Lend Academy Network Forum
Lending Club Discussion => Foliofn  LC => Topic started by: wlbsr on February 04, 2016, 01:43:33 AM

looking at IRR as a comparison to folio ytm. I am using the following as the stream of cash flows for the IRR...
[ask+accrued int] as the first amount and then each of the remaining payments less the 1% fee.
i am getting a fairly consistent .3 to.55 higher than the folio calculated ytm which i am thinking may have something to do with the fact that I am taking into account accrued interest in the first amount.... not sure if folio is?
my question is that every now and then i will come across a note where my calculated IRR is 1.5 to 2.0 hight than the folio ytm. when i look at these notes, i cant find anything out of the ordinary from the others that mostly seem in line with my calculations. any ideas? here is an example.
my IRR uses [66+.37] as the first cash flow and then 23 payments of [3.16.03] as the remaining
LoanId Par AskPrice AccInt Class Mat N Rate Bench IRR YTM Spread Date Pmt
581 38528044 100 66.00 0.37 6 36 23 8.67 8.39 9.42 7.72 1.03 20160127 3.16
https://www.lendingclub.com/foliofn/loanPerf.action?loan_id=38528044&order_id=52420245¬e_id=67624504 (https://www.lendingclub.com/foliofn/loanPerf.action?loan_id=38528044&order_id=52420245¬e_id=67624504)
what am i missing since my IRR is 9.42 while folio ytm is 7.72?

looking at IRR as a comparison to folio ytm. I am using the following as the stream of cash flows for the IRR...
[ask+accrued int] as the first amount and then each of the remaining payments less the 1% fee.
Accrued interest is NOT a cash flow. You are counting those dollars twice. Once when you accrue them, and once when the payment arrives. The cash flows when the payment is made.

Thanks  I think I inadvertently opened up two questions / issues.
1 as far as accrued interest, my logic was that since the payment is not prorated between buyer and seller for a purchase in the middle of a payment cycle and since the buyer gets the whole payment then the portion of the interest that accrued before the purchase seems like it would be an additional cash flow to the buyer. I was thinking in bonds prices are quoted without respect to accrued interest since its a moving target as interest accrues daily and the accrued interest was automatically added to the trade price at settlement. In bonds it called clean and dirty prices although i am not sure which is which. is LC/folio adding accrued interest at settlement to compensate seller or does seller need to include in their price? If accrues interest is not automatically compensated at settlement then it would seem that the accrued interest benefits the buyer and would be reflected in the ytm.
2  the accrued interest is only responsible for 40 or 50 bps in my IRR calculation so my original question / observation above was that in the note above the LC ytm and my IRR calculation are off much more than the variance cause by accrued interest. I only see this on say 3 or 4 per 1000 notes but curious if anyone can see any reason for this on the above note. thx. Bill

FRED: should have thought about it for 5 more minutes before replying. On second though I understand that cash flows are negative for purchase prices and then each of the remaining payments which are positive. accrued interest has nothing to do with a cash flow although it is being implicitly included in my return assuming LC is not adjusting back to the buyer at settlement.
Is LC adjusting at settlement? or is that on the sellers dime.
question 2 still stands.

I get 8.4% YTM. Please see the snapshots. Beginning investment is AskPrice; Starting date is 2/3/2016 when YTM was calculated (I assume YTM on folio is calculated on fly). The principla balance $65.98 on 1/12/16 is used as starting principal to calculated Interest and Principal portion of the future payments. Note, the last payment is smaller than $3.16.

thanks RaymondG  my code is assuming equal cash flow intervals which asnwers why some were in line with LC and others were not depending on the current date vs. the loan payment cycle dates. to what would you attribute the difference in XIRR at 8.4 and LC ytm at 7.72?

to what would you attribute the difference in XIRR at 8.4 and LC ytm at 7.72?
Sorry, I do not know.

Is LC adjusting at settlement? or is that on the sellers dime.
There is some information on the FAQs on the folio web pages.
When the monthly payment from the Borrower is disbursed to investors, accrued interest is paid to the holder of the Note. The person who holds the Note at the time payment is disbursed by Lending Club will receive their portion of the proceeds of that payment.
Having said that, don't think I have all the answers on this one. I tried some time ago to figure out exactly what calculation folio was doing to come up with their ytm, and I gave up. I look forward to you figuring it all out.

Anyone figured this out yet? Or, got explanations from Folio?

Anyone figured this out yet? Or, got explanations from Folio?
It's never made sense to me. It is easy to explain why it is wrong. List a loan at a 0% mark up. By definition, the yield should equal the rate on the loan. It doesn't.
That it one of the many things that pisses me off about Folio. I very rarely make use of it because of their crappiness.

partial answer opens up new questions
OK... a little more info. I find that some of the notes where my calculated IRR is higher than folio ytm is due to the payment being made is different from the original amortization. see this note or underlying loan:
https://www.lendingclub.com/foliofn/browseNotesLoanPerf.action?showfoliofn=true&loan_id=28272258&order_id=38251968¬e_id=58074749 (https://www.lendingclub.com/foliofn/browseNotesLoanPerf.action?showfoliofn=true&loan_id=28272258&order_id=38251968¬e_id=58074749)
the original amortized payment was 7.06 which is what I use to calculate my IRR (Yield). in Sept of 2015 the borrower started paying 10.14 per month which is the figure (along with a shorter number of payments) that folio uses to calculate its ytm.
I am not sure how or why the terms of the loan were changed in Sep. of 2015. My gut tells me to avoid loans where the payment amount has increased. In this case never late = True, there was one IGP payment which was after the payment amount was increased, and the fico score seems to have been constant to a slight increase during the life of the loan. I talked to LC and did not get an answer although they did promise to get back to me.

LC also told me that folio controlled the information available on the note trading platform, gave me their number, and said that they could possibly give me some insight into what info was displayed and not displayed on their platform. When I called folio they said that LC used a shell of their platform of which they had very little knowledge or control and referred me back to LC for specifics.
I had also planned to ask folio if they had any idea why my api requests had been pulling the exact same number of notes since Jan 27 but didn't get that far.

LC also told me that folio controlled the information available on the note trading platform, gave me their number, and said that they could possibly give me some insight into what info was displayed and not displayed on their platform. When I called folio they said that LC used a shell of their platform of which they had very little knowledge or control and referred me back to LC for specifics.
Chimp & pony show.

I'm very late to this party so my questions are of the noob category.
Accrued interest all goes to the buyer so it is up to the seller to incorporate that into their ask price.
When computing ytm there is no reason given the ask price, number of remaining payments, and existing payment amount to consider accrued interest at all. Correct?
Remember noob. You have to crawl before you walk ... :)

When computing ytm there is no reason given the ask price, number of remaining payments, and existing payment amount to consider accrued interest at all. Correct?
Correct.
Once you purchased a note from Folio, all that matter are the subsequent payments. If you sell the notes again, then the sale price and time also matter.

I have a unique and yes even perhaps bizarre way of computing ytm simply as an educational exercise (not a useful tool).
It only uses an online loan amortization schedule calculator such as:
http://www.bankrate.com/calculators/mortgages/amortizationcalculator.aspx (http://www.bankrate.com/calculators/mortgages/amortizationcalculator.aspx)
There is no "closed form" equation that will yield ytm given the set of parameters. Otherwise why bother.
So, again for educational purposes, enter the Folio asking amount (in pennies) as the loan amount (pennies needed for precision like 2500 for $25).
Enter the remaining number of months as the term.
Make a guess at the interest rate xx.xx% (when you finish this will be ytm).
Hit compute and the monthly payment will be computed and displayed.
Refine your interest rate guess so that eventually the monthly payment computed equals the existing monthly payment for the original loan.
Finally when you guess the interest rate that makes the new payment the same as the original then that interest rate is ytm.
Again, for the education of a complete noob, I ran this procedure for two loans.
One was pretty much dead on the same as the Folio provided and the other significantly higher.
Given the educational nature of the exercise is my procedure correct (meaning one of the Folio displayed ytm's is off badly) or is this my approach conceptually a bag of rubbish? Naturally I'd be curious as to what is wrong with the procedure of course.
TIA.

http://www.bankrate.com/calculators/mortgages/amortizationcalculator.aspx
The link you used implies equallyspaced payments (1month apart). When you purchase notes from Folio, this is not always the case.
If the note's payment is due , say, on 15th of the month, the YTM would vary if your purchased on the 14th (1 day prior to payment) vs. on the 16th (29 days before next payment).
I think RaymondG's XIRR Excel from previous page would give more accurate YTM. It gives you precise control on the time and amount of the payment stream.

The link you used implies equallyspaced payments (1month apart). When you purchase notes from Folio, this is not always the case.
If the note's payment is due , say, on 15th of the month, the YTM would vary if your purchased on the 14th (1 day prior to payment) vs. on the 16th (29 days before next payment).
I think RaymondG's XIRR Excel from previous page would give more accurate YTM. It gives you precise control on the time and amount of the payment stream.
Yes, I see. Time value of money. If you purchase 1 day before payment you get the entire month's interest but have not had the principal invested the entire month. XIRR allows one to account for just such examples. Guess it is up to the seller to incorporate accrued interest into the ask price every day to take this into account. I'll play around with RaymonG's XIRR method to get a feel for the process. Thanks to both!

I think RaymondG's XIRR Excel from previous page would give more accurate YTM. It gives you precise control on the time and amount of the payment stream.
Well, I played around with RaymondG's example and came up with possibly a third different YTM so here goes:
(https://forum.lendacademy.com/proxy.php?request=http%3A%2F%2Fi.imgur.com%2F3DGm0ce.png&hash=2360e96b3acd4fab7952e3ad124b3966)
Rows 2 through 8 are the givens. It doesn't matter if I buy the note or not the number of remaining payments, the monthly payment amount and the remaining payment due dates stay the same. So, in XIRR terms, the cash flow (payments net of fees) for sale is set (well actually very dependent only upon the borrower's future actions but that's a different problem).
What is the YTM (XIRR) of this cash flow if I buy it on 2/3/2016 and pay $66.00 for it? The Excel equation and result (9.95%) are on line 10.
It only differs from RaymondG's 8.40% result in that in his computation the last payment was calculated to be less than the others.
If that's true I clearly don't understand it and need a little help.
As a sanity check I computed in row 14 the interest rate as if it were a new loan with principal equal to the price I paid and 23 monthly payments where the payment amount is exactly the same as the original loan. Figured this should be close to the XIRR but a bit higher rate since this does not penalize me like XIRR does for the time value of money from 2/3/2016 and 2/12/2016. Line 13 annualizes the row 14 interest rate and is 9.98% and is close to the calculated XIRR. Seems to check out okay.
So, how bad did I mess this up? :o

It only differs from RaymondG's 8.40% result in that in his computation the last payment was calculated to be less than the others.
If that's true I clearly don't understand it and need a little help.
Please see notes in the attached files

I think RaymondG's XIRR Excel from previous page would give more accurate YTM. It gives you precise control on the time and amount of the payment stream.
Well, I played around with RaymondG's example and came up with possibly a third different YTM so here goes:
Nope. You did this funny thing where you computed a "rate" from the payment, and then attempt to apply this interest rate. This calculation is wrong. The interest rate was set when the loan was made. You can't just compute a different interest rate later. The borrower would certainly object!
My calculation matches the RaymondG calculation. I get 8.40%

If that's true I clearly don't understand it and need a little help.
Please see notes in the attached files
Thanks! I can now see the error of my ways. Couldn't figure out where your $65.98 came from. My mistake was to "assume" the note payments has been made according to the original loan amortization schedule. Looking back at the Folio listing I see that the borrower doubled up on the first payment. The principal amount owed on 1/12/16 was not $66.84 as it would have been if the borrower had simply paid per schedule on time, but the $65.98 you used (as shown in the Folio payment history table). Now I have at least a shot at getting the correct answer.
Nope. You did this funny thing where you computed a "rate" from the payment, and then attempt to apply this interest rate. This calculation is wrong. The interest rate was set when the loan was made. You can't just compute a different interest rate later. The borrower would certainly object!
My calculation matches the RaymondG calculation. I get 8.40%
The funny thing (lines 13, 14 and 15) played no part in my XIRR calculation. I did not mean to imply that the initial loan parameters were in any way magically changed. Just added it as a reasonableness test of the XIRR computation. Playing around to get a better understanding of what to expect. Of course if you start out with bad initial conditions, as I did, then the wrong answer will appear to be reasonable.
Good to know the correct answer is 8.40% so I can fix my error. Lesson learned.
Thanks to RaymondG and Fred93.

LoanId Par AskPrice AccInt Class Mat N Rate Bench IRR YTM Spread Date Pmt
581 38528044 100 66.00 0.37 6 36 23 8.67 8.39 9.42 7.72 1.03 20160127 3.16
https://www.lendingclub.com/foliofn/loanPerf.action?loan_id=38528044&order_id=52420245¬e_id=67624504 (https://www.lendingclub.com/foliofn/loanPerf.action?loan_id=38528044&order_id=52420245¬e_id=67624504)
Seems ironic that this "example" loan was prepaid in full on 2/25/2016 and with a hefty service fee of $0.67.
The issue date was 1/14/2015 so it was only a few weeks beyond the 12 months LC limits its fees for loans paid off early.
If I computed XIRR correctly on the completed loan :\ it was 7.95%. Without the service fee it would have been 8.70%.
These numbers presume the loan was not sold on Folio but continued to be held by its original owner.

wow. how did you catch that?
i started buying notes on folio in January, have purchased 42 so far and 2 of the 42 have been fully repaid. don't think this was one of them though.

wow. how did you catch that?
Quite by accident. I was just using the loan you started the thread with to learn how to calculate ytm.
Had to go back and look at the Folio listing to figure things out and there it was; all paid off.
Glad you didn't get left holding the bag with the quick pay off.

I have had 2 payoffs out of 42 loans in my first 2 months. whats the downside of an early payoff?
I wonder if this particular note being paid off was a coincidence. My original flag was based on an irregular ytm because of persistent payments at a level above the original amortized amount.

I have had 2 payoffs out of 42 loans in my first 2 months. whats the downside of an early payoff?
The downside of early payoff is that you have to invest again, and this reinvestment involves a DELAY. The money doesn't earn anything during this delay. There are several components of this delay, and most of them are out of your control.

I have had 2 payoffs out of 42 loans in my first 2 months. whats the downside of an early payoff?
If you purchased a nonCurrent note (i.e., In Grace, Late16, Late31), early payoff is very desirable. The earlier, the better!

I have had 2 payoffs out of 42 loans in my first 2 months. whats the downside of an early payoff?
The downside of early payoff is that you have to invest again, and this reinvestment involves a DELAY. The money doesn't earn anything during this delay. There are several components of this delay, and most of them are out of your control.
After writing this I realized I was thinking like an investor who purchases loans at the beginning.
If you buy on Folio, sometimes you buy below par. If you buy below par, and the payoff comes early, you win a higher return. (In this case yieldtoearlypayoff is higher than yieldtomaturity.) Conversely, if you buy above par, then early payoff produces a lower yield.
In all cases, reinvestment involves delay, which reduces your portfolio yield, because money is idle for a time while being recycled.

on folio, the reinvestment lag is not a big deal, probably worth the prepay bump if you are buying at a discount.
Any thoughts on maturity while we are at it? would a 36 mo loan with 24 months left be more desirable than a a new 36 mo loan at an equal ytm assuming no credit status degradation in the interim on the older note?
How about a 60 mo loan with 36 months left vs.a new 36 mo loan with all other credit history/info at ytms assumed to be equal.

Any thoughts on maturity while we are at it? would a 36 mo loan with 24 months left be more desirable than a a new 36 mo loan at an equal ytm assuming no credit status degradation in the interim on the older note?
How about a 60 mo loan with 36 months left vs.a new 36 mo loan with all other credit history/info at ytms assumed to be equal.
definitely 36 mo loan with 24 months left; and 60 mo loan with 36 months left, if the past payment history and credit history are good.

I think you all might be interested in the response I got from LendingClub when I emailed asking about the YTM calculation:
Additionally, I am unfamiliar XIRR and it is not what we utilize to calculate YTMs. Please note, we do not provide statistical support and I am unable to calculate the Yield to Maturity manually on your behalf.
To my knowledge you would not be able to use XIRR to arrive at our figures. Investors in the past that have successfully calculated our YTM have used Euler's method to compute the answer.
FACEPALM!
In case anyone is curious, Euler's Method is used to approximate differential equations.

I think you all might be interested in the response I got from LendingClub when I emailed asking about the YTM calculation:
Additionally, I am unfamiliar XIRR and it is not what we utilize to calculate YTMs. Please note, we do not provide statistical support and I am unable to calculate the Yield to Maturity manually on your behalf.
To my knowledge you would not be able to use XIRR to arrive at our figures. Investors in the past that have successfully calculated our YTM have used Euler's method to compute the answer.
In case anyone is curious, Euler's Method is used to approximate differential equations.
Thanks for the info. I'm pretty sure no one on this forum ever claimed to successfully compute the same Folio YTM as LC. There were plenty that tried. In 2016 I wrote some code that sold notes on Folio and I used the secant method to compute YTM. IIRC the result didn't match LC's but was reasonably close. I thought it produced the correct YTM and converged very quickly so I used it. See below for a very good description of the computation:
https://en.wikipedia.org/wiki/Internal_rate_of_return (https://en.wikipedia.org/wiki/Internal_rate_of_return)