Upstart and its risks as a lending institution

This is a long, pasted together thread that was triggered by a recent post about whether or not Upstart should be viewed a financial institution (essentially a bank), a hybrid entity that perhaps isn’t technically a bank but is highly dependent upon one bank (Cross River Bank), or is in fact something different. I’ve collected the pertinent posts below to make it easier. I found myself unsure of my own position, so I sent an email through Upstart’s website to investor relations, and within an hour had a response from its VP of investor relations. Nice touch, that…

From learninginvstor:

Folks, UPST had a good ER and their annual guidance was very good and surprised the market, that’s a fact.

Now, those people from our board who are heavily invested into UPST - just a friendly opinion - please understand fully the nature of the beast u are invested into and that’s a neo lender. The bottom line - this is a lending business (with AI edge) with a reselling of packaged loans to Wallstreet. Loans are predominantly personal unsecured focused on sub-prime (low credit scoring) clients. When there are/were/won’t be buyers for these loans they will have to take them on own balance sheet (see 2020). The bottom line profitability will evaporate as soon as the hits on the loan book will have to be reflected in P/L. If u have huge positions in a (neo) lending business - please study lending business in detail. Just a friendly advice.

My reply:

hi Learning: I think you have misunderstood upstart’s business. It keeps a small portfolio of loans on its books as it does r&d, but that is not significant in the larger picture. Other than that, it has no direct credit risk from the loans that pass through its platform – that is entirely held by the financial institutions that extend the loans, or those that ultimately buy the syndicated loans.

Upstart makes its money from fees charged to the banks, who rely on upstart’s software to make the decision on whether the loan applicant is solid.

Crazyczech’s comment:

I completely agree with Learning. It is absolutely true that Upstart relies on securitization markets. They avoid talking about this (perhaps for regulatory reasons) but it is a fact that a majority of their loans are sold to Cross River Bank. Those loans are packaged into structured credit instruments and sold off to investors. Anybody who invests in UPST should absolutely understand this dynamic. For anyone that thinks CRB keeps loans for themselves, do simple search on the internet for what their business is.

At this point, I thought I should get off my butt and do some actual due diligence, as I do in fact own upst and think very highly of their company and prospects. I composed a question to Upstart’s Investor Relations (link on their website) this morning, and within an hour I got a reply from their VP of Investor Relations. The entire text is below. I tried my best to paraphrase Learning’s characterization of Upstart’s business model, so apologies (@learning) if I didn’t do you justice.

My email to Upstart IR:

Further detail on your relationship with cross river bank
Dear IR:

I am a shareholder since mid-2021, and have ‘enjoyed’ quite a ride with your company. I have had a discussion with a fellow investor, and we understand the role of the securitization markets differently. Could you please give some clarity to the following:

It is my understanding that the bulk of loans (this excludes loans you keep on your books as you do r&d) end up on the balance sheet of the financial institution that extends them. Your role in those loans is simply the fee you collect from XYZ credit union as it evaluates the loan application.

In contrast, my colleague insists that the majority of those loans I describe above are in fact sold to Cross River Bank, where they are then packaged into structured credit instruments and sold on the securitization market. He quotes as support for this a section from your most recent 10q:

Cross River Bank, or CRB, a New Jersey-chartered community bank, originates a substantial majority of the loans on our platform. In the nine months ended September 30, 2020 and 2021, CRB originated approximately 72% and 58%, respectively, of the Transaction Volume, Number of Loans. CRB also accounts for a large portion of our revenues. In the nine months ended September 30, 2020 and 2021, fees received from CRB accounted for 65% and 59%, respectively, of our total revenue. CRB funds a certain portion of these originated loans by retaining them on its own balance sheet, and sells the remainder of the loans to us, which we in turn sell to institutional investors and to our warehouse trust special purpose entities. Our most recent commercial arrangement with CRB began on January 1, 2019 and has a term of four years with an automatic renewal provision for an additional two years following the initial four year term. Either party may choose to not renew by providing the other party 120 days??? notice prior to the end of the initial term or any renewal term. In addition, even during the term of our arrangement, CRB could choose to reduce the volume of Upstart-powered loans that it chooses to fund and retain on its balance sheet or to originate at all. We or CRB may terminate our arrangement immediately upon a material breach and failure to cure such breach within a cure period, if any representations or warranties are found to be false and such error is not cured within a cure period, bankruptcy or insolvency of either party, receipt of an order or judgement by a governmental entity, a material adverse effect, or a change of control whereby such party involved in such change of control provides 90 days??? notice to the other and payment of a termination fee of $450,000. If we are unable to continue to increase the number of other bank partners on our platform or if CRB or one of our other bank partners were to suspend, limit or cease their operations or otherwise terminate their relationship with us, our business, financial condition and results of operations would be adversely affected.

How do you see this relationship (relative weight of CRB vs new bank partners) evolving over the next 2 years?

Thank you.

And their reply was:

Thanks for your email and happy to clarify. Neither of you are precisely correct. We have 42 traditional bank partners that originate and hold loans, this makes up about 1/4 of our total volume. The other 3/4 of our volume is originated by one of two banks that act as conduits (including CRB). After origination, and seasoning for a few days, these loans are sold to a group of ~160 credit investors (hedge funds, mutual funds, etc). We also stand up a securitization, approximately every quarter, in order to provide liquidity to these credit investors. However only a minority of credit investor volume is ultimately securitized (largely opportunistically, such as over the last year when markets were particularly constructive) with most held for yield on their balance sheets. Hopefully that helps and please let me know if you have any other questions.



The main point that I think Learning was trying to make, and that I concur with is that Upstart is very dependent on funding from the credit markets. They market themselves as providers of AI to banks but in reality they are really providing credit rating and servicing services on loans that they originate (via CRB) and sell to credit markets. Credit markets are very fickle… they can be incredibly liquid at times like last year and incredibly illiquid like they were in March 2020 and many times of the past. The other 1/4 of their business - banks that retain loans on their balance sheet tend to be much more consistent with their credit appetite and may not create the same level of volatility in demand that you would see in the credit markets. For those of you who do not understand credit markets, especially markets for new unestablished category of credit like Upstart loans, think of equity markets for our stocks… one day they are super hot, next day nobody wants to talk about them - it is not as volatile with credit, lag times tend to be longer but you get the idea. Participants in the credit markets have infinite number of choices for how to invest their money from treasuries all the way to junkiest of junk bonds and everything inbetween. That is what drives this volatility. Banks on the other hand, especially the smaller ones that UPST targets, are primarily in the business of borrowing money from depositors and lending it to creditors, thus their appetite tends to be more consistent.

The main point here is that UPST is reliant on those wholesale credit markets to large extent. If they dry up, so will their revenue. Plus Upstart is reliant on companies like CRB to originate loans for them since this process can only be done by a company that has a banking license. This entire concept is new and regulators tend to be very funny when it comes to new concepts in finance. If anything happens to CMB or regulators start taking a closer look at CMB (and few others that do the same), it could materially impact Upstart’s business.

I hope I’ve made it somewhat clear. This is a complicated part of the financial word that most people have no idea exists. If all you take from this is that it is complicated and that there are many dependencies, that’s great. This is not an argument for or against investing in Upstart.


Thanks coldmountain for this diligence. Some observations:

  1. Given the high attention given to UPST on this board, it’s remarkable that apparently we still don’t seem to fully understand the mechanics of their business, at least as witnessed by the various board posts which coldmountain included up front.

  2. Their VP of Investor Relations clarified it somewhat but I’m still not sure that I get it. Some of his comments:

  • “After origination, and seasoning for a few days, these loans are sold to a group of ~160 credit investors”… (“seasoning”?)
  • “We also stand up a securitization, approximately every quarter, in order to provide liquidity to these credit investors” (“stand up?” “provide liquidity”?)
  • “However only a minority of credit investor volume is ultimately securitized…with most held for yield on their balance sheets” (how much is “a minority” and what does this sentence actually mean?)

I’m sure that my lack of knowledge of the “credit investment industry” (if that’s a valid term) contributes to this. My comments are intended to fall under “there’s no such thing as a stupid question”. I’d appreciate it if someone could translate the VP’s reply into something more accessible to the average person.

  1. It seems to me that the company could be doing a better job of explaining how this all works. If this very-focused board can’t quite get its mind around it, then I would not expect many others to.

  2. Lastly, and most importantly, is there anything about the VP’s reply that should affect how we view this company?



I’ll take a stab at this, RollerCoaster:

- “After origination, and seasoning for a few days, these loans are sold to a group of ~160 credit investors”… (“seasoning”?)

Seasoning means the loans are held in account to settle. There is usually a consideration period wherein either side of the transaction can declare it null and void. Seasoning to ensure loans are valid and will remain in effect reduces the chance that a few loans will be reverted to capital after a voided contract. (borrower chooses to reject the funds).

More here:

- “We also stand up a securitization, approximately every quarter, in order to provide liquidity to these credit investors” (“stand up?” “provide liquidity”?)

Securitization is a process where loans or other assets are bundled up for sale. This process involves numerical analysis of the products in the group, and the average characteristics. We have seen this before in the prime, near-prime, and sub-prime mortgage setting.

More on Securitization here:

- “However only a minority of credit investor volume is ultimately securitized…with most held for yield on their balance sheets” (how much is “a minority” and what does this sentence actually mean?)

This simply means that the holder (in almost all cases - a bank) holds the note and is waiting, hand out, for the contractually obliged payments each term (usually once per month).

Alright, now, what are the implications for UPST?

If the group of buyers loses appetite or does not have liquidity, the originating bank will have to hold. If that is not UPST, it is a condition that requires loans to be held for yield, as discussed above.

If UPST holds the loans (i.e. their R&D products which are still being tested or which have not had the sales funnel fully set up to sell), then UPST would be providing cash outlay for each loan and holding them for yield, just as the bank would.

So, let’s walk through some scenarios:

  1. The credit market dries up for an extended period of time.
  2. The UPST buyer pool decides to look elsewhere for products to purchase as securities.
  3. The UPST securities are not priced attractively for the buyers to be interested in the market

In scenario 1, loan originations would stack up and would be held on the books, being unable to be sold. This would be a market wide phenomenon, causing cash flow problems not only for UPST, but also for partner banks. Currently, the cash on deposit and reserves health of many banks is nowhere near the criteria to force this scenario.

For scenario 2, (which is really a subset of S3), UPST would be more and more reliant on loans held for yield with partner banks. Securitizations would fall as a proportion of loans originated and liquidity concerns for partner banks would be in focus.

Scenario 3 implies that loans are not performing as advertised. The KBRA reports would indicate that delinquencies are climbing and that defaults are out of spec. This impacts overall yield for securitized loans referenced, of course, but it also implies additional discounts are required to ensure attractive terms. A possible outcome in this scenario is erosion of fees due to UPST (loans aren’t worth as much, so UPST get squeezed on each new tranche of loans to be securitized). If this continues, more and more buyers may opt to no-bid for tranches listed for sale. This would be very bad for UPST, implying their model does not work as advertised.

Fee revenue would fall and new partnerships would slow based on a generally less desirable product (against it’s risk profile). UPST revenue, profit and growth would all slow as witnessed in Quarterly Reports.

There is no indication from the company that this is occurring. Quite the opposite based on the recent results this week.

Board analysis from the KBRA report would be a good review as there is a lot of information in those threads. Reviewing them from a post report, additional information perspective would be smart.

Before earnings:…

Before earnings:…


I just looked at S-1. Few points from there:

“Loans issued through our platform can be retained by our originating bank partners, distributed to our broad base of approximately 100 institutional investors and buyers that invest in Upstart-powered loans or funded by Upstart’s balance sheet. In the third quarter of 2020, 22% of the loans funded through our platform were retained by the originating bank and 76% of loans were purchased by institutional investors through our loan funding programs. Our institutional investors and buyers that participate in our loan funding programs, which include Goldman Sachs, PIMCO and funds managed by Morgan Stanley Investment Management, invest in Upstart- powered loans through whole loan purchases, purchases of pass-through certificates and investments in asset-backed securitizations. We enter into nonexclusive agreements with our whole loan purchasers and each of the grantor trust entities in our asset-backed securitizations, or ABS, under which our ABS investors benefit from our loan servicing capabilities .

Sounds complicated, isn’t it?

They still sell to Wallstreet 3/4 of their loans as of end 2021. I’m pretty sure that Wallstreet wizards don’t have any obligation to buy these loans from Upstart and will cease when it won’t be profitable for them.

Upstart balance sheet as of 30.09.2020

Total assets - around 310m
Out of that loans 122m

Operating cash flow for 9 months in 2020:

“Net cash used in operating activities was $52.8 million for the nine months ended September 30, 2020, which primarily consisted of a $84.0 million change in net operating assets and liabilities … The change in net operating assets and liabilities was mainly related to a $109.1 million increase in purchases of loans held for sale .” Does this mean that they had to buy these loans and put on own balance sheet in the downturn and were unable to find Wallstreet buyers at that time?

Then please check Note 3 to their financial statements - it provides details of their securitization practices with all the VIEs, trust structures, off balance sheet liabilities etc. It’s far from plain vanilla business.

Note 4 to financial statements describe fair value assessment of the loans, how they buy and sell loans etc.

Not even mentioning concentration risk to Credit Karma and Cross River Bank.

Again, guys, I am not saying that Upstart is a bad business, bad company or bad stock. My goal is to get attention of u - fellow board members - to risky parts of the business of this company. Indeed in August-November 2021 noone looked seriously at the risks because fundamental numbers were just stellar and SP soared. But when the price drops by 50-60% one wants to better understand the company one owns - this helps to have a conviction and hold thru tough times. Perhaps after thorough analysis one is fully happy holding or adding to Upstart - which is totally fine if one understands what one owns and sees much more pros to the investment thesis than cons.




Hopefully there will be an actual banker who weighs in on your questions, but here’s how I understand the answer from IR:

I apply for a $20k personal loan from Upstart. I sail through their model and have the loan approved in an instant and the money is in my account in two instants. I’m very happy. Meanwhile, let’s follow that loan. It can take one of three paths.

  1. A quarter of the time, my loan goes to live with one of the now 42 bank partners working directly with Upstart. That’s who will get my monthly payments and that’s who will suffer should I default. I’m on the balance sheet of one of those 42 banks.

  2. Three-quarters of the time, my loan is instead originated at Cross River Bank or the other conduit (whose name I don’t remember right now). But CRB is not into long-term relationships. They will sell my loan on to one of the 160 or so funds that IR mentioned.

But those funds have not originated my loan and don’t know me. They have to trust Cross River Bank, who is selling them the goods. To make that purchase feel like a safe deal to those 160 funds, they want the loan to be “seasoned” for a few days. That just means there’s enough time to make sure everything is as it appears with the loan. We experience “seasoning” when money coming into our bank accounts isn’t immediately part of our “available funds.” The last time I bought a house, I had to prove to the bank that the money for my downpayment had been in my account for the past six months. Or, if it was new money, I had to show where it came from and that source had to be acceptable to the mortgage lender. That is another kind of “seasoning.” Here’s an article about seasoning from Investopedia:

By keeping my loan for a few days, CRB makes sure the deal wasn’t a scam. That gives the lucky fund that buys it a bit more assurance that I am worth what they are paying for my loan.

  1. But whether my loan is moving in with one of the 42 banks or getting temporary housing at CRB before being sold to a fund, the lender who ends up with me doesn’t have a liquid asset. They have paid for my loan and will get back that money over time as I pay it off. That isn’t quite as much fun as cash that can be put to work right away–a “liquid” asset.

Because our friends at Upstart are generous folks, they want to help their partners have some liquid assets, too, so that they can get at least some benefit more quickly. The way they do that is by bundling up a bunch of their loans each quarter and selling the bunch of them as a package that can then be resold quickly and easily for cash. That is called securitization, and that’s the third path my loan might follow. This article explains how it works with mortgages, although it can be done with most any loan category:….

That practice, when used indiscriminately, can lead to problems as it did in the housing crisis of 2008. The securitized packages were bundled and re-bundled, sold and resold, so many times that those who bought them down the line never even knew what was in them. And because they were so easy to buy and sell and it was such a quick way to get money, lending standards went out the window and everyone was buying and selling junk until it all came tumbling down.

That’s why I was happy to read that only a minority of Upstart loans are securitized and sold in that way. In the majority of cases, the loans are held on the balance sheets of either the 42 banks who bought them directly or the 160 funds who got them after a few days of vacationing at Cross River Bank to make sure they don’t fall instantly apart.

All of those systems rely on the reputation of Upstart’s AI. When banks buy loans directly and experience fewer defaults than they had using FICO or other methods, they drop FICO and use Upstart more. They talk to their friends, join the referral network, and then there are more banks who want to buy direct from Upstart and fewer will need to spend a few days at Cross River Bank.

If Upstart continues to have lower default rates than other risk assessment tools, the 160 funds will also have more confidence and, unless there are rules about such things in the industry, maybe the early trip to CRB will not even be needed. The loans that pass through CRB must be (I think) more expensive to the funds, since CRB has to be compensated for holding onto a riskier pool of assets for the seasoning period.

I remember watching an interview sometime last year with Dave Girouard who was asked about the concentration risk with Cross River Bank. He didn’t seem to understand why people saw CRB as a concentration risk and said something to the effect of “Well, if for some reason they didn’t want to be the conduit anymore, we’d just find another place.” They (and the other bank) seem to be the Ellis Island of the loan’s journey in 3/4 of the cases. They go to CRB first, get checked out, and then they move on to the home where they will settle for good. The other 1/4 are pre-cleared by the bank that will hold the loan. As Upstart scales and brings on more banks, those percentages will shift.

Again, I am NOT a banker, nor do I play one on TV. I have not recently spent the night in a Holiday Inn Express. I may not have the above exactly right. But that’s how I understand the process.



These are all excellent answers. I would just add that majority of the loans that they currently retain on their balance sheet are most likely car loans that they are still trying to figure out. This was mentioned on the last earnings call.