sl75
Posts: 2,092
Likes: 1,245
|
Post by sl75 on Apr 29, 2019 7:34:11 GMT
I find this table on the "latest PF usage" most telling. Given the target is 125% . RS meeting the target latest PF usage requires that number to be 80% (i.e. 100/125). Not quite, that's mixing different concepts...
The percentages in the table you quote would be usage vs expectation. [Edit: or maybe not ... see later post]
The percentage that needs to be 80% to have an immediate 125% coverage for the cohort of loans would be the usage vs fee charged (i.e. the risk-based element of the fees charged to borrowers, and which accrue in the provision fund), which is very unlikely to be exactly 100% of expected losses as that provides no expected safety margin.
However, in order to build up 125% coverage over the long term, it's not necessary to have 125% fees charged...
e.g. considering a toy model where each cohort of loans is the same size and fully repaid before the next one starts: Cohort 1 charges borrowers 110% of expected losses. (giving 110% coverage - below target for now). Losses are as expected, so there's a 10% surplus. Cohort 2 charges borrowers 110% of expected losses. (giving 120% coverage due to cash surplus from cohort 1). Losses are as expected, so there's now a 20% surplus. Cohort 3 charges borrowers 105% of expected losses, but together with the surplus from the previous cohort, that gives 125% coverage which is on target. ... if losses remain on target on average, then coverage can be maintained simply by charging borrowers 100% of the expected losses.
Obviously the real environment has a lot more challenges than this to model to balance borrower fees against expected losses whilst also maintaining a competitive edge...
Given the competitive commercial environment, it's unlikely that RateSetter can afford to charge high enough fees to *immediately* build the coverage back up to 125%. This would likely cause the best borrowers who would otherwise have been part of that cohort to go to other competing loan providers, leading to higher losses than originally expected for that cohort, and failing to generate the boost to coverage that was expected.
Instead it's undoubtedly a fine balancing act, to make sure that in all reasonable outcomes coverage will remain sufficiently above 100% that lenders won't panic and pull their money out, whilst keeping contributions low enough that the best borrowers won't just go for a different competing loan quote.
|
|
ashtondav
Member of DD Central
Posts: 1,805
Likes: 1,087
|
Post by ashtondav on Apr 29, 2019 7:57:33 GMT
It remains troubling that despite unemployment being at its lowest for over 40 years the platforms I invest in are all showing signs of credit strain - FC, Z, RS etc. Any sign of a downturn and i’m Selling. And I think that may be soon.
|
|
r00lish67
Member of DD Central
Posts: 2,691
Likes: 4,048
|
Post by r00lish67 on Apr 29, 2019 8:20:46 GMT
It remains troubling that despite unemployment being at its lowest for over 40 years the platforms I invest in are all showing signs of credit strain - FC, Z, RS etc. Any sign of a downturn and i’m Selling. And I think that may be soon. Perhaps the platform credit strains are (one of) the first signs of a downturn?
|
|
jlend
Member of DD Central
Posts: 1,817
Likes: 1,444
|
Post by jlend on Apr 29, 2019 8:39:00 GMT
Just building on the post from sl75 We used to be able to see the expected PF cash remaining (surplus or deficit) from each annual cohort of loans. web.archive.org/web/20190322063941/Ratesetter.com/invest/statisticsI use to find the PF resources and PF usage tables the most useful as they showed the cash position of the PF for each annual cohort. The percentage figures were not of much use without the cash figures. The 2016 cohort is interesting as an example 1st March Update Expected Lifetime Loss 3.4% Latest projected 3.3% Actual to date 3.3% PF projected usage 106% PF projected surplus/deficit -1,131,261 This was a year when losses were less than expected but the PF usage was higher than expected resulting in a deficit. In the latest update the PF projected usage for this year has increased to 109% so the PF cash deficit for this year would have increased further. This cash surplus/deficit data was removed in the recent website changes, along with the downloadable loan detail spreadsheet and the more frequent updates to some of the statistics. 2017 is also interesting. The estimated loss at the point the loans were written was 3.2%. The latest projected loss is 3%. The latest projected lifetime use of the PF is 101%. So again even though the loss rate is less than expected, RS are still projecting a deficit in the PF for 2017 in the latest update. IMHO there is now less information to make informed decisions for lenders which is a shame.
|
|
sl75
Posts: 2,092
Likes: 1,245
|
Post by sl75 on Apr 29, 2019 9:25:32 GMT
Just building on the post from sl75 We used to be able to see the expected PF cash remaining (surplus or deficit) from each annual cohort of loans. Hmmm... considering the context of the more complete table, it's probably me who misinterpreted the (smaller) table when I saw wapping35 's screenshot, and that the % are indeed usage vs contribution [including future projected contributions for recent cohorts] rather than usage vs expectation.
The only other key additional point would seem to be that the current projections are lower than the actual usage to date (for 2015 and earlier cohorts), implying that additional recoveries are still expected which will ultimately reduce the net usage.
Comparing the wayback machine to the live site, I note that the pre-2015 cohort has had the projection change from 89% to 90% (whilst actual usage remains the same at 92%), implying that they no longer to expect eventually part of the money they previously considered recoverable, and the 2018 and later loans seem to be propping up the losses from 2015-2016. It's still seemingly "under control", but not as firmly under control as I'd imagined at the time of my earlier post.
I'd guess that the 2015 and 2016 cohorts would be the ones most affected by the unexpected outcome of the Brexit vote and the resulting uncertainties, whilst 2017 and later cohorts would have these uncertainties largely taken into account?
|
|
benaj
Member of DD Central
Posts: 4,884
Likes: 1,603
|
Post by benaj on Apr 29, 2019 9:33:58 GMT
The positive about the RS stat right now is the Positive PF cash. There are rooms for RS and Investors to maneuver.
|
|
jlend
Member of DD Central
Posts: 1,817
Likes: 1,444
|
Post by jlend on Apr 29, 2019 10:26:51 GMT
Just building on the recent post from sl75 I expect a large portion of the remaining pre 2015 loans have defaulted and are sitting in the PF. Hence the PF has already paid out and has covered these loans. RS may be getting small amounts back each month and may be able to sell some of these distressed loans on. We could get some indication of this if we could still download the loan spreadsheet which has been removed. There are not many loans left pre 2015 looking at the recent stats graph so any recovery does not make a huge difference in cash terms to the PF. I don't know whether the graph of loans outstanding includes those sitting in the PF that have not formally been written off. I think the 2018 cohort is the important one for RS 1. The remaining 2018 cohort loans make up circa 50% of the outstanding loan book 2. There was a circa projected 6m of PF surplus at the 1st March update for this cohort. The 2018 surplus will now have reduced based on the recent RS update which said the expected losses of this cohort have increased in the quarter. It is important for RS that this cohort does not get any worse as it is where the surplus is coming from. We do not know what the 2018 surplus is now as RS have removed it from the website. As for Brexit effect i really do not know. There is not much 2015 and 2016 money still to be repaid so this cohort probably does not matter so much. Putting aside Brexit, across the industry personal loans tend to default earlier in their life than secured property loans. The RS move into more secured property loans may be a good or bad thing, only time will tell. We use to be able to get an idea by looking at the loan spreadsheet, but this has now been removed so it wont be possible to assess this going forward. In any downturn it is difficult to know if RS will be more exposed to personal loan defaults or property loan defaults/falling property asset values. RS also now have inconsistent data on their new website. For example the PF on this page still says 118%. www.ratesetter.com/invest/investing-with-us/provision-fundweb.archive.org/web/20190429104650/https://www.ratesetter.com/invest/investing-with-us/provision-fundAlso the old data-hub page was not archived by the wayback machine so we cannot go back and see further historical data to look for any trends and material changes.
|
|
|
Post by davee39 on Apr 29, 2019 11:45:07 GMT
It remains troubling that despite unemployment being at its lowest for over 40 years the platforms I invest in are all showing signs of credit strain - FC, Z, RS etc. Any sign of a downturn and i’m Selling. And I think that may be soon. The unemployment figures are misleading. Increasing numbers are 'self employed' and struggling. Similarly minimum wage jobs may not be of great help to the indebted. I fully expect the provision fund to fail over the next five years since there is growing trade friction between the US and China and economic discontent across europe. With current low growth and high borrowing a recession is inevitable. I have been pulling out for several months.
|
|
sl75
Posts: 2,092
Likes: 1,245
|
Post by sl75 on Apr 29, 2019 12:40:02 GMT
2. There was a circa projected 6m of PF surplus at the 1st March update for this cohort. The 2018 surplus will now have reduced based on the recent RS update which said the expected losses of this cohort have increased in the quarter. It is important for RS that this cohort does not get any worse as it is where the surplus is coming from. We do not know what the 2018 surplus is now as RS have removed it from the website. As I understand it, the (just over) £6M surplus corresponded to the 16% by which the 84% projected usage was short of 100%...
... so the implied new figure with 90% projected usage would likely be closer to 10/16 * £6M, which is just under £3.8M.
There would of course be some variation from this "simple" calculation due to factors it doesn't take into consideration (e.g. reduced projected income due to early repayments).
Personally, I'd consider the more significant figure to be the 226% "capital coverage ratio"... and these two ratios would seem able to be interpolated (e.g. we should get all the capital back and half the interest we originally expected at 169% of expected losses), or even extrapolated!
|
|
|
Post by propman on Apr 29, 2019 15:46:50 GMT
Althis theorising is dependent on the accuracy of RS's projected defaults. As I have said far too often, there seems a worrying linkage between increased defaults on earlier cohorts and reduced expected defaults on more recent cohorts still too young to provide a good estimate of defaults. Clearly the best outcome by which RS can cover a deficit in the PF is for them to grow their current lending and increase the contribution made by new borrowers to meet the calls on earlier defaulted loans. This requires growing investors funds and a willingness for them to invest at lower rates. These are supported by appearing to have a stronger PF. As a result there is a strong incentive for RS to underestimate future defaults and over estimate future incoms and recovery of defaults.
Whether it is the capital buffer or the coverage ratio that matters most, this depends on expectation of investor action on implementing a PF Event. If you believe the majority will accept a lower return and stay put, RS could survive this amnd it would not be too dire for investors. However, if the majority of investors are shocked by the loss of expected interest and/or reduce their belief in RS estimates as a result of this occuring, then we can expect wholesale attempts to withdraw funds absorbing any money available to lend. This will result in the curtailing of new loans and no ability to raise additional funds for the PF. If this is believed to be more than a short term impact, then there will also be costs as they lay off the staff required to improve systems and deal with new loan applications. This is likely to lead to insolvency unless they can first raise sufficient funds from a capital raising to fund these costs (and potentially contribute to a new PF for new business). The administrator / credit manager brought in is likely to be less efficient than the current operations and so I would expect the net losses to investors increase (accepting lower price for selling defaulted loans and higher charges) and so a higher proportion of interest to be lost than initially expected.
Of course anyone not interested in lending on RS and for whom recovering capital is the only requirement would still best assess the PF from the Capital Buffer.
|
|
wapping35
Member of DD Central
Posts: 385
Likes: 210
|
Post by wapping35 on Apr 29, 2019 17:19:55 GMT
RateSetter , In October 2018 you indicated the following regarding having an independent audit of the PF default expectations and coverage ratio's/ future PF income/default recoveries etc. "The assets of the Provision Fund are audited each year as part of our statutory financial audit and the operations of the Provision Fund are audited as part our rolling internal audit programme. In the future, we also plan to have independent annual audits of the future contracted income and future expected losses."
Given the PF target ratio changes and the fall in the coverage ratio can I please ask when that "in the future" audit can be expected, it is after all over 6 months since that commitment was communicated. Thanks
|
|
|
Post by RateSetter on May 5, 2019 19:51:33 GMT
Thank you for all your comments. We’d like to round up and respond to the above questions and we also plan to reflect these themes in our next RateSetter blog on this topic. With projected loss rates around the same as expected loss rates for each year of lending, why isn’t the Coverage Ratio closer to 125%?
The key driver is a change in our forecast of future Provision Fund inflows – more borrowers than initially estimated are repaying their loans early (this theme is referenced in the Quarterly Update RateSetter Notices on 1 February 2019 and 15 November 2018). While this is a positive sign for credit risk, it reduces future payments into the Provision Fund. In some years of lending, this can more than offset reductions in loss forecasts, resulting in a higher Provision Fund usage figure for those years. Remember, at RateSetter you are investing in the performance of the whole active loan portfolio rather than the performance of individual loans or years. This means that every investor enjoys the stability of the entire RateSetter portfolio. As our consumer loans have terms of up to 5 years, a longer time-series of data helps us to refine our forecasting of future Provision Fund inflows. RateSetter now has almost 10 years of data and this enhances our ability to make accurate inflow forecasts. Why do previous years of lending have a Provision Fund projected usage greater than 100% and more recent years have projected usage below 100%?
Once again, the key factor is the change in our forecast of in future Provision Fund inflows which leads to a higher Provision Fund usage. The key factor is not deterioration in loan performance - as can be seen in the performance data, in most years of lending the expected lifetime loss rate and the latest projected lifetime loss rate are the same or not far apart, showing that in general the loans have performed in line with our expectations. Why is the expected loss for 2019 YTD lower than the same figure for 2018 lending?
We have revised our pricing for consumer loans which has led to a different expected loss for 2019 YTD than the expected loss for 2018. Is the Provision Fund audited?
There are three important components to the Interest Coverage Ratio: - Cash balance
- Expected Future Inflows
- Future Outflows
The audit and control of these components is in two parts: - Cash in the Provision Fund is independently audited on an annual basis as part of RateSetter’s financial audit.
- Expected Future Inflows and Outflows are not auditable in the same way as they are projections of future performance. RateSetter’s projections rely on modelling methodologies rather than tallying of amounts in an account. The projections are controlled by our Risk Analytics team, reporting into the Chief Credit Officer who is accountable to the Executive Credit Committee and the independent Board Risk Committee. We are considering the potential for projections to be externally analysed in future.
|
|
jlend
Member of DD Central
Posts: 1,817
Likes: 1,444
|
Post by jlend on May 6, 2019 9:54:03 GMT
I do wonder whether spreading the PF contributions over the lifetime of loans rather than upfront was a good idea.
A lot of us had our doubts at the time the change was made, even with the benefits assumed by RS.
It is interesting that the PF has been impacted not only by higher than expected defaults in some years but also by a rise in early repayments in some years.
It does beg the question whether it is worth considering a higher proportion of PF contributions up front, at least for some loans where the expected risk of early repayment or default is higher than average.
After 10 years RS has some statistics they can base this on. In the past we might have got some idea by looking at the loan book, but this cannot be downloaded any more.
|
|
wapping35
Member of DD Central
Posts: 385
Likes: 210
|
Post by wapping35 on May 6, 2019 10:38:23 GMT
Sounds like RS is not so keen on an external audit anymore. This does not really surprise me. I know the Banks are not too keen on external independent audits either (and indeed the audits of all company DB pension fund future assets and future liabilities). But auditors could clearly do the job. i.e. Quote from Oct 2018: In the future, we also plan to have independent annual audits of the future contracted income and future expected losses." May 5th 2019: We are considering the potential for projections to be externally analysed in future. Planning to do something versus considering to do something.... ======== In the end I asked the question back in October 2018 since RS wanted to know why I (and some other) long term investors have been drawing down their investment (since June 2016 in my case) and they wanted to know what would change my investment approach and an external audit was the answer for me at least. RateSetter if you do ever engage an external independent audit of those future PF projections (and of course the audit is favourable) I would certainly return and invest more, since it would provide far better visibility of the risk profile. I suspect however we won't see an external audit in the foreseeable future, to be honest what surprised me was the Oct 2018 commitment which now seems to be heavily watered down. Thanks for the answer.
|
|
r00lish67
Member of DD Central
Posts: 2,691
Likes: 4,048
|
Post by r00lish67 on May 6, 2019 10:50:15 GMT
I have a possibly stupid question (I may be misinterpreting something).
Expected Future provision fund inflows are currently £25.6m
Expected Future losses are £34.6m.
Does that not imply that the PF actual cash is going to continue to decrease on their own projections? I note it did decrease by £1m in the last month, despite the loan base increasing by £11m.
Regardless of whether I've misunderstood something in the above, I remain concerned about RS. The simple fact is that in terms of PF cash they actually hold in their hands, that number went down significantly last month (-7%) whilst the capital amount of loans it has to cover increased by 1.3%.
Rather than the 112% figure, which is derived largely using future projections (easily tweakable when the results aren't liked), I prefer to simply divide PF cash by loans under management. That ratio has gone down from 1.67% to 1.54% in a single month. Unfortunately I don't have the stats for further back, but I'm willing to bet it was wayyy higher.
|
|