Trends in Principal Charge Offs For Completed Loans at Lending Club

Many of us here are obsessed with defaults. We hate to see defaults and we are always looking for ways to minimize them. But not all defaults are created equal.

When you invest $25 into a new loan you have your entire principal amount at risk initially. But as soon as you start receiving payments you reduce your potential losses. I have some loans from August 2009 where I have received more in payments than my original principal, so even if the loan defaults now I cannot lose money.

But what about the entire Lending Club portfolio? How are defaults impacting principal losses over the duration of a loan? Well reader Bryce Mason wondered the same thing and he produced this attractive chart below.

[Update March 6: Bryce realized a small error in his calculations that has altered the chart slightly. The lines of the chart still end up in the same place but have a slightly different slope that is a little steeper at the beginning].

Bryce has analyzed all completed loans from the very beginning of Lending Club through the first quarter of 2009. He has graphed the amount of principal that was written off over the life of the loans, separating out the information by quarter.

Principal Write Offs Improving

It comes as no surprise that the first three quarters were the worst ones. Lending Club was just getting going and all these borrowers went through the thick of the financial crisis with these loans. And while the improvement hasn’t been consistent, you can see that Lending Club has certainly improved. The latest quarter, Q1 2009, which is not quite complete yet, is easily the best quarter thus far.

Here are some points to keep in mind when looking at this chart:

  1. These are all 36 month loans.
  2. Lending Club’s quiet period lasted from April until October 2008 and while they did issue loans during this time the volume was pretty low.
  3. In October 2008 Lending Club changed their underwriting rules significantly increasing the minimum FICO score from 640 to 660.

What was most interesting to me about this chart is that the lines were more linear than I expected.  I would have thought that principal write-off would be higher in the first year and then start to taper off after year 2. But as you can see in Q1-2009 the line has gone steadily up although the good news is that this is by far the best quarter.

I appreciate Bryce sending me this chart and allowing me to publish it here for all the readers to see. Let me know what you think in the comments.

 

Subscribe
Notify of
28 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Bryce M.
Bryce M.
Mar. 5, 2012 4:56 pm

First a note on method:

I want to disclose how I obtained the month of stopped payment, because it is imputed. I took the interest rate and the percentage of remaining principal and compared it to an amortization table I made in Excel that had all remaining principal values at the end of each month (1 – 36) for every interest rate possible (0.0500 to 23.0000). Thus I am assuming everyone follows the amortization schedule and not pay ahead. Finding the closest remaining principal in the Excel sheet given the interest rate, I pulled out the month of default. Of course, someone could have paid ahead and then defaulted, so it’s just an approximation, but I bet it is very close to reality.

I was also surprised about the continuing forfeiture past one year. We all hear the marketing departments saying loans are “seasoned” after a while, but I’d argue from the graphic that by the time there is very little probability of default, there is also very little at stake. Seasoning is useless for three-year loans, in my opinion.

Finally, I spent a good deal of time this past weekend trying to predict the loss percentage (among those loans that experienced a loss). I conclude that it is difficult to predict with just the information Lending Club provides. The best factors I could find were external (i.e., state-level economic variables).

Chris F
Chris F
Mar. 5, 2012 5:31 pm

@Bryce Very interesting. Thanks for the putting that info together. I think you mentioned earlier that you stick primarily to 3 year notes. I’m curious why that is. It seems that 60 month notes are too new to Lending Club (I believe they started offering May 2010) to have evidence pro or con, but curious if you had seen any trends that lead you to believe the rate of defaults will be higher on the 5 year notes?

I initially skewed toward 3 year notes, but found it hard to find enough to fill my portfolio.

Bryce M
Bryce M
Mar. 5, 2012 5:49 pm

@ Chris F. – I only invest in what I know, or at least that for which I feel I have an edge. As soon as we have 500-1000 5-year loans completed, I will begin modeling those. I have not developed forward-looking indicators that give any indication of current, or more-recent performance levels, but it’s on my mind. I suspect that it would require constant data pulls, unless LC was so kind as to put payment history in their loan history (so we can see if things ever went late, and then healed themselves). I’m not sure it’s worth it, though. Macro factors probably have more to do with it than any underlying borrower behavior changes.

Michael
Mar. 5, 2012 8:31 pm

Great vintages. I believe this is the first time I have seen these for Lending Club! Prosper has these for their platforms in one of the their power point presentations. The thing both companies share is the vintages are improving YoY for the most part.

Dan B
Dan B
Mar. 5, 2012 11:22 pm

Peter…………I mentioned this in another post a few weeks back but I’d like to know what percentage of Lending Club’s 5 year loans originated from Mar 2010 to Mar 2011 are currently on a “payment plan”. I assume you speak to them from time to time & might be able to find out?

Bryce M.
Bryce M.
Mar. 5, 2012 11:40 pm

Pretty sure that anyone can find that out, as “payment plan” is one of the statuses in the historical loan extract data set. I’ll look at it and let you know Dan. Even if they show up as “current” it would not be hard to identify them because they would be behind the amortization schedule.

Dan B
Dan B
Mar. 6, 2012 1:54 am

Peter…………Wow, 36? Ok, so under 1%?? Really? May I get a second opinion?…………………May I conduct an audit?

Let me explain. Only 30% of my personal account is invested in 5 yr. notes. But I’ve been selling “current” payment plan notes on an almost weekly basis for more than a year & a half now. And this is from a 650-750 note portfolio. So if the above number of 36 is correct than one of the following two scenarios should be true. Either almost all my sales haven’t been 5 year notes……………or I’ve owned an enormous percentage of those 36 five year payment plan notes in my comparatively small portfolio. I don’t know but both scenarios seem a bit far fetched.

Dan B
Dan B
Mar. 6, 2012 6:23 am

Peter………..I understand what you’re saying & have no doubt that you’re looking at the info accurately.

I just checked on Lendstats & it lists 138 5 yr loans as being on a payment plan. That’s actually more curious than my initial suspicion because it means that from Mar 2010 to Mar 2011 there are from what you’re stating only 36 of 4890 on a payment plan……………but from Apr. 2011 to Dec 2011 (9 months) there are another 102 on a payment plan from the most recent batch of +/-6600. Curious, since some of these loans are literally a few months old.
I know, you’re thinking that I forgot about the Jan 2012 through today notes, but apparently none of the 2012 notes are on a payment plan. Also curious. Why doesn’t this sound right ? Or is it just me?

Louis Lamoureux
Louis Lamoureux
Mar. 6, 2012 9:15 am

It’s possible that 5 year loans that were on payment plan, went “off” payment plan either in a good way by catching up or in a bad way by defaulting. So at this very point in time according to Lendstats, there are 138 in “payment plan”. But there could be a bunch more that were formerly in “payment plan”.
Dan, it’s probably too early for the 2012 notes to go into payment plan, For my notes, I noticed, they didn’t go into payment plan until they got into the 31 – 120 days late category and by anecdotal evidence I’m going to say around 60 days in.
Lou

Dan B
Dan B
Mar. 6, 2012 10:38 am

Lou……….Sure, I understand that I’m looking at a snapshot & some notes may have defaulted. But notes can in fact be “current” & on a payment plan simultaneously. Obviously I can’t comment on your specific note experiences, but I can assure you that notes CAN go into payment plan even without ever being late at all. I’ve seen seen notes go into payment plan without ever making any payments. So what you’re suggesting may intuitively make sense but just isn’t the way things work with Lending Club payment plan notes.

Mike
Mike
Mar. 6, 2012 9:27 pm

One has to wonder about a borrower who has to go on a payment plan before even one payment is made! I think I would sell that loan…maybe to Dan B.

Nice work here, Bryce, but I’m having a hard time with the graph, probably because I’m stupid. If a loan defaults at the 0 month mark, before one payment is made, isn’t 100% of your principal forfeited? The graph appears to me to look like the earlier the loan defaults, the less principal you lose. What am I missing, besides a brain for math?

Bryce M.
Bryce M.
Mar. 6, 2012 9:46 pm

– The denominator is the entire principal issued in that quarter, over all loans.

Dan B
Dan B
Mar. 6, 2012 11:32 pm

Mike……….I bet you’d love to do that to me, but wanting to do something & actually doing something are 2 different things. Suffice it to say that the “buying” & “selling” of notes is very much a zero sum game played out in shark infested waters. And in those waters I’m a great white who has done things like what you’re suggesting more times than you can imagine. So I welcome your efforts in trying to put one over me. Your chances of success/survival are about as small as if you were to physically jump into those waters while bleeding from an orifice 🙂

Mike
Mike
Mar. 7, 2012 5:58 am

So the take home message is that LC loans are not statistically more likely to default during any particular quarter, and that those default rates have been improving over time? That implies that there is no speedboating period the first few months after the loan is issued.

@ Dan, fortunately I’m a very good swimmer, and haven’t suffered any bleeding at all. In fact, I have been able to jettison many ‘current’ loans from my portfolio that mysteriously change status shortly after the trade settles. Maybe you have purchased some of these, and maybe you haven’t, but at least the information necessary to do this is available to those who look for it, unlike your ‘method’, and as long as I can sell these ‘current’ loans I’ll be a happy camper.

Dan B
Dan B
Mar. 7, 2012 9:52 pm

My previous comments on this thread were deleted by Peter because he feels he has the right to determine what is & what isn’t “civil discourse”. I disagree with censorship in general & disagree strongly with his belief that he has the right to simply delete what he sees fit WITHOUT even informing the rest of the readers & contributors here that a specific posters post has been deleted. Without even that minor courtesy blog readers & contributors should have serious doubts whether “free & unrestricted” discussions are occurring here.

How are we to know whether it is 1 or 10 or 50 posts that are being deleted here routinely without any of us even knowing about it. This is simply unacceptable & I encourage others to voice their displeasure to Peter about this “invisible censorship” that he has instituted. That is assuming that this post stays up here long enough for any of you to read it before being deleted.

Walter
Walter
Mar. 7, 2012 10:26 pm

Lending Club had a webcast on a similar issue back in December 2009. The conclusion of the presenter was that the default curve would drop back instead of flattening out (and NAR would rise back up). Bryan’s chart is showing % of principal charged off, so perhaps these are slightly different concepts. I had only been investing for a few months and I remember thinking that the presentation was informative. I’m not sure how you can access it now.

https://blog.lendingclub.com/2009/12/07/upcoming-webcast-the-borrower-repayment-curve/

Scott Langmack, noted author, analyst, and peer lending investor, will explain his view of the “Borrower Repayment Cycle”. Scott will discuss why in his experience a majority of defaults come in the first 18 months of a loan’s life.

Mr. Langmack believes this phenomenon can cause many investors to feel as if they are in worse shape than they are. “Most investors see a steady decline in Net Annualized Return from months 6 to 18, which is disappointing if you don’t understand that returns comes back up after that,” says Langmack.

In addition, Scott will explain how investors can smooth their returns and increase their stated Net Annualized Return by building a diversified portfolio over a period of months.

The Borrower Repayment Curve:
Why 6% Net Annualized Return today can be on track for 9.5% overall return

Dan B
Dan B
Mar. 7, 2012 11:30 pm

Peter…………That’s fine, but at least do the right thing & leave an indicator where a post by whatever author has been deleted…………rather than just vanishing it as if nothing was ever posted. Is that so much to ask?

As far as being the publisher & responsible & all, that’s wonderful. I hope that you keep taking that position in the years to come so that you will be providing your legal counsel ample employment. As a publisher who does not allow unrestricted discussion, you are therefore claiming responsibility for everything that makes it on to the site.

Dan B
Dan B
Mar. 7, 2012 11:41 pm

Walter……….I attended 2 of those Webcasts that Lending Club had in 2009. They were nothing but infomercials for the beginning investors. I submitted a number of questions in each of the 2 sessions that were ignored because the concerns I brought up were, I can only conclude, uncomfortable for them to address.

As for that investor Scott Langback, his conclusions were highly suspect. If in fact the returns would improve after the 18 month period as he suggested, then Lendstats wouldn’t be showing the real world average ROIs of those loans from 2009 -2010 to have remained stuck in the 4.8-6% range where they have essentially been in the last year. Those returns are passable, but they’re not going to be rising in this lifetime.

Lou lamoureux
Lou lamoureux
Mar. 8, 2012 7:44 pm

Peter, I think dan has a point, a short message saying ” this post was deleted for violating the site’s terms of use” wouldn’t hurt.
Lou

Dan B
Dan B
Mar. 9, 2012 2:02 am

Lou…….Thanks for supporting the principle of what I’m saying here. What annoys me the most about this matter is that imo Peter is lowering the threshold as to what constitutes “getting personal” to what I consider to be ridiculous levels. Of course everyone will have to either take my word on this or not, because the comment itself was deleted. IMO far harsher comments that were very personal in nature have been allowed on this blog in the recent past & I can only speak for myself by stating that I’ve not been in the least bit wounded by them. We’re all adults here & should be able to take & dish out the occasional jab. It is, after all, a part of everyday life.

Walter
Walter
Mar. 9, 2012 6:23 pm

@Peter/Dan
I did not mean to imply that I believe that the webcast info is still valid, only that at the time it came across as helpful. I can’t remember how/where he got his information to make his conclusions.