All market data is from FactSet and FE Analytics unless otherwise stated. Past performance is not a guide to future returns.
Regular readers will no doubt be surprised to see us weighing in on the banking crisis, given Evenlode’s exclusive focus on companies with high and persistent unlevered return on investment capital. This has always precluded bank ownership by our strategies, but given the materiality of the situation, we felt it important that we help our clients understand our thinking and the relevance of the events to our portfolio and our investment philosophy.
The crisis was started by the bank run at Silicon Valley Bank (SVB) which we tentatively blame on one of our perennial bugbears, corporate underspending. Management compensation was heavily linked to return on equity, which drove 100% of annual bonus and 50% of the long-term incentive programme. The resulting focus on earnings management appears to have led to two fatal decisions. Firstly, in 2020-2021 SVB’s client base of start-ups and venture capitalists experienced a massive surge of inbound funding, and the bank chose to buy long-tenor securities with this gusher of $128bn in new deposits (end 2021 vs. end 2019). The decision to buy long-duration rather than short-duration bonds yielded an additional $18m in income or 1% of FY21 net profits (according to Bloomberg reporting[i]), at the cost of building in a tremendous duration mismatch in the balance sheet. Secondly, management chose to stop hedging their interest rate risk at the absolute trough of US rates in 2021-2022, precisely the wrong moment. The motives for doing this aren’t clear to us but we wonder if earnings were part of it. This problem was exacerbated by the exceptionally concentrated and uninsured depositor base: the Wall Street Journal[ii] estimated that 90% of its $173bn deposits (as of end 2022) were held in just 37,000 accounts, for an average balance of $4.2m per account, compared to the median US account balance of $5,300 according to the 2019 Fed Survey of Consumer Finances[iii]. As many will be aware, US deposits have been advertised since 2008 as being insured up to $250,000 by the Federal Deposit Insurance Corporation (FDIC).
This combination of factors brought down SVB as the rapid rise in interest rates in 2022-2023 left their large book of US treasuries and agencies (mortgage-backed debt issued by Fannie Mae and Freddie Mac) worth about 17% less than the carrying value when marked to market (as of 31 December 2022), while deposits had naturally retained their full value. While the securities book was still ‘money good’ (it will pay out 100 cents on the dollar, its coupons and principles guaranteed by the US Government) interest rate rises meant that investors could replicate that cashflow stream at much lower cost using newer issuance, so prices had to come down to set those cashflows to equal value. These unrealised losses were more than SVB’s common equity capital. When news got out that credit rating agencies were on the case, those 37,000 accounts with deposits notionally at risk all ran for the exits.
The losses on SVB’s $92bn security book, while large, were readily quantifiable. This is not at all true of the $74bn loan book. Had the loan book been diversified and well underwritten, we’d expect that other banks would have purchased part or whole of SVB on the weekend of the deposit run. The US authorities’ inability so far to find a buyer for the book indicates concern about underwriting quality. SVB is now widely known to have offered below-market rates and to have lent against collateral which is likely to be highly levered to startup valuations. In happier years, SVB enjoyed substantial fee income from sources such as equity warrants in startups packaged in with loan deals, from investment banking to VCs, and access to prestigious VCs for its $10bn fund of funds. There is a natural concern that routine lending activity was used as a ‘loss leader’ to drive these attractive but more cyclical fee income streams. As this goes to press, US authorities have finally found a taker for the loan book at a 23% discount worth $16.5bn, which will be borne by the FDIC and its members.
We find this case study interesting as it highlights some longstanding Evenlode priorities which are readily applicable to our very different investment universe. Firstly, executive alignment and willingness to sacrifice near-term profits really matter. Secondly, the importance of sound capital allocation, disciplined risk management, and prudent leverage: all of these are just as applicable to non-financials as they are to banks. Thirdly, diversification: SVB had a singular focus on a cyclical end market, and consistently doubled down on that focus as it grew its loan book and depositor base. Lastly, pricing power: SVB enjoyed close relationships with the venture community, but still had to offer loans priced well below its competition to keep their business. This highlights the importance of investing in industries where customer decisions are not fundamentally driven by price.
At this stage it is hard to say if the banking situation marks the start of a deterioration in the wider economy. SVB was as we said uniquely focused on one region and one industry and had the highest proportion of uninsured depositors in the US system according to S&P Global[iv]. The same is true for the other two recent US bank failures, Signature (New York real estate and crypto) and Silvergate (crypto), while Credit Suisse had been suffering from massive deposit outflows for a long time. The other troubled US bank, First Republic, was in many ways similar to SVB in that it was highly focused on proving a niche service to affluent Californians, specifically the provision of very large mortgage loans (‘jumbo non-conforming’, in banking jargon, as they cannot be sold to Fannie or Freddie). We do not pretend to know if these failures are an isolated disaster or if they are harbingers of wider problems.
Come what may, though, it has some interesting consequences for our portfolio. Firstly, it’s clear that the financial world is effectively ‘underinsured’ after a decade of near-zero interest rates and three decades of declining rates. A generation of financiers have learned that hedging interest rate risk is a waste of time. The reversion on this tide should be a positive tailwind for the derivatives exchanges CME and LSE (both held in the portfolio), which dominate the markets for interest rate futures and swaps respectively. Secondly, the cost of capital for many start-ups had been lowered to near zero by low rates which allowed investors to make a directional bet on technological valuations. We had already spoken to many companies who were interested in acquiring valuable intellectual property from start-ups, but abundant capital allowed their owners to hold out for exorbitant prices despite being unprofitable, or even pre-revenue. Our expectation is that asking prices are on their way down.
Finally, we do have indirect exposure to the US smaller banking sector via our holding Jack Henry, which sells software and payments services to them. Its share price has reacted negatively, for two reasons. One is the concern that bank IT spending will be pushed out given banks’ desire to build capital via retained earnings. This is natural and not a major concern, as demand is being deferred not destroyed. The other is that government aid will disproportionately help the larger banks, which will hurt Jack Henry as it focuses on ‘community’ banks, defined as banks with less than $50bn in assets (SVB was many multiples of this size). While this is a risk, and not a new one, our position size accounts for this. We believe it is politically very unlikely that the US government will want to buttress its larger banks at the expense of the communities, particularly given the much more dispersed risk profile of the community banks.
Along with duration, the recent crisis also revealed another mismatch – that between antiquated banking IT infrastructure and modern consumer communication technology. The Wall Street Journal reported that the proximate cause of SVB’s failure was the Fed’s inability to send large payments after 4pm Pacific Time on a Friday[v]. This was fine in the 1970s, but with the client base communicating via Twitter and text message, it’s now possible for $42bn of deposits to be pulled in one day. Most private banks have similarly aged infrastructure. Jack Henry is gradually shepherding its clients in the painful transition from the obsolete technologies of mainframes and COBOL language to cloud-hosted, flexible software solutions. This ‘deferred’ spend by banks now looks more urgent than ever.
We continue to watch the situation and believe our portfolio remains well placed to weather the potentially volatile macroeconomic situation in 2023.
We look forward to being in touch again soon.
Chris (E), James, Cristina and the Evenlode team
27 March 2023