Impact of loan quality on results

We needed to further pursue the question raised in the last post about being unlucky in the selection of notes. We know that the rate of charged off loans is related to the grade (hence interest rate) assigned to the loan. The higher quality loans (A-Grade) have a lower default rate than the lower quality loans (G-Grade). The question that naturally occurs is to what extent does the mixture of loan quality impact my overall charge off rate.

The following chart begins to address this. It represents all the notes issued in 2015 divided into two groups. “My Notes” includes these notes that are in my account, and “All Notes” includes all the others.

	Count		Percent
Grade	All Notes	My Notes	All Notes	My Notes
A	73,196	140	17.5%	5.9%
B	117,349	257	28.0%	10.8%
C	120,088	479	28.7%	20.2%
D	62,042	612	14.8%	25.8%
E	34,506	442	8.2%	18.6%
F	9,471	346	2.3%	14.6%
G	2,068	99	0.5%	4.2%
	418,720	2,375

Clearly the distribution of my notes is skewed toward the riskier notes and that alone could account for the higher incidence of charged off that I experienced. In an effort to further analyze this impact, I looked at the 4 months with the highest number of notes at a specific grade.

Month	Grade	Count		Charged Off		Percent
15_07	D	6,748	254	1,118	47	16.6%	18.5%
15_07	C	13,104	194	1,495	17	11.4%	8.8%
15_07	E	3,899	173	791	30	20.3%	17.3%
15_08	C	10,019	124	1,000	12	10.0%	9.7%
Total		33,770	745	4,404	106	13.0%	14.2%

This shows that in July, 2015, I acquired 254 new Grade D notes that had a charge off rate off 18.5% so far. This compares to a charge off rate of 16.6% for all Grade D notes issued that month. Too small of a difference to be significant. When the 4 months are combined, the difference in charged off rate is still pretty comparable.

This wasn’t conclusive enough, so I generated another chart. This one shows the charged off rate by grade for all notes compared to mine. In some grades, I had a better experience and in other, worse. Unfortunately for me the poorer performances were in the risker grades where I had greater exposure.

	Count		Charged Off		Percent
Grade	All Notes	My Notes	All Notes	My Notes	All Notes	My Notes
A	73,196	140	2,028	2	0.5%	0.1%
B	117,349	257	7,234	18	1.7%	0.8%
C	120,088	479	13,246	58	3.2%	2.4%
D	62,042	612	10,392	111	2.5%	4.7%
E	34,506	442	7,370	90	1.8%	3.8%
F	9,471	346	2,769	116	0.7%	4.9%
G	2,068	99	713	38	0.2%	1.6%
	418,720	2,375	43,752	433	10.4%	18.2%

All of this charge off analysis led me to the conclusion that using automated investing did not work well for me, and that it was wise to stop it last December.

Thus far I have focused my attention on charged off rates. Absolutely, charge offs are a significant factor in overall performance. But there are several problems inherent in its use. I’ll deal with this and the development of a better measure in my next post. Follow me on Twitter, @billlanke, for notice as to when that occurs.