Longer DD with many numbers, all data from 10-Qs for Q2 2023 or 2022
Interest Income increases its share of total revenue, given large portfolio expansions (+33% loan principal, +108% interest bearing assets), slightly higher NII. Non-interest income, however, seems more organically grown. Except the line “Loan origination and sales” (used to be 43% of net revenue in H1 2022, of which 70%), which captures fair value adjustments of loans originated or held in the relevant period, as well as gains on sales. In H1 2023, Gains on sales were aggregate $6.34m, while on page 17 cumulative fair value adjustments were shown as $260m (difference in cumulated fair value adjustments Dec-22, Jun-23) in the same period. There is another table on p 35, which indicates that impact on earnings of those adjustments were also captured in “interest income” and others were $91m for loans and $115m in total for the same period. These numbers don’t add up – where does the rest come from? Interestingly, despite way higher sales volume in 2022, gains were actually down quite a bit, despite them suffering some losses selling home loans in 22.
Business Model problems:
A large part of SoFis valuation is the unplanned high interest income, which is generated by forcefully holding loans for longer terms on the balance sheet (sales volume plummeted 86% yoy). The high percentage of charge-offs (3% on personal loans (1.25% in 2022) and 18% for credit cards (11.4% in 2022)) shows that underlying loan quality can’t possibly be that great. Most likely, those unsecured loans were advanced for refinancing (credit cards, pay now pay later, etc.) debt which very well can become toxic should financial conditions continue to tighten or economic conditions to start shifting. BAC’s credit card segment shows <1% charge-off rates on 10.68% yield, which is bizarrely lower than SoFi’s personal loans. For comparison, BAC’s yield on personal/consumer loans was a mere 4.35% in H1 2023. Given those circumstances, one might wonder how shitty those loans actually are…
What’s more, the market currently thinks that the re-commencement of student loan repayments is going to boost SoFi’s business, as many previously federal borrowers are going to refi with them. Just two problems; first, SoFi currently earns on average 4.93% interest on student loans. Given that SOFR and now 2yr treasuries are already at 5%, this should mean that fair value adjustments of student loans should actually be negative, when they aren't on their balance sheet. Compare that to SLM, which charges 7% on federally backed loans, and 10%+ on average on private loans (Maybe SoFi's loan quality is better or the originated more during pandemic years, but they hardly offer any data on their borrowers – which SLM does). Moreover, SoFi leverages approximately half its student loans via warehousing at a 564-691 bps rate, which does not make sense whatsoever. The only explanation is, that they have a very segregated portfolio of highly yielding (prop. 9%+) risky loans and a few high-quality low yielding ones, otherwise back-leveraging does not really make sense at higher borrowing costs than yields. Noted, that they’ve reduced their drawable commitment and got that facility probably in 0% SOFR times. But even if assumed, that the other half is funded by deposits costing 3.8-4.6 %, they really don’t make any money. Looking ahead, funding costs (warehouse and deposits) are likely to keep increasing at least for a year, while chargeable interest rates on student loans will be effectively capped by a 5.5% federal rate for eligible people. One might argue, “Well what about falling rates, having loans yielding 4.9% is then really good”, well they're all going to refinance, same old problem back when we had wonderful CDOs etc.
What‘s your thoughts on this – currently have no exposure, because I don‘t really know whether the above is actually valid.
Leave a Reply