Test case index explained

Following is my first test case report that will contain the performance measure “index”. The general concept is to compute what the lender expects during normal pay back. This varies depending on the status and how the note was acquired. But all need the monthly payment due to pay off the loan over its initial term. This is computed from the amount invested, interest rate, and term. I’ve noticed the monthly payments on some notes varying a bit (particularly the first payment), so we’re starting with an approximation.

There is also an issue with when the payments are made. Sometimes the date is adjusted. So, I’ve computed the number of days from the issue of the note and today’s date. I then attempt to compute the number of months payments that should have been made (which could be off by one). For notes originally purchased on the Lending Club platform the calculations are made from the issue date. For those purchased on the secondary market, FOLIOfn, the date of purchase is used. This results in the expected payments to date. The index is the ratio of this versus the amount paid so far.

While looking at some specific notes, several anomalies were discovered. Some people paid more than the amount due or paid ahead. This resulted in an index of over 100. I arbitrarily set a maximum index at 100. Some appeared to be significantly short, but it was a timing issue on when the payment was received. It became obvious that the approximation for the number of expected payments to date could cause a significant variance in the index, particularly in the first few months. So, I made an adjustment by looking at the number of payments expected and one less than this. I then used the higher of the two index values. The results of these adjustments will cause most notes to have an index of 100 early on in their life cycle.

For fully paid notes, the index would drop below 100 if they were paid off early. One example was a 3-year note paid off with about 9 months remaining. Its index dropped to 97.1 because of the loss of the interest that would have been received over the next 9 months.

Description	Current	Other	Fully Paid	Charged Off	Index
Test 1 End in 2017	4%	22%	74%	1%	96.2
Test 2 End in 12 months	0%	88%	13%	0%	98.1
Test 3 LC Picks	0%	95%	3%	3%	99.5
Test 4 Zip Grade	0%	70%	0%	30%	99.9
Test 5 By Percent	0%	93%	3%	5%	99.7
Test 6 By Funded	0%	80%	0%	20%	100.0
Test 7 Random	0%	73%	3%	25%	98.6

Clearly, it’s too early to draw conclusions. I’m also concerned that the index for test cases involving purchased notes is being misstated by using averages. I’ll be working on this problem next week. Follow me on twitter (@billlanke) and I’ll tweet when I post.