the higher the score of a note the less likely the loan is to default
How do you relate profitability with the default risk? In theory, the interest rate should already track with the default risk. Taking greater risk is rewarded with higher returns and conversely investing in lower default risk should have lower returns. You may be able to arbitrate if you can beat LC's algorithm, but they've been getting better with time.
Is your method an improvement on LC risk ratings? Or is the portfolio score giving a risk adjusted expected return?
I ask because I've toyed with doing something similar myself and I'm curious how you went about it.
Yea, it was quite the evolution along a slow road to settle on the best strategy...
I agree that if LC was perfect the interest rate would track with the default rate. By asking a lot of questions over the years to their customer service you realize that not all loans are treated equally. They also have a vested interest to process as many loans as they can to the market so they can make their money in platform fees etc... Speed reduces due diligence based on statistics.
Given all of that, there are holes. Initially, I thought it would be more. But the algorithm spits out about 2-4% of all the loans on the market annually. These are the loans that are represented by a score of 66+. Loans with lower scores can still be viable investments but those loans have indicators that they carry a higher default risk than the higher scoring loans. Once you are buying in the 66+ group you are more confident that you will have a lower default rate (<2% based on the amount of cash defaulted) and the performance of your portfolio now depends on the avg interest rates that you are purchasing (highest LCPicks score with the highest interest rate available). I realized I could get my portfolio to my target ANAR I wanted much quicker/easier once I was more confident in what I was buying. So I started with risk mitigation first and maximizing my interest rate second.
My personal account was the initial guinea pig account with all the learning inefficiencies in it and I have 372 loans with only 19 defaults. A second, very interesting finding was once I switched methods the loans that I had default were actually defaulting less than previous loans. Aka, paying back 20 of the 25 vs 5 of the 25. Makes a real difference in performance over time not to mention new/continued purchasing power of future loans.
Hope this helps.