The Lending Club Investment Portfolio

Avatar
Nillia Ekoue
Avatar
, and
Posted on Apr 16, 2020

Objective

As a group of data scientists, our goal was to investigate the potential to invest in peer-to-peer loans. We wanted to find out what defines good loans and bad loans and construct loan portfolios with advanced return-risk profiles.

What is the Lending Club?

The Lending Club is one of the first peer-to-peer lending services and is still a major player in the marketplace. Instead of borrowing from a bank, borrowers can borrow from “investors” directly. It is often less expensive and potentially easier to qualify for.

How does it work?

The Lending Club offers individual and business loans. Individual loans range from $1,000 to $40,000 principal and have terms of 3 or 5 years. Borrower interest rates range from APR ( Annual Percentage Rate) 6.16% to 35.89%, depending on credit score, credit history, and past borrowing record. Investors can reserve “Notes” in increments as low as $25, which is quite an incentive for small investors. Diversified loan portfolios are expected to earn annual returns between 4% and 6%.

Data Analysis

As prospective lenders choosing from available borrowers, we started with a dataset of loans with 2.2 millions of observations and over 150 variables covering the years 2007 -2014. We narrowed the scope of our analysis by focusing on features only visible to investors and by downsampling our data. While computing return and default rates, we defined and analyzed some of the risks of the loan faced by investors.

Loans Risks for Investors

Based on the first analyses, we categorized the following types of loans as lower-risk and high-risk loans, considering that investors may want to minimize their risk exposure.

  • Low-risk loans: Fully Paid
  • High-risk Loans: Defaulted, Charged off, Settlement, In Grace Period, and Late Payments. 

The ratio of default  vs fully paid loans

Profit/Loss by term and grade

The longer the term of the loans is, the more likely investors are to lose or earn money. However shorter terms are safer investments (with smaller rewards).
Additionally, the below graph displays that the grade is a significant factor that affects the returns of the loans.

Risk Factors

Upon performing a random forest model on loans with only features visible to investors, the results of features by importance showed that "debt-to-income" ratios, states the borrowers are located at, purposes of loans, and FICO score (consumer credit score) are significant indicators of risks of loans. The loans with high numbers of these indicators were more likely to be categorized as "default" loans. 

  

Portfolio building

Based on the above analysis, we created a selection of portfolios that considered features that signal risks of loans and filter them to offer a range of risks and returns by what investors are looking for. We targeted loans that were fully paid to understand each of the scenarios.

Portfolio Performances Compared

The above graph shows that risky (diversifying into high-interest loans) strategy often yields significantly higher returns than safer strategies, but the next question is: how can we mitigate the additional risk?

Modeling Default Likelihood of Low-Grade Loans

In order to increase the accuracy of the above portfolio, we stack 7 binary classification models to predict defaulted loans. The important metric for this training was precision. We wanted to reduce false positives in predicting fully paid loans. The below chart shows testing precision for different years. 

Machine learning enhanced portfolio

By focusing on higher risk loans (grade D and below) and predicting their likelihood of default using various machine learning models, we successfully increased the returns of the portfolios. The resulting portfolio performance improved by 10%. The below chart compares "risky" portfolio includes predicted loans vs non-predicted loans by the model.

Conclusion

We recommend an investment strategy of diversifying overall, with the more low-risk loans selected by machine learning. Our future work would encompass model improvement and parameter tuning of the model.

 

Github: https://github.com/kisakiwata/capstone_nit-k

About Authors

Avatar

Kisaki Watanabe

Data Scientist with strong consulting experiences in data analytics/visualization and risk management, serving for industries ranging from social networking service, game, pharmaceutical, media, and advertising. Advanced skills in fraud investigation and trend projection/analysis with tools such as Tableau,...
View all posts by Kisaki Watanabe >
Nillia Ekoue

Nillia Ekoue

Nillia graduated from Fairfield University with a Master's degree in Mathematics. Her background includes different exposure levels to Economics, Finance, and Mathematics. Her interests are in Healthcare, Education, Retail, and Finance and Insurance services.
View all posts by Nillia Ekoue >

Leave a Comment

No comments found.

View Posts by Categories


Our Recent Popular Posts


View Posts by Tags

#python #trainwithnycdsa 2019 airbnb Alex Baransky alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus API Application artist aws beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep Bundles California Cancer Research capstone Career Career Day citibike clustering Coding Course Demo Course Report D3.js data Data Analyst data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization Deep Learning Demo Day Discount dplyr employer networking feature engineering Finance Financial Data Science Flask gbm Get Hired ggplot2 googleVis Hadoop higgs boson Hiring hiring partner events Hiring Partners Industry Experts Instructor Blog Instructor Interview Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter lasso regression Lead Data Scienctist Lead Data Scientist leaflet linear regression Logistic Regression machine learning Maps matplotlib Medical Research Meet the team meetup Networking neural network Neural networks New Courses nlp NYC NYC Data Science nyc data science academy NYC Open Data NYCDSA NYCDSA Alumni Online Online Bootcamp Online Training Open Data painter pandas Part-time Portfolio Development prediction Prework Programming PwC python python machine learning python scrapy python web scraping python webscraping Python Workshop R R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn Selenium sentiment analysis Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau team TensorFlow Testimonial tf-idf Top Data Science Bootcamp twitter visualization web scraping Weekend Course What to expect word cloud word2vec XGBoost yelp