Value of previously late notes

[joshmer]

Just a question for anyone out there to throw some thoughts around.   I recently started to reprice my notes that have had a late payment before and are now current.   I figured that 95% of principle balance would be a good place to start with notes that have a downward trending FICO, but I'm been surprised to see the large portion of my notes selling quickly.   Has anyone done an analysis the value of a note going late at least one time?  Outside of lending club's cure rates I didn't see a post on true statistics around notes having further payment issues.

Thanks!

[yojoakak]

There's plenty of never-late notes for sale right around par so I'm not sure why anyone would bother taking a chance on a 5% discount.

Ah well. The Wisdom of the Market I guess.

[Fred]

Some statistics on Current notes on FOLIOfn as of tonight (4/20):

I recently started to reprice my notes that have had a late payment before and are now current.

Your notes will have status=Current, but neverlate=0, so the first 3 rows on the table above apply.  If you are not desperate, you should follow the pricing in avgMarkup column.

[rawraw]

Is that how your model determines the selling price for FolioFN strategy?

[merits]

generally, I don't think it is easy to sell now-current notes with a premium.
The average above is the average by listing, not the average sale price of the market.

[Fred]

Is that how your model determines the selling price for FolioFN strategy?

Actually, I used the statistics to determine which notes to buy on FOLIOfn.

[Ran]

Is that how your model determines the selling price for FolioFN strategy?

Actually, I used the statistics to determine which notes to buy on FOLIOfn.

Can not believe you still have positive returns by doing this. I took a snapshot of Foliofn listings two months ago and what I found out is that the loan default rates are MUCH higher than discount.

[Fred]

I took a snapshot of Foliofn listings two months ago and what I found out is that the loan default rates are MUCH higher than discount.

Care to elaborate more, specifically on how you measure default rates on FOLIO-purchased notes?  How much different are they compared to the LC-published default rates?

[core]

Fred those statistics are nice to look at, but I don't see how examining what's currently listed can give you confident purchasing data.  You're only looking at what hasn't sold, and none of that which sold quickly.  Also it seems to me that your data would be seriously skewed by just a few people who list things at unrealistically high prices with no intention of selling the notes except to fools.  This increases your max markup, increases the avg, (MAYBE changes your purchasing, depending on how you have it set up) but the real market hasn't changed one bit.

[Fred]

Fred those statistics are nice to look at, but I don't see how examining what's currently listed can give you confident purchasing data. 

As an investor, I define my problems as: given the available notes on FOLIO now, which ones should I buy?

I am basically making a statistical bet based on LC-published default rates.  Once I buy a FOLIO note, the note's default rate is essentially reset for me (i.e., I take the memoryless Markovian assumption).   The spread between FOLIO-purchased notes default rates vs. LC-published default rates would be my profit.

BTW, I currently do not buy Current notes on FOLIOfn (never late or otherwise).

[Fred]

This increases your max markup, increases the avg, (MAYBE changes your purchasing, depending on how you have it set up) but the real market hasn't changed one bit.

I guess as a buyer, the column I am interested in the most would be the minMarkup.  If I were in the business of buying Current notes on FOLIOfn, I would seriously consider buying the Current note with 18.63% discount.  That one was 5 std-dev lower than the average.

[core]

As an investor, I define my problems as: given the available notes on FOLIO now, which ones should I buy?

I'd just add two words to that at the end:  "if any".  If they're all overpriced you don't want to buy them, even if that means it leaves you with undesired uninvested cash.  Also your goal probably shouldn't be to minimize available cash as quickly as possible regardless of what the current market is doing.  A better deal may appear in 10 minutes or tomorrow.  A whale may be gobbling up things today, driving up prices.  Another whale might be liquidating his entire portfolio for -20% tomorrow.

I guess as a buyer, the column I am interested in the most would be the minMarkup.

Certainly.

If I were in the business of buying Current notes on FOLIOfn, I would seriously consider buying the Current note with 18.63% discount.

The one with the recent collection note that says "RTP"?  Or the one where it says he's employed by the solar company where the execs were just taken out of the building in handcuffs and all employees were terminated?  There might be a reason it's so cheap.  Sure, I'd consider buying it, just the same as I'd consider buying anything else that's discounted somewhat.

That one was 5 std-dev lower than the average.

Again, average and standard deviation are absolutely meaningless here when you have all the noise at the top end that can come and go depending on a single person's whim.  If whale guy takes down all his +69.99% listings in a certain class that you are tracking, then your std dev suddenly goes down, and so does your mean.  How valid is that?  Not at all.

To me it seems a lot like looking at a NASDAQ level 2 quote or similar and saying to yourself, "Gee, there's this $20 stock but I see that some guy is willing to sell it for $900 a share.  That's interesting.  (Or is it?) Stock B doesn't have any $900 offers though.  Let's see, I'll just compare the averages and std deviations of stale GTC orders for available stock symbols, do some waving of the hands, and..."  ...and waste a fair amount of good brainpower and CPU cycles on a useless exercise.  Or at least that's the way it seems to me.

[Fred]

If I were in the business of buying Current notes on FOLIOfn, I would seriously consider buying the Current note with 18.63% discount.

The one with the recent collection note that says "RTP"?  Or the one where it says he's employed by the solar company where the execs were just taken out of the building in handcuffs and all employees were terminated?  There might be a reason it's so cheap.  Sure, I'd consider buying it, just the same as I'd consider buying anything else that's discounted somewhat.

Ha .. ha ...  It's funny when you make things up.

The borrower was actually a B2 grade with 10+ years employment with USAF in San Antonio with verified income. Of course, this does not tell the whole picture about his latest credit risks.

One of the hidden gems of statistical analysis is for notes from the same loan.  These notes are essentially of the same credit risk -- they are from the same borrower.  Once you make a go decision on the loan-level, having information about the cheapest note is definitely advantageous.

There are a few notes from that USAF loan, with markup up to 5.85% -- definitely, we do not want this note.

That one was 5 std-dev lower than the average.

Again, average and standard deviation are absolutely meaningless here when you have all the noise at the top end that can come and go depending on a single person's whim.  If whale guy takes down all his +69.99% listings in a certain class that you are tracking, then your std dev suddenly goes down, and so does your mean.  How valid is that?  Not at all.

I beg to differ.  If you only take a one-time snapshot, yes that would be a concern.  However, having collected & analyzed FOLIOfn data for many months, I definitely see niches and patterns to take advantage of.

If you didn't want to take that USAF Current note at 18% discount, that's fine.  Like I said, I would have seriously considered it.

[Bohb Daishi]

Just a question for anyone out there to throw some thoughts around.   I recently started to reprice my notes that have had a late payment before and are now current.   I figured that 95% of principle balance would be a good place to start with notes that have a downward trending FICO, but I'm been surprised to see the large portion of my notes selling quickly.

Your notes must be ridiculously good to sell at a -5% discount. I actively trade in late-to-current notes, and I'm lucky if I can even get -10% when they go current. Anything better than that and I'm stuck holding onto notes for months on end.

That one was 5 std-dev lower than the average.

Again, average and standard deviation are absolutely meaningless here when you have all the noise at the top end that can come and go depending on a single person's whim.  If whale guy takes down all his +69.99% listings in a certain class that you are tracking, then your std dev suddenly goes down, and so does your mean.  How valid is that?  Not at all.

I beg to differ.  If you only take a one-time snapshot, yes that would be a concern. However, having collected & analyzed FOLIOfn data for many months, I definitely see niches and patterns to take advantage of.

If you didn't want to take that USAF Current note at 18% discount, that's fine.  Like I said, I would have seriously considered it.

Your numbers will still be incorrect. While in theory it's possible that everything you see are all fairly priced, in practice it's much different. Lending Club refuses to share any actual Folio sales data, so the only thing you can base your numbers on are the prices of the notes that have not sold yet. There's a reason why they haven't sold. They are either too expensive, or the buyers who would be willing to purchase them have already used up all their available cash purchasing other notes that had better discounts.

I have plenty of investment-grade notes that are always on the market at a +7% discount, just waiting for someone to buy them. I don't expect these to get sold any time soon, but if they do then it's gravy. But given that many people price their notes in such a way, how do you filter this strategy?

The only true data set that you can go off is what price notes actually sold at (i.e. the bid price), not what they are currently priced at (i.e. the offer price). Everything is worth only what a purchaser is willing to pay for it.

[Fred]

While in theory it's possible that everything you see are all fairly priced, ...

As I mentioned before, my trades are based on the spread of default rates.  This is akin to trading CDS.   IMO, for this strategy fair pricing is less of an issue than the size & likelihood of the spread.  In fact, this strategy will not work if all notes are priced fairly, because this implies zero spread.

I have plenty of investment-grade notes that are always on the market at a +7% discount, just waiting for someone to buy them. I don't expect these to get sold any time soon, but if they do then it's gravy. But given that many people price their notes in such a way, how do you filter this strategy?

If I am desperate to sell, I could take a peek at the market and try to beat the most-discounted note by a little.  Say, my competitor is selling the note at 8% discount, I could sell mine at 8.1% discount.

Everything is worth only what a purchaser is willing to pay for it.

Agree.  FOLIO is essentially a buyers' market.  Once you've run your analysis, you can cherry-pick the notes to buy, if any.  I've bought a lot more notes than I've sold. This all can be done programmatically.  If you run this strategy throughout the day, you can take advantage of fleeting opportunities.