You are now being redirected to Orchard Platform Markets, LLC.
If you aren’t redirected in a few seconds click here.
You are now being redirected to Orchard Indexes.
If you aren’t redirected in a few seconds click here.

Subscribe to our Newsletter

Login

  • Investor Portal

  • Originator Database

  • Manager Database

If you forgot your login email, please contact your Orchard Account Manager or support@orchardplatform.com.

Understanding the Drivers of Loan Performance: Loan Age

quantifying_impact

Note: throughout this post, we are displaying the quarterly average charge-off rate to reduce noisiness in the monthly data.

 

Beginning in Q1 2016, we started to see an uptick in charge-off rates in the U.S. Unsecured Consumer Lending sector. This charge-off rate was stable at approximately 0.40% for over three years before spiking to a peak of nearly 1% in Q1 2017. In this piece, we argue that this increase in charge-off rates is primarily due to aging of the loan portfolio and that deteriorations in borrower performance play a smaller secondary role in the changes we observe.

Charge-off Rates

While an investigation of the above graph tells the story of significant deterioration in borrower performance in 2016 and 2017, we believe this misses several key points.

 

The charge-off rate is calculated by summing the charged off balances and dividing by the beginning principal balance in a particular month. While an investigation of the above graph tells the story of significant deterioration in borrower performance in 2016 and 2017, we believe this misses several key points. Charge-off rates differ significantly across dimensions of borrower and loan characteristics, and therefore any changes in these underlying characteristics over time will have an effect on monthly charge-off rates. One typical example would be borrower FICO scores—if average borrower FICO scores were lower for a given month, we would expect higher charge-off rates in that month.

While most people are familiar with the effects that changing FICO scores, interest rates, or debt-to-income (DTI) ratios would be expected to have on charge-off rates, there is generally less understanding of the relationship between a loan’s age and charge-off rates. Average charge-off rates are strongly impacted by a loan’s age, rising sharply for the first twelve to eighteen months before flattening to a stable rate. As in the example of changing FICO scores, changes in the distribution of loan ages over time would also be expected to impact charge-off rates. In particular, as the outstanding loan population ages, we would expect to see increasing charge-off rates.

Charge-off Rates by Age

From 2012 through 2015, the distribution of loan ages was relatively stable. Loan originations were increasing steadily over this period, which meant that the majority of loans outstanding at any point tended to be recent originations. The average loan age was approximately 8 months (7.9 – 8.3 months) for every quarter in 2012 – 2015. In 2016, however, poor early results from the 2015 vintage accompanied by negative news stories coming out of some of the largest online lenders caused a drop in investor interest, resulting in a steep reduction in loan originations. During this period, there was a notable aging effect within the loan portfolio, with average loan ages rising to 13.4 months in Q2 2017. Note in the graph above that loans aged 8 months have charge-off rates of approximately 0.6%, while loans aged 13 months have charge-off rates of approximately 1.05%, meaning that this aging would lead to significant upward pressure in portfolio-level charge-off rates.

Balance Distribution by Age

While this aging did play a significant role in the increase in charge-off rates in 2016 and 2017, deterioration in underlying loan performance did still play a role. We discussed some of this deterioration in detail a few weeks ago in our blog post on Lending Club’s 2015 and 2016 vintage performance. In the graph below, we plot the quarterly charge-off rate separately for loan ages 1-24. These graphs show that even after controlling for the effects of loan age, we do still see increases in charge-off rates in 2016-2017.

Charge-off by Quarter

We would like to propose a way of analyzing the changes in charge-off rates that controls for the aging effects identified above. This would give us a clearer way of identifying trends in borrower performance by removing the confounding role that age plays on the portfolio. In the graph below we present a calculation of hypothetical charge-off rates if the distribution of loan ages in Q2 2012 was held constant. Specifically, for each period we calculate the charge-off rate for each loan age, 1 month – 60 months. Then, instead of determining the overall charge-off rate based on the actual loan age distribution, we use the distribution of loan ages from Q2 2012. This controls for the effects of aging in our calculation of charge-off rates. We can see that charge-off rates still rise in 2016 under this calculation—as we mentioned, borrower performance did deteriorate—but the ascent is nowhere near as steep as shown initially. We believe this is a better measure of underlying borrower performance.

Charge-off adjusted

In the graph below we plot both the actual and adjusted charge-off rates by quarter, allowing for easier comparison. Notably, both of these rates tracked each other extremely closely from Q2 2012 – Q4 2015, after which they began to diverge. This close tracking is due to the fact that, as shown earlier, there was little change in the distribution of loan ages until Q1 2016, when new originations began to drop. This graph makes it clear that the largest part of the increase in charge-off rates over the last 6 quarters was due to an aging loan population, and that borrower performance deterioration played a smaller secondary role.

Charge-off by Quarter adjusted

When analyzing loan performance, it can be difficult to determine what the underlying drivers are for any changes you see. Many factors—interest rates, FICO scores, loan term, loan age, DTI, and more—can have strong impacts on the performance of pools of loans. Perhaps more importantly, these factors are all interrelated, and changes in one variable are generally related to changes in another. We believe that it is important to perform deep analysis like this on the underlying drivers of performance, as initial impressions can sometimes be very misleading. Finally, we would like to note that while this post focused on the impact of loan age on charge-off rates, a very similar analysis could assess the impact on prepayment rates, which display a similar relationship to loan age as do charge-off rates.