Order Complete; All Notes Issued... (2) Filters & Workflow

O'Shaunessey Dam from Wapama Falls at Hetch Hetchy, April 2010

Let me jump right into 'what' I do in the 5-8 minutes I spend each time I login to LendingClub, the 'why' will follow.

• On the Account Summary page, I first click on the "Invest" link which brings up the Build a Portfolio page.
• After a short delay, the page is populated with the count and what's possible from currently available notes: the total number, the three automatic portfolio selection buttons, each with their predicted rate of return - a conservative, moderate and an aggressive option:

On the right there is a "More Options" button where you can dial in your preferred rate and get a basket of notes recommended. If you wish the entire available amount could be invested - all in one fell swoop! I, however, prefer to do my own filtering. I click on the "Open" link in the Filter Notes box along the left column.

• I have a saved filter that I select which causes the number of notes to shrink to about 20% of the total number available. More on my filter below.
• One of the filter items that I include but leave without a value is the Funding Progress slider. Depending on the amount of time I have in any given login session, I slide this over to 50%, if I'm not rushed and 80% if I only have a minute or two. This filter reduces the resulting number of notes to just a handful that already have lenders.
• I click on the number of notes available to actually see a list. Strangely, the notes are presented in some random order. I click the "Rate" link twice to sort the short-list into descending interest rate order.

Finally, starting with the note offering the highest rate, I begin my manual analysis.

The Filter

Having scoured through the Statistics pages at LendingClub, running Excel Pivot Tables on the downloaded data and through my own learning from non-performing loans in my trial account, I've come to rely on the following filter criteria for hand picking notes in my IRA.

• "Reviewed by LendingClub" - I only look at notes that have been approved. It is my perception that almost 25% of notes that are initially presented do not get approved - since I am investing time on each loan, I find this filter worthwhile. Committed funds to unapproved notes are returned to the available funds pool after a few days. This filter may not be as important when using automated investment tools.
• "Interest Rate" - having set a goal to have a portfolio yield above 10%, I exclude grade A loans which yield below 10%. Based on statistical analysis of yields for each grade, I used to filter out grade F notes too, but it appears there was an anomaly in F grade loans in the early days. You can see the change in borrower selection criteria has had a marked effect on the overall default rate which, as of this writing, is 2.11% annualized across the entire LendingClub platform.

• "Loan Purpose" - This filter has been derived from the pivot table exercise, the statistics page and my own experience with non-performing notes in my trial account. The filter I use only shows me notes with the following stated purpose: Refinancing credit card; Consolidate debt; Renewable energy financing; Wedding expenses; Paying for dream vacation; and Other. A word on the two odd appearing purposes - Wedding and Vacation. My initial instinct was to refrain, but statistically, the default rate on these two is extremely low - counter intuitive at first blush. Right?
• "Term" - my initial pass picks 60-month notes only. Two reasons: higher yield and lower fees for same class of borrower and 2 years longer earning period for same time spent on selection. However, if in a given note picking session I don't find enough notes to invest in, I'll turn off the 36-month filter and take a look at those notes too.
• "Exclude Loans already invested in" - this one is obvious.
• Finally, "Funding Progress" - I try not to buy into notes that have not had at least 50% funding commitment, mostly because by the time the notes get this level of commitment other lenders have asked questions and received answers, the broad lender sentiment is visible and you know the note will close in the next few days causing your funds to begin working sooner.

Once this saved filter is applied and results updated, the original set of notes gets whittled down to a very manageable set that I process individually.

The Manual Analysis

With a list of only about a dozen or twenty notes sorted in descending rate order, I begin my final selection. Starting at the top, clicking on the loan title, which is a hot-spot even if it doesn't look like one, the load summary appears in a drop-down. I then scan for the following:

1. Eyeball the ratio of Monthly Payments/Gross Income - I try not to go too far north of 10% on a 60-month note, with a little more latitude on the 36-month, the theory being: lend to borrowers who can afford to service the debt. Unfortunately, this is not a filter that can be programmed today. I look for the green check mark next to Gross Income, if one exists, my lending confidence goes up since the income has been verified.
2. Scan through the Revolving Credit Balance and Line Utilization - filters do exist to handle these criteria - but I've noticed that the credit reports are often inaccurate and lump debts together and can skew the filter. If utilization is high (>80%) I'll take a closer look if other aspects are favorable, if not, I may pass on the note.
3. Length of employment - again, it's possible to filter on this, perhaps I should. I typically do not lend to those employed less than 1 year - unless they are employed by a known main-stream employer.
4. Loan Description and Q&A - I'll walk away if there is no loan description AND there are no answers. Sometimes notes get funding without these - perhaps through automated funding, I avoid them. In case the note is for debt consolidation - I'll look for the kind of debt the borrower is funding. I pass on the note if I see the borrower being irresponsible, funding low rate debt with a higher rate LendingClub note in the name of convenience. I've seen 9% loans being repaid with 15% notes. I also scan for any issues with borrower attitude.

Determining size of loan fraction

This is a very nebulous item. I've noticed that the amount I invest in a note depends on my mood, level of comfort with the statistics and the Q&A. The only numerical guidance I sometimes use is to determine the average per lender investment in the note so far. Unfortunately, determining that figure is labor intensive so I look at it only if I feel a strong urge to. To get that figure, one must look at the 'complete listing' - a link is available just below the loan description section.

In Conclusion:

As I wrote this blog, I went through a real-time note buying session and you can see that I was able to buy 6 notes worth $425 that met my criteria. Click on the next link and you can see what my methods have allowed me to achieve over the last seven months, a  reasonable placement rate, varying between 102 notes in March and 39 in April yielding expected rates between 14.49% in February and 16.12% in July.

Finally, reiterating the caveat: my filters and methods represent my personal appetite for risk - I guarantee that yours will be different. Do share your thoughts and experience. I hope this has been helpful. 

Here's to low default rates and happy peer-to-peer lending!