Andrew Zeng, Stanford University
On March 10th, 2023, Silicon Valley Bank, a premier financial institution in the startup world, suddenly collapsed. The federal government soon seized its assets, and as its clients tried, increasingly in vain, to withdraw their deposits, panic soon set in. Wealthy, platformed individuals began demanding that the government bail out the bank and compensate them, while news anchors and financial analysts predicted a ripple effect that would irreversibly end the dominance of Silicon Valley.
And as those analysts struggled to present a narrative of what had driven the bank’s collapse, they almost universally settled upon one theme. In an interview with CNBC, Keith Fitz-Gerald, principal of the Fitz-Gerald Group, lamented that “the greed and avarice that has long been present in Silicon Valley has come home to roost.” Shaun Davies, the research director for the Burridge Center for Finance, proclaimed the failure as “eerily reminiscent” of the 2008 housing crisis, a sentiment echoed by many other experts. And some, like the news anchor Sara Eisen, pointed their fingers at venture capitalists like Peter Thiel, who urged their startups to withdraw their money from the bank, as the principal culprits.
But while the crises of the first twenty years of the 21st century were driven largely by corporate greed, the collapse of Silicon Valley Bank teaches a wholly different lesson: that of the consequences of poor planning and the growing importance of risk management. It is true that Silicon Valley Bank’s leadership was recklessly greedy, and it is true that it may have yet recovered had Thiel and others not compounded the bank run. But it is critical to note that the bank was already in an incredibly precarious position owing to deep failures in risk management and future planning, even before the sequence of events that led to its collapse. Indeed, Silicon Valley’s gross failures in risk management were what ultimately caused its downfall.
The History of the Collapse of Silicon Valley Bank
The story of Silicon Valley Bank’s collapse begins with the deregulation of the financial sector in May 2018. In response to complaints that the Dodd-Frank Act passed in the wake of the 2008 crash was overly “one size fits all,” the Trump administration put forward the Economic Growth, Regulatory Relief, and Consumer Protection Act (EGRRCPA). The EGRRCPA set the definition of a “systematically important bank” (SIB) to include only banks with $250 billion or more in assets, with all other banks subjected to lighter scrutiny and liquidity requirements. The logic was that the collapse of a smaller, regional bank like Silicon Valley Bank would not lead to a crisis on the level of 2008, but would be relatively trivial.
The act had been forcefully lobbied for by the Silicon Valley Bank leadership, as less regulation would mean higher potential profits. In testimony to Congress in 2015, SVB CEO Greg Becker argued for the passage of the act by citing the bank’s “strong risk management practices,” having made “meaningful investments [in] our risk systems, hired additional highly skilled risk professionals, and established a stand-alone, independent Risk Committee,” before concluding that “SVB’s business model and risk profile does not pose systemic risk.” Thus, subjecting the bank to the oversight reserved for banks with more assets would be an “unnecessary burden” with no real purpose.
But while Becker and other executives purported to rate risk management highly, their subsequent actions betrayed that façade. The immediate impact of the passage of the EGRRCPA was a massive reduction in oversight over the bank, with some interesting side effects. The Dodd-Frank Act had previously required mid-sized banks to report a metric called the Liquidity Coverage Ratio (LCR). As described by the Bank for International Settlements, the LCR is “designed to ensure that banks hold a sufficient reserve of high-quality liquid assets (HQLA) to allow them to survive a period of significant liquidity stress.” Under the Dodd-Frank Act, Silicon Valley Bank would have been required to maintain a certain LCR to hedge against a bank run, but with the EGRRCPA, there was no longer such a constraint. By the end of 2022, Silicon Valley Bank had an LCR of just 75%, substantially below the threshold for larger banks. Such a low LCR meant that in the event of a bank run, the bank would experience significantly more destabilization (which it, of course, did).
In 2021, a Federal Reserve review of the bank found glaring problems in its risk management strategies, with subsequent investigations revealing that senior management did not implement basic risk management practices, that the bank was unable to self-identify weaknesses and proactively manage risks, and that its internal audit department was ineffective. Meetings between members of the Fed and the bank’s leadership revealed that the bank believed that higher interest rates would help its financial situation, which was clearly a false assumption. Nevertheless, while it increased its oversight of the bank, the Fed failed to sufficiently incentivize the bank to fix its problems.
In the midst of this process, the bank’s Chief Risk Officer (CRO), Laura Izurieta, stepped down in April 2022 and sold all of her stock in the bank. For the next eight months, the bank’s leadership, which had once proclaimed its belief in the importance of risk, left the position vacant, so that as the Fed increased interest rates, there was no CRO to correctly assess the likely effects on the bank’s financial situation. Additionally, the “stress tests” which Becker had once proudly cited to Congress went away as soon as they were no longer required by regulators—in the years preceding the crisis, the bank was never subjected to a single stress test.
All of this predictably came to a head when the bank, in need of fresh money, sold off some of the treasury bonds it held for below their market price. Treasury bonds are generally considered to be safe, but only if they are held for their full duration; if they are sold off before then, their owners will likely receive some money, but far less than the market value. When Silicon Valley Bank sold off its bonds, it looked like a sign of desperation; it had, after all, been willing to lose billions of dollars in order to raise some money. As a result, depositors who feared a crisis ironically caused one by withdrawing their money en masse. As the bank struggled against the liquidity crisis, Thiel and other entrepreneurs advised their startups to withdraw their money from the bank as well, and as its clients all attempted to withdraw, the bank collapsed.
Lessons for the Future
In late April 2023, the Federal Reserve released a report detailing its view on the causes of Silicon Valley Bank’s fall. The report identified failures in risk management as one of the key drivers of the crisis and made several recommendations. First, supervisors would have to consider “how to develop a more robust understanding of the risks banks face” by collecting more data on liquidity and running more tests. Next, supervisors needed to strengthen resilience, with many of the points proposed in line with Nassim Nicholas Taleb’s “anti-fragility” model. And finally, supervisors needed to act more decisively and avoid using consensus-based frameworks of leadership.
All of this is well and good—the report is exemplary of the transparency needed to properly respond to a banking crisis of such a magnitude, and it was especially commendable in its genuine attempts at soul-searching within the Fed itself. But it becomes problematic when one begins (as some analysts have) to view the recommendations made within the report as guarantors against such a collapse in the future. In reality, there will be a growing amount of uncertainty and disruption as the pace of technological advancement increases and as geopolitical risks continue to evolve. So while the measures proposed are a good start, much more rigorous methods are needed.
Because what’s scary about the Silicon Valley Bank collapse is that the individual risks that cascaded to bring it down seemed trivial. There are plenty of companies that go some time without a good CRO, and the bank’s avoidance of stress tests might not have been so bad had it been financially sound. But the combination of all of these risks led it to reach a tipping point in March 2023. Once that tipping point was reached, a previously trustworthy institution would meet an unceremonious end.
What specific rigorous methods are needed is not immediately apparent. It might involve some combination of guarding against “anti-fragility,” or integrating outside, non-expert voices into the risk management process, or commissioning the creation of dozens of future narratives in order to have a better chance of hedging against uncertainty. What is clear is that the collapse of Silicon Valley Bank has exposed the necessity for drastic improvements to risk management. Unfortunately, corporations will not be able to engineer themselves out of these risks using a simple step-by-step blueprint. Instead, it will become increasingly vital to undergo rigorous risk management and to invest much time and resources into planning for the future.
In the gripping 1995 French social commentary La Haine, Hubert Koundé recounts the story of a man falling off a skyscraper that parallels the fall of SVB. “On his way down past each floor,” he explains, “he kept saying to reassure himself: So far so good… So far so good… So far so good,” until the ground suddenly swells up to take him. “How you fall doesn’t matter,” Hubert concludes. “It’s how you land.” And as you fall past each successive story, perhaps things look OK, and the view is nice—but it’s only as the ground swells up under you that collapse becomes imminent.