Post a Comment On: Transplant Athlete

Update: Now that Lending Club has raised their loan limit, these numbers have all changed. So, please don't rely on these results. You can follow the methodology to perform your own analysis on the new data.
Once again, past performance is not indicative of future performance, invest at your own risk, etc. etc.
I was looking at the data and wondering how best to analyze 14,000 records. I decided to look at charge-offs and defaults (since that's what we're trying to avoid). Most people assume that smaller loans are more likely to be paid back. Slicing and dicing the data into ranges, (for example: 0 to 1000, 1001 to 2000, 2001 to 3000, etc.) and counting the number of loans that defaulted or were charged off in each of those ranges, we get a graph that looks like this:
[Image]And the data (via OPENOFFICE Frequency Function):

Countrange 151000 302000 523000 404000 735000 536000 397000 768000 249000 7010000 2511000 4112000 813000 1214000 5315000 1916000 317000 818000 319000 3220000 821000 222000 323000 1224000 4525000 These results are pretty intriguing. If our hypothesis is that smaller loans tend to be repaid, this chart seems to indicate the opposite. Before you go investing in loans of $21,001 to $23,000 though, we need to check the fully paid and current loans to see what that chart looks like.
Slicing and Dicing, we get this chart:
[Image]And the data:

Countrange 1511000 5572000 8193000 8454000 12635000 10176000 7277000 10148000 5169000 131810000 28111000 73512000 31913000 28114000 77415000 33716000 13117000 22618000 10419000 50720000 7221000 6722000 3423000 12424000 46225000 The distributions are similar which probably says more about the loans on Lending Club than the performance of those loans. I think there's some useful information here, but let's keep digging to see what we find.