Harmoney

[Cool Bear]

Have updated the summary.pdf document with quite a few characteristic by grade chart sets where they make sense. Some of these appear to offer some hints of how the Harmoney grading process may be working.

At this stage I'll take a break from making new charts unless someone asks for something specific or I think of something useful - if I've missed anything, let me know.

I'll put up the final summary document and unique.csv and raw.csv sometime late tonight. If you find any errors or have problems with these let me know.

Time to digest some of this info

Myles, I have a bit of time on my hands today and decided to study your summary report in detail as well as play with the unique.csv file (which more than double my own data set). Your report is really helpful.

One item that puzzle me is your income ratio chart on page 13. The total number of charge-offs for the unique.csv is 1028 out of 22252 loans. The total charge-offs shown in your chart is only about 650+. Wonder if there is some error somewhere. Can you have a look please?

Much appreciated, CB

[myles]

One item that puzzle me is your income ratio chart on page 13. The total number of charge-offs for the unique.csv is 1028 out of 22252 loans. The total charge-offs shown in your chart is only about 650+. Wonder if there is some error somewhere. Can you have a look please?

Good pickup CB - typo of 'Dept Sold' instead of 'Debt Sold' - now corrected so just download the pdf again.

Makes a significant difference to that graph, and a quite interesting one too!

[Cool Bear]

Good pickup CB - typo of 'Dept Sold' instead of 'Debt Sold' - now corrected so just download the pdf again.

Makes a significant difference to that graph, and a quite interesting one too!

Thanks Myles.

Interesting indeed. We would have thought that higher % would leads to higher defaults. Maybe it is due to the F grades where the loan limits (and thus monthly repayment amounts) are low. When you can spare the time, can you do the same chart again with just A,B,C,D and E, leaving out the F loans. Thanks again

[myles]

Colour change just so their is no confusion. Looks very much like the original which didn't have the 'Debt Sold' loans (probably a lot of F's in those 'Debt Sold'?):

Attachment 10124
In general the same shape though. Can't think of any obvious reasons why it is like it is?

Perhaps as you suggest, higher grade loans are only given out to those with lower income ratios, hence some influence even down a few more grades?

[myles]

Ran it for just ABC grades and still a similar shape, though low number of defaults so a bit 'jagged'? Beets me why?

I'd bumped my Income Ratio test up to 20% (from 15%) when loans dropped off a while back - never changed it back, but I've bumped it to 21% now

[Cool Bear]

Colour change just so their is no confusion. Looks very much like the original which didn't have the 'Debt Sold' loans (probably a lot of F's in those 'Debt Sold'?):

Attachment 10124
In general the same shape though. Can't think of any obvious reasons why it is like it is?

Perhaps as you suggest, higher grade loans are only given out to those with lower income ratios, hence some influence even down a few more grades?

Thanks again Myles

[Cool Bear]

Ran it for just ABC grades and still a similar shape, though low number of defaults so a bit 'jagged'? Beets me why?

I'd bumped my Income Ratio test up to 20% (from 15%) when loans dropped off a while back - never changed it back, but I've bumped it to 21% now

I think if we do not invest in F, we might as well forget about this repayment to income ratio. Like in your case, why stop at 21%? The total default rate (from your latest chart) for income ratio of the over 22% as a group is only about 3.4% which is about the same or less than your 0 to 21%

[Cool Bear]

Colour change just so their is no confusion. Looks very much like the original which didn't have the 'Debt Sold' loans (probably a lot of F's in those 'Debt Sold'?):

Attachment 10124
In general the same shape though. Can't think of any obvious reasons why it is like it is?

Perhaps as you suggest, higher grade loans are only given out to those with lower income ratios, hence some influence even down a few more grades?

Would be it too much to ask if you could run one for each grade, just to see the difference?

[myles]

My thoughts are that anything over 0.25 are a bit 'random' and numbers are perhaps a little low to be significant.

Leaving out 0.22-0.24 should reduce defaults if only by a little. Plus I think it just makes sense to limit to around that 20% range.

Otherwise agree - doesn't appear to be an overly useful metric for loan selection.

[myles]

Attachment 10125
Attachment 10126
Attachment 10127
Attachment 10128
Attachment 10129