Marketplace lenders have taken a lot of hits in the last year, but can they keep the battleship afloat? (Image source: Threadless, Wallpapertube)

Flying into the headwinds of P2P lending

Peer-to-peer lending (P2P, also called marketplace lending) has accounted for the largest proportion of funds raised in fintech over the last few years: it was the right product for the right market at the right time. But there are a number of signs that an industry shakeout is likely in the making, as many of the macro trends that have proved to be a boon are starting to reverse out.

The fuel for the P2P fire: retreat of banks, low rates, and good credit environment

In the period following the great credit crunch of 2008, a sizable drop in the availability of credit left a void for other players to fulfill. Traditional lenders like banks were busy fighting for their survival, with many of their lending practices in the boom years to blame. As a result, banks put a pause to giving out loans while they tried to figure out what went wrong, thereby making it harder for consumers and businesses alike to borrow. The scale of this impact is visible in the chart below, which exhibits a slight dip in total US consumer credit.

Total US consumer credit (excluding mortgages), in trillions of dollars (Source: Federal Reserve)

Don’t let the significance of that drop go over your head: this line represents ALL credit outstanding at the moment (except for mortgages, which would skew the picture given the size of those loans). In order for it to drop like that, it means that more credit was lost than created. Almost no new credit was originated (you don’t get a loan, and you don’t get a loan…nobody gets a loan!) and outstanding loans were paid off or charged off. That’s a very unusual and ominous kink and it makes me think of the trendline for global human population, which has almost always moved up. The last time there was downtick, the Black Plague wiped out half of Europe.

Both dips were caused by the bubonic plague. Also, while “insignificant” from an evolutionary standpoint, it wiped out 30–60% of Europe’s population at the time (Source: infogr.am/the_black_death-7)

This view is even more dramatic when you look at the trendline for revolving debt (let’s use this as a proxy for credit card debt). While not a perfect comparison to peer-to-peer (P2P) loans (since P2P loans have a set term length often around 3–5 years vs. credit cards that are supposed to be paid off monthly), both appeal to similar borrower bases and P2P loans have been pitched in the past as credit card alternatives.

Revolving US consumer credit, in trillions of dollars (Source: Federal Reserve)

In addition, lenders benefited from a low interest rate environment, providing them cheap cost of funds even without a deposit base. And they’ve benefited from a favorable credit cycle, where charge-offs have managed to stay low.

To recap: P2P lenders have done well because they stepped into a historic gap, they’ve benefited from low cost of funds, and the credit cycle since the last financial crunch has been favorable.


But the honeymoon’s over: all that good stuff listed above? Yeah, no…

Interest rates are starting to creep up and Fed Chairwoman Janet Yellen has indicated further increases are likely to come. While that may be great for central bankers to “restock” their fiscal policy arsenal and ensure greater moderation in growth, it means that the cost of business will go up for lenders. In order to loan out money, these firms have to borrow (or hold) it from somewhere else, and it’s not easy to offset this impact by simply charging borrowers more (e.g., due to competitive pressure from traditional and alternative lenders, regulations like usury rate caps, etc.).

1 year US Treasury rate, based on rate from end of period (Source: Federal Reserve)

While we are still at historic lows, this is the highest rate environment that these P2P lenders have faced (about 2–3x what we’ve seen the last 7-8 years), and it will only get more expensive.

More worryingly, charge-offs are starting to tick up. Doesn’t matter if you look at all consumer credit or credit cards, both show a slight increase. Bad debt is a double whammy: not only do you lose interest earnings today, you also lose the original capital, preventing you from earning any future earnings and offsetting your income on other loans.

While charge-offs for total consumer credit have declined materially since the dark days of 2009, when the economy was in the dumps, it’s notable that the trend has hit an inflection point at the start of 2016.

Charge-off rate for all consumer credit (Source: Federal Reserve)

This is also true when just looking at charge-offs for credit cards. Again, bad debt levels peaked in 2009, got under control in the following 5–6 years (in fact, they show even better results and discipline than pre-recession figures!), but are starting to rise.

Charge-off rate for US credit cards (Source: Federal Reserve)

So that’s the overall US picture. What about P2P lenders? Let’s take Lending Club, a leader in the space with around $9 billion in originations in the last year and in operation since 2007. I’ve compared their charge-offs with for a 7 year period stretching from 2007–14 with Federal Reserve data, again both for all loans and credit cards only.*

Comparison of charge-off rates across Lending Club, all US consumer loans, and US credit cards, from 2007–2014 (Sources: Lending Club, Federal Reserve)

It’s probably only worth looking at Lending Club’s performance post-2010. It takes a few years (and many thousands of working hours of damn smart people) to create a solid, reliable, and predictive credit scoring model, which ensures that issued loans are going to more creditworthy individuals. But even if we chalk up the first 3–4 years to a learning curve, a “stabilized” 2010–2014 Lending Club is posting charge-off rates more than 2x that of credit cards.

Maybe this isn’t a fair comparison since borrowers on these platforms may be less creditworthy than the borrower base that allows balances to revolve on credit cards. But during this period, the average interest rate charged by Lending Club was 14.25%, while the average credit card charged 15.07% in 2014. So you’re getting less income with greater risk of loss. No bueno.


Don’t get me wrong, I don’t believe the P2P industry will disappear. Despite the challenges these players have faced in the last year (layoffs at Prosper and Avant, the very public dismissal of Lending Club’s CEO, etc.), I’m fairly positive these marketplace lenders will remain a key part of the credit ecosystem in the coming years. However, there remains meaningful room to improve and an economic downturn will make consolidation and narrowed focus a likely outcome, with platforms choosing to merge, sell themselves to traditional lenders, or divest non-core businesses to ensure they can live to fight another day.


*Methodology note (super boring clarification on data — feel free to stop reading)

TL;DR: The 2012–2014 data is the “cleanest” view of Lending Club’s loan charge-offs, as it’s just enough time for the newest loans to show loss behavior while exhibiting a long enough track record to be confident that we’re not penalizing them for early learning lessons.

The reason I chose the 2007–2014 timeline is to ensure we are comparing a relatively “mature” Lending Club loan portfolio to US figures. When you issue a new loan, defaults are relatively low in the first year, but tend to quickly normalize in the following year, i.e., 18–24 months after the loan had been issued. This is especially true since Lending Club’s loans often have a maturity of 3–5 years, so while there will be further losses, in the outer years the impact will be more muted. So by stopping the clock at the end of 2014 and looking at the data today, there’s more seasoning and accuracy in the loss data, otherwise the numbers would likely be a little low.

By this same logic, Lending Club’s 2012–2014 data should have relatively few “mistake” loans from their early years when they were still learning the ropes. Even a 5 year loan from 2007 is already off their books, and again, most of the problematic borrowers would have likely defaulted somewhere between years 1–3 of the loan, not year 4 or 5.

So by stopping the clock at the end of 2014 and looking at the data today, there’s more seasoning and accuracy in the loss data, otherwise the numbers would likely be a little low. By this same logic, Lending Club’s 2012–2014 data should have relatively few “mistake” loans from their early years when they were still learning the ropes. Even a 5 year loan from 2007 is already off their books, and again, most of the problematic borrowers would have likely defaulted somewhere between years 1–3 of the loan, not year 4 or 5.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.