In one of my most popular posts from last year, and one of my first posts ever, I shared with everyone how I was planning on investing in my Lending Club account. Once I opened my Prosper account in May, I share the same details for that account. As I mentioned in my 2014 goals post, crunching numbers and over-analyzing my peer to peer lending accounts is a hobby of mine, which translates nicely to putting some significant time into developing my Lending Club and Prosper investment criteria and strategy.
Now that we are into the New Year, it is time to share my investment criteria, the strategy and methodology I use to develop those criteria, where I do my data filtering, and lastly I will share some simple, but effective criteria for those who aren’t having much luck finding notes. However, due to the length of the post, I will be sharing all of this information in a short three-part series of posts.
Where to research investment criteria?
For a couple of years, there was only one robust and prominent peer to peer lending statistics site. Lendstats.com was where everyone went to research filters, snoop on other lenders (for Prosper), and dive into data with a wide range of filtering options. It was at Lendstats that I first cut my teeth on determining filters for my original Lending Club account.
However, as with any growing and developing industry, new competitors and better resources emerged. Of the new sites, Nickel Steamroller (NSR) became the dominant player for the data nerds after a period of absence from the founder of Lendstats led to some stale data. Since NSR came onto the scene, I’ve been making use of both sites, but over the last year or so, have exclusively used NSR for my filtering and investment research needs. However, for a long time, NSR was almost exclusively limited to Lending Club investors and lacked the ability to do research for Prosper. Prosper-stats.com filled this need with a robust platform that exclusively focused on Prosper.
But the improvements haven’t stopped there, as just recently Michael, the founder of NSR, and Rocco, the founder of Prosper Stats, joined forces and released Nickel Steamroller 2.0. This incredible new site has combined the best features of both sites enabling users to research both Lending Club and Prosper filters, perform portfolio analysis on their existing accounts, and find a tremendous amount of other resources including economic and other peer to peer lending related charts. The primary areas of NSR that will be discussed today are the back testing/filters portion of the site with both Lending Club and Prosper. These pages are found under the Analytics drop down seen in the image below:
I strongly suggest that anyone and everyone interested in peer to peer lending, with either Lending Club or Prosper check out NSR 2.0. For additional information on the functionality of the site, be sure to check out the release information from NSR’s blog, and two reviews done by Lend Academy and Lending Memo.
As I mentioned above, a lot of time and thought have gone into developing my filters, and while they can never guarantee future results, they can at least be a starting point. Ultimately, I am not a statistician and my inefficient methods have certainly resulted in positive results for me personally, but should not be relied upon exclusively. So with that in mind, let’s dive right in and see what I will be using.
My 2014 Lending Club Investment Criteria
While I am not the riskiest investor in the space, I attempt to seek a balance between high returns and low defaults. Just like last year I have two Lending Club criteria, but unlike last year, I have segregated mine into 36 and 60-month filters. Each filter is fairly similar with just a few small tweaks between the two. While I will get more into the process of developing these filters in the next part of my investment criteria/strategy series, I will provide a couple of pieces of context to the filters below.
First and foremost is the equal weighting of all loans at $1,000 per loan. The permalinks below to my filters on NSR will have this selection marked. As someone who does not have the means to invest in whole loans, this provides a better snapshot of what returns I might be able to expect with my filters as it removes some of the distortion from different sized notes at different credit grades. Secondly, when filtering, I generally am only looking at notes issued a year or so prior and before. This “ages” the filter and helps remove some of the return distortion provided by the massive influx of notes in the last year or so as Lending Club and Prosper have exploded in popularity.
Lending Club Primary Filter: 36-Month Notes
- Loan Grade: C, D, and E only
- Public Records: Zero
- Inquiries last 6 months: Zero
- Open Accounts: Greater than or equal to 10
- Purpose: Credit Card Refinancing and Debt Consolidation only
- States: All excluding FL and NV
- Annual Income: Greater than $75,000
- Employment Length: 10+ years
- Home Ownership: Own or Mortgage
- Loan Amount: Greater than $10,000
- Term: 36-Months only
All said I have a total of 11 criteria, including the term criteria, for my 36-month filter. As of 2/13/14, this filter projects a return on investment (ROI) of 12.89% for loans issued through 12/31/2012. These notes have an average age of 24.6 months and a default rate of 2.74%. Here is a link to NSR showing my filter and the corresponding up-to-date results: Lending Club 36-Month Filter.
Lending Club Primary Filter: 60-Month Notes
- Loan Grade: C, D, E, F, and G
- Public Records: Zero
- Inquiries last six months: Zero
- Open Accounts: Greater than or equal to 10
- Purpose: Credit Card Refinancing and Debt Consolidation only
- States: All excluding FL and NV
- Annual Income: Greater than $75,000
- Employment Length: 10+ years
- Home Ownership: Own or Mortgage
- Loan Amount: Greater than $10,000
- FICO Score: Greater than or equal to 669
- Term: 60-Months only
For my 60-month filter, I have a total of 12 criteria, including the term criteria. As of 2/13/14, this filter projects a ROI of 16.03% for loans issued through 12/31/2012. These notes have an average age of 22.8 months and a default rate of 1.76%. Here is a link to NSR showing my filter and the corresponding up-to-date results: Lending Club 60-Month Filter.
My 2014 Prosper Investment Criteria
Unlike with Lending Club, I only utilize one filter for Prosper. This is primarily due to my lack of confidence in Prosper’s ability to appropriately rate and evaluate 60-month notes. I will get more into my evaluation and thought process behind this in my next post discussing how I arrived at my various criteria and my research methodology and strategy. Again, like with my Lending Club filters, I look at loans issued prior to the end of 2012 and with an equal weighting of $1,000 per note across all notes.
Prosper Primary Filter: 36-Month Notes
- Prosper Rating: C, D, and E
- Income Range: All ranges greater than $50,000
- Public Records: Zero
- Inquiries last six months: Zero
- Public Records last 10 years: Zero
- Open Credit Lines: Greater than or equal to four
- Current Delinquencies: Between zero and one
- Months Employed: Greater than or equal to 60 months
- Term: 36-Months only
All said I have a total of nine criteria for my Prosper filter, including the term criteria. As of 2/13/14, this filter projects a ROI of 17.69% for loans issued through 12/31/2012. These notes have an average age of 30.2 months and a default rate of 6.03%. Here is a link to NSR showing my filter and the corresponding up-to-date results: Prosper 36-Month Filter.
So there we have it, my three primary filters that cover all of my peer to peer lending accounts. Stay tuned over the next couple weeks as I dive further into my methodology and strategy when researching filters. Again, there are limitations to my methods (which I’m sure a few stats geeks could readily point out) but they’ve been working for me for a few years. Additionally, for those who are looking for less complicated filters, come back and check out my simple filters for both Lending Club and Prosper when I post those next week. In the meantime, feel free to check out my Lending Club and Prosper pages and see where my accounts currently stand!
Update: My Simple Lending Club and Prosper Filters
Have you finalized your filters for Lending Club or Prosper? What have you found effective? Still concerned about investing or opening a peer to peer lending account?
Flickr: canyon289
Thanks for sharing your filter criteria. I just back-tested your LC 36 months criteria on PeerCube. Your criteria shows higher ROI and lower Default than all loans benchmark. But …
There are only 1,474 total loans issued in past 8 years combined. And only 259 loans among them have fully matured.
Unless you invested more than $25 per notes, you couldn’t have invested more than $1,975 in 2011, $9,700 in 2012, $18,625 in 2013 assuming you invested in all notes issued in that year.
I am not sure that there is enough data to support your expectation of high return low default or you have a scalable strategy as portfolio size grows while maintaining sufficient diversification.
Statistics is a funny math, you should also worry about what it doesn’t tell you in addition to what it is telling you.
First off, thanks for stopping by and commenting Anil. I appreciate you checking out my filters and providing your input. You’ve done a great amount of research on various criteria, etc., and I’ve enjoyed reading a lot of what you’ve published on the subject.
Addressing my first Lending Club filter which you’ve back tested on PeerCube, I understand it is of a limited nature. While certainly less than 300 completed notes is not a huge amount, there is a sizable amount of research demonstrating the overall decay in returns for 36-month loans, regardless of filter. In terms of diversification, having any more than 200 or so notes provides ample diversification in a portfolio, with exponentially diminishing returns the more notes you have. Since I am not relying on one filter, concerns for diversification further becomes a moot point. I agree that the ability to scale is important, so to truly evaluate that, you will need to combine both filters I am using, and look at the number of notes becoming available. In 2014 there have already been over 200 notes to meet both of my Lending Club criteria combined, more than enough for my needs.
If I recall correctly, you prefer to use a smaller portfolio of notes with a larger investment in each?
Thanks again for stopping by Anil, much appreciated!
W2R,
Great article on peer to peer lending. I have really wanted to make the jump into this kind of investment but feel like I should build up my stock portfolio for awhile first. I have bookmarked this as future reference for when I finally decide to take the leap. At what point in a new investors timeline do you think venturing into peer to peer lending would be a good idea?
I can’t say exactly at what point it is right for you, but there are some general guidelines such as no more than 10% of your investable assets. As a base recommendation, I would suggest you don’t begin investing unless you can put at least $2,500 into an account. This will get you to 100 notes, which will be a strong base level of diversification to protect you against a loss of capital. If I were to start from scratch today, I would not start investing in P2P lending unless I could put $5,000 in as 200 notes reduces the variance in returns to almost ensure a positive return.
Thanks for stopping by Div-asaur!
Yes, this is the post I’ve been waiting for! Thanks so much for sharing your strategies!
Not a problem Mr. 1500! I think you will also be interested in the simple filters I’ll be sharing as well.
Thanks for sharing those research sites. I used lendstats a long time ago, but haven’t really felt the need to revisit. I remember CA used to be the state everyone wanted to avoid, now it looks like FL and NV. You have to wonder if it will change again in the future and if that filter criteria is even meaningful… I ought to investigate sometime.
Thanks for the information!
Thanks for stopping by CI! If you want a great read on the states, check out Lending Memo’s post on Redlining. I was going to reference it when I discussed how I came up with my filters, but here is a sneak preview. Ultimately, while plenty of states are below average, I only wanted to eliminate the worst two overall without filters.
I have concern with Nickel Steamroller Lending Club data’s correctness . I am switching to use lendstats. Their query results are very different. I trust lendstats because their filter fields mapping to lending club’s official filter while Nickel’s not. And lendstats data is closer to reality. Nickel’s loss rate is much less than lendstats.
I’m actually a bit confused by some of your statements Paul. Both Lendstats and NSR use the exact same loss factors, and NSR offers the same fields, and more, than those that are offered at Lendstats. Additionally, the ROI on NSR might seem lower, but it takes into consideration the fees paid by an investor whereas Lendstats does not. I just ran a couple of test filters and found the data to be quite comparable with the biggest difference in ROI being a result of the effects fees have on your returns.
I would encourage you to use whichever site you choose, but I would offer to you that NSR does not have incorrect data. If anything, I have much more faith in the validity of NSR’s data than that of Lendstats.
I appreciate you stopping by and expressing your opinion Paul.
Thanks for you reply. Actually the results I got from the 2 sites are very different. If you test result is comparable, there might be something wrong with my query. Here I’d like to consult you about some column’s meaning on Nickel. For “Open Accounts” on Nickel, should it be equivalent column to “Open Credit Lines” on Lendingclub? And what’s the equivalent column on Nickel to “Total Credit Lines” on Lendingclub?
Paul, I now understand where you were finding your discrepancies, yet want to again establish that the underlying data itself is very much complete at NSR.
You are correct that Open Accounts on NSR matches the Open Credit Lines on Lending Club. As for Total Credit Lines, NSR does not currently offer that as a filter. Having communicated with Michael several times, I can certainly inquire as to the possibility of adding that attribute.
For a full description of all the loan attributes (independent of which statistics site you use), check out the Data Dictionary from Lending Club found at the bottom of the linked page.
Thank you for clarifying your filtering questions. If you have any other questions, just let me know!
I looked on NSR today and note that “Total Accounts” is under the Credit tab. I believe that corresponds to Lendstats’s “Total Credit Lines”.
Thanks for the heads up on that B7!
Paul S, a few things. We use different loss factors. You can adjust these on the “Results Options” tab to make them match Lend Stats. To my knowledge the loss estimates on Lend Stats have not changed in years. Additionally we use a different ROI calculation. I’m obviously partial, but I feel the data on our site is the most accurate on the web.
The credit line information is located under the “Credit” tab. This weekend I added a new feature to pull the full LC data dictionary description in and display it when you hover over the filter. But other than that, the filter name is derived directly from the column name on Lending Club’s more recent export format. It might be worth pointing out too, LC has multiple data exports. We are using the latest which has the most complete data. Yet another difference, LendStats by default excludes principal that LC funded. We include it because we are trying to get the asset classes’ performance.
If you are concerned about our data, I would say download lending club data export, and get the count of charged off loans for 2007 (all completed). Compare that number to what NSR reports, and then what LendStats reports. You will find Lend Stats seems 5 loans short for 2007. Reason unknown. That should clear up your concerns.
Michael, thank you for stopping by an providing some explanations to some of Paul’s questions. Much appreciated!
Paul, lendstats.com stopped updating long time ago. Is it alive now?
RaymondG, I think it is “relatively” still alive, but in my opinion, not reliable as a result inconsistent care taking. At this point I use NSR exclusively for my filtering/backtesting needs.
Excellent post. I am about to dive into prosper, and I had initally thought that $2500 would be my starting point, though I agree that more ($5k) would be more beneficial. That being said, should I start with my $2500 or wait until next year when I can throw $5k at prosper, as you’ve recommended.?
My $5k recommendation is primarily there as a representation for the amount that people should commit to their P2P lending account(s). If you are committed to growing your Prosper account, then I would encourage you to go ahead and get started. With the reinvestment of payments alone, you should end the year with well over 100 notes (reasonably diversified), and will be on your way towards a fully diversified account. Adding another $2,500 to the account in the next year will certainly carry you over the top.
Thanks for stopping by and I hope I was able to help!
Thanks for the reply and reinforcing my plan.
W2R,
Thanks for sharing your filter criteria. I’m a fellow NSR nerd here. I have spent many of the last 6ish weeks honing my filters as I started new Roth IRAs for me and my girlfriend on LC.
I gotta say – your 60 month looks decent with ROI and loss, but I think you could do a bit better for your 36 month filter. I haven’t made a 36-month filter myself yet, but my gut tells me that with some prodding, you could drive down loss to 2% or less, drive up ROI, and maybe even drive up your total number of matching loans a bit, with some experimentation and loosening of screws.
If you want, I’ll give you one of my filters that I use through an automated service. (I strongly dislike filtering and browsing for notes manually… hopefully one day LC will step up their interface game)
Neil, I’d be happy to take a look at your 36-month filter and see what you’ve come up with. Feel free to send me a note through my contact page and we can go from there. Out of curiosity, which automated investment service are you using?
Important to note, no 60-month loans have reach full maturity at this point. So, given similar risk factors, they most likely have not reached their full default rate. Secondly, there are a few filter criteria that improve the performance of both 36 and 60-month notes. Unfortunately, this criteria limit the history of the filter to a couple of years or less, which is not long enough to have a full view of the stabilized returns.
I appreciate you stopping by Neil, and best of luck with your two new Roth IRA accounts!
Thanks for posting this! I’m waiting for my funds to be available so I can start and this strategy is the easiest to implement so I can get started ASAP! I’m more comfortable with a $25 note just because I can swallow that easier than $1000 default so that’s the only adjustment I’m going to make. Thanks again! I look forward to reading more of your blog!
Good luck with your investment! I hope you have much success with it and be sure to let me know how it turns out!
I can’t seem to find all of these filters in lending club.
They are all in Lending Club, however some of the actual filter selections can only be done through the API.
I am curious what your 2015 investment criteria will be? This post was 1 year ago. Any changes?
Peter, I’ll be adding a post this coming month that shares my 2015 investment criteria. Keep a look out for it!