|
Post by dan1 on Aug 25, 2018 8:31:08 GMT
There was mention of a 30% rate of non-perfoming loans in the Crisis of confidence in the platform thread. I wanted to compile the figures for myself based on the All active and past loans. Values by the "Status" and "Defaulted" columns (i.e. those provided by FS):  The "Loan Active" and "Loan Completed" broken down by days overdue (days past the Expected End):  Calculating the performing (active loans less than 3 months overdue and completed loans with full payment of capital and interest), and non-performing (active loans overdue by at least 3 months, partially recovered loans and unredeemed loans):  The Non-Performing Loan (NPL) Ratio of the outstanding loan book is 36%.
|
|
|
Post by dan1 on Aug 25, 2018 8:35:25 GMT
For completeness, here are the figures for number of loans as opposed to value. Most relevant for investors placing £x in each and every loan. Values by the "Status" and "Defaulted" columns (i.e. those provided by FS):  The "Loan Active" and "Loan Completed" broken down by days overdue (days past the Expected End):  Calculating the performing (active loans less than 3 months overdue and completed loans with full payment of capital and interest), and non-performing (active loans overdue by at least 3 months, partially recovered loans and unredeemed loans):  The number of non-performing loans expressed as a percentage of the number of outstanding loans is 23%.
|
|
arby
Member of DD Central
Posts: 910
Likes: 959
|
Post by arby on Aug 25, 2018 9:16:23 GMT
Is this provided by FS or have you scraped the info from the 'All active and past loans' pages and aggregated yourself? If so, that's amazing. Thank you. This is what I was hoping to find rather than the FS front page of £266m lent, ~£860k lost which always massively frustrates me!
|
|
|
Post by dan1 on Aug 25, 2018 9:33:26 GMT
Is this provided by FS or have you scraped the info from the 'All active and past loans' pages and aggregated yourself? If so, that's amazing. Thank you. This is what I was hoping to find rather than the FS front page of £266m lent, ~£860k lost which always massively frustrates me! Yup, downloading each and every loan on the All past loans page, tiresome. If you sum all of those loans (excluding the cancelled loans) you get £261m not the £266m on the statistics page. Anyone know why?
|
|
arby
Member of DD Central
Posts: 910
Likes: 959
|
Post by arby on Aug 25, 2018 9:37:58 GMT
Is this provided by FS or have you scraped the info from the 'All active and past loans' pages and aggregated yourself? If so, that's amazing. Thank you. This is what I was hoping to find rather than the FS front page of £266m lent, ~£860k lost which always massively frustrates me! Yup, downloading each and every loan on the All past loans page, tiresome. If you sum all of those loans (excluding the cancelled loans) you get £261m not the £266m on the statistics page. Anyone know why? And you're going to provide a weekly update? Good man!  It's a real shame that it isn't available as a bulk download, but we all know why. Thank you again for providing it. It will be useful to compare the percentage defaulted (by number of loans) vs our own portfolio; see if our due diligence actually pays off!
|
|
|
Post by dan1 on Aug 25, 2018 9:39:55 GMT
I'll try and analyse the interest paid and outstanding at some point but, it is a bank holiday weekend and the sun is shining
|
|
bg
Member of DD Central
Posts: 1,368
Likes: 1,929
|
Post by bg on Aug 25, 2018 10:16:52 GMT
Is this provided by FS or have you scraped the info from the 'All active and past loans' pages and aggregated yourself? If so, that's amazing. Thank you. This is what I was hoping to find rather than the FS front page of £266m lent, ~£860k lost which always massively frustrates me! Yup, downloading each and every loan on the All past loans page, tiresome. If you sum all of those loans (excluding the cancelled loans) you get £261m not the £266m on the statistics page. Anyone know why? I actually get total loans of 262,357,303 I don't agree with the classification of any loan that repaid with all capital and interest more than 3 months late as non-performing. A full repayment is a good loan in my book, i'm not sure any investors in the loan would be too unhappy. There's a big difference between that and a capital write off which is really all we are looking to avoid.
|
|
SteveT
Member of DD Central
Posts: 6,865
Likes: 7,904
|
Post by SteveT on Aug 25, 2018 12:34:41 GMT
What I’ve often wondered (but never been bored enough to work out) is how much the FS figures are flattered by renewal loans, especially where the renewal interest is rolled up into an increased LTV or funded by a further loan / tranche.
Consider, for example, a £1m loan at 50% LTV that renews 3 times, increasing the loan each time for the accrued interest, but then eventually defaults. AIUI, the first 3 loans would forever be reported by FS as completed, fully performing loans (totalling something like £3.25m), more than overshadowing the eventual circa £1.2m default. In practical terms, it was really a single 2+ year loan where the borrower never paid a penny.
Powerboats spring to mind, and even the disastrous Whitehaven scheme claims 3 fully performing Completed loans!
|
|
kielbasa
Member of DD Central
Posts: 245
Likes: 411
|
Post by kielbasa on Aug 25, 2018 14:41:25 GMT
What I’ve often wondered (but never been bored enough to work out) is how much the FS figures are flattered by renewal loans, especially where the renewal interest is rolled up into an increased LTV or funded by a further loan / tranche. Consider, for example, a £1m loan at 50% LTV that renews 3 times, increasing the loan each time for the accrued interest, but then eventually defaults. AIUI, the first 3 loans would forever be reported by FS as completed, fully performing loans (totalling something like £3.25m), more than overshadowing the eventual circa £1.2m default. In practical terms, it was really a single 2+ year loan where the borrower never paid a penny. Powerboats spring to mind, and even the disastrous Whitehaven scheme claims 3 fully performing Completed loans! Good comment! I would happily contribute to a whip-round to pay someone to do the legwork to look at how each security "performs".
|
|
|
Post by dan1 on Aug 25, 2018 18:34:26 GMT
To alert that I've replaced the non-performing loan tables in the two OPs. The NPL Rate should of course be calculated from all loans. I've also added the NPL Rate of the outstanding loan book.
|
|
|
Post by dan1 on Aug 25, 2018 18:36:00 GMT
Yup, downloading each and every loan on the All past loans page, tiresome. If you sum all of those loans (excluding the cancelled loans) you get £261m not the £266m on the statistics page. Anyone know why? I actually get total loans of 262,357,303 I don't agree with the classification of any loan that repaid with all capital and interest more than 3 months late as non-performing. A full repayment is a good loan in my book, i'm not sure any investors in the loan would be too unhappy. There's a big difference between that and a capital write off which is really all we are looking to avoid. I calculate the total loans by copy and paste of the All active and past loans page into Excel, delete the cancelled loans, then sum the remaining loans. My total is £261,421,732.78. Is this how you do it because if it is it suggests that we don't see the same loans? Regardless, neither total is that near to the value stated on the Loan statistics page of £266,118,577 (correct to the end of July 2018). Maybe there are private loans that we don't have access to? I chose 90 days to classify loans as non-performing because this appears to be a common cut-off point. To be clear, I am not suggesting that all non-performing loans are bad (vice versa, I would not suggest all performing loans are good) but the probability that all capital and accrued interest will be paid is reduced the longer the loan is overdue. The easiest to grasp summary of non-performing loans that I could find... www.investopedia.com/terms/n/nonperformingloan.aspIn terms of: 1. The statistics above will over-estimate the NPLs because they don't take into account sums received from the borrower prior to the loan completing. 2. I assume "capitalized" refers to increasing the outstanding capital to pay for interest (as per the comment from SteveT), in which case the statistics above will under-estimate the NPLs. 3. At which point FS (should) default the loan. I would guess the under-estimate of 2. far outweighs the over-estimate of 1. and therefore I consider the NPL Rate as optimistic.
|
|
littleoldlady
Member of DD Central
Running down all platforms due to age
Posts: 2,970
Likes: 1,797
|
Post by littleoldlady on Aug 25, 2018 19:22:09 GMT
Consider a portfolio of one year loans all at 12% interest paid at the end of the year. It is simple arithmetic that if 30% by value of a portfolio default at an average loss of 28% then this will wipe out the interest on the performing loans. Or if 20% of loans default with a 48% loss, or if 40% default at 18% etc. None of these scenarios looks outlandish to me.
|
|
|
Post by dan1 on Aug 25, 2018 20:13:52 GMT
The chart below shows the location of capital by loan status (active, completed, defaulted) over the entire life of the platform. It's designed to show how the loan book has expanded over time. The loan book has retracted by a little over £4m from the peak, but this may be due to increased funds allocated to loans in the process of filling (it's impossible(?) to know the value of new funds as distinct from renewed funds - i.e. loans in the process of renewing where investors have checked Renew?). I posted a similar chart in a loan thread as a line chart whereas this is a stacked area chart.  ...and by number of loans 
|
|
bg
Member of DD Central
Posts: 1,368
Likes: 1,929
|
Post by bg on Aug 25, 2018 21:21:27 GMT
I actually get total loans of 262,357,303 I don't agree with the classification of any loan that repaid with all capital and interest more than 3 months late as non-performing. A full repayment is a good loan in my book, i'm not sure any investors in the loan would be too unhappy. There's a big difference between that and a capital write off which is really all we are looking to avoid. I calculate the total loans by copy and paste of the All active and past loans page into Excel, delete the cancelled loans, then sum the remaining loans. My total is £261,421,732.78. Is this how you do it because if it is it suggests that we don't see the same loans? Regardless, neither total is that near to the value stated on the Loan statistics page of £266,118,577 (correct to the end of July 2018). Maybe there are private loans that we don't have access to? I chose 90 days to classify loans as non-performing because this appears to be a common cut-off point. To be clear, I am not suggesting that all non-performing loans are bad (vice versa, I would not suggest all performing loans are good) but the probability that all capital and accrued interest will be paid is reduced the longer the loan is overdue. The easiest to grasp summary of non-performing loans that I could find... www.investopedia.com/terms/n/nonperformingloan.aspIn terms of: 1. The statistics above will over-estimate the NPLs because they don't take into account sums received from the borrower prior to the loan completing. 2. I assume "capitalized" refers to increasing the outstanding capital to pay for interest (as per the comment from SteveT ), in which case the statistics above will under-estimate the NPLs. 3. At which point FS (should) default the loan. I would guess the under-estimate of 2. far outweighs the over-estimate of 1. and therefore I consider the NPL Rate as optimistic. I have a script that runs every 15 mins that pulls in data for all loans, completed loans, late loans, > 3 months late, > 6 months late, > 1 year late amongst other things. The results are saved in a database. Yes 90 days late is the IMF definition of a NPL. In general though for dev/bridge loans many organisations use a longer time frame but it is fairly subjective. For unsecured business loans and the like (like FC loans), loans that are 3 months late are a massive worry. If the business is in trouble, they can't cover the interest and there are no assets then the risk of complete loss is very high. For a bridge loan, overrunning is very typical. Not being able to pay interest after 6 months does not impair the value of the asset...of course if the asset is very illiquid with a dubious valuation then the risk of default was high on day 1. If I wrote a bridge on a residential home in London at 50% LTV...then I really wouldn't be too worried if they were a few months late in refinancing. If I had lent against a performance, high spec powerboat then I would be worried as there is absolutely no market for such a boat and I would never have lent in the first place. My point was that if a loan had repaid in full then it can't be considered non-performing, even if it paid 3 months late. The loans to worry about are those that result in an actual loss.
|
|
arby
Member of DD Central
Posts: 910
Likes: 959
|
Post by arby on Aug 26, 2018 9:04:10 GMT
My point was that if a loan had repaid in full then it can't be considered non-performing, even if it paid 3 months late. The loans to worry about are those that result in an actual loss. I mostly agree with you; my only point to consider would be measuring how long it took an average before they paid to see if there's a 'point of no return', then we apply that to the current 'late' book and make a judgement on that. Obviously there's no single defined point of no return and is instead a probability distribution, but that's how we work out which of those <3 month later loans will likely result in a loss. Out of interest, why run your data scraping script every 15 minutes? I could imagine why for the secondary market, but on the all loans I'm a bit confused. I haven't been in the FS game for too long so there's a good chance I haven't spotted every advantage that can be gained, which you're also not required to share! 
|
|