This is 清华创业者网's Tencent Weibo homepage. Follow now!

清华创业者网 收听





Repeat borrowers, analogous to repeat customers in other industries, provide a healthy indicator that alternative lending is here to stay. Marketplace-funded loans have proven to be an attractive asset class from the investors’ point of view but, it would not be a true success unless it were benefitting the borrowers as well.


In this post, we’ll take a look at repeat borrowers and how their profiles change between the times of their initial and succeeding loans. Assuming they performed well on their first loans, we would expect to see borrowers getting better deals on subsequent loans. Specifically, to keep it consistent for all repeat borrowers, we compared their first loan to their second loan using Prosper’s public listing data.


First, we compare the average loan profiles between the repeat borrowers’ loans:


These statistics are promising. Borrowers are able to secure larger loans with lower interest rates. Financially, they seem to be doing better too. Their income has increased on average around 5% and their FICO has increased by about 9 points. How is this reflected in their loan ratings? Below, we compare the distribution of ratings side by side:


Consistent with our earlier findings, we can see a shift towards less risky ratings for the borrower’s second loan. In particular, the percentage of AA loans has doubled compared to their first loans. Although this shows that the distribution has shifted towards lower risk ratings, it does not show how borrower’s transitioned to this new distribution. To see this, we can look at a data visualization technique called a heat map.


This matrix shows the movement of loan ratings from the borrower’s first loan to their second. Their initial rating is plotted along the left while their new rating is plotted along the bottom. To read the plot, start with their first loan rating on the left and find their second loan rating on the bottom. Each row sums up to 100% and the darker colored squares show a higher density of borrowers. For example, of the borrower’s who received an AA rating their first time (reading left to right), roughly 60% received an AA rating the second time, 20% received A, 10% received B, etc.


There are a few key points we can infer from this graph:


We see that the movement of ratings is more of a distribution rather than a strict upgrade.


If borrowers all kept the same rating, we would see a dark green line along the diagonal. However, here we are seeing an even distribution with the concentration slightly more towards the top-left of the graph. This indicates that the borrowers have been given less riskier ratings the second time.