Gotcha, and on the other side - the actual default would be the total # of defaults / cumulative # of days of all the loans.
How do you add up paid off / charge offs? Is it taken to current date, or date of repayment/default?
Printable View
Have a read of the Harmoney and Lending Club definitions - I linked to them above.
No cumulative # of days, just number of days / 365.25.
Defaults are all 'Charge Off' + 'Debt Sold' vs everything except 'Cancelled' loans.
The final date was the query I was asking about before - Harmoney don't appear to do the same as Lending Club. Lending Club takes it back an arbitrary 120 days, Harmoney take from first to last (that's my interpretation of what their wording). So first loan start date to last loan payment date is what I'm using (I actually had max(start date) - min(start date), but have 'fixed' that to max(last payment date) - min(start date) and it has shifted the scatter pretty much under the line in lower risk grades - high risk grades are off the line (i.e. lower default rate on high risk grades then estimated).
It seems a pretty rudimentary way of calculating it, but it appears to be the way it is done...
Putting this up a bit early. It includes the two updated csv files that were put up today. The data set summary on page 4 should be correct for the unique.csv file which is what all of the charts are based on.
I've dropped off the last chart as it seemed to be causing some confusion, will revisit it at a later time.
Please note the warning in the summary about comparing values to annualised values.
I can claim the largest loan :) There is a story behind it - it wasn't meant to be - moral of the story - don't get distracted when purchasing loans. I've given up sweating over it, if it defaults it will hurt a bit, but not too much now :)
summary.pdf
unique.csv
raw.csv
I've deliberately not compressed the csv files (they aren't overly big), just to avoid problems. These are big enough that they may cause some spreadsheet software to 'chug' so be prepared for that.
If you find any errors in the data or corrections to the numbers in the summary, please let me know and I'll correct them when I can.
If there are charts that you think would be useful to everyone and would like me to try to generate, I'm happy to do that within reason.
If you find the secret to selecting the perfect loans, please share.
If you see a loan for a Caravan, you might be onto a good thing ;)
Enjoy!
As its currently proposed harmoney could end end up caught up in this too
https://www.interest.co.nz/personal-...erson-test-and
I cannot find a definition of who they define as "high-cost lenders"
However if we just look at interest and fees alone in the combined data set there 11 Unique loans where interest paid to date is greater then amount invested - Key point to rember is is only what has been paid todate - the real number will be alot higher once loans reach full term. Also as proposed the 100% cap includes all fees as well as default charges
"Interest and fees on high-cost loans will be limited to 100% of the amount borrowed (the loan principal). Thus if an individual borrows $500, they will never have to pay the lender back more than $1000, including all fees and interest, the Government says. This will only apply to "high-cost lenders" with the aim being to prevent unmanageable debt and financial hardship from accumulating large debts from a small loan.The idea is that even if the borrower defaults, they would repay no more than twice the original loan principal, including interest, default interest, and all fees. "
heard on the radio (but havnt seen in writing) that high cost was defined as interest rates greater than 50% per annum.
Highest at harmoney was 39%, dont know if that is the case now..
Fantastic effort Myles. Lots of top notch analysis which will help shed valuable insights, to help optimise our loan selections, even should the economy does head towards more turbulent times.
Am comparatively an excel hack compared to your DB charting skills, but heres a lone graph that some may find useful.
It tracks lifecycle stats by month of initiation. Stats are all as a % of initial capital invested. eg int% = total interest earned to date for mmmyy loan/ total investment made in mmmyy
Also the "arrears" figure I've taken is my own calculation of "principal at risk" (being the entire principal outstanding of loans more then half a payment in arrears). In addition I don't rely on HM judgement on arrears. Happy to share how I calculate if requested.
Second graph shows remaining principal in $ value. Use this as a measure of future potential (eg potential for further interest / further defaults). Months with less then 10% principal outstanding should have minimal change to Int/Charge off.
Thirdly is a measure of clear interest %. Being Interest earned less principal in arrears less defaults (no HM fee). Note it is not annualised, but over the period of the investment (which given the high early repayment would probably work out as annual anyway!).
Attachment 10071
Once again, many thanks Myles for your time and effort and producing a fantastic top-notch and very professional report.
E4 with partial pp, just up.
Previously would have mulled this over, and on a day with few loans would have taken it. Thanks to Myles's data will leave this one :)
Attachment 10072