I was tooling around on Eric’s Credit Community last week when I came across the interesting chart above. It is an analysis of all the late payments on mature loans on Prosper. Basically, it analyzes all the loans that have reached maturity (those loans that were originated prior to Dec 17, 2007) and looks at the spread of those loans that actually went late.
This is very useful information to have for p2p investors so you can gain some understanding of the likelihood of a loan going late once it has reached a certain age. By analyzing close to 7,000 loans that went late we can get some idea as to the pattern (these were all three year loans). The raw data for this chart is here. Based on this data, we can see that by 18 months 79% of the loans that were going to go late were already late. At 12 months that number is 63%.
The Unknown Factors
There are several unknowns when looking at this data that need to be pointed out:
- The biggest unknown is that most of these borrowers endured the worst economic crisis in 75 years with unemployment roughly doubling in a little over a year. This would obviously have had an impact on late payments and might have skewed this data somewhat.
- There is no breakdown on the different grade of loans. It is quite possible that a portfolio filled mainly with higher grade loans will have a different pattern to one focused on lower grade loans. Eric’s does provide some analysis on late loans of various grades here but there is less detail than in the chart.
- With new underwriting standards now at Prosper, in place since July 2009, there have been far fewer defaults and fewer late payments. What impact these new underwriting standards have had on the monthly breakdown of late loans is yet to be determined.
- We don’t know whether a loan that was late actually defaulted or became current again and was fully paid off.
The Key Takeaway
The key point that I see from this graph is that it is likely in your own portfolio of three year loans that you will receive more late payers in the first 12 months than you will in the middle or last 12 months. Obviously the number will vary depending on the individual portfolio but we can surmise this. You will likely suffer more defaults early rather than later in the lifecycle of a loan.
This doesn’t mean you are out of the woods by the 12 month mark. But if you are getting discouraged with a few defaults and a subsequent drop in your Net Annualized Return, you can find some reassurance with this data. One final point to keep in mind: there is a time lag between a loan going late and it actually defaulting. Both Prosper and Lending Club give borrowers 120 days to get their loan current, so no defaults will hit your account until at least month five on any of your loans.
If you believe this to be an average sample a strategy worth trying might be to buy loans that are 18+ month old at par on the secondary market.
But on the other hand, the loans traded at the secondary might be the bad apples if the seller figured out to identify statistical warning signs.
Good point. I have not spent much time in the secondary market but that could be a strategy worth trying. Before I did though, I would do a great deal of exploration and getting my feet wet before committing to a strategy like this.
Wow, the figures are almost obscene. So Prosper had 7000 loans go late ? Out of how many loans that were issued? Even if only 70% of those late loans turned into defaults, that’s 4900 defaults. And I don’t it’s a stretch to say that 70% is graciously understating reality.
I haven’t seen a chart for Lending Club but I’d be willing to bet that the first 6 months or so would have much fewer “lates”……….in addition to there being much much fewer “lates” period.
Dan, it is even worse than you thought. The early days of Prosper were fraught with all kinds of problems. If you go to Prosper’s performance page for the time period in question here you will find out that of the 16,989 loans issued from 11/1/05 – 12/17/07 a total of 6,703 were charged off (39.45%). So we can assume that most of the late payers turned into defaults.
Eric’s Credit Community doesn’t have the exact same stats for Lending Club but we know that while defaults were high in the beginning they never reached the levels at Prosper. I was not part of it back then, but looking at the data it is amazing that Prosper survived after such an unsuccessful beginning. I give them credit, though, since relaunch defaults and late payers are way down.
The chart surprises me. I would’ve thought it would have been more of a bell shaped curve, with most defaults occurring toward the middle of the loan term. Your data shows that there is no speedboating, or honeymoon period, at least on Prosper. In fact, it’s just the opposite. The highest risk of default is within the first year or so. That tells me many borrowers knew they were going to default on their loans even before they got them. Most of them do this by declaring bankruptcy after one or a few payments. I’d like to see a similar chart based on LC’s data. Thanks for the great work!
Mike, I hope to share a Lending Club chart at some point, but my guess is that it is similar. It seems that there is a percentage of people who are trying to game the system by taking out a loan with no intention of ever paying. If everyone intended to pay, it would be more of a bell shaped curve but unfortunately that is not reality. I will be doing a post next week on how to try and filter out the scammers before you invest in any loans.
Sounds like that post will be helpful for a lot of us ‘investors’. Looking forward to it, and thanks in advance.