Author Topic: 36 months vs 60 months  (Read 19321 times)

Fred93

  • Hero Member
  • *****
  • Posts: 2245
    • View Profile
Re: 36 months vs 60 months
« Reply #15 on: November 09, 2015, 02:22:30 PM »
The longer loans earn a higher interest rate, but have more defaults.

Net, they appear to earn a higher return, HOWEVER, we only know for certain during GOOD TIMES.  The 60 month loans weren't around during the 2008/9 recession.  I suspect that the longer loans will do more poorly during the next bad times.  It seems likely that longer loans are just another form of risk.  It is important to understand that the historical record of "all loans so far" is biased because the longer loans were added after the big recession.

AnilG

  • Hero Member
  • *****
  • Posts: 1123
    • View Profile
    • PeerCube
Re: 36 months vs 60 months
« Reply #16 on: November 09, 2015, 03:13:27 PM »
The 60 month loans might have lower net return than 36 month loans for same vintage. The 60 month loans were first issued on Lending Club in May 2010 so the first vintage has just started to mature in 2015. It might not be unique to Lending Club as 60 month loans on Prosper also show similar results.

Lending Club Loan Performance from https://www.peercube.com/histperf/perfbyattr/term

VintageLoan LengthOrigination PeriodLoan CountInterest RateLoss RateNAR
20103601/2010 - 12/20109,15611.31%6.62%5.70%
20106005/2010 - 12/20103,38113.83%14.58%4.49%

Prosper Loan Performance from https://www.peercube.com/histperf/prosperperfbyattr/Term

VintageLoan LengthOrigination PeriodLoan CountBorrower RateLoss RateNAR
20113601/2011 - 12/20119,65123.61%14.14%14.38%
20116001/2011 - 12/20111,03020.16%18.67%11.08%
« Last Edit: November 09, 2015, 03:22:15 PM by AnilG »
---
Anil Gupta
PeerCube Thoughts blog https://www.peercube.com/blog
PeerCube https://www.peercube.com

Rob L

  • Hero Member
  • *****
  • Posts: 2133
    • View Profile
Re: 36 months vs 60 months
« Reply #17 on: November 09, 2015, 04:32:46 PM »
I can't help but compare LC loans to bonds in general. The longer the term the higher the interest rate risk. For the sake of argument lets say sometime in the future LC increases interest rates. Investors with 60 month loans are stuck at the lower rate while those with 36 month loans are able to invest at the higher rates sooner. If you think interest rates are going up then the 36 month loans are preferable. If you think they are going down then 60 month loans are preferable. If you think you can forecast interest rates you are probably wrong (though the Fed's muddy boots all over the bond market have made this a more tractable problem).

There's also the liquidity factor to consider. With 36 month loans everything is over in 3 years; good, bad and ugly. All loans turn to cash by that point. With 60 month loans it takes 5 years for all loans to turn to cash. This may trump a few basis points of NAR given an individual's particular circumstances. Meanwhile Folio is neither a deep nor liquid market. Someday, hopefully sooner rather than later, this will no longer be true. I look forward to the day when there is a robust, liquid and deep secondary market for LC notes with continuous bid/ask quotes, sizes, and narrow spreads. LC could go a long way towards making this happen with an API and historical data to encourage it. If not LC then someone else will do this and IMHO gain a significant competitive advantage. It is something that needs to happen.

Meanwhile, considering my personal situation, I'm sticking with 36 month loans. Three years and done is important to me.

RaymondG

  • Full Member
  • ***
  • Posts: 247
    • View Profile
    • Email
Re: 36 months vs 60 months
« Reply #18 on: November 10, 2015, 12:59:26 AM »
Lending Club Loan Performance from https://www.peercube.com/histperf/perfbyattr/term

VintageLoan LengthOrigination PeriodLoan CountInterest RateLoss RateNAR
20103601/2010 - 12/20109,15611.31%6.62%5.70%
20106005/2010 - 12/20103,38113.83%14.58%4.49%

The results from NSRPlatForm is different. https://www.nsrplatform.com/#!/stats/lendingclub
VintageLoan LengthOrigination PeriodLoan CountInterest RateLoss RateROI
20103601/2010 - 12/20109,15611.68%4.99%5.87%
20106005/2010 - 12/20103,38114.11%7.10%6.40%


Fred

  • Hero Member
  • *****
  • Posts: 1421
    • View Profile
Re: 36 months vs 60 months
« Reply #19 on: November 10, 2015, 03:56:19 AM »

AnilG

  • Hero Member
  • *****
  • Posts: 1123
    • View Profile
    • PeerCube
Re: 36 months vs 60 months
« Reply #20 on: November 10, 2015, 12:59:43 PM »
I don't know how NSR defines and estimates Loss Rate. Reviewing the chart posted by Fred and loss rate numbers published by Lending Club, I will guess NSR might be under-estimating the loss rate. The Lending Club Cumulative CO Rates spreadsheet provides lot more information including split by credit grade https://www.dropbox.com/s/0l5tijquoma4k3y/Cumulative%20CO%20Rates%20Projected%20Loss%20Curves%20v15_OCT2015.xlsx?dl=0,

VintageTermLoss Rate
2010Q1366.37%
2010Q2366.48%
2010Q26015.68%
2010Q3365.35%
2010Q36013.47%
2010Q4365.54%
2010Q46012.58%



Lending Club Loan Performance from https://www.peercube.com/histperf/perfbyattr/term

VintageLoan LengthOrigination PeriodLoan CountInterest RateLoss RateNAR
20103601/2010 - 12/20109,15611.31%6.62%5.70%
20106005/2010 - 12/20103,38113.83%14.58%4.49%

The results from NSRPlatForm is different. https://www.nsrplatform.com/#!/stats/lendingclub
VintageLoan LengthOrigination PeriodLoan CountInterest RateLoss RateROI
20103601/2010 - 12/20109,15611.68%4.99%5.87%
20106005/2010 - 12/20103,38114.11%7.10%6.40%
« Last Edit: November 10, 2015, 01:02:04 PM by AnilG »
---
Anil Gupta
PeerCube Thoughts blog https://www.peercube.com/blog
PeerCube https://www.peercube.com

RaymondG

  • Full Member
  • ***
  • Posts: 247
    • View Profile
    • Email
Re: 36 months vs 60 months
« Reply #21 on: November 10, 2015, 09:41:19 PM »
The Lending Club Cumulative CO Rates spreadsheet provides lot more information including split by credit grade https://www.dropbox.com/s/0l5tijquoma4k3y/Cumulative%20CO%20Rates%20Projected%20Loss%20Curves%20v15_OCT2015.xlsx?dl=0,

Thank you very much for the analysis. Here are some questions regarding to the chart/data that split by credit grade.
1. How was the loss rate of each month calculated? Is it [$ charge-offs in month-x] / [$ principal in BOM], or [$ charge-offs in month-x] / [$ total invested in the beginning]
2. Total Loss rate might not be the simple sum of loss rates in each months, depending on how they were calculated. If they were equal to [$ charge-offs in month-x] / [$ principal in BOM], these loss rates of each month have different basis because of decreasing principal, charge-offs, and prepayments.
« Last Edit: November 10, 2015, 09:43:11 PM by RaymondG »

AnilG

  • Hero Member
  • *****
  • Posts: 1123
    • View Profile
    • PeerCube
Re: 36 months vs 60 months
« Reply #22 on: November 11, 2015, 12:40:41 PM »
It wasn't my analysis. The spreadsheet is published by Lending Club. The Glossary and Disclaimers tab describe more on the calculations performed. The Data tab gives raw data.

Gross Credit Loss   Total Dollar Loss/Original Loan Amount
Net Credit Loss   Net Dollar Loss/Original Loan Amount, accounts for recoveries
Cumulative Net-Charge Off Rate   Equals the gross dollars charged off divided by the original funded amount

Cumulative Net Charge-Off Rates
Cumulative net charge off rates are calculated by dividing the gross dollars charged off by the total original loan amount funded.  Historical charge off rates are not intended to predict future charge off rates.  Actual charge off rates experienced by any individual portfolio may be impacted by, among other things, the size and diversity of the portfolio, its exposure to particular Notes or groups of Notes, as well as macroeconomic conditions.

The Lending Club Cumulative CO Rates spreadsheet provides lot more information including split by credit grade https://www.dropbox.com/s/0l5tijquoma4k3y/Cumulative%20CO%20Rates%20Projected%20Loss%20Curves%20v15_OCT2015.xlsx?dl=0,

Thank you very much for the analysis. Here are some questions regarding to the chart/data that split by credit grade.
1. How was the loss rate of each month calculated? Is it [$ charge-offs in month-x] / [$ principal in BOM], or [$ charge-offs in month-x] / [$ total invested in the beginning]
2. Total Loss rate might not be the simple sum of loss rates in each months, depending on how they were calculated. If they were equal to [$ charge-offs in month-x] / [$ principal in BOM], these loss rates of each month have different basis because of decreasing principal, charge-offs, and prepayments.
« Last Edit: November 11, 2015, 12:43:11 PM by AnilG »
---
Anil Gupta
PeerCube Thoughts blog https://www.peercube.com/blog
PeerCube https://www.peercube.com

RaymondG

  • Full Member
  • ***
  • Posts: 247
    • View Profile
    • Email
Re: 36 months vs 60 months
« Reply #23 on: November 11, 2015, 11:12:29 PM »
@AnilG - My bad. I should have read the Glossary and Disclaimers tab. Thanks.

Back to the Loss Rates. I wonder if the Loss Rate 14.58% for 60Month loans of 2010 Vintage is cumulative loss rate over 60 month. If that's the case, then average over 5 year is about 3% per year. Or, average over 36 months when most losses are expected, the averaged loss rate is 4.86%. So estimating the net return in a simplified way, in the first 3 years,

36 months : 11.31% - 6.62%/3 -1% = 8.1%
60 months : 13.83% - 14.58%/3 -1% = 8%  - but remember, it assumed remaining principal would suffer little loss in the following 2 years

Look at the cumulative charge-off tables in The Lending Club Cumulative CO Rates spreadsheet in attachment. 60-month has materially larger Gross Credit Loss (cumulative loss rate), for example 15.19% for D, but its Annualized Credit Loss is better. The overall Annualized Credit Loss of 60-month loans looks close to 4%.

Anyway, the point is the loss of 60-month loans is not that scary like what my first impression was when saw the table below.


Lending Club Loan Performance from https://www.peercube.com/histperf/perfbyattr/term

VintageLoan LengthOrigination PeriodLoan CountInterest RateLoss RateNAR
20103601/2010 - 12/20109,15611.31%6.62%5.70%
20106005/2010 - 12/20103,38113.83%14.58%4.49%
« Last Edit: November 11, 2015, 11:36:34 PM by RaymondG »

jennrod12

  • Full Member
  • ***
  • Posts: 134
    • View Profile
    • Email
Re: 36 months vs 60 months
« Reply #24 on: November 12, 2015, 02:02:47 PM »
For comparison purposes, you could take an example of 15 years and figure out the cumulative loss of three 60-month loads vs. the cumulative loss of five 36-month loans. 

Of course this would be hypothetical, since a lot could change over 15 years, but I was just thinking of a simple way to compare apples to apples without having to guestimate the differences between performance of early and late months of the loan.

If I got a 36-month loan instead of a 60-month, I'd have to do something with my money in those last 24-months, which would probably be to invest in another 36-month loan....  So do I have more exposure to loss with the 36-month loans in that scenario?

Jenn

AnilG

  • Hero Member
  • *****
  • Posts: 1123
    • View Profile
    • PeerCube
Re: 36 months vs 60 months
« Reply #25 on: November 12, 2015, 05:23:12 PM »
Jenn,

You have the right approach to comparing two investments of different length. How to compare two investments of different length is of considerable interest in academic research. There is no consensus on appropriate method of comparison. One way to do comparison is to assume that you re-invested in 36 month loans for another 24 months (2 years) after the first 36 month term and then compare the ending portfolio value with same investment in 60 month loans. This method assumes that you are able to reinvest at same interest rate after 36 months.

The Annualized return method is not an appropriate method to compare two investments when there is variation in loss rate and return from one year to next. In such situations, the annualized method will give you incorrect value for your portfolio value at the end of specific investment horizon. It is the same reason the prospectus from mutual funds and other investments show the value of hypothetical $10,000 investment over time when comparing their performance with benchmark.

For comparison purposes, you could take an example of 15 years and figure out the cumulative loss of three 60-month loads vs. the cumulative loss of five 36-month loans. 

Of course this would be hypothetical, since a lot could change over 15 years, but I was just thinking of a simple way to compare apples to apples without having to guestimate the differences between performance of early and late months of the loan.

If I got a 36-month loan instead of a 60-month, I'd have to do something with my money in those last 24-months, which would probably be to invest in another 36-month loan....  So do I have more exposure to loss with the 36-month loans in that scenario?

Jenn
---
Anil Gupta
PeerCube Thoughts blog https://www.peercube.com/blog
PeerCube https://www.peercube.com

Lovinglifestyle

  • Hero Member
  • *****
  • Posts: 902
    • View Profile
    • Email
Re: 36 months vs 60 months
« Reply #26 on: November 12, 2015, 05:56:31 PM »
Ha.  And the LC Prospectus is where, again??   :o

Jenn,

You have the right approach to comparing two investments of different length. How to compare two investments of different length is of considerable interest in academic research. There is no consensus on appropriate method of comparison. One way to do comparison is to assume that you re-invested in 36 month loans for another 24 months (2 years) after the first 36 month term and then compare the ending portfolio value with same investment in 60 month loans. This method assumes that you are able to reinvest at same interest rate after 36 months.

The Annualized return method is not an appropriate method to compare two investments when there is variation in loss rate and return from one year to next. In such situations, the annualized method will give you incorrect value for your portfolio value at the end of specific investment horizon. It is the same reason the prospectus from mutual funds and other investments show the value of hypothetical $10,000 investment over time when comparing their performance with benchmark.

For comparison purposes, you could take an example of 15 years and figure out the cumulative loss of three 60-month loads vs. the cumulative loss of five 36-month loans. 

Of course this would be hypothetical, since a lot could change over 15 years, but I was just thinking of a simple way to compare apples to apples without having to guestimate the differences between performance of early and late months of the loan.

If I got a 36-month loan instead of a 60-month, I'd have to do something with my money in those last 24-months, which would probably be to invest in another 36-month loan....  So do I have more exposure to loss with the 36-month loans in that scenario?

Jenn

RaymondG

  • Full Member
  • ***
  • Posts: 247
    • View Profile
    • Email
Re: 36 months vs 60 months
« Reply #27 on: November 16, 2015, 12:28:25 AM »
I have completed an analysis in a setup that assumes the account has been invested for long enough time. In this perfect setup, the investor will purchase fix amount new notes and withdraw fix amount cash every month, and so the total investment will always be the same over months. Please see the chart for comparison between purchasing 36month and 60month. The Loss distribution is assumed to be:

The Lending Club Cumulative CO Rates spreadsheet provides lot more information including split by credit grade https://www.dropbox.com/s/0l5tijquoma4k3y/Cumulative%20CO%20Rates%20Projected%20Loss%20Curves%20v15_OCT2015.xlsx?dl=0,

The worksheet is shared in my google drive:
https://drive.google.com/file/d/0B0PqZljirBQDNmVla1NmQkhkZFk/view?usp=sharing
« Last Edit: November 16, 2015, 12:30:34 AM by RaymondG »

hoggy1

  • Sr. Member
  • ****
  • Posts: 401
    • View Profile
    • Email
Re: 36 months vs 60 months
« Reply #28 on: November 17, 2015, 04:02:23 PM »
I don't know how NSR defines and estimates Loss Rate. Reviewing the chart posted by Fred and loss rate numbers published by Lending Club, I will guess NSR might be under-estimating the loss rate. The Lending Club Cumulative CO Rates spreadsheet provides lot more information including split by credit grade https://www.dropbox.com/s/0l5tijquoma4k3y/Cumulative%20CO%20Rates%20Projected%20Loss%20Curves%20v15_OCT2015.xlsx?dl=0,


Hi Anil,

Where did you get this file? Is it in your dropbox acct or LC's? How did you find it? Does LC publish someplace an index of public data? Or did you ask LC to provide the data?

This is a very important file. It uses a pivotTable so you can not only look at the breakout by term but by Grade as well. Been wanting this data for a long time.

Steve

Steve
« Last Edit: November 18, 2015, 05:00:10 AM by hoggy1 »
Steve

PhilGD

  • Full Member
  • ***
  • Posts: 150
    • View Profile
    • Email
Re: 36 months vs 60 months
« Reply #29 on: November 17, 2015, 04:59:11 PM »
I don't know how NSR defines and estimates Loss Rate. Reviewing the chart posted by Fred and loss rate numbers published by Lending Club, I will guess NSR might be under-estimating the loss rate. The Lending Club Cumulative CO Rates spreadsheet provides lot more information including split by credit grade https://www.dropbox.com/s/0l5tijquoma4k3y/Cumulative%20CO%20Rates%20Projected%20Loss%20Curves%20v15_OCT2015.xlsx?dl=0,


Hi Anal,

Where did you get this file? Is it in your dropbox acct or LC's? How did you find it? Does LC publish someplace an index of public data? Or did you ask LC to provide the data?

This is a very important file. It uses a pivotTable so you can not only look at the breakout by term but by Grade as well. Be wanting this data for a long time.

Steve

Steve


LC doesn't advertise its location, but you can find this file and more published monthly:

http://additionalstatistics.lendingclub.com/