PeerCube: Founders' Loan Filter for Lending Club

Since the release of PeerCube about two weeks ago, I have learnt a lot from the feedback and input of PeerCube users. Please keep your comments coming ...

Steep Learning Curve with P2P Lending

As seasoned P2P lenders, sometimes we forget how much of a learning curve is there for a beginner peer to peer lender. While the new lenders may be familiar with stock/equity market, the knowledge of debt/bond market and credit report data analysis is non-existent or at best rudimentary.

During development of PeerCube, we were concerned about our decision to not include loan parameters that are accessible only with authentication at Lending Club site. But after almost two weeks of usage on PeerCube, we can observe that even the twenty nine parameters we presented are overwhelming and confusing for new peer to peer lending investors.

I believe we can help such lenders through education and simplifying loan selection process. Here, I see the benefits of education initiatives for new peer to peer lenders that Peter Renton has taken with his Social Lending site. At PeerCube, we want to do our part in lowering the barriers to loan selection and investing by beginner P2P lenders.

Starter Pre-tested Loan Filters

Typically, it is very difficult to get started with a blank piece of paper. The same behavior we noticed at PeerCube. While the seasoned lenders are missing the ability to use advanced parameters and perform various calculations, the beginner lenders are finding the task of prioritizing and selecting from over two dozen available parameters insurmountable.

We would like to help beginner lenders overcome the "blank paper" barrier by providing a set of starter and pre-tested loan filters. We also would like to offer guidance for developing suitable lending strategies. To meet these objectives, we have decided to release the loan filter Mrs. RT and I use for selecting loans on Lending Club. In future posts, I will also discuss how we developed and tested this filter while educating the readers on various filter parameters.

Even though quite a few P2P Lending Blogs disclose their loan selection strategy online, the beginner lenders may not be aware of such information. We believe compiling such information at one place and in easy to use format will go a long way in helping and growing P2P lending community. To achieve this vision, we will be reaching out and requesting permission from such bloggers to convert their published loan selection strategies for use on PeerCube. These filters will be offered to PeerCube users as starter loan filters that they can modify and save for their personal use.

Founders' Loan Filter (FF)

Some users may have already noticed a new green button "Founders' Filter" on PeerCube Filter page. Clicking this button will take users to filter parameters page with appropriate parameters already selected. Users can either use this filter as-it-is or modify to their liking and save the modified version for future use.

We have been developing and refining the Founders' Filter (FF) for almost a year. The published version is our latest revision, the 21st. The FF is based on our analysis of historical loan data from Lending Club and subsequent revisions were made to exclude the loans that didn't match our expectations of quality loans that appeared in the results of original filter.

Most investment strategies caution that past performance may not be representative of future results. This caution also applies to FF. We believe that FF offers a conservative lending strategy on Lending Club but we can not be sure of its suitability for every user. Please use FF as the starting point and at your own discretion.

FF Lending Criteria

Listed below are the conditions that FF uses to filter loans available on Lending Club platform.

• Loan Amount Requested: $5,675 or higher,
• Loan Grade B, C, and D. Sometime we include grade A loans when very few loans show up in FF results. Primarily filter selects B grade loans,
• Only 36 months Loan Length,
• Debt to Income ratio: 14% or less,
• No Delinquencies in last two years,
• At least 36 months since last delinquency,
• No Public Records,
• Revolving Line Utilization: 52% or less,
• Inquiries in last 6 months: One or less, and
• Employment Length: 2 years or higher.

The FF loan filter doesn't exclude loans by Loan Purpose, Home Ownership, FICO Range and Location.

FF Loans Status

The first chart below show the status of loans historically that meet FF lending criteria. The status for the whole population of loans in Lending Club historical loan database is also shown in second chart for comparison.

I need to caution again here that past performance may not be representative of future results. Also, FF loan selection strategy may not offer the sufficient diversification for a Lending Club notes portfolio as only 986 loans met FF criteria out of total 72,015 loans as of September 10, 2012 in historical loan database.

In future blog posts, I will discuss how I developed FF lending criteria and also how performance stacks up with time.

Mrs. RT: From Lending Club Night to PeerCube Night

As we’ve been using PeerCube to select loans, our Lending Club “date” night has ended. Now Mr. RT and I can review loans during our own free time, exchange comments virtually and invest loans without discussing about each loan together. Ahhhh, I miss those Lending Club nights…

The first checkpoint: Filter Loans

Now this is how I select loans using PeerCube. I can create a loan filter, save it and use it again and again to find “quality” loans using my own filter. The loan parameters I care about first and most are the following:

• Interest Rate (exclude low interest),
• Loan Grade (exclude both too good and too bad ones),
• Debt to Income Ratio,
• Delinquencies (no delinquency),
• Public Record (no public record),
• Revolving Line Utilization,
• Inquiries, and
• FICO Range (exclude both too good and too bad ones).

I set these loan parameters as I like. I don’t differentiate borrower’s location, home ownership, loan purpose, and length of employment at this stage. I select all options for these parameters.

The second checkpoint: Filtered Results

By looking at the filtered results, I tend to notice and avoid loans for business purpose and with short employment length. But my filter is pretty selective already and usually only a handful of quality loans are selected. So, I review details of loan one by one anyway.

• Loan purpose – I’m hesitant to invest on business purpose loan or self-employed borrowers. But Mr. RT would rather invest on such loans.
• Open Credit Lines – I avoid the borrower with too many open credit lines.
• Earliest Credit Line – Based on this information, I can sometimes see the link between borrower’s age, loan purpose and financial situation.
• Balance among Loan Amount, Monthly Income, Revolving Credit Balance, Debt to Income Ratio – By looking at those numbers, I try to determine if a borrower is capable of paying back a loan on time. Because living cost is so varied from region to region, I wish I knew average income based on the place where a borrower lives. That would be a useful information when selecting a loan. Mr. RT, can you add this feature to PeerCube?

The third checkpoint: Link to Lending Club

Once a loan has passed the second checkpoint, I check the detailed loan information found on Lending Club site by clicking Invest button on PeerCube. I’m interested in reviewing the following information:

• Monthly Payment – This is, again, a good input to see borrower’s capability of pay back.
• Current Employer – I tend to trust borrowers working for established company, schools, hospitals, state, and government.
• Questions and Answers.

Next Step

Finally, I make decision on the loan and go back to PeerCube, rate the loan and leave my comment.

I wait for Mr. RT to review the loans I picked. On Loan Review page, he could review loans with highest peer rating or loans with most peer opinions. Currently he finds my rated loans easily. In the future, when PeerCube becomes popular (!?), he may not be able to find my rated loans. Mr. RT, how about a feature that a user can follow a particular user’s ratings and opinions?

After developing PeerCube, our Lending Club night has actually evolved into PeerCube night. Instead of discussing about loans, we discuss about bugs and new feature of PeerCube that can help collaborate further and speed up loan selection process.

AG's Rant

Last week, Peter wrote a post on how interest is calculated on P2P loans. Though the article is good for beginner P2P lenders, the best part of article is the discussion in the comment about how pre-payment of a loan can give lenders negative return after Lending Club's 1% service fee. Mrs RT already wrote about her displeasure when one of our loan was prepaid so early that we didn't get any return on that note. Being laid back, I chalked it up to cost of doing business.

I reluctantly agree with the argument that there should be a pre-payment penalty for borrowers though it should be very nominal. But I also feel like that P2P lenders are starting to act like 'blood-sucking corporations and financial institutions focusing on maximizing ROI.' What is the difference between them and us? I believe unlike corporations, we still have a soul and compassion for human beings.

I am big fan of Derek Sivers and such discussions remind me of this  animation.

I miss the mob from Derek Sivers on Vimeo.

Introducing PeerCube: Smart Lending with Peer Insights for Lending Club

I am very excited to announce the launch of PeerCube, a web service platform to assist peer to peer lenders make collaborative and speedier decisions during loan selection process. PeerCube is currently available to all lenders on Lending Club P2P lending platform. We plan to roll out similar features for Prosper Marketplace in the future.

I feel it is important for me to share what PeerCube offers before I talk about how PeerCube came about.

What PeerCube offers?

I will be upfront. PeerCube is optimized for how I and Mrs. RT make loan selection and lending decisions. I hope that PeerCube provides value to fellow lenders, too. We focused on two areas of loan filtering and loan review.

Personalized Loan Filters

• Ability to create, load and delete custom loan filters available at one location.
• Load a saved filter, adjust any of the 29 parameters, and save the resulting filter as a new filter.
• Filtered results showing loans meeting a filter criteria, high level information about the filtered loans and a Detail button to view the loan details, funding progress and peer insights.

Collaborative Decision Making

• Review top peer rated loans, your recent reviewed loans, and loans with most peer opinions at one location.
• Rate loans, share your opinion and review peer rating and opinions

A few screen captures from PeerCube website are shown at the end of this blog post for your review.

How PeerCube came about?

The credit for the idea seed that grew into PeerCube goes to Mrs. RT. As she wrote in her blog post Filtering the Currently Available Loans, she wasn't impressed with the loan selection process through Lending Club website as well as through CSV download that I was using.

She thought both processes were slow and took too much of time commitment versus the amount to be invested. She suggested I look into automating the loan selection process with couple of her requirements in mind:

1. Avoid downloading the CSV file for currently available loans and inserting the custom filters on the columns to filter loans every time we decide to review available loans for investment.
2. Avoid the need to save the filter criteria in a separate document that is referred to create custom filters, every time we download new CSV file for currently available loans.

Personally, I was also feeling that we were spending too much time for the amount we were investing each time. With the original requirements from Mrs. RT, I created a couple of web pages for our home server that automatically downloaded the currently available loan data from Lending Club, a loan filter, and result page that showed the filtered loans.

Mrs. RT likes to select loans and decide on investing together as mentioned in her post Let's do Lending Club night!. When we couldn't hold Lending Club night several times due to scheduling conflict, it was clear we needed "virtual" Lending Club nights to continue investing. So a new requirement of collaborative decision making and opinion sharing was tacked on to the original project.

Mrs. RT was so excited with peer to peer lending and Lending Club that she couldn't stop talking about it with her friends, doctor, hair dresser and everybody else who will listen. Naturally, some people expressed interest in learning more and requested access to our project. Such requests along with complexity of using VPN to access home server remotely, it became clear that this web service needed to be hosted at a web hosting company with public access. So a new requirement of user authentication and management was tacked on to the original project requirements. And, the final result is PeerCube as it stands today at www.peercube.com.

As a fellow peer to peer lender, all readers are invited to use this free service. Your feedback and suggestions for improvement and new features are most welcome through blog comments, Contact Us page at this blog and Contact page at PeerCube.

Screen Capture #1: Create, load and delete custom loan filters

Screen Capture #2: Adjust any of the 29 loan parameters

Screen Capture #3: Filtered Results

Screen Capture #4: Loan Review

Screen Capture #5: Peer Insights

Disclaimer: PeerCube is not endorsed by, owned by or affiliated with Lending Club, Prosper, and any other peer to peer lending platform.

Mrs. RT: Avoiding Early Pay-offs at Lending Club

Note: After a week long hiatus, we are back. Mrs. RT once again wanted to share her opinion. She has become quite a proponent of peer to peer lending. She can't resist talking about it at social gatherings.

Later this week, I will have some very exciting news to share with you. Stay tuned ...

The Panic of early pay-off

The presidential campaign is getting hotter and hotter these days. So, today’s topic is how bad political extremes are…. Oops! This is about peer to peer lending. Let me change the gear.

The other day, Mr. RT said that one of his loan at Lending Club was fully paid off. What!? No, we were supposed to keep on collecting interest for five years. Wasn't that the term of that loan? Did we lose money because of fees? I became a little panicked. “Calm down, Mrs. RT”, said Mr. RT. Then, I realized I was talking about losing 25 cent fee. It’s good to have analytical person as partner.

The Good and Bad feeling of early pay-off

That was how I found lenders could pay off early without pay-off penalty. This is good to know!
In our own situation, we paid off our mortgage before the loan term ended. Our interest was so low that it didn't make sense to hang on to the mortgage balance as it was not beneficial for the tax deduction anymore. Moreover, we can save future interest. That’s a great feeling to become loan free! So, I can imagine this borrower, who paid off our loan far in advance, must feel the same way I do.

However, it’s contradicting thought, I know, as lender, I’m looking for more “reliable” loans, so that I can collect money during the loan term and count on estimated return. Hmm..., that’s how all the banks and mortgage lenders are thinking…? I feel I am becoming greedy like them.

Avoiding early pay-off

If I read the loan descriptions like these for 3 year loan, there’s no way I will lend money to those “responsible” borrowers:

“I have a great job, make a 6 figure salary and can pay this back in about a year; if it came down to it a few months,”

“I can even afford to pay off this loan in less than 2 years if necessary.”

Sorry, you are too good for me! I became curious to know what characteristics people who paid off early have in common.

Imitating Mr. RT

I downloaded historical loan data from Lending Club site. I limited my analysis to the loans issued in 2012 and that were fully paid off. There were 536 loans issued in 2012 that were already fully paid off as per the data available on September 3rd. Because there are only two loan terms, 36 months and 60 months, these borrowers are way ahead of the schedule to pay back.

Further, I excluded partially funded loans. In the end, there were 454 loans that were paid off and issued this year. What I found interesting and think it would be good for me to consider selecting loans are following:

• 398 loans or 87.7% of loans that were paid off early were for amount $20,000 and less.
• 364 such loans or 80.2% of loans were issued to borrowers with credit grade from A to C.
• 29% and 27% of borrowers that paid off loans early were with credit grade A and B respectively.
• 20% of such loans were issued to borrowers with credit grade A for amount less than $10,000.

Key Takeaway

I defer to Mr. RT for detailed analysis in the future. But what I concluded for my lending criteria is to avoid good credit borrowers for low loan amount because they may pay off early and cause the expected return of portfolio to be lower.

Now I will have to be careful in selecting both extremely good and bad credit borrowers.