P2P Lending: what worked and what did not

TLDR: if you ask me, most did

When I started this blog four years ago, I planned to dedicate ample space to p2p lending. I had been experimenting with investments in that space for three years and, at that time, I was starting to experience the first, unfortunate, negative surprises. Other blogs were mainly providing enthusiastic reviews and promises of regular double-digit returns that required a minimal amount of ‘work’ (passive income yay!!); few were having a sober discussion about p2p, exhibiting positive aspects but also possible risks. The market was frothy and the biggest reason was the generous incentives, in terms of cashback, p2p platforms were paying. I thought a critical analysis would provide value for readers…and value for me, thanks to the cashback system.

A lot has happened since. If you are curious about my past writing, posts are all classified under the category called…”p2p lending”; there are too many to link, the one that got the most success was this one.

A long time ago, in a galaxy far far away

The p2p industry faced its first reality check in 2019, when some platforms scams halted investor redemptions and ultimately disappeared with their money. COVID inflicted a second, close-to-mortal blow a year later when small businesses had to shut down and consumers in Emerging Markets lost their jobs and, consequently, their ability to repay their loans. This year, the Russian invasion of Ukraine represented a sucker punch to a sector that was slowly recovering, given platforms’ heavy exposure to the aggressor, the victim and the countries around them.

P2p lending was born on the premise of disrupting how banks lend money, closing the spread between deposit and loan rates, but also extending loans to entities that could not previously have access to them; it was ultimately a bet on things getting better, a more fluid, interconnected and borderless financial world facilitated by frictionless data collection feeding predicting models that could only be dreamed of when banks built their infrastructure and frameworks.

We did not exactly end up there.

What did not work

Last month Benn Eifert wrote a great post on bullshit in investing. Here is a section where he talks about yields in crypto but can be used for p2p lending as well:

yield has to come from somewhere. If you can’t understand in simple terms where yield comes from, what risk you are being compensated for bearing, then the yield is likely not sustainable.

The biggest misconception about p2p is around the advertised high yield they offer: some people read 12% and think guaranteed 12%, others read 12% and think scam. I never properly wrote my theory on where that yield comes from, a failure that put me (and this blog) in a bad light for both parties: on one side I was daring to be sceptical about some platforms, I was having fun staying poor, on the other the blog has a cashback page, the capital sin that makes any of my arguments invalid.

The yield is a reflection of the ‘distance’ between lenders and borrowers. Banks are typically the actor that facilitates the movement of capital from lenders to borrowers but they do not always engage in this activity. A bank’s decision to whether lend or not is based on a risk assessment; sometimes that assessment does not exactly make sense. I inquired my bank for a mortgage when my wife was pregnant and in-between jobs: if I wanted her name on the mortgage, the risk according to the bank would go up and therefore the applied spread. They explained the reason to me, which sure enough did not consider the most likely scenario, i.e. that my wife would find a job once my daughter will be at the nursery. This is just an example of how banks’ risk management departments see the world in black and white while everything out there is just shades of grey. Amex is bothering me every other month to increase the limit on my card to £8k/month but when I tried to switch to a new, basic card from Capital One (so that I could claim a new welcome bonus in two years with Amex) they rejected my request, I was too risky according to their assessment. They looked at the same data Amex is looking at…

There is ample space for someone with a better risk model some common sense to jump in and fill the gaps. Banks often do not lend to clients that are ‘too small’: from people that want to borrow against their investments (thank god Interactive Brokers!) to small Real Estate developers (welcome EstateGuru), up to niches like litigation funding (howdy AxiaFunder). Yields are a reflection of borrowers’ appetite and competition; if you are the only lender in town, you can charge more.

High spreads between lender and borrower rates are a requirement to compensate for banks’ high-cost base: physical shops on every high street corner, layers and layers of legacy IT infrastructure, the salary of the guy that chaperone the client from the branch door to the ATM. “Re-build” the business in a leaner and more efficient way and more juice will flow to lenders.

Did you ever check credit card rates? Dudes have always borrowed at crazy rates…but it was hard to notice because credit card companies had all the interest (pun intended) to keep rates as hidden as possible. If you want to go on holiday tomorrow and you do not have the money, what you check is the final monthly cost of borrowing to make it happen: £250/month, can I afford that? yes, ok let’s go. If you are in that mindset, you do not rationalize it like “wait, if I save £150/month, I can take the same holiday next year at a way cheaper price”. Borrowers do not check the interest costs in % terms, just how much is the final cost compared to their budget (more often than not, they do not even check that and simply assume they could pay it anyway).

Ultimately, high rates compensate for high default rates: fools yield, as someone calls it. It means that if you charge 20% interest to a pool of borrowers, your final return will be lower than that because some borrowers will default. Usually, your return will be way lower than 20%, in the single-digit area. This is because you are lending to someone that got rejected everywhere else and that rejection cannot always be explained with banks using too conservative risk management algos.

For all the above reasons, I think that 7% – 8% returns can be reasonably achieved if the investor manages to create a sort of “beta exposure” to the sector. This brings us to the second failure, aggregators like Mintos. There was an enormous amount of platforms at the peak of the p2p craze, and still, there are way too many now after closures, defaults and exposed scams. To achieve that beta, an adequate exposure to loans issued by hopefully legit originators that represents a fair sector, type and geographical diversification, investors have to dedicate an extensive amount of time to research. Aggregators would solve this problem by internalising the research and offering a one-stop shop for investors. Unfortunately, no one so far has succeeded in this. I stopped researching for new platforms more than a year ago and several bloggers in the space that I respect have moved on; it might be there is a good aggregator out there that emerged recently but I would doubt it.

The third failure is the disruptors who became…disruptee. Zopa in UK and Lending Club in the US were the OG of p2p lending and both are now banks. I see this more as an indication that, when done right, the business can stand on its own legs (yields are legit); so much so that some platforms decided to keep yields for themselves than pass them on to investors. This represents THE risk for retail investors, meaning platforms are “using” them to test and validate the product and, once ready, they either keep profits for themselves or turn to bigger pools of capital that require less effort from the platform side and are fine with lower returns (think family offices, hedge funds in the credit space, maybe some enlightened pension fund). It’s like adverse selection in insurance: the good platforms you found will disappear and only the scammy ones remain. FellowFinance was definitely not the best platform I invested but their returns were decent: they also decided recently to become a bank…

When I read the Multicoin Capital post on The Generalist I realised one of the many reasons why I fail at this game: when things get tough, I lose interest. I have a limited amount of time to dedicate to research (these days, almost none considering the kiddos) and after some bad experiences (Viventor and DoFinance) I got allured by other shiny objects like crypto and strategic asset allocation models. The decision was not rational, I simply stop reading about one topic and move to others. This way, part of the knowledge I built is wasted and the time dedicated lost. Now that the tide is low and many platforms are naked, should be the moment to find the good ones. This failure is 100% on me. I always leave (a big) part of the research process to other bloggers/writers, the fact that most of the ones I know abandoned themselves p2p lending did not help either.

What worked

These are the platform I am currently using: Twino, Viventor, ViaInvest, LinkedFinance, FinBee, Bondora, DoFinance, EstateGuru, InRento (was EvoEstate), FellowFinance, Kuflink, Mintos and Iuvo. In the past, I used LandlordInvest, Investly and Cashare and I closed those investments with a profit.

Given that any platform can go sayonara from one day to another and I cannot immediately liquidate my investments, it is hard to say if a profit that got re-invested within a platform can be considered a profit. Also, defaulted loans can still generate future cash flow through the recovery process. I can only draw some partial conclusions here.

Out of the 16 platforms I used, 2 got in serious trouble: Viventor and DoFinance. That’s a 12.5% fail ratio, not great. To be fair, I wrote in the past about platforms that I thought were scams and turned out to be scams…worth some moral points at least? The good news is that my strategy of withdrawing regularly 50% of profits generated by each platform lessened the loss: I have at risk ‘only’ 64% of my original investment in Viventor and 50% in DoFinance. On the positive side, they are still making payments so that % will decrease further in the future.

I wrote extensively in the past about how I consider Bondora very shady at best. It is still in the active platform category because it takes time to unwind my investment; as of writing, I have left only 15% of my initial commitment to withdraw plus a small, so far paper, profit. Same situation, different reason, with FellowFinance: since they became a neo-bank and stopped to issue new loans, I started to withdraw.

Mintos and Iuvo are in a ‘grey zone’: some of the originators on their platforms went belly up and the final performance will depend on how effective and efficient their recovery process is. Those originators got in trouble because of Covid and the Ukraine war, two events that can be genuinely defined as black swan-ish. Ish because those are also risks quite common when investing in Emerging Markets (in my professional experience I had to deal with defaults in Argentina multiple times, Lebanon, Ecuador, Egypt, Greece and most likely others that now do not come to my mind…I know the drill). The main issue is that the platforms did not provide enough diversification, so that those cases could be easily absorbed by gains in other loans.

All the others are in a good-to-great position. Linked Finance had an impressively low amount of defaults for a platform lending to small businesses that had zero revenues for most of 2020 due to Covid. EstateGuru and Kuflink did not generate a loss so far on any single loan for any of their investors (although I had a default on EG more than 5 years ago and the recovery process is not concluded yet). Twino, ViaInvest and FinBee did so well that I was able to withdraw 50% or more of my initial investment while the capital invested are 50% higher than what I put in (i.e. 66% of the capital that is now generating interests is ‘house money’). FinBee is actually a great empirical example of not-so-fool yield: I got many defaults there but the initial yield was so high that the overall return is still above 15% (gross of the withholding tax).

The Future

When I started to look at p2p lending I knew it was a high risk / (ideally) high reward investment. I believed and still believe the yield is coming from legit risks, not Ponzi shite. There was a clear transparency issue, this is why to get the p2p beta I diversified a lot along multiple axes: platform, loan type, geography and time. I wanted it to be passive, but I knew there was (unfortunately) a lower bound on how passive it could be.

I consider the losses I listed above nothing more than the tuition to the investment university. I knew there would be some. I paid my tuition in stocks, stock picking, shorting stocks, IPOs, bonds, distressed debt, FX, options, technical analysis trading…you name it (I did not include crypto only because it came after). This is why even in the most optimistic phase, even when I was reading other bloggers gloating over their (paper again) profits, I never diverted a huge %s of my portfolio to the space.

In a natural process of elimination, month after month bad platforms get more irrelevant while the good ones get closer to the point where they will generate new cashflows out of past profits. And those will be truly passive income. I am now diverting all the profits into IB because I have to reach the point where they will, finally, let me trade US-listed ETFs. But I hope in the not-so-far future to get the chance to look at new, promising platforms like Landex.

Twino, ViaInvest and Mintos recently became regulated platforms. How much more secure they are is still to be tested (financial regulators missed so many scams that is not even useful to list names) but we are definitely going in the right direction. However myopic, is better to have another set of eyes looking at those platforms than not; at least there is a well-defined group of people that might face real consequences in case something funky happens, not CEOs that turned out to be just fake LinkedIn profiles. Plus, each platform is now part of the EU investor compensation scheme, a program that ensures a minimum level of compensation per investor of EUR 20.000 (this works only in case of fraud or other malpractices, not if the investor buys shitty loans).

Godspeed?

There are still multiple storms at the horizon that can capsize the p2p boat for good.

Do you remember the Great Financial Crisis? A contributing factor was the conflict of interest for banks issuing mortgages that were not kept on their books; they were generating risk and then passing it to investors. P2p lending suffers the same problem and it will never go away. The silver lining here is that banking activity was run by agents that were concerned about next year’s bonus, not the long-term health of the bank. Here the owner of the p2p platform should have an interest in the platform long-term. But the possibility of reckless lending just to hoard commissions is high.

Another conflict, this time for investors, is between compounding and default risk. In simple terms, compounding and default risk are both positively correlated to the amount of time capital stays invested: what’s the optimal amount of capital to risk per unit of time? When I started to invest in p2p, I thought that default risk was inversely correlated to time: the longer a platform was operating, the less the chance it was a scam. Viventor proved me wrong (I am not sure how much the change of ownership was a factor in the platform default). If we are dealing with a “Taleb’s turkey”, $XIV type of investment, regular rebalancing, i.e. taking profits out, is the optimal strategy. If we are dealing with a legit investment, the more we re-invest profits, the faster we compound. At this stage, I prefer to leave some profits on the table and withdraw regularly than going all-in, even on regulated platforms.

The more things go well, the higher the chance other platforms will close to retail investors. I might be right about p2p lending and still be the one crying…although it will be a bit like this:

In this case, not re-investing profits would have been a big mistake.

One of the main reasons I decided to invest in p2p lending is that (I thought) it represented a risk uncorrelated to traditional stocks and bonds. In the ensuing years, reading about Crisis Alpha and Downside Diversification and the empirical evidence of the Covid-crisis made me change my mind. Some p2p loans are more uncorrelated to the global economic cycle than others: EstateGuru, LandEx, AxiaFunder, InRento presents more idiosyncratic risks than global ones. In general, p2p diversification comes more from its lack of liquidity and timely mark-to-market valuations a la Private Equity than real uncorrelated dynamics. A “black box” like Twino or ViaInvest has printed a streak of years with a 10% return or above that can almost equal Lebron reaching the NBA Finals (7 years and counting vs 8) but a lot is happening behind their guaranteed returns.

Lastly, I cannot avoid mentioning that my cashback plan failed eminently. I received way way more slander for publishing the links than actual clicks. A little lesson for the SEO expert I’ll never be: go big or do not even bother.

What I am reading now:

Follow me on Twitter @nprotasoni