Soliciting feedback for my strategy

[SLCPaladin]

Let me first start off by saying that I am not a data-head. I am an RN at a hospital, who happens to love investing. I'm not overly technical, but I grasp most, if not all, of everything on the forum. And I have had success with marketplace lending in general. I've invested with LC and Prosper for about 5 years now and my returns have slowly declined from about 12% to 7.5% now. I am not an accredited investor, though I am close. I'm perplexed with how, and if, to continue investing.

My skill set is probably different than some of those on this forum who run advanced analytics, portfolio analysis, and who automate tools and build scripts for back-testing and modeling. I do none of this. I probably could if I invested the time into learning, but my extra time is usually invested into my career working long hours at the hospital, or playing with my young son. Here's how I basically operate: I use LC's filters to give a first-pass and then I spend about 5-10 minutes every day or so and select notes to purchase based on some pre-screening from filters.

With that background, I'm soliciting insight to where I have room for improvement, or any blind spots. Here are some of my filter criteria:

1) Exclude sand states (Arizona, California, Florida, Nevada)
2) Max DTI of 20%
3) Min length of employment of 5 yrs
4) Homeowner
5) Exclude the following loan purposes: medical, business, invest in learning/training, wedding, vacation, moving
6) Monthly income >$4000
7) Inquiries < 2

For the last 6-9 months, I have shifted to the front of the spectrum, only investing in A,B, and C grade notes. I was very heavy in the lower end as I chased yield. My returns have suffered for this as LC delinquencies and charge-offs have been ugly as of late for this part.

At this point, I'm just looking for input from what I may be missing in terms of strategy or new ideas I haven't considered. Most of what I filter on is stuff I picked up from this forum, from Peter, and others. I wonder though if this is still working. Any thoughtful advice is appreciated.

[SLCPaladin]

As an addendum to my original post, I have some lingering questions about things I occasionally do when picking notes. I wonder if there is anything to this, or if I am just spinning my wheels, or worse, am I being counter-productive? I would describe myself as an advanced beginner when it comes to data and back testing.

1) I occasionally will just "go by my gut" when selecting. For example, if I have $225 dollars to allocate for the day, and my filter returns 30 notes, I sometimes just sort by the highest% funding and use that as a starting point. I feel like there is something to the "wisdom of the crowd" approach (James Suroweicki writes about this), so my belief is that if a loan is more rapidly funded, it might have done so for a reason, even if it's not obvious to me.

2) All things being equal, I tend to want to fund loans with a lower dollar amount than a higher dollar amount. I like loans of about $4k to about $15k, but I really like loans under $10k the most. My rationale is that these lower dollar amount loans are easier to service.

3) I tend to dislike loan amounts that go for the max allowed. For instance, I tend to stay away from loans that are issued with $35k or $40k. I am leery of people who take the max amount that is offered to them. Maybe I am missing out by doing this, but it just feels irresponsible to take the max allowed instead of what you need. Instead, I like loan amounts that are close to the revolving credit balance as it suggests to me a consolidation of debt rather than increasing debt load.

4) I tend to give preference to certain professions over others. I like government jobs, teachers, nurses, STEM jobs because I think they either have good job security, or a good market demand for them. There is no exact science for this, but it's just sort of an art. I stay away from sales, marketing, or jobs that strike me as having higher volatility in turnover.

5) I stay away from loans that have "N/A" or "Owner" or misspellings.

6) I tend not to loan to A1 notes with the lowest interest rate. I like A-grade notes with 6.5 or 6.9% because I feel like that gives me more of a cushion for defaults.

Again, I am completely open to any and all feedback. Just looking for blind spots or things I might explore further as I refine my strategy.

[jz451]

The first thing I would do if you haven't done so already is create an account for NSR and input your payment info (only way to use more than three criteria filters at a time) and play around with the criteria you want to filter until you have some filters you feel comfortable using. For me I have filters broken down for each grade.

Here's the criteria I'de play around with.

Year (start/end date of analysis)
Subgrade
FICO score
Term
Home Ownership
Income
Revolving Balance
DTI
# Inquiries
Employment Length

Analyzing the filter you provided starting in 2014 it yields  7.1% and a loss rate of 2.9% w/ avg age of 19.4 months. So it yields a lower rate than what you had in the past, but your loss rate is going to go down as well, with a break down by year of the following:

            Yield       Avg Rate    Loss Rate    Months
2014    7.63%    11.19%     3.23%         32.09
2015    6.37%    10.01%     2.85%         19.86
2016    7.07%    10.13%     2.37%         10.13

I combined grades A,B,C together, but I would really say that the yield overall will increase if you filter grades separately and weight each grade according to what you want, for example you want more C loans so say a weighting of 20% A, 30% B, 50% C.

In the end it is up to you what you want to do and would suggest using NSR to play around with the filters until you find something yo are happy with.

[jz451]

After seeing you posted again I'll create a new post in response to that.

1. I don't invest on the primary platform but occssionally look, and using it for a model I am making. I look at the loans with the highest funding since they will be funded quicker. Doesn't mean they are better than loans with less funding, just a way to get the loans into the portfolio quicker.

2. Using the filter you provided that I looked at in my first post, the highest yielding loan amounts were the highest. $28k and greater yielded over 7.6% vs the average of 7.1% and just glancing over it the smaller amounts $10k and below yielded less than 7% ranging from 6.19% to 6.75%. So I would either ignore loan size or invest in bigger loans.

3. Same as above.

4. I ignore job title since it doesn't tell me anything as people input a million different titles and I know nothing about them. I exclude if it says N/A, same for year of employment as they usually go together.

5. Same as above

6.  I would do the same, I've looked into it and A4-A5 are probably the best bet for a hgiher yield of A loans, maybe A3, but not too many.

[fliphusker]

Like JZ I no longer invest on the primary but when I did my backtesting led to a filter similar to yours.  With a few exceptions.

1) Exclude sand states (Arizona, California, Florida, Nevada)--This is still included in my FOLIO filter.
2) Max DTI of 20%-- Was the same. 
3) Min length of employment of 5 yrs--If I remember, backtesting showed lil correlation.
4) Homeowner--I only took notes that OWN or Mortgage.  Oddly enough people who owned their homes did have a bit more chance to default.  Was not big enough for me to be overly concerned.
5) Exclude the following loan purposes: medical, business, invest in learning/training, wedding, vacation, moving-- I only invested in CC and refi.
6) Monthly income >$4000--Mine was $60k/year.
7) Inquiries < 2-- Mine was set to 0.  Anything above my backtesting showed a spike. 
FICO was a must over 680. 
I did only invest in notes that were over 12%
Another filter I used was public records.  That was a big huge red flag for me.  If they did it before, did they learn anything from it?  Well no, they have got themselves into another credit crunch.

Here are quick notes from my 9 chargeoffs.  (Yes ridiculous sample size I know.)
1st -- System engineer $9k, 6 years, rent, 1 public record.  Bought on primary market bad filter from LR
2nd --Unloader, $3300, 10 years, Own, 1 INQ.  FOLIO
3rd -- Program coordinator, Own, $6k, 8 years employment.  FOLIO
4th -- NA job length, NA title, Own, $3300, 1 pub record.  Primary, another bad LR filter
5th -- NA job length, NA title, $3500, Mort, 2 INQ.  FOLIO
6th and 7th -- Cop, 10 years, $7k.  FOLIO met ever filter from primary as used to use primary filter on FOLIO.  Which is silly.
8th -- Teacher, 10 years, Mort, $6800.  FOLIO and met all my primary filters.
9th -- Head custodian, 10 years, $2100, 1 pub record, 1 INQ.  Another LR bad filter. 

So after almost a year I never had loan (out of 80.  I know small sample) that went belly up with my own filter. 

[SLCPaladin]

Thank you fliphusker and jz451 for your insight. I have run back testing on NSR platform, which is where most of my filters were derived from in the first place. It has been a long time since I have spent any amount of time and I think it is now time to revisit this. I think a glaring error that I made is that I went by ROI instead of loss rate. My suspicion is that some of the back testing I originally did is not really relevant to future results because the calculated back test yield is materially different from future yield for any given note grade, especially above C-grade notes.

[jz451]

That only leaves Treasury bonds since there are forums about stocks/bonds/etfs/mutual funds. Even basic investing is hard since you need to do research to know if it's a good buy or not.

I'd find an investment that I felt comfortable enough with that I didn't need to commit huge chunks of text to an internet forum in order to solicit feedback from other people currently doing the same in order to make themselves feel likewise - but I am a very, very lazy person...