Thursday, June 28, 2012

Lending Club Loan Length - Credit Grades and Default

Request to my Readers: I am enjoying sharing and discussing with you my findings from analysis of Lending Club historical loan data file. I would like to find out what other topics about Lending Club may interest you. Please take a few seconds to answer the Poll posted in the right sidebar.

You may recall a few weeks ago, Michael at Nickel Steamroller wrote a very interesting blog post Balancing Your Portfolio by Loan Term - 36/60 Month using the spread he painstakingly recorded for currently available loans on Lending Club. Let's compare his findings with analysis of Loan Term in my previous post Lending Club Loan Length: Best 60 month spread with B4 Grade Loan. Let's see if I can answer some of his questions.
"The higher interest rate [for 60 month loan] is to reflect the additional time you have to put your money at risk."
This statement is true but Lending Club goes about it in a round-about fashion. Lending Club reduces the Credit Grade of 60 month loan instead of increasing the interest rate while maintaining the Credit Grade. I assume most investors, as I do, compare the loans of same Credit Grade when selecting loans for investment. So, they don't necessarily pick higher returns with 60 month loans but potentially lower default risk for such loans. If you want higher interest rate/return with 60 month loan, look for such loans with Credit Grade that are within 5 or 6 credit sub-grade from your target Credit Grade criteria for 36 month loan selection.
"If you see the spike for D1 there was actually no spread on the currently listings. There was only one D1 [60 month loan] on the system (Member Loan ID 1333671). Not sure why this was."
As mentioned in my previous post, Lending Club's underwriting process skips D1 credit grade for 60 month loans. The only way of having a D1 grade credit loan is to change the loan amount. This is also the reason why there was no spread listed on D1 loan as Lending Club can't give D1 grade to a loan just by changing the loan length. The chart below shows the loan volume in 2010 and 2011 for all Credit Grades and Loan Lengths. As the chart shows, there is a discontinuity in loan volume at Credit Grade D1 for 60 month loans. Also, there are very few 60 month loans with Credit Grade A.


Credit Grade and Default Rate

Before I review the defaults as a function of credit grade, readers need to keep in mind that the historical loan data file was captured at the end of April 2012. So only loans listed in the first four months of 2010 have aged two years.

The chart below shows the charged-off and default for 36 month and 60 month loans. The 60 month loans listed during the second quarter of 2010 are the oldest loans. As the chart shows for Q2 of 2010, except for Grade D loans, the default rate is higher for 60 month loans. With further aging, I expect that default rate for 60 month Grade E and F loans may be lower considering the high number of 36 month loans with late payments in these grades. The similar patterns are observed for 2010 and 2011.


I made following observations from the chart for loans listed in 2010. Please share any additional insights you may have through comments.
  • For A grade loans, the default rate for 60 month loans is higher than that for 36 month loans. For B and C grade loans, the default rate follows similar pattern. In aggregate, the default rate for A, B, and C grade loans is 6.16% for 60 month term versus 4.34% for 36 month term.
  • The pattern reverses for D, E, F, and G grade loans. The default rate for 60 month loan is lower than that for 36 month loans. In aggregate, the default rate for D, E, F, and G grade loans is 9.064% for 60 month term versus 10.669% for 36 month term.
  • This reversal of pattern is most probably due to a large difference in loan volume with 36 month and 60 month terms. For example, 19.98% of grade A, B and C loans were of 60 month term compared to 49.32% of grade D, E, F, and G loans of 60 month term. The chart below shows number of 36 and 60 month loans for last three quarters of 2010.

It appears that Lending Club's underwriting process does well in matching the default risk between 36 month loans and 60 month loans using the credit grade of the loans. It remains to be seen how much more the default rate for 60 month loans will rise because such loans listed in Q2 of 2010 have completed only 2/5th (40%) of maturity term compared to 2/3rd (66%) of maturity term for 36 month loans.

Key Takeaways

  • Lending Club's underwriting practice to downgrade the Credit Grade of 60 month loan to adjust for extra risk is at best confusing for lenders. A practice of maintaining Credit Grade but increasing the interest rate on 60 month loan would be much more clear as most lenders on its platform understand higher risk deserves higher return.
  • Lending Club may need to provide resources that explain and educate lenders on credit risk evaluation. Otherwise, similar to NSR, lenders may turn away from longer maturity loans.
  • Comparing default rate of 60 month loan with 36 month loan for the same Credit Grade is like comparing oranges to apples. Lenders may need to shift credit grade 5 or 6 steps for 60 month loan to get a fair comparison of default rate with 36 month loan.

Once again, Peter Renton wrote an interesting post Why Aren't Banks Offering P2P Lending and Other Questions in response to an email from his reader. I just wanted to add couple of points.
  • Web Bank is an Utah based industrial bank specializing in private label credit cards and financial services. Their business model differs from traditional banks that offer services to consumers. Instead Web bank offers services to businesses who want to roll out their own financial services products such as from Dell, Circuit City and other retailers. Checkout Web Bank's products and service page for more information.
  • It is very typical of established businesses to ignore the emerging trends in their marketplace until too late. The banks are no exception. A Harvard professor Clayton Christensen did extensive research on this topic and published a very popular book The Innovator's Dilemma: The Revolutionary Book That Will Change the Way You Do Business.

Interested in understanding data analytics field, consider reading How to Measure Anything: Finding the Value of Intangibles in Business and Super Crunchers: Why Thinking-By-Numbers is the New Way To Be Smart.

4 comments:

  1. Good stuff. It would be interesting to back out the shift (or use a 5 grade shift as test case) and re-run your analysis/charts, since that is the takeaway coming from your analysis.

    ReplyDelete
    Replies
    1. Conor

      Thanks for the comment. I agree I should have shifted the grade and re-run the chart. But I was being lazy knowing that pretty soon I may be redoing some of this analysis with new data file at the end of the quarter. So, I just ball-parked the observation by looking at chart.

      Thanks for keeping me honest.

      Anil

      Delete
  2. Rather than shifting the credit grade, perhaps it would make the most sense to take a leaf from the p2pxml book and compare based on interest rates directly.

    ReplyDelete
    Replies
    1. As my previous post indicates credit grade and interest rate are interchangeable when it comes to loan length. Both 36 month and 60 month loans in same credit grade have same interest rate. I did create charts for interest rates instead of credit grades but they came out exactly same and offered no new insight. You offered good thought as it may easier to shift interest rate than credit grade in Tableau. I will try it out. Thanks.

      Delete