What would happen if...

Last night I was on a panel at the WeBank event at Nesta in London. It was a debate on the future of peer-to-peer lending and social finance. The panellists included Giles Andrews CEO of Zopa, and we had a lively debate, me playing the role of antagonist of course. I always seem to get cast in that role, as I was saying to a few people afterwards, the token banker with an innovation job who "doesn't get it".

I always laugh, of course. Being a bank innovator is quite a different thing to being one in a start-up, or even one in a well running company with some years of history such as Zopa. In exchange for our greater access to resources, we are also subject to far more significant constraints. We have to run what I call the "Can we get it up" test: an assessment of whether we have the political capital to actually make something happen or not. The more unusual and unprecedented something is, the more political capital we need.

In a start-up, by contrast, if you want to do something, you just do it.

Anyway, I was at this panel, and both the other P2P companies (the other was as yet un-launched Kubera Money) said something very interesting: they have cordial relationships with the regulators, and expected that to continue. In other words, they were expecting fairly light touch supervision.

Now, don't forget these are outfits running on fee income. They aren't going to have any fat in their operations, and so it is probably just as well they don't have the compliance load a fully supervised bank would have.

The regulator is also very, very busy, of course, particularly now.

But I can't help but wonder what would happen if one of the big four banks here in the UK rang up the FSA to tell them that they were entering peer to peer in a substantial way.

I simply cannot imagine that a light touch supervisory regime would be the result. Perhaps they'd go even further than the U.S has already gone on this topic.

Now this brings me back to the main point I raised on the panel. P2P is either disruptive to banks or it isn't. In order to be so, it needs to serve customers we can't, at a price lower than we can, or else be more convenient or have some other operational advantage which is unattainable for banks. Credit crunch aside, I don't really see how P2P has any of those things, frankly.

As Giles said yesterday at the panel, though, "we are absolutely going after share from the big banks". So these are not new customers going to P2P, then.

So rather than a disruptive, under the radar insinuation into the business which P2P might win, what we have here is a long drawn out competitive battle for share. That's a pity, in my view.

If they grow too big, banks will respond. The response will trigger a regulatory avalanche which banks are already set up to manage, but smaller P2P companies aren't.

And during all of this, they'll be engaged in a ruinous price war with bankers on fees and other charges. Costs going up, but revenue going down.

Not a pretty picture.

Comments

[John Wilson]

James, I very much enjoyed the event last night, especially the panel discussion. As I commented here

http://greatapps.blogspot.com/2009/01/webank-case-for-peer-to-peer-lending.html

you certainly seemed to have been cast as the pantomime villain for the "show".

I felt quite strongly that Giles was being disingenuous in making interest rate comparisons between Zopa and banks. When depositing with a bank you are transferring loan default risk to them. Losses are borne by the bank's shareholders and depositors also benefit from deposit protection schemes in the event of bank default. In Zopa you retain this risk, since it accepts no risk as a marketplace [other than negligence in credit scoring and vetting borrowers one presumes]. Hence some element of the difference between the rate paid by a bank and that received by Zopa lenders has to compensate for that. In conversation after the event, their Giles admitted to me that whilst the rate differential is about 6% following the sharp fall in bank deposit rates [Zopa lenders are averaging 9% less 1% Zopa fee versus average savings rates of 2%], back in the summer it was about a 2% differential.

What makes this latter figure astounding to me is that it indicates that Zopa lenders are evidently not figuring in an allowance for borrower default. Whilst average historic loss rates may be only 0.2% across all lenders, some lenders will have lost considerably more if they were the ones that had lent to defaulting borrowers. More significantly, whilst Zopa claim that their credit screening process rejects a considerable percentage of borrower applicants and keeps them clear of sub-prime loans, I suspect that the deterioration in the economy is going to push up their default rates in line with the experiences of banks on similar tranches of unsecured personal debt.

My assertion regarding default risk being overlooked by lenders was further validated when I enquired about whether Zopa would consider offering credit protection insurance and Giles advised that it had been offered but there had been minimal interest in the product. Perhaps people are being overly seduced by the touchy-feely aspect of "social lending" and become too trusting or are ignorant of the risks.

Whilst savers are undoubtedly complaining about the pitiful rates currently offered on deposits, in the current environment I suspect that many people are most concerned about return of capital than return on capital, at least temporarily.

[Colin Henderson]

Re 'long drawn out competitive battle for share".

The assumption behind that premise is that cost of capital is the same for a bank, as it is for the individual lenders in a P2P operation. [Note the individual lenders in P2P could be ndividuals or institutions in my view]

There are several factors at play:
- acquisition costs
- processing costs
- customer service costs
- technology costs
- overhead costs [eg innovation groups :-) ]

For each of those factors there are a set of facts, and assumptions. I know something of Zopa, Prosper, Lending Club, and a lot of CommunityLend's prospective model. Specifics aside it is a relatively safe argument that the cost structures associated with the five points above are not the same in a Bank or P2P lender.

In one sense you are right, that a Bank could weigh in at a loss in order to kill off he P2P lenders. But the mechanics of return on investment required by shareholders including the government who require [presumably] the bank to return to normal service as soon as possible would work against that approach.

The thrust of the argument presented is that if the situation warranted the bank would merely enter the market and sail happily off into the sunset. My contention is that is impossible due to the costs associated with the above. I may have to blog separately to cover details of what I am thinking of, but as one simple example what bank would be able to lever social media and forums for self help versus falling back on the 1,000 person call centre?

The notion of P2P is predicated on more than the P2P part for lending - it is predicated on a web 2.0 internet model without the trappings of traditional bank costs.

The defence rests.

[Giles Andrews]

@ James

I enjoyed the debate, hope you did too. I thought you made good points. We certainly aren't complacent about regulation and compliance but I think you are right to raise it as an issue.

@ John Wilson

I don't usually respond to posts but not sure I like being called disingenuous.

If you are going to quote my conversation with you please do so accurately and don't make misleading comparisons. All Zopa returns are quoted after Zopa fees so the current 9.1% is 9.1%. The returns before last summer showed a differential of 2%+ which lenders must have felt was appropriate at the time. Those returns are locked in for the period of the loan so the differential return from those loans will have risen significantly as interest rates have fallen. Risks of bad debt may also have increased in the same period but the question you should be asking is whether they have risen by as much as the interest rate differential has risen. Lenders' pricing reflects market forces of supply and demand as much as their perception of risk and the supply and demand equation has altered significantly.

Re creditor insurance, this product offers no protection to lenders but pays out to borrowers in the event their circumstances change (death, sickness and unemployment typically). Actuarial evidence suggests that insured loans actually provide a worse return to lenders than uninsured ones, presumably because the risk profile of those selecting the product is worse than those not.

You are correct in saying that our borrowers did not take up the product when we offered it - perhaps because they were of lower risk? The debate has now moved on somewhat with the competition commission likely to ban several versions of the product shortly and many high profile lenders withdrawing it only this week http://www.guardian.co.uk/money/2009/jan/20/ppi-sales-pulled, presumably in advance of being made to.

[JamesD]

John,

Zopa considered offering to lenders a default insurance product. During discussions it received a poor reception because it would only have been available for well diversified loan books. Since lenders at Zopa who are prudent use at least Zopa's guidance on an interest rate premium to add to allow for default risk that made the product worthless unless default rates rose to well above Zopa's guidance. It's possible that they will. Those who could have gained from the product were the ones with low diversification, who wouldn't have met the requirements to buy it. Zopa's easy lend screen asks the lender to specify a target rate after fee and bad debt allowance so it's likely that most lenders are including those bad debt allowances in the rates being offered to borrowers.

Rates to lenders in the 36 month markets are currently around 6-6.3% after fee and bad debt allowance. That tells you how new lender money is pricing things: memory of interest rates that used to be available from savings accounts for much of last year.

When it comes to rates, here's my experience:

Month Avg Rate Amount lent
2008 Aug 16.80% 980
2008 Sep 17.59% 720
2008 Oct 15.95% 530
2008 Nov 14.10% 120
2008 Dec 10.86% 1200
2009 Jan 10.74% 1010

These are rates before fee and bad debt allowance and include some Listings loans. For the August to November period my Markets (no Listings) rate after applying the fee and bad debt allowance was 12.7%. It'll be about 8-8.5% for the last two months but I haven't calculated it.

As you may know, Zopa will only generate a personal quotation if there's enough money in the market to fund it and will only accept an application after generating a personal quotation. Applications are always fully funded because the money for them has been reserved in advance.

The times when higher rates are available are when there is insufficient funding in the market to meet demand for loans, and hence some people aren't able to get a quote for the amount they want to borrow. At these times the cheapest offers in the markets run out of funding and offers at higher rates are included, increasing the rate for the prospective applicant and the lenders even when there is sufficient funding for the desired loan amount. That was happening quite regularly in the August to October period, didn't happen in December and has only happened on a small number of days so far this year, when loans above £14,000 weren't available in the 60 month C and Y markets even at the most well funded part of the day and dropped to a couple of thousand or less at times. So my lending shows two distinct periods: August to November when I was looking to maximise returns and compromising amount lent, then December and January when I was compromising rates for volume, but still with a fairly high rate threshold. During the early period my maximum exposure per borrower requirement was the significant limit on amount lent, during the latter it's the minimum target rate.

I seek to cream off the top of the available rates, so Zopa's numbers are more representative of the whole market.

[Thomas]

I should talk about how bank's capital cost advantage would only hold to the extent that they borrow short-term and lend long-term etc.

But I, and others, had this discussion in much much greater length here.
http://www.interfluidity.com/posts/1225607671.shtml

@Giles - Aren't u using http://technorati.com/ to track & reply to every high profile blog mention of Zopa?

[James Gardner]

@Colin: Presumably, your argument is based on the fact that for all five cost drivers you mentioned, the differential is substantial - perhaps even close to free - for P2P lenders.

My contention is that I doubt it would remain so if they had to compete like for like with a bank.

For example, I'd imagine if a call centre was avaialble to a P2P lender, they would choose to use it, especially if you want to get to scale, and the customers you reach aren't all the comfortable with forums et al. A bank in P2p would offer a call centre, therefore, a P2P lender would have to do so also...

[John Wilson]

@Colin

I was actually referring to points made by Giles during the debate and thereafter, as I stated in the "offending" paragraph.

My point remains that in quoting rates against those offered by banks for savings, you aren't comparing similar products, since a bank savings product and related rate includes deposit protection and avoidance of potential loss of capital arising from bad debts. Hence the gross differential overstates the rates gap. Consequently, my thanks to @Giles for providing a "cleaner" rate of approx 6% for comparison.

As recently as mid-December, HBOS were still offering 6% for 12 month deposits which challenges the notion that Zopa necessarily offers better rates. Obviously I readily acknowledge that the dramatic slump in the base rate has altered the picture considerably in the last few weeks.

As for credit protection, this is evidently a terminology misunderstanding between us. I am referring to protection for lenders i.e. insurance against bad debts and not cover for borrowers. I'd be curious how prices for such protection would compare with your estimates, bearing in mind that an Insurer would be liable for losses, unlike Zopa who don't financially stand behind their estimates for reasons I understand, namely that you are operating a market.

[Bruce Davis]

James, yes it was a difficult role to play in a debate with an audience split between the rationalists and idealists! Thanks for provoking the discussion and stopping us talking to ourselves (the p2p irony?)

the comments to your post have something of deja vu about them in that many of the ex-egg colleagues with whom we founded zopa (about which they were profoundly cynical it has to be said for the reasons you give, John et al).

As I rather long windedly put it in the debate itself, what zopa does is align a social experience with cultural ideals of self and place expressed through money - money being a cultural medium as it has always been in history. Zopa's return is always realistic (my bad debts are well below my expectations of my return which just forms a kind of bonus every year).

I am still working on all the different ways (www.oikonomics.typepad.com) that people create their own zopa on zopa - of which rational economic individuals are only one (minority) constituency.

I recommend to some of the commentators to try tackling Simmel's Philosophy of Money (written in 1904 but prescient of today) if you haven't already.

[Thomas Barker]

It might be useful to draw a distinction between p2p credit systems [like Zopa] and p2p monetary systems [like Scred.com or RipplePay].

A p2p credit system allows independent parties to reach bargains over the future distribution of money across an existing monetary system. [Here's $400, now wire me $120 for the next 4 months.]

Systems like Scred allow parties to act *as* banks and emit their own circulating credit.

(I haven't read Simmel, but I really enjoyed Money Unmade, which references it quite a bit.)

[mietwagen]

Sehr gute Seite. Ich habe es zu den Favoriten.