Free Updates

Exclusive content to your inbox for FREE!

The Case for P2P Lending as a Fixed-Income Asset Class Part 2

by Peter Renton on July 2, 2013

The following article is the second of a two part series by Matt Shibata, CFA, looking at P2P lending as a fixed income asset class from the perspective of an investment professional. The first article is hereMatt is a Portfolio Manager at Morling Financial Advisors. He invests in P2P loans via retail channels and advises clients who invest in P2P (and other asset classes) via institutional channels. Matt can be contacted at matt@morling.com.

Credit Risk

Just as Michael Milken found that junk bond rates in the 1980’s were sufficiently high to buffer against the losses from defaults, I see the same favorable dynamic in P2P lending today. P2P has higher rates, defaults, and total returns than many other types of fixed-income today. Drilling down within P2P lending, we find that higher-grade notes have fewer defaults, while “riskier” notes have higher returns (again, because the higher rates are more than sufficient to cover the losses from higher defaults).

While some commentators may dismiss P2P as an asset class, due to the higher default rates, we know that price (or yield) is the arbiter of risk and so we need to examine whether P2P compensates investors for the risk taken. Whether we define risk as an unknown outcome or the probability of loss, we need to quantify it. For simplicity’s sake, lets run a portfolio of notes through several scenarios.

Best Case: We could assume the status quo continues, which is to say consumer finances are improving and defaults remain low. Loss-adjusted returns may average 4-5% on a high-grade portfolio, 8-9% for a broad-based portfolio, and 11-13% for a low-grade portfolio.

Worst Case: In a worst-case scenario, á la 2008-2009 recession which saw unemployment surge and consumer balance sheets decimated, we learned a couple of things. If you look at data on all Lending Club notes, as well as independent analysis on sites like NickelSteamroller.com or LendStats.com, the only year in which total originations had a negative return was 2007. Of course, there was some pretty horrific performance out of loans originated during the fall and winter of 2008-2009. I do not have space to detail the composition, timing, and other more granular aspects of the recession performance, but this data can easily be pulled from the aforementioned sources.

Prosper notes had significantly worse losses during the recession, although I believe much of this is attributable to their early practice of allowing participants to set rates. This taught us that skilled and professional underwriting is essential to maintaining the integrity of P2P product. Furthermore, it supports (but does not prove) that originator-set rates/prices ensure P2P investing is a profitable experience.

Anyone can create their own stress-test by assuming higher default rates, but I am content to use the historical stress-test of the 2008-2009 recession. I should mention that it would be prudent for anyone using a historical or hypothetical stress-test to take into account the many changes made on Lending Club’s platform since 2007, some of which may improve returns and some which might reduce them.

Similar to interest rate risk, the credit risk of P2P loans compares favorably with other fixed-income assets. It is possible the US could experience another recession, perhaps even a deep one. If you only buy the riskiest notes during the worst months, you are definitely hosed (that’s a sophisticated institutional term). However, this would be hard to do even if you tried; if you diversify across grades and vintages, P2P is an attractive asset class.

Below are the monthly return distributions of all Lending Club notes, since inception, sorted by grade:

Monthly Return Distributions - LC Notes by Grade

And here are the monthly return distributions for all Lending Club loans, since inception, relative to high yield bonds and the S&P 500 over the same timeframe. Notice the x-axis scale when comparing the two charts…

Monthly Returns Distributions - Multiple Asset Classes

One last consideration is to understand where your credit risk exposure is. Typically, an ABS sponsor establishes a separate entity, called a special purpose vehicle (SPV), to buy and hold the loans and issue securities to investors. The SPV is considered “bankruptcy remote” from the sponsor, which means the loans are protected should the sponsor go bankrupt.

Prosper has established an SPV, called Prosper Funding. Prosper Marketplace (the service side) originates the loans, transfers them to Prosper Funding, who then issues the notes to investors. So investor assets are held in Prosper Funding, which is distinct from Prosper Marketplace; thus, in a hypothetical Prosper Marketplace bankruptcy, creditors should not be able to recover the loans/notes from investors.

Lending Club, on the other hand, holds the loans/notes on its balance sheet. According to the most recent prospectus, if Lending Club went bankrupt, then note holders would be considered unsecured creditors of Lending Club (not of the borrowers). Basically, investors are lending money to Lending Club, who originates and owns the loans and promises to pass the principal and interest payments back to the investors. So investors do not technically own the notes, as Lending Club states in its prospectus, and it is important to note that you are ultimately exposed to Lending Club’s credit risk (as well as the borrowers’). Given the current financial position of Lending Club, I am not too concerned about the lack of an SPV, but it is worth mentioning and investors should monitor Lending Club’s balance sheet regularly.

Interest Rate Risk

Since all P2P loans are short-term (3 or 5 years), are fully amortizing, and have relatively high coupon rates, duration (a measure of interest rate risk) on any given note is extremely low. I believe that the duration for any given note may even overstate many notes’ price sensitivity to interest rates, since the secondary market is not yet mature. It would be interesting to analyze Foliofn data for the past two months and see if secondary market prices dropped significantly in response to the rally in rates. Regardless, the interest rate risk is miniscule, which is ideal for times like these.

However, even if price fluctuations from rising rates are minimal, owning fixed-income in a rising rate environment creates an opportunity cost. This is because the rates on new loans will be higher than the rates on the loans you own. While I have read some research showing that 36-month loans have less credit risk than 60-month loans, I have not seen anyone mention the fact that there is a much lower opportunity cost to holding a 36-month loan, should interest rates rise. If rates rise, I assume the P2P originators will raise their lending rates. A shorter loan term will return your principal more quickly, which means you can reinvest it at higher rates sooner. Lastly, like many asset-backed securities (ABS), P2P loans have an average life that is shorter than their stated maturity, due to the borrower prepayment option. This further reduces the opportunity cost if rates rise (although this would be called “call risk” in a falling rate environment).

Looking around the rest of the fixed-income market, we find 2-year Treasuries yielding .26%, most 2-year investment-grade corporate bonds below 1%, and 2-year high-yield corporates in the  4-5% range. A 36-month P2P note has a similar duration to a 2 year bullet bond, but offers a much higher yield than all of these:

Yield Curve

P2P Risks

Beyond the risks common to all fixed income, the two P2P-specific risks I see involve the withdrawal of P2P’s competitive advantages: solid underwriting/risk-management and originator-set rates.

As we saw with investment bankers during the dotcom bubble and mortgage bankers during the housing boom, it is always dangerous to incentivize the origination of investment products, especially if the originators do not intend to hold product on their books. Although the P2P originators collect a 1% servicing fee, the majority of their revenue comes from loan origination fees. It is important to recognize this fact and monitor any changes to the underwriting standards.

Secondly, as noted previously, the fact that the originators are setting the prices/yields and shielding the offering prices from market prices mean that excess returns are available. If this changes in any way, I would be very cautious as this is the crux of the P2P value proposition.

Conclusion

In summary, I do think P2P is a viable fixed-income asset class, with a unique advantage over similar ABS and high-yield debt. Given current dynamics, I think most fixed-income portfolios should include an allocation to P2P loans. Interest rate risk is almost non-existent and there is more than enough yield to cover the credit risk, except under the most dire scenarios. That said, I believe the asset class proved itself through the recent recession, relative to other debt. Investability (especially in size) remains an issue, although that is another blog post. Nevertheless, for this reason, P2P investing is an asset class in which individual investors can outperform the larger institutional investors due to the small size of the current P2P market. Of course, I suggest they borrow some of the tools from traditional fixed-income investment management to refine their strategies.

{ 16 comments… read them below or add one }

Rob Leonard July 2, 2013 at 1:42 pm

This two part series is terrific. Very insightful. Usually I like to take the other side of at least something, but find myself in fundamental agreement with just about everything here, and find plenty of food for thought on themes that had not occurred to me.Thanks!

Reply

Peter Renton July 2, 2013 at 8:57 pm

Thanks Rob. I completely agree. I wanted to publish these articles because there are several points that the author brings up that have never been mentioned on the blog before.

Reply

Matt S July 5, 2013 at 10:09 am

Thank you for the kind words, Rob.

Reply

B. Mason July 2, 2013 at 3:51 pm

It’s a good series, and forms the basis of how I feel about the asset class. One suggestion to the reader is to note that the author uses all loans through December 2012 to state returns and variance. Given that losses occur throughout the term and the huge acceleration of growth in LC originations (meaning that we are overweight in young loans), it would be a better presentation of returns and variance to focus on more seasoned notes. This choice most likely mutes the variance and overstates the returns a little bit.

Reply

Peter Renton July 2, 2013 at 9:12 pm

Hi Bryce, Good to have you back from down under. Hope your trip was great.

Including newer notes most likely overstates returns somewhat but the difference in the second chart showing the distribution of all loan grades versus S&P/High Yield Bonds is so stark it would have no material impact. And I think that comparison is most important here.

Reply

Matt S July 5, 2013 at 10:14 am

Mason brings up a good point that the rapid volume growth creates some caveats when evaluating the raw data (along with the short history too). As Peter mentions, from a high-level perspective, I think the raw data is sufficient. I didn’t cut and dice the data at all, just for the sake of simplicity and transparency.

However, when you are investing and fighting for every basis point of return, you will want to be much more nuanced, which is why I do use services like Mason’s P2P-picks.com.

Reply

suedwein July 3, 2013 at 3:42 pm

I very much agree with the author’s assessment around P2P being a viable and interesting asset class. I feel very differently about one of the arguments presented though. The chart comparing monthly LC return distributions to junk bonds and the S&P is misleading, as it seems to suggest that LC provides much less volatile returns in comparison to those other asset classes. Not only does the chart understate default rates based on as noted in a previous comment. More importantly, it seems to be comparing market returns for junk bond and equity indices to cash-based return analysis based on actual defaults in a broad LC portfolio. That would seem to be negating any interest rate risk or changes in actual or expected default as well as overall shifts in investor risk tolerance, which are prime factors driving the monthly return distribution of the comparators. If you measured junk bond returns on the same basis as the LC distribution above – solely looking at underlying bonds’ payments/defaults while ignoring any price action – you’d see a much more compressed distribution of returns as well.
If the chart above were an accurate reflection of the risk/return characteristics of the P2P asset class, it would render P2P easily the most desirable building block for any portfolio, by a long shot.

Reply

Matt S July 5, 2013 at 9:51 am

You are astute to note that the P2P return distribution uses what you call “cash-based” returns, rather than price returns. If I had used monthly total return data for P2P notes, to include price movements from credit spreads and changes in interest rates, I would need to get this data from the secondary market. To be honest, I have not used FolioFN much or evaluated historical pricing data, but I do think it would be extremely interesting to see how note pricing on the secondary market moved during times of stress (such as the past 3 summers). My intuition is that credit spreads would not widen significantly, due to the immaturity of the market (mainly, it is relatively small compared to primary market (which is why I did not use it), not integrated with other financial markets, and has high transaction costs). Also, many notes trade at a premium, which would have the effect of pulling performance forward in time and would at least partially offset any credit spread widening. I estimate most 3-year notes to have durations of 1 to 1.5 years, so the interest rate risk and volatility from rate movements should be extremely small (especially now, since the Fed is holding down the front end of the curve).

However, if anyone disagrees with the above rationale and wants to look solely at what suedwein calls “cash-based returns,” then you could just compare the yield-to-maturity of P2P vs other fixed-income. Instead of a distribution, you’d get a single point as suedwein mentions, which we refer to as yield-to-maturity. I included a yield curve in the post for this reason, so that readers could make an apples-to-apples comparison using standardized metrics. I’d also note that the bond curves are based on stated YTM, versus estimates of loss-adjusted YTM for the P2P curves.

Your question obviously brings up a few more questions and could be a long and interesting discussion, but I hope the above addresses the main thrust of your comment.

Reply

Dan B July 4, 2013 at 3:32 am

Although I don’t disagree with the main premise behind this series, I’m a little uncomfortable with a bit of the reasoning here. For starters the author seems to be suggesting that ’08-09 scenario is the “worst case” scenario that investors could face. I’m not sure if I would be so bold in making that assessment considering that there are many who can convincingly argue that we didn’t really fix most of the problems leading to that situation & that we essentially kicked the can down the road by (in many cases effectively) artificially re-inflating the economy…………….thereby potentially leading to an even more severe correction i.e. recession, down the not too distant road. You can see evidence that supports this argument if you look at the rather rapid recovery of commercial & some residential real estate in prime areas of the country. So I think that to state that ’08-09 is a worst case scenario…………………is a bit unwise.

Secondly, though I agree with the author that few can argue that p2p, as an asset class, fared comparatively well during the ’08-09 recession, I’m not sure how much can be gleaned from that performance when the next recession hits, Why would I say this? Because for starters we have no idea whether the next recession will be 1 year from now or 5 or 10 years from now. Normally, not knowing the time frame wouldn’t matter that much & wouldn’t prevent a responsible assessment……………..but in this case I think it does very much matter. If the next recession were to start next year, then we could probably come up with some pretty decent estimates based on what we know of the borrower from this year & the past few. But if the next recession hits 5 or maybe 10 years from now, it’s very possible that the profile of the 2014-2016 borrower & so on going forward will be somewhat different compared to today & potentially very different compared to the ’07-09 borrower. The 07-09 borrower was, after all, a very early adopter (if you will) He/she was using something that very few were even aware existed. It’s easy to forget because we here are all “early adopters” too. Bottom line is that I wouldn’t bet money whether the average person from that group was a better or worse credit risk than someone from this year…………………much less years from now. Just something to keep in mind, if you so choose.

Reply

Matt S July 5, 2013 at 10:08 am

Points taken. Anything can happen in the future and nobody has a crystal ball. Investing is often making probabilistic decisions, based on incomplete information (about the future). This post details the case for P2P lending as an investment today, relative to other fixed-income.

I did mention in the above post that: “Anyone can create their own stress-test by assuming higher default rates, but I am content to use the historical stress-test of the 2008-2009 recession.”

Reply

John Dowding July 4, 2013 at 9:15 am

I also appreciate and enjoyed the articles, but I had a question/comment on this paragraph:

While I have read some research showing that 36-month loans have less credit risk than 60-month loans, I have not seen anyone mention the fact that there is a much lower opportunity cost to holding a 36-month loan, should interest rates rise. If rates rise, I assume the P2P originators will raise their lending rates. A shorter loan term will return your principal more quickly, which means you can reinvest it at higher rates sooner.

One point that took me forever to grok is that P2P loans are not like bonds, where all the principal is returned at the end of the term. Rather, principal is returned continually through the lifetime of the loan. Given that, is the rate of principal return really that different between 36 and 60 month loans? If rate rises, I will be reinvesting *a portion* of my principal in higher rate loans immediately, just perhaps a modestly higher portion from 36 month loans. Maybe it’s just the “much lower opportunity cost” that I am quibbling about.

Any flaws in this reasoning that I am missing?

Reply

Matt S July 5, 2013 at 9:58 am

Yes, that is a fair quibble John :) “much lower” is subjective. I meant much lower in percentage terms, not absolute terms. The difference between a duration of 2 and 3 is only 1, but in percentage terms it is 50%. Just another one of the phenomena we’re seeing with extremely low yields.

One additional note to readers, which John brings up is that amortized loans repay principal throughout the life of the loan. However, principal and interest are not paid evenly; much more interest is paid initially and much more principal later. I would do a google image search for “loan amortization chart” for a visual.

Reply

Matt S July 5, 2013 at 10:17 am

Thank you for all for taking time to read, as well as challenge my assumptions and conclusions here in the comments section. I have tried to address each comment so far. At the end of the day, this post is more of a primer and many of you may have more advanced questions which we can delve into deeper in future posts.

Reply

Peter Renton July 5, 2013 at 5:01 pm

Thanks Matt, I appreciate you being so responsive here to the comments. This has been an interesting discussion and I look forward to you addressing some of the more advanced questions here in the future.

Reply

Mike S July 18, 2013 at 11:13 am

Very insightful article – I have never thought to compare P2P notes so directly to bonds before. As an active individual investor in P2P since 2009, I have definitely seen the secondary market landscape shift over the last few months, as originations skyrocket. Institutional investors are apparently gorging on 2nd market P2P notes, as the Folio platform has nearly dried up with everything except notes that are obviously troubled (late, on a restructured payment schedule, etc). Even when a good note can be found on Folio, if you wait more then 40 seconds to order it, it’s likely already been bought up by someone (or something). This is especially true now that both LC and Prosper have the “premium” service where their staff manages a customer’s portfolio of notes for a percentage fee.

There’s no way for me to prove it, but it seems fairly obvious that it isn’t a fair playing field out there… I highly doubt the institutional customers are using the same clunky, slow Folio platform I am forced to use, nor are the staff members who are out buying notes for their premium retail customers. I wouldn’t be at all surprised if they had a direct API to get better access to the platform. The legality of that seems murky at best (look at all the red tape when it comes to the fairness of equity market platforms), but it will take regulators and the SEC years to catch up with this industry. That’s what’s most troubling about it, and is its greatest risk to a fixed-income portfolio, if you ask me.

Reply

Peter Renton July 19, 2013 at 5:12 pm

Mike, It will no doubt surprise you to learn that nearly every institutional investor I have spoken with are not using the secondary market at all…to buy or sell. The increase in activity in the secondary is, in my opinion, due to the third party tools available now. Tools like InterestRadar.com make it very easy to buy notes on Folio with detailed criteria. I also know of some home grown tools where people have automated the note buying on Folio.

There is no API to access the Folio platform but several enterprising developers have created screen scraping technology to allow some level of automation.

Reply

Leave a Comment

Notify me of followup comments via e-mail. You can also subscribe without commenting.

{ 1 trackback }

Previous post:

Next post:

Real Time Analytics