|
Post by moneyball on Sept 19, 2014 20:32:28 GMT
Something that I have been thinking about lately and further sparked by another thread on diversification, is the coverage ratio of the provision fund over the estimated claims against it.
I know RS have been deliberately trying to increase this for at least a number of months (if not years) and generally, I welcome this. However, I was curious on whether RS have a particular target in mind for this ratio and furthermore, what would individual lenders like to see?
At the time of writing, its 2.26. My gut feeling is that a ratio of 3.0 would be a reassuring long term goal but cant provide any reasoning behind this. Do people think this is too low?....or even too high? I believe I read somewhere recently that Zopa (who have a ratio around 1.2 but I could be wrong on that) are rightly or wrongly, happy with their current coverage. Their approach does seem to based on the aggressive reach for 'best buy' status which may tie in with this but could very well not be the full story.
Any thoughts?
|
|
|
Post by cautious on Sept 20, 2014 6:58:42 GMT
Hello moneyball,
Take a look at the 'provision fund' thread which also covers this topic.
You are right on Zopas lower coverage rate.
If you read their reasoning it leads me to the conclusion that their credit checks are perhaps more diligent than RS because they are pegging the provision fund level to that of their bad debt experience........plus their bad debt figures are readily available and I check them weekly.
RS seem to place money into their provision fund 'just in case'.
I'm not saying one is better than the other, but to be able to compare figures would be nice.......after all its our risk.
|
|
|
Post by westonkevRS on Sept 20, 2014 7:07:31 GMT
This is a good topic, and I'd be interested to know what the "ideal" coverage is. Of course this depends on the bad debt estimates being right, but also the level of balances being covered. RS show both.
But I've been in financial services for over 20 years and worked through enough recessions to know that future estimates based on historical performance can quickly go out of the window. So as in my simplistic view, the bigger the fund the more robust through the economic cycle, irrespective of what clever or prudent methods are used to estimate losses or stress test a portfolio.
Kevin.
|
|
|
Post by mattr on Sept 20, 2014 9:00:10 GMT
I'm generally inclined towards the bigger is better theory too. Except that the bigger fund is of course coming at a cost to someone. There's a zero sum game here of:-
Fees and interest payable by borrowers = RS income + Provision Fund Contribution + Lender Income
This goes back to what I was saying on an earlier thread about transparency. While RS is OK at this, we really don't have much of a clue about that apportionment process, and that impacts our income as lenders. I'm trusting RS not to kill the goose that lays golden eggs by either getting too greedy or too reckless (either is bad news for us). Longer term, I think RS should be looking to be more transparent. RS could of course play hardball and say "That's up to us, we're not letting our competitors see under the bonnet, if you don't like the rates go elsewhere." That would certainly be a traditional response to a consumer question, but I hope RS is better than that.
|
|
jcb208
Member of DD Central
Posts: 820
Likes: 606
|
Post by jcb208 on Sept 21, 2014 9:10:49 GMT
The provision fund is fast approaching 8 million and like many others I hope they continue to build on this.Not only will this help in hard times but allow ratesetter to continue growth due to the more cautious investors thinking this business is now a pretty safe place to invest money
|
|
gnasher
Member of DD Central
Posts: 207
Likes: 146
|
Post by gnasher on Sept 21, 2014 10:12:53 GMT
I am in the bigger is better camp too. I lend on several p2x platforms and RS is my boring and safe option (where boring and safe is of course good). As time goes by it is getting a higher proportion of my new money.
The very essence of RS for me is that there is no DD, and I can expect to get every penny back, and not suffer the tax implications of higher rates plus defaults. I am very happy to get a slightly lower rate if that is what it takes to build a bigger fund.
|
|
|
Post by elljay on Sept 21, 2014 10:50:23 GMT
and I can expect to get every penny back Fingers crossed!
|
|
|
Post by geoffrey on Sept 21, 2014 14:04:25 GMT
As it *appears* that RS is lending at slightly higher rates of interest than ZP, it makes sense that the cover ration of the Provisions Fund should be higher on RS, since a higher rate must translate into higher defaults. Having a large PF does mean that RS can gradually explore the territory of loans at slightly higher rates, using its statistical experience to make informed estimates about what the balance between higher rates and higher defaults is likely to be. As long as this is done gradually and is assiduously monitored, there is enough spare capacity in the PF to ensure that the percentage going into it can be adjusted in the light of changing circumstances.
The real test of the PF will come when the BoE interest rate begins to rise off its rock bottom low. As many borrowers will also have mortgages and credit card loans, this will begin to squeeze them (not nice). Then we'll all be very thankful for RS's cautious approach to the PF!
|
|
Investor
Member of DD Central
Posts: 662
Likes: 590
|
Post by Investor on Sept 21, 2014 18:30:56 GMT
BoE interest rate rise decision in August was split 7-2. Have a feeling that if we see a 6-3 split this time round we would be seeing rates rise this side of the General Election. Fairly certain that any initial rise will be in 0.25 increments to give the mortgage payers time to start adjusting to the change.
|
|
|
Post by westonkevRS on Sept 21, 2014 18:53:11 GMT
Just to counter Geoffrey ascertion that RateSetter should have a higher coverage ratio if the portfolio included higher risk loans. This wouldn't necessarily be true, as long as the expected loss calculation already reflected this. A higher expected loss number would reduce the coverage ratio, so to still maintain a higher coverage would imply a significantly larger Provision Fund in absolute pounds and as a ratio of outstanding balances.
Of course this requires an acccurate forecast of future losses (or as accurate as any forecast can be). And for the record, RateSetter's portfolio isn't (IMHO) especially higher risk. In fact read Equifax's report quotes in our brochure and they state RateSetter has the lowest default rate of any loan provider that subscribes data to them.
Kevin.
|
|
|
Post by geoffrey on Sept 25, 2014 5:31:12 GMT
Just to counter Geoffrey ascertion that RateSetter should have a higher coverage ratio if the portfolio included higher risk loans. This wouldn't necessarily be true, as long as the expected loss calculation already reflected this. A higher expected loss number would reduce the coverage ratio, so to still maintain a higher coverage would imply a significantly larger Provision Fund in absolute pounds and as a ratio of outstanding balances. Yes, the definition of cover ratio would be the amount in the fund divided by the predicted default totals for the current loan book. So everything depends on the accuracy of the prediction. My contention is that the higher the average loan rate, the higher the uncertainty in the predictions, so the higher the cover ratio needs to be to account for the uncertainty. I'm sure RS's expertise goes a long way towards mitigating these risks. I do think that RS is very right to be cautious, though, given that macro-economic uncertainties could play a huge role in whether borrowers can continue to afford repayments. Anyone remember the surge in rates just prior to the ejection of sterling from the European Exchange Rate Mechanism? Unthinkable (now) BoE rates in the teens, and a huge housing slump...
|
|
sl75
Posts: 2,092
Likes: 1,245
|
Post by sl75 on Sept 25, 2014 11:54:28 GMT
One might also consider that the necessary coverage ratio for a provision/safeguard fund would depend in part on how conservative the initial default estimates had been...
e.g. if one company bases their default estimates on an average case (defaults likely to be about this level in an average year) and another bases their default estimates on a 95 percentile case (defaults likely to be at or below this level for 19 of every 20 years), the required coverage ratio would be rather different.
|
|
|
Post by westonkevRS on Sept 26, 2014 20:21:43 GMT
The future is, by it's nature, inherently unpredictable. We can forecast based on previous performance, but that's just rear window analysis. At the end of the day you have to know that at some point in the future performance will be a lot worse than your previous averages. You just gotta make sure you've enough gas in the tank when that inevitability occurs.
|
|
|
Post by westonkevRS on Nov 18, 2014 11:21:09 GMT
In response to both a customer email and a RateSetter blog comment, we've published externally the response on how the coverage ratio is calculated and how it can be volatile on a month-to-month basis. Hopefully this volatility will reduce as volumes grow, but also the coverage grow as the portfolio matures and the prudent " straight line" assumption of first year defaults to the remaining term dissipate: www.ratesetter.com/Blog/Article/Robust_protection_or_a_flawed_guarantee_#comment-1699050959" Thank you for the question and thank you for being a RateSetter lender.
The anticipated bad debt was originally based on loan approval rates and the expected similar prime portfolio’s “run books” supplied by Equifax and CallCredit. RateSetter as a young company based it’s early forecasts on this analysis, together with reference agencies expertise and RateSetter analysis on application quality.
Now that RateSetter has started to have sufficient volume the expected loss is based on actual loan performance. This is calculated on a 12-month approved loan cohort and 12-months subsequent performance, in line with traditional credit scoring and expected loss forecast methodologies used by financial services firms. This is calculated for every open loan in real time based on it's risk profile, it's amortisation through the loan and outstanding balance. The actual segmented expected loss rates are calculated monthly, where one monthly cohort is removed from the equation and a new month is included. This can cause some volatility in the calculations if the next new month (from 12 months ago) enjoyed higher bad rates than the month removed. This was the reason the expected loss figure and hence the coverage dropped in September and October.
I would say that the estimated forecast methodology does include a number of prudent conservative estimates. This includes “smoothing” the bad debt by credit score segments and assuming equal bad debt through the lifetime of the loan when in reality bad debt drops considerably as loans mature. The first 12 months of a loan cohort segment is always the worst for losses, and hence the expected loss quoted is a maximum value that might actually be much lower when the cohort has fully amortised. There were no sudden bad debts realised and the actual cash in the Provision Fund continued to grow. This is real; and not forecasts which can always be open to scrutiny. As I write we have £9.3m in the fund. Note that as recently as November 2013 the coverage ratio was 190, and has increased as the debts haven’t realised, money is left in the fund and new Provision Fund contributions are made. Our philosophy is to grow the Provision Fund to bad debt coverage so that we are strong through future economic head winds when they inevitably come" Kevin.
|
|
c88dnf
Member of DD Central
Posts: 364
Likes: 266
|
Post by c88dnf on Nov 19, 2014 17:04:20 GMT
In response to both a customer email and a RateSetter blog comment, we've published externally the response on how the coverage ratio is calculated and how it can be volatile on a month-to-month basis. Well done Ratesetter. If only every P2P provider was as open with both policies and their trading figures!
|
|