### Author Topic: What does 9 month loss estimate REALLY mean?  (Read 6122 times)

#### rj2

• Jr. Member
• Posts: 85
##### What does 9 month loss estimate REALLY mean?
« on: August 11, 2016, 01:47:37 AM »
LC's ANAR is based on a nine month loss estimate.

I have questions about that, if I have \$100 capital in a 31 day late note, is the \$24 nine month estimate:

1. Is that present value, \$24 today? Or future value, \$24 in nine months--so worth less today

2. Is that the estimate for a note that is 31 days late? Or the average late note, more likely around 75 days late?

Any way to get the value of a note based on whether it is 30 vs 60 vs 90 days late? Obviously 60 days late is worth more than 90 but where to get that?

I used to not care but the differences are important in understanding what a late note is worth if you list it for sale in folio.

Money invested in my portfolio is earning about a half percent a month by my conservative estimate based on my allocation, though LC claims it's more. I want to calculate what discount is the break even on holding a non performing note in hopes of payment versus selling and immediately reinvesting the cash.

If it's \$24 nine months from now, it's worth \$22.92 today and with the 1% fee worth selling at \$22.69. If that's already present value it's now worth selling below 24.24... But that also depends on whether it's \$24 the day it goes late, or for the average of all 31+ day late loans.

I'm willing to hold the late loans if the expected value is greater than what I can get on folio, but how to calculate that?

Let's overlook things like changes in credit score and focus just on the length of delinquency for now.

Seems hard to arrive at the expected value necessary to determine what price in folio is profitable to sell.
« Last Edit: August 11, 2016, 01:55:02 AM by rj2 »

#### Fred

• Hero Member
• Posts: 1421
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #1 on: August 11, 2016, 04:19:52 AM »
1. Is that present value, \$24 today? Or future value, \$24 in nine months--so worth less today

The numbers are unitless; they are probabilities.  The Late31-120 note has a 76% probability of being charged-off in the next 9 month; or 24% of being recovered.

If you multiply the probability with the outstanding principal, then the resulting amount is neither present nor future value.  It is the expected amount you will receive in the next 9 months.

2. Is that the estimate for a note that is 31 days late? Or the average late note, more likely around 75 days late?

The calculation is based on loan status, which covers from 31-day to 120-day late.  So the estimate so refer to notes closer to the 75 days rather than exactly 31 days.

Any way to get the value of a note based on whether it is 30 vs 60 vs 90 days late?

Yes, go to https://www.lendingclub.com/info/demand-and-credit-profile.action

#### Rob L

• Hero Member
• Posts: 2108
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #2 on: August 11, 2016, 10:38:36 AM »

Any way to get the value of a note based on whether it is 30 vs 60 vs 90 days late?

Yes, go to https://www.lendingclub.com/info/demand-and-credit-profile.action

Just to add a bit more, the loan migration chart shown by Fred is for all loans lumped together.
Separate charts for each term (36-60) and each loan grade A .. G given its term would be of more help when evaluating a single loan.
That's a lot of charts of course (2 x 7 = 14).

If I understood your theme you would like to know the expected YTM of a note given the performance history of similar notes in the past.
Think that is a complicated question with a lot more moving parts than grade and term. Maybe a few individuals and third parties have answered it. I haven't (yet).

#### rj2

• Jr. Member
• Posts: 85
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #3 on: August 15, 2016, 01:07:01 AM »
I'm trying to come up with a rule of thumb for what a late note is worth on average, ignoring the individual properties of the loan other than how late it is.

I'm not trying to come up with a method of valuing an individual loan. Rather, I would use this rule of thumb as a starting point and make adjustments from there. For the purpose of this rule of thumb let's assume that any factors from the original listing are already accounted for in the grade and interest rate. A poor assumption, but a good one for making rules of thumb.

There are two main things we know about late loans that seem particularly relevant: how late they are, and what has happened to credit score since the loan was issued.

Let's leave credit score aside for now. I'd like to nail down lateness before factoring in score. Let's assume score is unchanged OR has only changed because this loan itself is reporting late.

So now we have these rates of loan migration from LC:

67% at 31 days late
82% at 61 days late
85% at 91 days late

These are the percentages of loans that default over a 9 month period. That's some good information but it isn't enough to calculate the present value of these loans.

Does that mean that the remainder of the loans become current having paid back all accrued interest?

And for those that default, how much money is recovered? I'm aware that some defaulted/charged off loans actually see some amount of recovery.

Example: I have a loan that has just turned 61 days late, with credit score unchanged so far from the original score. My loan has exactly \$10 in outstanding interest and principal.

From the above, what is the expected present value?

Is it \$1.80 based only on the 82% number?

Is it slightly higher due to the prospect of recovery?

Is it slightly lower because of the prospect that it's going to become temporarily current but then default anyway, just, more than nine months from now?

I'm on the LC platform for the long term. I could sell my note on the folio but I would do this only if I can receive more for it than my expected value from holding it. If I can get a higher return by holding it, I'd prefer that.

So what's the rule of thumb, what's the price below which you make more money holding the loan than selling?
« Last Edit: August 15, 2016, 01:11:12 AM by rj2 »

#### Fred93

• Hero Member
• Posts: 2221
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #4 on: August 15, 2016, 02:32:10 AM »
So now we have these rates of loan migration from LC:

67% at 31 days late
82% at 61 days late
85% at 91 days late

These are the percentages of loans that default over a 9 month period.

That's not quite right.

If you click on the "learn more about this chart" link right below the chart, LC provides more explanation.  The numbers in the chart are average net loss fractions.  They even give you the formula.
« Last Edit: August 15, 2016, 02:33:59 AM by Fred93 »

#### rj2

• Jr. Member
• Posts: 85
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #5 on: August 15, 2016, 03:44:44 AM »
Ah, so then those numbers are less useful because we don't know over what time period those fractional payments are received and therefore don't know how to discount those future payments back to present dollars.

For my account I estimate that I'm earning 7 to 8 percent on money I invest. So using 8% as a discount rate, if my \$10 late note is going to pay \$1.80 at some future time, to bring that back to present dollars we need to know over what time.

If I'm getting that\$1.80 in three months it's worth \$1.76 now but if I won't be getting that money for nine months it's only worth \$1.70, assuming I reinvest it at my account average return of 8%.

That's a big difference when valuing notes in folio. It's a 3.5% difference in the break even discount rate, meaning the discount at which you expect the same return by either selling the note or holding it (ignoring variance drag).

#### rj2

• Jr. Member
• Posts: 85
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #6 on: August 15, 2016, 12:59:19 PM »

Well I guess we could assume that the payments are received on average after half the 9 month period. In that case assuming an 8% return we should discount the fractional recovery amounts by about 3% to bring them to present value.

So the correct discount rates on average are:

31 days 69%
61 days 84%
91 days 88%

That's the present value of those future fractional payments with a lot of assumptions, like credit score is unchanged.

#### Fred93

• Hero Member
• Posts: 2221
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #7 on: August 15, 2016, 01:43:44 PM »
Well I guess we could assume that the payments are received on average after half the 9 month period. In that case assuming an 8% return we should discount the fractional recovery amounts by about 3% to bring them to present value.

The "9 month" time frame was chosen simply because after 9 months almost nothing is ever recovered.  Most recoveries occur soon after the money is due.  As time elapses, the chance of recovering anything goes down.

#### rj2

• Jr. Member
• Posts: 85
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #8 on: August 15, 2016, 03:35:58 PM »
Sure, I understand that. I'm trying to figure out how to value these late loans. I'm trying to find a good way to do it with a rule of thumb, price them in batches.

I don't want to dig into the details of a bad loan worth \$1.80 to see if it's really worth \$1.76 or \$1.84, wet are talking about a nickel here or there.

But if I had a way to roughly price them correctly with a simple rule I could list them en masse and hopefully eventually automate that.

#### rj2

• Jr. Member
• Posts: 85
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #9 on: August 15, 2016, 03:37:44 PM »
The takeaway from this thread is that the LC ANAR numbers need to be discounted about 3% to account for the time value of money, since they describe payments that will not be received for 3 to 9 months.

#### Rob L

• Hero Member
• Posts: 2108
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #10 on: August 15, 2016, 03:52:11 PM »
The "9 month" time frame was chosen simply because after 9 months almost nothing is ever recovered.  Most recoveries occur soon after the money is due.  As time elapses, the chance of recovering anything goes down.

Not to split hairs but I think they chose 9 months such that the whole smash covers a year and there exists a final trailing three month period after the last possible charge off to accommodate any possible recoveries:

Quote
"Net charge offs” are equal to the total principal amount charged off less any funds recovered within three months of the last day of the applicable nine month period. For example, if a loan was in “Grace Period” in January 2015 ,the principal amount of \$1,000 was charged off in September 2015, and \$100 was recovered in October 2015, the net charge offs for that loan would be \$900."

Anyway, the 9 months is just a conveniently picked rather arbitrarily long time as you say.

#### rj2

• Jr. Member
• Posts: 85
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #11 on: August 15, 2016, 09:01:17 PM »

I wonder why LC doesn't include time value of money in their ANAR. They value the loan by the amount that they expect to recover, without discounting that amount back to present value. As a result, they over-state the value of all accounts on LC, as measured by ANAR.

#### Fred

• Hero Member
• Posts: 1421
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #12 on: August 17, 2016, 04:35:49 AM »
I wonder why LC doesn't include time value of money in their ANAR.

At what rate should LC discount their ANAR?  If it is at 1% or lower, the difference between discounted vs. non-discounted would be very small.

Besides, the discount rate would be a yet-another topic of contention.

#### rj2

• Jr. Member
• Posts: 85
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #13 on: August 17, 2016, 11:46:03 PM »
They can let the investor set the discount rate. It should be set to the expected return. They can use the platform average return as a default and let the investor modify it.

#### Fred

• Hero Member
• Posts: 1421
##### Re: What does 9 month loss estimate REALLY mean?
« Reply #14 on: August 18, 2016, 11:02:23 AM »
They can let the investor set the discount rate. It should be set to the expected return. They can use the platform average return as a default and let the investor modify it.

Are you saying, "discount rate should be set to the expected return?"

If we are talking about NPV, this is the wrong rate to use.  Discount rate should be set at the investor's cost of capital -- should be much lower the LC's expected return; otherwise, it does not make sense to invest in LC notes.