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


  • Investor Portal

  • Originator Database

  • Manager Database

If you forgot your login email, please contact your Orchard Account Manager or

Curve Analysis of Lending Club 2015 and 2016 Vintages


In our monthly and quarterly reports, we frequently comment on the observed underperformance of LendingClub’s 2015 and 2016 vintages when compared to earlier vintages—particularly the 2013 and 2014 vintages which have offered both higher returns and lower charge-off rates over their first 12 to 24 months on book (MOB). In this analysis, we take a closer look at this trend and offer some possible reasons for the apparent underperformance of more recent vintages.

Rising Default Rates in Recent Vintages

Let’s start by looking at the Conditional Default Rate (CDR) and Charge-off vintage curves for 36-month and 60-month term loans from 2010 to 2016.


Granted it’s still early days for the 2016 vintage, but the first thing we note is that the 2016 vintage is defaulting at a higher rate than all previous vintages for 36-month loans.

36-month CDR_vintage

36-month Charge-off Vintage Curves

2016 is also trending higher than the 2015 vintage for 60-month loans, and the 2015 vintage is currently experiencing a higher rate of defaults than all other vintages at the same MOB.

60-month CDR Curves_vintage

60-month Charge-off Vintage Curves

One important thing to note is that higher CDRs are not, in and of themselves, necessarily bad. Higher CDRs accompanied by similarly higher interest rates can still result in an acceptable rate of return for the given level of risk. However, in this instance, from 2013 through the first half of 2016 we’ve seen falling interest rates at LendingClub. Anecdotally, we attribute this to two primary, interrelated factors: increasing competition and a need to continually attract borrowers to meet investor demand. So, while rates haven’t fallen drastically when accompanied by increasing defaults, it is putting significant pressure on returns for the 2015 and 2016 vintages.

Weighted Average Interest Rate by Month

It’s also important not to jump to conclusions. One might assume the increase in defaults is a result of LendingClub originating more loans to lower quality borrowers, or a loosening of credit standards to meet investor demand. But when we look broadly at their weighted average FICO scores we don’t see that (below). In fact, 2015 FICO scores are comparable to what we see in 2013 and 2014. There was a larger drop from 2011 into 2013, around 15 points but, using FICO as a proxy for credit quality, we just don’t see a significant decline in 2015.

Weighted Average FICO

While not the primary focus of this post, we’ve also been looking into delinquency rates reported by commercial banks and published by the Federal Reserve.


There’s clearly been an uptick in consumer loan and credit card delinquencies over the 2015 to 2016 period.

All of this is to say that the increase in delinquencies is not unique to LendingClub loans. It appears to be a broader trend across consumer credit.


Net Annualized Return (NAR) Grade Curves

Next, we use our Net Annualized Return Grade Curves to look at loan performance in more detail. These curves show the cumulative return-to-date in a given month, across loan grades, and broken out by Vintage and Term. Month-zero returns are stratified by Grade with the highest risk, lowest grade loans earning the highest rate of return. Over time, the performance converges to a tight range. Everyone knows this pattern. The higher the risk, the higher the expected return. But what is interesting, and what this type of chart is great at illustrating, is that over the life of a loan—and as loans start to charge-off—the realized return falls.

NAR Curves_vintage

If a lender does a good job pricing the loans—if things proceed as expected—you end up with a graph that looks something like the 2013 60-month graph. In this case, the G grade loans might end up yielding 3-5% higher than the A grade loans, and there is a clear separation between grades with very little crossover of one group over another.

However, if you look closely at the 60-month graph for 2014, you can see that G grade loans crossover all other grades. This is an indication that the charge-offs are higher than expected for that grade. This is most noticeable in 2015. Both 36- and 60-month, F and G grades are significantly underperforming. For 2015 36-month loans, after about 20 months, F and G grades are already in negative territory regarding realized returns to date. And we expect they’ll likely continue to fall slightly from there.

Most notably, in the 36-month 2016 vintage loans—after only around eight months of seasoning—the G grade loans have already fallen to be the worst performing grade in that vintage, also on a path towards negative returns.


Charge-off and Prepayment Grade Curves

Our Charge-off Grade Curves visualize the cumulative Charge-off Rate across loan grades, broken out by Vintage and Term. Similar to our return analysis, if things proceed as expected, there will be clear stratification between loan grades as charge-offs increase over time. In this series of charts we are primarily concerned with two features; how high the charge-off rate is and how quickly a vintage reaches that rate.

For example, the 2014 vintage, F and G grade 36-month loans are charging off well below 30% (~22%-24%). Comparatively, 2015 vintage F and G grade 36-month loans are already charging off near or above 30%, with 12 months less seasoning than the 2014 vintage. The 2016 vintage is still young but appears to be tracking 2015. 60-month loans display similar characteristics, although the 2016 vintage does appear to be charging off at a slightly lower rate than 2015.

Charge-off Grade Curves_Vintage

Prepayment Grade Curves visualize the rate that borrowers prepay loans across loan grades, and are broken out by Vintage and Term. If things proceed as expected, there will be clear stratification between loan grades, but they will be tightly packed. In this series of charts, we are also concerned with how high the prepayment rate is and how quickly a vintage reaches that rate.

Prepayment Grade Curves_vintage

In this series, you can see that before 2015, LendingClub loan prepayment curves have been relatively similar across vintages (peaking around 30%). Beginning with 2015 the curve appears to flatten sooner and at a slightly lower rate that previous vintages. This is most noticeable in the 2016 vintage of both 36- and 60-month loans. The curves of all grades of both of these loan types appear to have flattened out at 20 MOBs just under 20%. In normal market conditions, slower prepayment speeds would be a positive for loan performance (higher interest yield), but in this case, the lower prepayment rate corresponds to higher charge-off rates for 2014-2016 having a severe effect on returns as shown in the NAR charts.


Taken together, this set of charts actually may say more about the health of consumer credit in general, rather than anything specific to LendingClub. The trends of increasing charge-offs/defaults combined with slowing prepayments support a broader narrative of a potential deterioration in consumer credit and fits with reports about traditional lenders who have been boosting reserves against potential losses in anticipation of a turn in the credit cycle.

For LendingClub’s 2015 and 2016 vintages, lower interest rates, driven by high investor demand for these loans, accompanied by a worsening of the broader credit markets has led to a particularly bad return profile for these vintages. However, so far for 2017 we see an increase in interest rates and average FICO score.


2017 Rate_FICO

While the change in FICO is not particularly meaningful, it is good to see rising interest rates that aren’t driven by lowering quality of loans.


  • Is it possible to ask you to run your analysis with more recent data?