How Useful Are Borrower Guarantees In Peer-To-Peer Lending?
Borrowers provide personal and even corporate guarantees to say that their own personal or other business, including cash and their residences, can potentially be called upon to pay off a business debt. The business debt might be a small business loan or a property loan, such as a property- or a residential buy-to-let mortgage.
Borrowers don't always have a lot of other wealth, though. Also their personal financial situation can deteriorate after the business loan is taken out. On top of that, a court judge isn't always going to agree that a debt can be paid off using personalor might set inconvenient conditions for lenders when doing so.
Our chief risk modeller, who has modelled for virtually every bank and building society under the sun, set the scene on guarantees when 4thWay was founded in 2014. The bottom line is that we have taken a sceptical view of guarantees inas our starting point and any platforms that claim guarantees will substantially improve results for lenders will need to prove it.
Basically, guarantees may or may not have value. The value they have at each P2P lending platform might only be discerned from their results over time.
A lot of time has past since then, so I'm looking at a variety of platforms to see how much difference guarantees have made. Below, I've also shared other thoughts not specifically about guarantees, which are shown in italics.
LendingCrowd, small business loans
Note that LendingCrowd has temporarily paused its P2P lending to focus on government-backed pandemic loans. Even so, we can investigate how well its guarantees have performed.
- 12/100 guaranteed loans have turned bad at some point.
- 7/100 non-guaranteed loans have turned bad at some point.
Based on all historical loans that started over 18 months ago, 12 out of every 100 guaranteed loans has turned bad at some point, while it's just seven out of every 100 non-guaranteed loans. So borrower guarantees have not reduced the chance of LendingCrowd* (read review) borrowers missing lots of payments. Quite the reverse.
Too few loans have been approved to make those precise figures highly reliable, but there are enough to show that it's likely that guaranteed loans have genuinely been more likely to turn bad.
- 69% of bad-debt amounts have typically not been recovered on guaranteed loans.
- 61% of bad-debt amounts have typically not been recovered on non-guaranteed loans.
In terms of recoveries, guarantees again haven't proven their worth. The proportion of bad debt that has not been recovered is more or less equal.
Taking a simple average, 69% of the initial bad debt is not recovered, versus 61% for non-guaranteed. However, there are too few bad loans to make this particular figure statistically reliable in any way, so there might be no difference.
In researching this article, I happened to notice that LendingCrowd used to lose lenders around 80% of any bad debts after recoveries, which is just about okay for these kinds of loans. More recently, this has come down to 60%, which is about as good as it gets for small business loans. It's a very good improvement.
- 9.02% lending interest rates, before bad debts, has typically been paid on guaranteed loans.
- 8.48% lending interest rates have been paid on non-guaranteed loans.
On a simple average of lending interest rates, guaranteed loans earned just half a percentage point more, at 9.02% versus 8.48%. This will not have been sufficient to make up the difference in its bad-debt results.
The slight difference does indicate that LendingCrowd tends to take guarantees on its riskier loans, which explains the higher proportion of loans turning bad. Hidden from these figures could be a higher recovery of bad debt on guaranteed loans using the guarantees. They could in hidden ways be proving their worth.
Note that the lending interest rates are earned every year, albeit on a shrinking loan amount as the loan gradually repaid, whereas bad debts occur just once over the full life of the loans, and then parts of those bad debts are recovered. You therefore can't directly compare the lifetime bad-debt rate shown above to the annual lending rate. And doing so is beyond the scope of this article.
Conclusions on LendingCrowd's borrower guarantees
The spread between lending rates has been far from sufficient to make up for the far higher proportion of guaranteed loans that turn bad. For LendingCrowd, then, non-guaranteed loans have served lenders considerably better.
Lenders do need to spread their money across lots of loans as a priority over-and-above selecting non-guaranteed loans. In addition, LendingCrowd might well have already adjusted its interest rates or loan-approval processes as a result of the disparity I've described.
If you're looking to lend primarily in safer LendingCrowd loans, you'll still do better to focus primarily on its A and A+ grade loans than on its non-guaranteed loans. This is a more reliable way to lower your risks and you have more choice of loans.
Funding Circle, small business loans
As with LendingCrowd, Funding Circle has temporarily paused its P2P lending to focus on government-backed pandemic loans.
- 9/100 guaranteed loans have turned bad at some point.
- 9/100 non-guaranteed loans have turned bad at some point.
These figures are looking solely at Funding Circle's small business loans (so I mean excluding its abandoned property-) that started 18 months prior to their last data feed, which was, sadly, way back in June 2018.
- 75% of bad-debt amounts have typically not recovered on guaranteed loans.
- 85% of bad-debt amounts have typically not recovered on non-guaranteed loans.
Recoveries on guaranteed loans have been noticeably higher, with an extra 10% of the outstanding bad debt being recovered. This is what you'd hope for and probably expect, since it's the main point of guarantees.
Funding Circle's loan book is many dozens of times larger than LendingCrowd, making these statistics more meaningful, although there's still room for the figures to converge a bit more given enough time and more loans.
75%-85% of all bad debt being non-recovered is certainly okay and within normal bounds for these kinds of loans, but not best in class.
As with LendingCrowd, 9/100 loans turning bad at Funding Circle doesn't mean 9% of the pounds you're lending will turn bad and be lost every year. The loans take years to repay – and it's nine loans in ten that at some point in time suffer some default. To hammer the point home a bit more, close to 50% of the total lent in bad loans was written off.
So you can expect to lose roughly half your money on the loans that turn bad, before interest earned. The actual average losses over the life of all the loans was therefore more likely to have been around 4.5% at Funding Circle.
- 9.72% lending interest rates, before bad debts, has typically been paid on guaranteed loans.
- 10.80% lending interest rates have been paid on non-guaranteed loans.
On interest rates, non-guaranteed loans have earned over one percentage point more than guaranteed loans. On an, this is going to work out as more or less sufficient additional compensation for the somewhat worse recovery on loans that turn bad.
Conclusions on Funding Circle's borrower guarantees
Funding Circle's figures are professional looking, unsurprisingly, as it has had tens of thousands of loans upon which to learn from and improve its results, backed up by a very large team. Funding Circle has been adept at pricing its loans (i.e. setting interest rates and also borrower grades) from quite early on. It seems to understand its borrowers very well.
You can't choose your own loans at Funding Circle. But the above results show that lenders using Funding Circle needn't be alarmed if you're allocated a lot of non-guaranteed loans or vice versa. The difference between the two is quite small and you are paid higher lending rates to compensate for the additional risk in non-guaranteed loans.
Rebuildingsociety, small business loans
- 32/100 guaranteed loans have turned bad at some point.
- 21/100 non-guaranteed loans have turned bad at some point.
Rebuildingsociety* does high interest-rate, high bad-debt loans, so its bad debts are high. The proportion of guaranteed loans that suffer some bad debt, before any recoveries are factored in, is far higher than non-guaranteed. Again, I'm looking at Rebuildingsociety's more mature loans only and excluding more recent ones.
Rebuildingsociety's loan book is also small and therefore the disparity between the two figures could shrink somewhat over time – or expand further.
- 93% of bad-debt amounts have typically not recovered on guaranteed loans.
- 93% of bad-debt amounts have typically not recovered on non-guaranteed loans.
That's not a typo, both guaranteed and non-guaranteed loans have had the same results so far.
This article isn't about comparing P2P lending companies' results, but showing you examples of how guarantees have performed. However, if you're wondering, it's not a surprise to see that the bad-debt amounts recovered are worse with Rebuildingsociety than LendingCrowd and Funding Circle, since Rebuildingsociety is focusing at the riskier end of the business-borrower market. To that extent, its recoveries are not worrying, but rather expected.
What's really interesting to see is that, when you factor in the interest earned on the loans that turned bad at some point, the guaranteed loans have performed considerably better.
When you deduct the interest paid to lenders from the outstanding bad-debt amount, non-guaranteed loan amounts that turned bad have made a net loss of 66% for lenders, on average. Meanwhile, for guaranteed loans, the net loss on bad debts is just 35%. So, in this regard, guaranteed loans have certainly performed a great deal better.
Note again that it's not respectively 66% and 35% of the entire loan that is a net loss. It's 66% and 35% of the amounts that were outstanding on the loans at the point they became bad debts.
- 16.80% lending interest rates, before bad debts, has typically been paid on guaranteed loans.
- 17.09% lending interest rates have been paid on non-guaranteed loans.
On Rebuildingsociety's lending interest rates, these have been identical for both guaranteed and non-guaranteed loans, to within a few tenths of a percentage point.
This is interesting, as it tells us something about the large amount of interest earned on guaranteed loans that turned bad. That interest was either due to borrowers struggling on to pay for longer to avoid facing the guarantee. Or it shows borrowers managed to pay a lot more additional and penalty interest after the loan turns bad, compared to borrowers who haven't given a legal guarantee. Other data in Rebuildingsociety's loanbook feed indicates that it's a mix of both reasons, but that it's mostly from the additional/penalty interest after the loans have turned bad.
…To continue the side theme I've been writing about in italics, Rebuildingsociety has paid out £4.1 million in interest on its more mature loans and has outstanding bad debts of £2.8 million on the same loans. This just further hammers home that annual lending rates can't be compared to the number of loans that turn bad. Otherwise, Rebuildingsociety's lending rates of 17% would clearly be insufficient to cover 30/100 loans turning bad. I think the pound amounts we're talking about make that clear.
Conclusions on Rebuildingsociety's borrower guarantees
Rebuildingsociety described to us in interviews that “personal guarantees depend on the individual, but they are valuable”. Certainly, in terms of interest, including penalty interest, received, the guarantees appear to be showing their worth. Yet Rebuildingsociety needs to watch that the spread between the number of loans that turn bad between guaranteed and non-guaranteed loans doesn't widen.
Assetz, property-backed small business loans
As with LendingCrowd and Funding Circle, Assetzis focusing on government-backed loans and is not available for new P2P lending at present. It's just the typical situation with small business lending at the moment…Perhaps, on reflection, this article on borrower guarantees was ill-timed!
- 11/100 guaranteed loans have turned bad at some point.
- 22/100 non-guaranteed loans have turned bad at some point.
So Assetz Assetz * (read review) has approved far fewer non-guaranteed loans, so the spread might shrink over time. But it's not likely to completely disappear, barring changes to Assetz 's processes.'s non-guaranteed loans have been nearly twice as likely to turn bad.
Assetz's loans are all secured against real property, which is far better than a guarantee when it comes to recovering losses. Lenders can therefore much more easily swallow a higher proportion of loans turning bad at some point than they could with Funding Circle or LendingCrowd.
- 30% of the total lent on loans that turned bad have not been recovered on guaranteed loans.
- 17% of the total lent on loans that turned bad have not been recovered on non-guaranteed loans.
If you've kept up so far and have just drunk a very strong coffee, you might have noticed that the last statistic was slightly different to the other P2P lending companies. With LendingCrowd, Funding Circle and Rebuildingsociety, I've shown you the typical proportion of debt not recovered on loans that turned bad versus the amount that turned bad in the first place.
In contrast, with Assetz , I'm showing you the bad debt not recovered versus the total lent, i.e. including the part of the loan that was repaid and did not turn bad at any point. The reason is simply that I don't have all the same comparable figures to hand for Assetz at the moment. But it's still sufficient for our analysis here.
Here, you can see that non-guaranteed loans have considerably less outstanding. It's possible that some of this difference is due to the non-guaranteed loans in Assetz 's historical loan book being more mature, i.e. a greater proportion of them are even older and therefore Assetz has had more time to recover even more cash from the borrowers.
Also, since less than 82 loans in total have turned bad, and just 20 non-guaranteed loans have turned bad, more time is probably need for these figures to settle where they truly belong.
- 8.15% lending interest rates, before bad debts, has typically been paid on guaranteed loans.
- 7.76% lending interest rates have been paid on non-guaranteed loans.
There's basically no difference in lending rates between the two types of loans.
Conclusions on Assetz's borrower guarantees
Non-guaranteed loans have turned bad twice as often as guaranteed loans and lending rates are basically the same. This might, however, be offset by the non-recovered bad-debt amounts; for non-guaranteed loans, those remaining bad debts might be potentially half as much as with the guaranteed loans.
More time and more loans are needed to clarify the performance of Assetz's guarantees some more. Assetz does the most bespoke types of loans of all the providers I've covered today. And, after stricter P2P lending regulations in the past year or two, its loan-approval processes have also been tightened up.
You've also got P2P lending companies like Proplend* (read review), CrowdProperty (read review) and Kuflink* (read review) that have either zero or near zero bad debts, based again on loans issued at least 18 months ago. These companies have a mix of guaranteed and non-guaranteed loans, so both types of loans have performed the same.
Naturally, none of the P2P lending companies I've covered today rely solely on guarantees to save lenders from an overall loss, so you can't assume that other factors aren't impacting these results. The next step in analysis like this would be to control for other factors, such as whether there was a first oron property . But the loan books will have to be much, much (much!) larger before we could pick the impact of guarantees apart to that degree.
I hope you've learned from this, as there are several lessons, including:
- Don't consider guarantees to have any value until proven.
- Spreading your money around across lots of loans is usually going to be more important than looking to lend just in the better performing guaranteed (or non-guaranteed) loans.
- When selecting loans, you'll probably get a wider selection and more reliable results by selecting using other criteria, such as the borrower grade. (At least when a P2P lending company is good at grading.)
- And the data clearly shows that you need to look beyond guarantees to see the overall quality of all the P2P lending company's lending processes, its borrowers, and ability to recover bad debt.
Which P2P Lending Sites Offer Protection?
Independent opinion: the opinions expressed are those of the author(s) and not held by 4thWay. 4thWay is not regulated by the ESMA or the FCA, and does not provide personalised advice. The material is for general information and education purposes only and not intended to incite you to lend.
All the specialists and researchers who conduct research and write articles for 4thWay are subject to 4thWay's Editorial Code of Practice. For more, please see 4thWay's terms and conditions.
The 4thWay® PLUS Ratings are calculations developed by professional risk modellers (someone who models risks for the banks), experienced investors and a debt specialist from one of the major consultancy firms. They measure the interest you earn against the risk of suffering losses from borrowers being unable to repay their loans in scenarios up to a serious recession and a major property crash. They assume you spread your money across hundreds or thousands of loans, and continue lending until all your loans are repaid. They assume you lend across 6-12 rated P2P lending accounts or, and measure your overall performance across all of them, not against individual performances.
*Commission and impartial research: our service is free to you. We already show dozens of P2P lending companies in our accurate comparison tables and we keep adding more as soon as they provide us with enough details. We receive compensation from Assetz How we earn money fairly with your help., Kuflink, LendingCrowd, Proplend and Rebuildingsociety, and other P2P lending companies not mentioned above when you click through from our website and open accounts with them. We vigorously ensure that this doesn't affect our editorial independence. Read