covered for defaults by the shield?
|
|
r00lish67
Member of DD Central
Posts: 2,691
Likes: 4,048
|
Post by r00lish67 on Oct 19, 2019 16:10:11 GMT
They are, but not as much as one might think, and when I've stated my assumption previously that any reclaims via insurance are already baked into the forecasts, LW's Matthew hasn't corrected it. is it not the case that institutional lenders money is not covered for defaults by the shield? I think you are correct, but in any case all of LWs statistics seem to exclude those loans (frequent refs to each number being totals "covered by the shield"). Surprised we haven't seen any comment from Matthew as yet, would be interesting to know his view on all of this. Particularly pertinent since for some reason the next stats update is only scheduled 2 months away rather than the usual 1 (20th Dec). Not sure why that is.
|
|
benaj
Member of DD Central
Posts: 4,857
Likes: 1,591
|
Post by benaj on Oct 19, 2019 19:31:52 GMT
"Next update 20/10/2019"? so it will be up tomorrow right?
|
|