Harmoney

[whitt]

and I've got one with a ratio just above 4% - this was on 10/04/2017 which had a large number of loans going through that day.

4% that very good and confirms my thoughts something has changed recently. Now these stats are more in line with what I had hoped with Autolend, lets hope it continues.

[RMJH]

4% that very good and confirms my thoughts something has changed recently. Now these stats are more in line with what I had hoped with Autolend, lets hope it continues.

5% here but much better and trending down. Will transfer more cash in if this continues. I haven't had to do that in six months.

[myles]

One Month in:

Invested: $21,150
Loans: 313
Auto-Lend: 76
Avg. Interest: 22.13% (Weighted)
Avg. Exp. Return: 14.80% (Weighted less Default+Fees+Tax)

Expected Monthly Return to deal with: $561.46 (Principal + Interest : excludes payoffs)
XIRR: 32.06% (no doubt meaningless at this point in time)

Grade and Term graphs:

Grade+Term.png

I've significantly changed my Auto-Lend criteria range now based on the below Interest less Default graph:

Interest-Default.png

Going to try to significantly increase the D's and E's, but there are not that many to pick from, so this will take some time.

I'm hoping to double up over the next month...

[alistar_mid]

I've significantly changed my Auto-Lend criteria range now based on the below Interest less Default graph:

Interest-Default.png

Going to try to significantly increase the D's and E's, but there are not that many to pick from, so this will take some time.

I'm hoping to double up over the next month...

Its not as simple as that, the chance to default is per year, and each individual loan has an average length of time before it defaults, where you would receive none or some of the interest payments until that point.

Also when you model what would happen if default rates spiked, the value point shifts from the DEF end to the ABC end.

I have done extensive modelling on this.. I did find that the sweet point is in the E for the current market conditions, should the economy tank and default rates spike, then you would want more ABC's.

I do think most A's suck the interest rate is too low and they are often paid back early. Likewise a lot of the F's - 12% expected annual default rate? no thanks Jeff.

Fwiw this is my distribution, I have 800+ loans, average loan value probably $90 (ie 4 notes most of the time, sometimes 2 or 3)

hamroney dist april 17.JPG

[myles]

Take some of your points, but I think the graph is a good indicator based on what Harmoney expect/model. (Interest rate is yearly as is default rate, per loan or across all loans). I'd like to know what the average actual loan length is for 36 and 60mth loans.

If there were a sudden increase in job losses then this would effect default rates across all risk grades, potentially high income earners could be effected more (e.g. IT workers in the past).

There is a bit of data out there of what happened with Lending Club in and around 2008, p2p in general fared very well.

My Risk Grades graph is looking more like yours already, and suspect that's about where it will end up. B5 are just to good to pass up and I really need to be across a fair set of ranges to keep up with the turnover.

What are your thoughts on what would happen to p2p lending if the Housing market took a big hit (lots of noise in Australia at the moment that his might be close, maybe not a tank, but a significant slow drop)? Could it happen here, it would hurt an awful lot of NZ'ers?

[myles]

Two links that might be useful to consider for grade selection:

http://www.lendingmemo.com/p2p-lendi...n-performance/

Some of the comments don't actually match the grade block comparison numbers (D's are on top, not A's in 2009). The summary at the end holds a very important point.

http://www.lendingmemo.com/risk-grad...-club-prosper/

Risk tolerance isn't the same for everyone...

[Probably mentioned in there somewhere but if it's not: Small investments are typically much more risky/volatile than larger, diverse investments.]

[alistar_mid]

Take some of your points, but I think the graph is a good indicator based on what Harmoney expect/model. (Interest rate is yearly as is default rate, per loan or across all loans). I'd like to know what the average actual loan length is for 36 and 60mth loans.

If there were a sudden increase in job losses then this would effect default rates across all risk grades, potentially high income earners could be effected more (e.g. IT workers in the past).

There is a bit of data out there of what happened with Lending Club in and around 2008, p2p in general fared very well.

My Risk Grades graph is looking more like yours already, and suspect that's about where it will end up. B5 are just to good to pass up and I really need to be across a fair set of ranges to keep up with the turnover.

What are your thoughts on what would happen to p2p lending if the Housing market took a big hit (lots of noise in Australia at the moment that his might be close, maybe not a tank, but a significant slow drop)? Could it happen here, it would hurt an awful lot of NZ'ers?

yeah even though i looked into quite a bit I came to the same conclusion you did - that D/E's are the sweet spot

That lending club article is one that i reference a lot when I explain to people it (harmoney) should be reasonably safe in a recession - it shows whats kinda obvious the higher risk loans are fine when the economy is good and defaults are low, but if conditions where to change and defaults increase then in theory those will go first. Thats why I haven't gone all out on C - F's. I am cutting back on B's, but like you said B5's do represent a lot of value.

Hmm should the housing market crash?

I guess then that would coincide with interest rates going up, so people may not be able to afford their mortgages, and would prioritize a mortgage over a p2p debt, so I guess we would see defaults increase.
For those without ties to the property market who have p2p lending, I guess they would suffer too, housing crash probably = less economic confidence, higher unemployment, some p2p loan holders default

Overall housing market crash = not good, would probably have a negative effect on p2p lending

[myles]

Overall housing market crash = not good, would probably have a negative effect on p2p lending

Yeah, not good

One thing I see a lot of people say that they include in their auto-lend (and manual selection I suspect) is 'Owned/Paying Mortgage' - thinking if the borrower can pay a Mortgage they can pay an additional loan. I started out that way, but have now come to the conclusion that a good spread into Rentals is a good choice, exactly because of this reason - if the Housing market drops, my thinking is the 'Owned/Paying Mortgage' group would likely be the higher defaulters, whilst rent would likely drop making the 'Renting' group less likely to default and hence a good diversification choice in this situation?

At the end of the day, I guess the bigger spread you have across most/all types of criteria the more stable your return will be, not necessarily the best return though, if you have the time to ride out any major dips, leaning towards higher risk will give much higher returns.

[permutation]

....
this is my distribution, I have 800+ loans, average loan value probably $90 (ie 4 notes most of the time, sometimes 2 or 3)

hamroney dist april 17.JPG

Analyzing your graph, in my experience of 1000+ loans, Your B,C,D proportions are great very similar to mine.

But I am getting rid of all A,E,F grades over the next few months and place all the repayments into B,C,D so my spread will look B35, C40, D25. B+C=75% B+D=60% C+D=65%. Expect a RAR of about 17%.

My current default number from 1000+ all time loans is (15) A-D 2/15; E+F 13/15 ouch!!

[myles]

My current default number from 1000+ all time loans is (15) A-D 2/15; E+F 13/15 ouch!!

Is that really an ouch!? If your loans are equal amounts, that default rate only represents 1.5%, E's and F's are worth much more than that in the additional Interest, yes? I guess it depends how many E+F that 13 comes from?