Clarifying SoFi's "Tricky Accounting" and Insights from My Conversation with SoFi CFO Chris Lapointe:
A quick apology to my readers. I have three huge projects winding down by the end of the year at my 9-5 job. I’ve been putting in much longer and more focused hours there over the past couple months and it’s only going to get worse between now and Christmas. I try to write at least every two weeks, but I do not know if that will be possible in December. Once the new year rolls around I should get back to a more steady cadence. Hopefully you are all ok bearing with me during this time. I would also really like to launch a YouTube channel at the beginning of the year, but have not reached the paid subscriber level I think would justify the time necessary. If you want to see more content from me, that’s how you can make it happen. 100% voluntary as always and I share all my content either way.
Data Driven Investing is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
SoFi Sold Loan Gain on Sale
The first purpose of this post is to set the record straight on SoFi’s sold loans and fair market values. There was lot of chatter when the 10-Q came out that SoFi is somehow being dishonest by reporting the gain from loan sales at 105.1%. Here is the disclosure from the 10-Q.
SoFi sold $15.098M in loans in 3Q23. What has everyone up in arms is the fact that the cash they received up front was exactly $15.098M and the entire 5.1% gain was from "servicing assets recognized." People are concerned that SoFi is trying to pull the wool over investors' eyes and pull off some tricky accounting so they can claim they have 105.1% gain on sale margin when really their loans are only worth 100% of unpaid principal.
That is not what is happening here. SoFi services these loans. That means the borrowers actually keep paying SoFi the money. SoFi then takes a fixed payment and then pays the rest to the company who bought the loans. The contractually obligated fixed payments are what is included in the "servicing assets recognized" above. In this case, the partner who bought the loans did not want to pay out as much up front, but is willing to give a larger portion of each payment going forward to SoFi to service the loan. The $747k, which is the entirety of the gain, is recognized as the present value of all those future guaranteed payments based on a discounted cash flow model.
How is that present value calculated? The weighted average life of loans for these loans is 1.5 years. If you wanted to back out the value of each payment, you could use a 36-month discounted cash flow model. The variables that would go into that DCF would be the total loan balance remaining, the same prepayment rate and default rate as is available in their fair value measurements (copied here for convenience) and for a discount rate you use the 2-year treasury rate on the day of the sale.
That's a lot of accounting jargon, so let's use a real world example that's more understandable. You lend $200 to your friend Joe. He agrees to pay you $20 a month for a year, so $240 total. You immediately go to your other friend Jill and tell her that if she pays you $200 now, you'll pay her $19 a month for the next year, which is $228 total. You now have the original $200 back, and every month Joe pays you $20, you pocket one dollar, and give the remaining $19 to Jill. That dollar that you are pocketing every month is the servicing right. It's guaranteed you get that money as long as Joe pays his loan, and you already got your original $200 back. That's exactly what SoFi did.
It is true that if Joe stops paying you then you don't get the extra dollar. But imagine you had 1000 Joes you were loaning to and you knew that the average default rate was 46 Joes per year, and you make sure that you account for that when you figure out how much you are going to make.
The servicing assets calculation reported above takes into account that some of the loans will go delinquent and stop being paid. The loss of those funds because of delinquency is included in the 5.1% gain on sale margin (GOSM).
Now, onto fair values. The fair values always have the servicing rights built into their fair value. That’s the way they’ve always been reported. Furthermore, every single time they’ve reported a sale, they’ve always included the servicing right in what they were reporting. That means that yes, in fact, SoFi sold their loans at a higher value than what they currently have marked on their book. Also, since then, they closed the $375M securitization with Blackrock and reported that the GOSM on that securitization was 5.0%. So, to recap, they have sold about $490M of loans above their current fair value marks.
On Capital Ratios
While I was talking to Chris, I asked a few other questions as well. The first was where they plan on operating their capital ratio long term. This has been a question I’ve been pondering as the company moves towards its capital constraints. To give a little bit of context, these are quarterly risk-weighted assets (RWA) for the company. The green bar is the range of expected capital ratio where I expect them to want to be long term. The red line is where they would be to hit the minimum ratio.
RWA has been expanding at about $2.5B/qtr since 2Q22. The concern here is that they get into the top of the green range and then there is a recession and there is capital destruction and the red line comes down to meet the blue bars and SoFi is forced to liquidate assets at unfavorable prices or regulators shut them down.
I asked Chris if he could help me understand where they want to operate long term. His response is that they determine where their operating limit would be by performing a stress test. They assume a severe recession based on Moody’s S4 scenario and make sure that in that type of recession they will stay above the regulatory minimums with the increased losses of that deep economic recession.
For those that aren’t aware of what those scenarios mean (a group that included me a month ago), Moody’s does multiple scenarios of how things might play out in the economy moving forward. The S4 scenario is a deep recession, worse than what most people think of when they think of a hard landing. How likely do they think this scenario is? In their own words, “In this recession scenario, there is a 96% probability that the economy will perform better, broadly speaking, and a 4% probability that it will perform worse.” You have to pay to get the actual scenarios, but they have published their July 2017 report on their website, which gives an idea of how bad conditions would have had to get to hit their S4 scenario. It’s pretty dire. The 2017 scenario includes a deep recession in Europe, a cumulative 21% drop in the housing market between 2017 and 2019, peak-to-trough GDP decline worse than the Great Recession (which was the worst recession since the Great Depression), unemployment would have peaked at 10% in 2019 and stayed above 9% until 2021. You get the picture. Considering the economy was chugging along just fine in 2017, this would have been an incredible reversal. I do not know what the current S4 scenario looks like, but given the current macro uncertainty, I’d assume it’s even worse than that.
SoFi’s game plan assures that even in a recession that is worse than the worst recession in modern times, they will stay above the regulatory minimums. Chris did not say exactly what that means in terms of the lower limit they want to hit on that key capital ratio, and I assume it will move slightly as the S4 scenario gets worse or better, but he said it works out to a few hundred basis points. So I think my 12.5%-14% range is pretty good.
One last thing to remember is that the green bar and red line move every quarter based on their total risk-based capital, which moves roughly in line with the change in SoFi's tangible book value. Last quarter their capital increased by $75. This moved the green bar up from $23.3B-$26.1B to the current range of $23.9B-$26.8B. It also moved the red line up from $31.12B to $31.85B. Assuming that they actually hit profitability as guided for in Q4, I expect them to add another $100M-$120M in total risk-based capital next quarter, which would move the green bar and red line up further. They are guiding for an additional $300M-$500M in tangible book value growth next year as well, which would boost the top of the green bar up another $2.4B if they hit the low end of that guidance.
Tech Platform Transparency
There was a question that I actually put on the Investor Q&A that I expected them to answer about the tech platform that they didn’t actually answer. It was the top voted question by both total votes and shares represented that was not answered (humble brag). Fortunately I asked Chris the same question on the phone and he did answer it there. The question was:
Can the company give more public KPIs on the tech platform like transactions processed, dollars processed? Also breakdown number of products by type (PRP, banking core, Konecta, etc.) and a breakdown of clients by business type (neobank, B2B, platform, large FIs, etc.)?
Chris’ answer was that they plan on adding additional detail, metrics, and reported KPIs as the tech platform becomes a bigger part of the revenue share. While that might not have been the answer I was hoping for (which would have been “Yes, and we’ll start reporting this quarter”), it is something of which they are cognizant. He talked about how they want to give investors and analysts more information to be able to better understand that part of the business and be able to model it better.
Will Deposit Growth Limit Member Growth?
Finally, we also talked about the point which will come some time in the future when deposits are enough to cover the entire loan book. I asked if they would restrict member growth once their deposit base was large enough to cover all of their assets. He said there are plenty of assets they could balance against those incoming deposits that would give a healthy return and keep them above their minimum capital ratios. He also said in that hypothetical scenario they would not limit member growth just because deposits are outpacing the growth of the loan book.
This was important to me because I’m more concerned with product adoption and secular growth than maximizing profits. One of the reasons I invest in SoFi is specifically because they think about this industry different than incumbents. SoFi doesn’t view members as tools to maximize return on equity (ROE). This is part of the reason that LendingClub, for example, is basically flat on the deposit front recently. Their loan book isn’t growing, so they see no need to add deposits, which cost them 4.65% APY.
While it is true that deposits are a liability for a bank, a sticky banking customer that comes along with them, to me, is way more important than limiting accounts and customers to maximize ROE on the existing loan book. As long as the LTV/CAC is high, I’d rather have additive profit, even if it means sacrificing a little bit of ROE to get there. As SoFi continues to expand their product suite, additional members in the ecosystem will be a boon to future cross selling. This is evidence that SoFi has their priorities straight. The long term value of adding people into the ecosystem is more important than maximizing returns in the short term.
Thank you to everyone who has become a paid subscriber. I am now at 70% of my Tier 2 goal. I plan on ramping my content output as paid subscribers increase. Here are the content tiers:
Tier 1 – About 2 articles/month on the Substack. This is the level I’m comfortable at for now with what I’m making.
Tier 2 – I’ll start a YouTube channel and do at least one video every other week.
Tier 3 – A minimum of 4 articles/month on the Substack
Tier 4 – Add a weekly Twitter Spaces where I can engage better with followers and answer questions people may have.
Disclosures: I have long positions in SOFI and LC.
The information contained in this article is for informational purposes only. You should not construe any such information as legal, tax, investment, financial, or other advice. None of the information in this article constitutes a solicitation, recommendation, endorsement, or offer by the author, its affiliates or any related third party provider to buy or sell any securities or other financial instruments in any jurisdiction in which such solicitation, recommendation, endorsement, or offer would be unlawful under the securities laws of such jurisdiction.