SoFi Or Discover: Which Digital Bank To Buy In 2023? (NYSE:DFS)

The grey line between FinTechs and legacy digital banks has been blurring in the last 18 years. The likes of SoFi Technologies (NASDAQ:SOFI) and LendingClub (LC) obtained their banking charter in the last 18 months and have pivoted to retaining unsecured consumer lending assets on their balance sheet as well as competing aggressively for deposits.

Investors sometimes forget that this is all been done before and is very similar to the existing business model of incumbent digital banks such as Discover Financial Services (NYSE:DFS).

In this article, I will compare and contrast the DFS and SOFI and conclude which is the better investment for 2023.

Background

DFS

Discover Financial Services is a digital banking and payment services company. Discover is the United States’ third most popular credit card and fourth-largest payments network by purchase volume. The company issues the Discover card, America’s cash rewards pioneer, and offers private student loans, personal loans, home loans, checking and savings accounts, and certificates of deposit through its banking business.

Discover also operates the Discover Global Network comprised of Discover Network, with millions of merchant and cash access locations; PULSE a leading ATM/debit network and Diners Club International, a global payments network with acceptance around the world.

SoFi

SoFi is a financial services platform that was founded in 2011 and initially focused on the private student loan market by providing student loan refinancing options. Since its founding, SoFi has expanded its lending strategy to offer home loans, personal loans, and credit cards. The largest loan assets comprise unsecured personal lending. The Company conducts its business through three reportable segments: Lending, Technology Platform, and Financial Services (including deposit taking and other non-lending financial products, such as money management and investment product offerings).

SoFi Technologies division (enabled through the acquisition of Galileo and Technisys S.A.) operates as a platform-as-a-service for a variety of financial service providers, providing the infrastructure to facilitate core client-facing and backend capabilities, such as account setup, account funding, direct deposit, authorizations, and processing, payments functionality, and check account balance features.

SoFi’s operating divisions are illustrated below:

It is important to note that the majority of the “contribution profit” (a non-GAAP measure equal to revenue less direct expenses) is derived from the lending division. This is currently SoFi’s core engine of revenue and ultimately profits.

The majority of the organic revenue growth (excluding the Technisys S.A. acquisition) is driven by the Lending segment. And the growth in the Lending segment is predominantly driven by growth in the balance sheet from the retention of unsecured personal lending as well as hedging income negating the impact of rapidly rising interest rates.

The Balance Sheet

DFS

Discover’s loan assets predominantly comprise Cards, Student Loans, and Personal Loans as represented below:

Total loans have increased 17.2% year-on-year and the NIM of the loans is ~11%.

DFS holds these loans to maturity as such it reserves loans in line with the lifetime expected methodology as per CECL accounting standard. Note DFS is focused on predominantly prime customers with FICO scores >660.

SoFi

SoFi’s loan balances as of Q3’2022 are shown below:

SoFi total loan assets are currently at ~$11 billion and have grown by over 80% year on year driven by personal loans or in absolute terms by over $5 billion (compared to DFS which has grown by 17.2% year on year and ~$15 billion in absolute terms).

SoFi’s loan yields and FICO metrics are reflected below:

However, important to note that the majority of SoFi’s assets are classified as Held-For-Sale from an accounting perspective. This means that SoFi does not reserve for expected lifetime losses (as per the CECL methodology) which I perceive as particularly risky with a looming recession approaching.

This issue was debated at length in my earlier article (with many of the bulls pushing back on my argument), but I remain firm in my view that SoFi is unlikely to sell the personal loans in the near term given that investors’ demand is clearly waning (see LC’s pre-announcement warning) and importantly if SoFi sells these loans, the associated NII will dissipate which essentially undermines the 50% growth narrative.

Finally, whilst SoFi has plenty of capital to fuel growth, its capital ratios in the banking subsidiary are reducing rapidly. The key Tier 1 Leverage capital ratio has reduced from ~34% in Q3 to ~17% quarter on quarter. The math suggests that growth in assets will be constrained unless SoFi starts generating significant organic GAAP earnings in the banking subsidiary and/or recapitalize the subsidiary (and unclear how much capacity there is in the other divisions to recapitalize the banking subsidiary).

This issue may come to a head if a recession ensues and SoFi will be forced to recognize material loan losses. As noted above, SoFi is completely unreserved and as such if the tide goes out, SoFi could be seen to be swimming naked.

Valuation

DFS

DFS price to tangible book value is represented below:

As you can see, DFS is currently trading just above 2x tangible book value. In normal macroeconomic conditions, it trades at 2.5x to 3x. However, it is essentially an unsecured personal lender, so the share price drawdowns are sharp when economic concerns (such as the recent pandemic or economic downturns) surface. In fact, I periodically purchase puts on DFS as a quasi-macro hedge when I am concerned about recession prospects.

Another useful perspective is looking at PE ratios. Currently, Discovery is trading at a ~6.5x PE ratio given macro uncertainties. The E part also incorporates incremental CECL reserves, so the real PE is likely much lower. The stock currently signals an impending recession but obviously a mild one and nowhere near the 2008/2009 global financial crisis. In normal times, one would expect DFS to trade at a range of 9 to 12 PE ratio.

Capital allocation for DFS and share buybacks are key for total shareholders’ return and as you can see from the below chart, DFS has been reducing the share count consistently even though it currently suspended its share buyback program.

SoFi

I tend to think of SoFi as two separate businesses comprising the banking business and the technologies division. In fact, I believe it makes absolute sense for the management team to spin off SoFi Technologies to a separate listed entity as there are very limited synergies in place (that couldn’t be solved by a tight Service Level Agreement). SoFi banking could then be run effectively like a bank whereas the technologies division can attract a completely different class of shareholders. At the moment, it is neither here nor there which I think leads to a mini “conglomerate” discount and frankly only causes confusion for investors.

The below chart illustrates the price per tangible book value of SoFi:

As can be seen from above, SoFi’s price to tangible book value has been higher than 6 at its peak, whilst currently at 1.72x. As noted above, it is very challenging to ascribe a credible valuation for SoFi.

For a start, the banking subsidiary is growing quickly in recent quarters but it is not clear (as yet) whether this growth is sustainable given capital constraints and whether SoFi will be able to compound the book value rapidly which is imperative for its success. Banks’ profitability is directly proportionate to the size of their book value.

It is also not clear to me whether SoFi’s strategy is to hold loans for much longer (maturity) or sell these after 6 months. If it intends to sell the loans, then it will lose the benefit of associated NII. If it intends to hold to or near maturity, then by right, it should also book a CECL provision.

The economic incentive to maximize profitability suggests that SoFi should be holding the loans to maturity on its balance sheet. This is far more profitable than selling the loans and I suspect that is where the business model will land, otherwise, SoFi will not deliver strong profitability.

Additionally, in this type of business, scale matters. DFS holds ~10x loan assets compared to SoFi and runs with an operating efficiency of ~40%. As a comparison, $LC with only $3.6 billion of unsecured personal lending is generating an efficiency ratio of ~60%. Again, it is not clear what unit economics SoFi is generating but with ~$7 billion of unsecured lending (and no CECL or tax payable), I would have expected it to generate strong profitability on a GAAP basis (which is currently not the case). This suggests that its cost structure is somewhat bloated and the ROE generated and book value accretion may not be as remarkable as some would desire to credibly justify a high multiple book value or PE ratio.

In short, there are many unanswered questions at the moment with SoFi, and will need to wait for the next few quarters for some of these answers to be revealed.

Final Thoughts

In my last article on SoFi, one of the bulls invited me to highlight any items that the SoFi management team promised and did not deliver. My response has been “that I am more concerned about what they delivered that they didn’t promise!”.

I am not sure that the grande plan was for SoFi to become a predominantly unsecured personal lender, it looks to me like it evolved over recent times. Do not get me wrong, I think that the unsecured personal lending space is very attractive but strategies need to be tightly executed.

Banks in this space need to be focused on operating efficiency and credit risk management. The returns need to be very high (e.g. >25-30% ROE on a post-tax GAAP basis) to compensate for the risks. Scale matters and growing assets and book value is a key enabler to achieving this scale.

DFS has been doing this for decades and has a clear playbook to manage through consumer credit-driven downturns. SoFi’s credentials in credit risk management remain very much unproven to date. SoFi is a very late entrant into this space yet has been rapidly building a large balance of unsecured lending with a recession looming (and no reserves). This could end in tears if a serious recession ensues.

Whilst I expect SoFi share price to recover strongly if a soft landing is confirmed, I do not find the risk/reward equation sufficiently attractive to rate this as a buy. At the same time, DFS with a PE of ~6 to 7 historically has been a strong buy. I may initiate a long position in the near term.

On the flip side, I do believe the SoFi Technologies division is a much more unique and attractive business model in the long term. I wish that SoFi would spin it off as a separately listed company as a pure play on the “AWS for Finance”.

I still remain largely on the sidelines when it comes to SOFI and I much prefer LC as a pure play on a growing unsecured lender. LC is profitable on a GAAP basis, has market-leading cost efficiency, proven credit risk credentials, and currently trading at a discount to tangible book value.