Too innovative to fail

Subprime comes to Silicon Valley.

The cradle of innovation.

There are, generally speaking, two arguments in favor of the actions that federal regulators took to contain the fallout after Silicon Valley Bank failed last week.¹ 

The first is that a bailout was necessary to prevent a much wider crisis. In the time after SVB’s collapse and before its depositors were fully backstopped by the FDIC, we were teetering towards a full-blown financial panic. If the run on SVB proved “successful”—in that the people at the front of the outgoing wire queue got their money back and the people at the back of the line got blown up—there could be a lot more runs. 

Though we’ll never know if this would’ve happened,² regulators evidently deemed it both plausible enough and plausibly catastrophic enough to intervene. Notably, in this telling, it doesn’t matter who the players in this crisis were, or who’s fault the fire was. Everything was about to burn down—call the fire department now, and investigate for arson later. Questions about how we punish bad actors, avoid moral hazard, prevent future crises, or ethically justify the bailout—save the general principle that widespread suffering is bad and we should try to avoid that—are all secondary.³ Promote the general welfare; promote maximum employment; bail us out.

The second argument in favor of the rescue is more specific to Silicon Valley. Thousands of startups banked at SVB; when the bank failed, these companies could no longer access their money. If their deposits were locked up for days or weeks—or if they outright evaporated along with SVB’s equity—these companies couldn’t continue operating. They’d have to furlough employees, fire people, or shut down entirely. It would be, as YC CEO Garry Tan put it, an extinction level event for startups.

Though Tan and others making this argument often cited the number of people who’d lose their jobs because of this—with estimates ranging as high as 120,000—the argument clearly hangs on more than economic impact of those job losses alone. In January of this year, 3.8 million people quit their jobs, and another 2.3 million were let go. Tens of thousands of additional layoffs is, quite literally, a rounding error. Moreover, professional and business services’ layoffs and firings increased by more than 120,000 in January alone, from 417,000 in December of 2022 to 580,000 in January. If anything, the VC class has cheered that uptick in layoffs, including another 10,000 announced by Meta just this week, as a necessary and prudent correction.⁴

The acute danger of this crisis, then, wasn’t that some people would be fired, though that would indeed be bad for those people; it was who would be fired. As Tan suggests in an interview with CNBC, tech startups are our future—our future employers, our future engines of growth, our future competitive advantage, and the future ramparts of our national defense. And it’s our duty to protect our future.

It’s an evocative argument, akin to saying “our children are our future.” And who can argue against protecting the children?

But Tan’s appeal has one clear but unsaid implication—that Silicon Valley deserves special treatment. That we, as tech startups, are a special class of employer. That we matter more to the country and to the “general welfare” than other companies. That we, like our children, deserve extraordinary protection.⁵

Admitting this openly, however, presents a pretty serious complication for Silicon Valley. If tech startups (whatever that actually means) were deemed special, and were given, say, access to exclusive services, the typical social contract for that sort of recognition, as is the case for banks that are officially designated too big to fail, is regulation that prevents abusing those services—regulation that most of Silicon Valley’s boosters would say is anathema to innovation. Tech startups and the financial entities that fund them wouldn’t just be kneecapped by additional scrutiny; they, in fact, need to be less regulated to thrive. Companies like Uber and Airbnb needed to operate in legal gray areas; VCs need to be able to deploy capital quickly, and in huge concentrated piles around immature but promising innovations, to speedrun technologies through what would otherwise be a slow and ineffective evolution. 

Maybe! These things may all be true! Startups may be too uniquely important to allow them to die en masse; for an armchair economist like me, it certainly seems like the tech industry is—or at least, has historically been—a strategic asset for the United States (and I guess, if we want to be weepy about it, for humanity). And Silicon Valley may not be Silicon Valley if there’s more regulatory oversight. Nor is this necessarily a logically inconsistent position! It’s possible for Silicon Valley to get a regulatory pass and a government backstop if things go wrong. 

But that’s almost certainly a politically untenable position, which likely explains why it’s only been quietly promoted over the last week. At some point, though, we’re going to have to confront this tension more directly. We might be able to avoid it through this crisis, because it can remain sufficiently hidden behind the arguable necessity of bailing out SVB anyway. But the next crisis—say, a widespread collapse of new VC funds or a crash in venture debt that threatens thousands of startups at once—may be more localized. And in that case, “we’re too innovative to fail” may be the best case a desperate tech industry has. 

In this way, SVB’s collapse really could be Silicon Valley’s Lehman moment, not because it leads to some immediate catastrophe, but because it catalyzes a dramatic regulatory shift. Like Lehman, SVB demonstrated how seemingly narrow bets—on mortgage-backed derivatives in the former case, on private tech startups in the latter one—can get leveraged into a national crisis. Though SVB has been rightly blamed for mismanaging their balance sheet, that mismanagement was made possible by the mismanaged bets of venture capitalists (who, for their part, would say that mismanagement was made possible by the mismanaged policies of the federal government). In the end, though, it likely won’t matter that much who’s right; there’s political hay to be made by reacting to the crash that happened and not the one that, hopefully, didn’t.

Like Lehman, SVB also deeply wounds an already teetering mythology about the industry it serves. Before 2008, Wall Street was the prestigious destination for elite college graduates; when I was in college, it was, if not a noble career, a defensible one. Following the financial crisis, tech took over the top cultural spot: it paid well, partied hard, and, unlike finance, “made the world a better place.” But after years of high-profile scandals, the image of the heroic founder—the misfit, the rebel, the dreamer who’s pushing the human race forward—has been replaced by one of a con artist or copycat opportunist. And venture capitalists, rather than emeritus entrepreneurs, are seen as unsophisticated podcast bros⁶ who play swarm ball with billions of dollars. Just as the exact cause of the crash may not matter, neither will the correctness of these characterizations, so long as they have political appeal. 

I don’t know. Maybe there’s a regulatory reckoning; maybe not. But on the question of if there should be one, I can’t shake the feeling that SVB, also like Lehman, was wrapped up in a subprime crisis of its own. And unlike Lehman, which was the huge cloud of deadly carbon monoxide that killed half the miners, SVB could just be the canary.

The arc of Lehman's implosion is well known: Banks aggressively made loads to unreliable borrowers, packaged those loans up into “diverse” derivatives, and sold them back and forth to one another. But all the loans were correlated, the derivatives were neither safe nor diversified, and eventually, all at once, the whole thing went nuclear.

Squint, and Silicon Valley's financial scaffolding has a lot of similar contours. Nearly every aspect of starting and running a company has been streamlined, automated, or offloaded to a SaaS startup that promises to do it for you. You can file incorporation paperwork, manage your books (including, apropos of nothing, spreading your deposits across different nine banks to increase your FDIC insurance limit nine times over), and run your entire back office in a few clicks. Though these could be very good things individually—just like making some creative loans might be good for helping make financial services more accessible!—in the aggregate, they encourage a lot of risky behavior. Specifically, they make it easy to start a lot of startups that are bad ideas.⁷ They make it easy to start startups that are good ideas too, but good ideas usually take a lot longer to IPO than bad ideas do to blow up.

This could all be fine, if these companies never made it out of dorm rooms and garages. But many do, and with brand name backers: YC, for example, invested in more than 500 companies in 2022, up 1,000 percent from their storied years when they were incubating companies like Airbnb, Stripe, and Dropbox. VCs, for whom no KPI matters as much as who else you have a term sheet from, keep following on; Silicon Valley’s collective terror of down rounds keeps ratcheting companies’ valuations higher. And the whole thing, like the subprime mortgage market, gets layered in “derivatives”—startups selling to each other, funds investing in founders who invest back into funds as LPs—that hide how concentrated everyone’s exposure is around a bunch of risky assets. 

This music, obviously, can stop. And when it does—as it did in 2000, and 2008, and seems to be now—maybe companies die, capitalism works as intended, and we all move on. But today’s widening crisis at least demonstrates that what happens in Silicon Valley’s financial institutions doesn’t stay in them.⁸ If we want to keep the tech ecosystem’s innovation engine going—if it really is too innovative to fail—we also need to make sure its financial engine doesn't become too risky to exist. 

---

Fair ways to fail

There was actually a third reason that people said that the SVB bailout was necessary: People and companies losing their deposits wasn’t fair. Startups who banked at SVB didn’t do anything wrong, unethical, or even (we all thought) risky; all they did was open a checking account. It wasn’t their fault; they were victims of mismanagement and an economic shock; they deserve to be made whole. 

Ok, yes, on one hand, these points seem obviously true.⁹ On the other hand, I think it’s worth at least briefly reflecting on why they seem so obviously true in this case, and not in others.

To state the obvious, unfair things happen to people all the time. Notably, tens of thousands of people have been laid off by large tech companies. They didn’t do anything wrong, unethical, or even risky—many of them took jobs at stable (we all thought) giants like Google, Facebook, and Salesforce. All they wanted was a job. They then lost it; it wasn’t their fault; they were victims of mismanagement or the same economic shock.

This is generally seen as natural and normal. Yes, people are supposed to get fired for poor performance, but sometimes, bad things happen to their employer.

However, companies going under because of SVB’s implosion is perverse. Startups are supposed to be fail because they build bad products or spend too much on ads or commit fraud, but not because bad things happen to their bank. And when that happens, we, apparently, have a moral obligation to correct it.

It’s striking how parallel the two situations are, and how differently so many people in tech respond to them. Is the argument about fairness just self-serving rhetoric? Is it because there’s a clear way to reverse SVB's collapse but not layoffs? Is it because layoffs are fairly common occurrences, and we stop seeing things that happen all the time as unfair?¹⁰

I don’t know. Think about it though. If you’re convinced that this was the series of unfortunate events that demands extralegal actions to correct—to make fair—why, among all the unfair tragedies that befall so many people, is this one special?

1. This story has been told and retold dozens of times by now, and I—a data blogger and accidental CTO who didn’t know what AFS or HTM meant until last week—won’t be able to summarize it nearly as well as they already have.

2. The banks that people were worried about got hammered by financial markets on Monday, which some people might say is proof that more runs were imminent. But it’s also possible that they got hammered because of Sunday’s announcement, not despite it. Don’t bail out SVB, and regional banks are overvalued, because they might blow up next. Bail out SVB, and regional banks are overvalued, because they’re about to lose their regulatory hall pass.

3. You could argue, as some have, that the mess was made worse by Very Online venture capitalists who, in their zeal to get the fire department to come, ran around lighting more fires. The VCs would say they were simply pointing out fires that were already lit. Unfortunately, if the world really was on brink, it doesn’t really matter who’s right. Either way, the most important thing is that the firefighters show up.

4. Of course, it’s possible that the different reactions are lazy hypocrisy and self-interested reasoning, and “the workers” are a rhetorical volleyball to be thought of or ignored, depending on how their interests align with that of my portfolio. It is, after all, difficult to get a man to argue on Twitter for something, when his carried interest depends on not arguing for it. 

5. Some may say, no, we just want the same fair and equal treatment that any industry would get in this situation. To which I’d ask, do you think Tan has a point? Most people—including me!—would struggle to see his argument as a complete non sequitur. 

6. According to Forbes, four of the top 51 VCs in the world are women. Three of the top 51 are named Peter.

7. There’s also very little personal risk in starting a company. If it goes well, you win big. If it doesn’t, you lost someone else’s money, made a lot of connections with important people, and will probably have an easier time getting your next job or raising money for your next startup. As I said before, this dynamic means that “hundreds of startups exist to be successful and to solve a problem, in that order. They’re trophies, status symbols, learning experiences, lottery tickets, lifelong dreams, and, somewhere down the list, a means for making a particular product.”

8. To reiterate, the point here isn’t if Silicon Valley caused the SVB to collapse or not. The point is that SVB was a bank that almost exclusively served the Silicon Valley ecosystem, and its collapse is having effects that go well beyond that ecosystem. 

9. Or not! Some people disagree! They say that startups put too much money in a single bank, didn’t vet that bank well enough, and ignored basic practices of financial governance. Which, to me, warps the concept of fairness beyond recognition. Just because someone could’ve conceivably avoided some bad thing doesn’t mean that bad thing is fair.

10. I would reflexively say that someone getting shot by a stray bullet was treated more unfairly by the universe than someone who died of cancer, but I’m not sure that makes any sense.