Whats new in the SVB collapse?
Introduction
$\rm\LARGE{N}$either bank runs nor regulatory failures are surprising.
Before the invention of deposit insurance and the establishment of FDIC in 1913, bank runs were quite prevalent. The creation of the Federal Reserve as a lender of last resort and FDIC as a deposit insurer were aimed at the ‘prevention of contagion’ and not necessarily to prevent bank runs.
Wait a minute; bank run causes contagion. - what am I missing?
Well, not necessarily - a bank run is not a sufficient condition for contagion. FDIC did not protect uninsured depositors for most bank runs as the contributing factors were idiosyncratic and not system-wide. But when the bank run signaled contagion risk, regulators did whatever it took to restore confidence in the banking system – for instance, FDIC backstopped the uninsured depositors and creditors in the 1984 Continental Illinois bank run because the bank was considered ‘too big to fail’.
Regulatory failures, too, have time and again plagued the banking sector and also the real economy. From the Savings and Loan crisis of the 1990s to the Subprime Mortgage Crisis of 2008, regulators have often failed to identify and preemptively solve a simmering issue, disrupting the real economy. A notable example of financial sector turmoil affecting the real economy is the 2007-2008 global financial crisis. The collapse of the US housing market and the ensuing banking crisis led to a severe global recession, with significant consequences for employment, trade, and investment worldwide.
So, is there anything new about the SVB bank run? This blog post aims to investigate the novelty aspects of the SVB bank run.
I ask: What was distinct and unprecedented in the SVB bank run?
Focusing on this bank run’s distinctive and novel aspects, I intend to bring the new beast under the spotlight.
Claim
My claim is that there are two key novelties:
i) Speed of withdrawals and the role of modern information dissemination technologies
ii) Emergence of new temporal category “too IMPORTANT to fail”. And yes, it is IMPORTANT, and not BIG
With a better understanding of the aforesaid novelties, we would be better informed in our responses to the crisis. In the next section, I will discuss each novelty in detail.
The Novelties
1. Speed of withdrawals and the role of modern information dissemination technologies
On Friday March 17, more than $40 Billion was withdrawn by the depositors of SVB in less than 24 hours. While the idiosyncrasies of the SVP depositors population —high concentration of >250k depositors, startup deposits well networked via the Venture Capital Firms, and sophisticated nature — certainly contributed to the firehose withdrawal, the speed of information dissemination and withdrawals were nonetheless unprecedented.
And I believe that the lightning-fast withdrawals in the wake of panic will now be a phenomenon, thanks to the advances in information dissemination technologies. Never in the past had everyone had access to the same news at the same time, that too coming directly from the ‘Principals’. The news ‘agents’, a.k.a the established media houses, are becoming less important each day as the news consumer can follow the news creator herself (the ‘Principal’) on various platforms. Holding or letting go of the information is no longer in the strategy set of media houses, and thereby they can not regulate information flow anymore. In this new state of the world, information emanating from credible ‘Principals’ reach directly to the consumer, and such information is deemed least imperfect and least incomplete by information consumers, and hence most actionable; such information gets the maximum audience, which in turn can trigger mass-scale individually rational behaviors. Unsurprisingly, someone like Peter Thiel can indeed trigger runs. I must clarify that by ‘trigger’, I do not insinuate ‘cause’ - there is a vast difference. The German invasion of Poland in 1939 is generally considered the trigger for World War 2, but the underlying causes of the war were many - from the Treaty of Versailles, German humiliation, failure of the League of Nations, the rise of totalitarian regimes, etc.
In an erstwhile informationally inefficient market, all else equal, even an insolvent bank could have gotten lucky and prevented bank runs if the uninsured depositors’ population was pretty diverse in its information access, and withdrawals weren’t possible with a click on a smartphone. If different people are reading different newspapers, and each newspaper has its own spin on the news, then bank runs can be delayed. In the interim, bank can raise capital to solve the solvency problem. Accomplishing such an outcome in 2023 and beyond is almost impossible - although SVB tried with Goldman Sachs on their side for capital raise and persuading Moody’s to defer the rating demotion. But as we all know, they failed.
As a side note, I wonder if there could be a scenario where the modern information infrastructure can prevent bank runs. The Nobel prize-winning Diamond-Dybvig model mathematically demonstrated the existence of multiple bank-run equilibria even when the bank was perfectly solvent, a.k.a. Panic-based bank runs. Panic-based runs happen due to misinformation and general negative sentiments not backed by fundamentals. Faster information dissemination could deter such a run as any misinformation would be called out in an evidence-backed manner and disseminated swiftly across the entire system.
The open internet and social media are formidable information dissemination tools that will be instrumental in coordinating collective action. However, the nature of such coordinated action would always be determined by individual rationality rather than the larger social good. Therefore, information dissemination would put withdrawals on steroids in case of a solvency-based bank run but will reduce the probability of panic-based runs by calling out misinformation.
2. Emergence of new temporal category “too IMPORTANT to fail”
Regional banks like SVB are not considered ‘systemically important’. They are subject to relaxed capital, liquidity, and risk management requirements and have lesser regulatory oversight. The fundamental idea of such triaging is to separate ‘too big to fail institutions’ as their size and connectedness are such that any problems with them could spiral into a contagion. Regulators employ stricter rules for such systemically important banks, but there is also an implicit commitment to protect them should they be in a mess. But if SVB was not a systemically important bank, why did FDIC end up rescuing the uninsured depositors? They are certainly not a ‘too big to fail’ bank. Shouldn’t the moral hazard argument for not backstopping SVB depositors dominate the regulators’ response? After all, the fear of bank runs is necessary to incentivize bank managers to manage their risks better. It does not make sense why regulators backstopped the uninsured depositors in the SVB case, or does it?
I believe that the regulators’ response was correct, as there was a real risk of contagion. SVB bank run was caused due to system-wide stress and wasn’t purely idiosyncratic. It is interesting how people arguing over the reasons behind the bank run miss the blatant signal provided by the regulators themselves of what they think was going on with SVB. The fact that FDIC backstopped a non-systemically important bank’s uninsured investors is a telltale acknowledgment by the regulators that SVB’s collapse has more to do with a system-wide issue than its firm-specific maladies. Bloomberg estimates suggest that US banks are sitting on unrealized losses of $620 Billion on held-to-maturity securities, primarily due to unprecedented interest rate hikes.
Granted, SVB did not manage its interest rate risk but you do not characterize a COVID death as a diabetes death, even though diabetes is an established co-morbidity for COVID - simply because the counterfactual is that the patient was not dying of diabetes in the absence of COVID.
Therefore, even though SVB is not ‘too BIG to fail’, given that the culprit is a steep interest rate hike that has led to massive unrealized losses across the banking sector, regulators can’t afford to let the uninsured depositors go bad. SVB is indeed ‘too IMPORTANT to fail’ bank at the current time. Had the bank run happened solely due to SVB-specific issues, there wouldn’t have been any contagion risk, and bailing out uninsured depositors would have been preposterous. Regulators did the right thing to mitigate the contagion risk, but by backing the depositors in full, they also acknowledged that the fundamental reason behind the stress was a system-wide phenomenon, and SVB being the weakest link, was the first to fall.
My hope for the future
While I prefer to refrain from scapegoat hunting, it is essential to recognize that the extant stress in the financial sector is, inter-alia, a failure of economic theory.
Let me explain - the steep hike in the federal reserve interest rate is the primary culprit, and with >5% federal funds effective rate, why would anyone want to keep their money in bank deposits that pay near zero nominal return, let alone inflation adjustment? Flight of deposits is not and should not be surprising at all; this will continue to happen irrespective of the risk management capabilities of the bank owing to the wide arbitrage.
This huge arbitrage emerged due to Jerome Powell’s unprecedented credit tightening, which in turn was because the federal reserve mischaracterized inflation as ‘transitory’ and was late in their inflation-curbing efforts. Such ‘transitory’ mischaracterization is a failure of our extant economic theories, assuming Federal Reserve is largely apolitical. Our economic theories have been good in ex-post analysis but have been terrible in predictions, which was proven once again. Furthermore, policy interventions should be subject to a higher level of testing before full deployment. Regulators should develop mechanisms to simulate and experiment with such extraordinary interventions - it will take a lot of creativity and innovation. Still, given the wide ramifications of these policies for financial and real sectors, we feel the need for field-tested policies.