Default Rates broken down by Credit Grade and DTI Buckets

Published by knix on November 21, 2007 02:24 am under prosper

This post is in response to a post I saw in the Prosper Lender forums requesting for stats on defaults vs. DTI. To obtain this data, I have queried all the loans that have been paid in full, defaulted, or repurchased from 2/1/2006 – 7/19/2007. The query ended on 7/19 because I wanted to give sufficient time for loans to default. Buckets have been broken down based on credit grade and the DTI range. The chart was cut off at 59.99% DTI because there was insufficient amount of data points but the raw data is in the tables below if you are interested in taking a look at them. Please bear in mind that these numbers do not include loans that are still in the repayment phase. That is why the numbers seem exaggerated. You can hover over each point to see the percent default rate.

Here is the raw data for the counts of each type of loan broken down by credit grade and DTI buckets:

AA Credit Grade	0-9.99	10-19.99	20-29.99	30-39.99	40-49.99	50-59.99	60-69.99	90-99.99	>100
Paid in Full	428	235	86	26	17	4	6	2	12
Defaulted	2	2	2	0	0	0	0	0	0
Repurchased	14	0	6	0	2	0	0	0	1
Total	444	237	94	26	19	4	6	2	13

A Credit Grade	0-9.99	10-19.99	20-29.99	30-39.99	40-49.99	50-59.99	60-69.99	70-79.99	80-89.99	>100
Paid in Full	150	175	74	41	16	10	2	2	2	13
Defaulted	4	8	4	8	4	0	0	0	0	0
Repurchased	6	4	6	0	2	0	0	0	0	0
Total	160	187	84	49	22	10	2	2	2	13

B Credit Grade	0-9.99	10-19.99	20-29.99	30-39.99	40-49.99	50-59.99	60-69.99	70-79.99	80-89.99	>100
Paid in Full	104	150	88	48	12	10	2	2	2	22
Defaulted	2	30	6	12	2	2	0	2	0	2
Repurchased	0	8	4	0	0	2	0	0	0	0
Total	106	188	98	60	14	14	2	4	2	24

C Credit Grade	0-9.99	10-19.99	20-29.99	30-39.99	40-49.99	50-59.99	60-69.99	70-79.99	80-89.99	90-99.99	>100
Paid in Full	114	164	120	48	14	6	6	2	2	2	14
Defaulted	10	36	24	6	6	6	2	0	0	2	0
Repurchased	0	0	2	0	0	0	0	0	0	0	0
Total	124	200	146	54	20	12	8	2	2	4	14

D Credit Grade	0-9.99	10-19.99	20-29.99	30-39.99	40-49.99	50-59.99	60-69.99	70-79.99	80-89.99	90-99.99	>100
Paid in Full	120	162	94	48	16	26	2	0	6	2	20
Defaulted	24	62	34	20	10	4	0	2	0	0	4
Repurchased	2	2	2	0	0	0	0	0	0	0	0
Total	146	226	130	68	26	30	2	2	6	2	24

E Credit Grade	0-9.99	10-19.99	20-29.99	30-39.99	40-49.99	50-59.99	60-69.99	70-79.99	80-89.99	90-99.99	>100
Paid in Full	71	134	80	32	14	6	2	2	4	2	4
Defaulted	74	162	50	20	12	4	4	0	0	0	6
Repurchased	2	8	2	0	0	0	0	0	0	0	2
Total	147	304	132	52	26	10	6	2	4	2	12

HR Credit Grade	0-9.99	10-19.99	20-29.99	30-39.99	40-49.99	50-59.99	70-79.99	80-89.99	>100
Paid in Full	110	62	38	12	8	0	0	0	2
Defaulted	310	180	74	32	8	8	2	2	10
Repurchased	18	8	2	2	0	0	0	0	0
Total	438	250	114	46	16	8	2	2	12

Although there cannot be any real concrete conclusions drawn from this analysis, due to the low number of loans thus far, I think it is safe to say that there is a general increase in default rates as one’s DTI increases. This increase in default rate is even more visible as credit grades get lower and lower. It is also interesting to note that rarely does the default rates of one credit grade overlap with that of another (based on the graph above with the maximum DTI being 59.55%). This means that more attention or weight should be given to credit instead of the borrower’s DTI.

You may also be interested in reading the following posts:

Glossary for status of loans:

Defaulted – The loan defaulted
Paid In Full – The load was paid in full
Repurchased – The loan was repurchased by Prosper

If you have any comments or suggestions to add to this list, please feel free to comment below.

3 Comments so far

Tom on November 22nd, 2007

Nice analysis. Thanks for doing this.
pjz on November 25th, 2007

It is hard to come to any conclusion on defaults vs DTI on Prosper because Prosper skewed the data by strongly recommending low DTI listings. Thus a larger percentage of low quality loans were funded at low DTI contributing to higher defaults for low DTI. A better analysis would be DTI versus extended credit date in each credit grade. That analysis will have to wait till more Prosper data is available.
knix on November 25th, 2007

Thanks for the comment pjz. You have a good point there on the skewed data. I definitely plan on doing some more analyses on defaulted loans as more data becomes available.