top of page

Peer Loan Investing Guide

Our techniques employ advanced analytics methods applied to massive amounts of historical data. Our analytics experts have developed studies which we believe allow us to better predict which peer to peer loans are most likely to be repaid in full. 

How does evaluate loans?

We have developed an algorithm which considers up to 15 pieces of data from a borrowers' application and credit history. Each data element is weighted based on its past correlation to loan defaults. This algorithm produces a score from 0 to 100 for each loan. A score of 50 or higher is considered investment quality and we believe these loans are significantly less likely to default. Our goal is to help you increase your Lending Club investing returns.

Can you tell me more about this algorithm?

We built our algorithm after studying over 1,000,000 peer to peer loans. Each available data element was studied in order to determine if it is correlated with loan default rates. After eliminating elements with no correlation, we determined which elements were correlated with each other, thus allowing us to remove or adjust the weighting of correlated elements. The final step was to use all of the information to determine the optimal relative weighting of each correlated data element in order to produce the algorithm for scoring new loan requests. 


What did you find?

We found some results that were logical and not surprising. We also found some elements where correlation was counter-intuitive. This means that the correlation we found was the opposite of what one might expect. For example, in some cases borrowers with higher amounts of outstanding debt are less likely to default.  

How much lower are default rates for highly rated loans?

In general, default rates for highly rated loans across all grades were approximately 7 percentage points lower than the average default rate across all loans on the platform. You should remember that this is based on historical data and we cannot guarantee that future loans will perform the same way. In addition, this is an average across all loans, so a portfolio will only include a limited number of loans which can result in significant variability from past averages. Every point lower the default rate is can increase your Lending Club returns by a point. 

Do you consider a borrower's credit score in evaluating their loan request?

No, we do not consider the borrowers credit score. Credit score is based entirely on a borrowers credit history. While this provides a quick and easy way for underwriters and lenders to evaluate borrowers, our research has shown that there are better indicators of a borrower's likelihood of repaying a loan. It is the underwriter's reliance on credit score to determine the grade (interest rate) that provides the opportunity for analysis such as ours to gain additional insight into the loans that are offered. 

You said that your research shows correlation between data elements and loan defaults. Does it matter that correlation is not necessarily causation?

No, it does not matter. We are not trying to determine why loans default. Our system merely looks for information that has some correlation to loan default likelihood and uses that relationship to evaluate future loans.

I already have a system for selecting loans. How can you help me?

We can help by providing you with our opinion of the loans you are considering. You may find that some of them are not necessarily as good as you think. In other cases our rating may validate your assessment and give you more confidence that you are making good choices. Remember, our researchers have analyzed over 1,000,000 loans. Most people do not have the time or technical ability to do this type of analysis.  

Are you ready to turbo charge your P2P lending portfolio?

bottom of page