Jun

15

As I've mentioned before, I am not a very good peer-to-peer lender.

Lending Club says that my account (adjusted for loans that are in default, late, etc.) is now worth 3.9% more than it was when I opened it in early 2016 and that my annualized return has been 2.87%, much lower than what they claim to be typical, which is >5%.

Stats on my loans are given below. I give the number of loans for each rating category from A to E, along with the number of "bad" loans in each category.

I define "bad" to be either "in grace period", "late 16-30 days", "late 31-120 days", or "charged off".

Occasionally you'll hear someone claim that you should just lend to the highest interest rate borrowers because the bad loans are relatively independent of rating. That was totally false in my case. I had literally zero bad loans in the "A" category, and only a reasonable 5% of my B loans were "bad". Meanwhile about 13% of my C and loans went bad, and 27% of my E and F loans went bad. "E and F" borrowers probably overlap a lot with "Ebay merchants" in my opinion!!!

Of course, when a loan goes bad, it's typically not a 100% loss, but believe me, it's pretty bad.

In retrospect I would have done much better by sticking to A and B loans, or even just A loans.

.

.

.

.

.

.

Stefanie Harvey writes: 

I have been playing on Lending Club for 5 years. I now choose all the loans I fund rather than "index"

- I only choose "A" loans
- Borrowers must have 2 years of employment and a credit score higher than 730
- I exclude any loan for "medical" or "relocation"
- My return is around 5%

Great discussion; Jeff thanks for posting that article (and agree - poorly edited book excerpts!)

I have done over 1000 loans and the A loans default at just under 2%. I lost more the first year I invested when I tried mixing risk. 

anonymous writes: 

Very interesting, thank you.

Maybe the whole idea of the D, E, and F loans with interest rates >20% is just a broken model. It's hard to see how those can have a good outcome for several reasons:

–If you're really strapped for cash, are you even going to be *able* to pay off the 20% loan, when the interest itself is so staggering?

–If you're really strapped for cash, are you even going to be *willing* to pay off the 20% loan, given that there's no collateral? Surely you'll pay on anything else first–mortgage, student loan, whatever. And your credit rating can't go much lower than it already is.

Seems like maybe the whole model — charge a higher rate to compensate for the higher risk of default–might be broken because of a Heisenberg effect–the rate itself affects the default rate.

Regarding loan selection, I read elsewhere some stats purporting to show that, counter-intuitively, one of the best credit risks is people who borrow to pay for a wedding.

Conversely, borrowing to "start a business" turned out to be a bad credit risk. The easy explanation was that such people would soon be quitting their steady-paying jobs.

anonymous writes: 

It brings to mind the payday loan business and makes one think that productive research could be done in that area, with the usual caveats.


Comments

Name

Email

Website

Speak your mind

2 Comments so far

  1. Ed on June 17, 2017 9:42 am

    Your experience matches Eric Frankenstein’s theory

  2. Ed on June 17, 2017 9:49 am

    If the platform does not allow pyramiding of interest payments (fees) they have removed what makes payday lending so profitable

Archives

Resources & Links

Search