A storm in a value store

Do founders really need a true "bank"?

This time last week, I hadn’t given Silicon Valley Bank more than 2 minutes of cumulative lifetime thought…

As I wrote a couple of weeks ago, I have newly-founded a company in the Data SaaS space:

Once more unto the breach...

As we are still so new, we are yet to raise funds and are unaffected by the Silicon Valley Bank situation - which has now been resolved, permanently so in the UK and hopefully soon, too, in the US. My hopes and thoughts are with the founders and investors, who found themselves in difficulty through no fault of their own 🙏.

Why did it happen?

The seeds of its demise were sown when it invested heavily in long-dated US government bonds, including those backed by mortgages. These were, for all intents and purposes, as safe as houses.

But bonds have an inverse relationship with interest rates; when rates rise, bond prices fall. So when the Federal Reserve started to hike rates rapidly to combat inflation, SVB’s bond portfolio started to lose significant value.

If SVB were able to hold those bonds for a number of years until they mature, then it would receive its capital back. However, as economic conditions soured over the last year, with tech companies particularly affected, many of the bank’s customers started drawing on their deposits.

SVB didn’t have enough cash on hand, and so it started selling some of its bonds at steep losses, spooking investors and customers.

It took just 48 hours between the time it disclosed that it had sold the assets and its collapse.¹

One of the things I and others observed is that part of the vulnerability of a typical bank to this situation is how complex its operations are: buying government bonds and having to hedge interest rates accordingly, lending, trading… it makes it very hard to guarantee deposits. A typical bank has <= 10% of deposits available to be withdrawn at any given time, due to the use of deposits as capital for other operations.

For startup founders, especially at early stages, we have very simple needs:

1. We need to be able to accept capital/store money somewhere (that is preferably safer than under our mattress).

2. We need to be able to use this store to either move money to a more fully-featured account where we can do payroll, pay for services etc or to be able to directly do it from there.

3. That’s really it.

We don’t need:

1. To borrow money, as we are funded by venture capital and angels and have little revenue to justify any debt.

2. Complex financial products.

3. Foreign Exchange: most of us are funded in USD and have all of our expenses in USD, too. If we are paid in other currencies, then Stripe etc can pay us in USD. If we need to do payroll in other currencies, then Deel/Remote/Oyster etc can convert our USD to the relevant currency.

What if we could have a “bank” - and it still has to be a bank, as any organisation that holds deposits has to be a “bank” and regulated accordingly - that simply holds deposits and perhaps makes money from "a" (and maybe not "the") central bank interest rate²? Today, this rate could be around 4%. The “bank” could take a small sliver of this rate and pass the rest on to depositors. It would never lend money, buy bonds or create complex financial instruments. It would always have >=100% of deposits available to be withdrawn. It would be impossible for it to become insolvent, as the main central banks can’t really become insolvent.

The MVP would only need to be able to accept deposits and allow founders to transfer money elsewhere. That’s it! It’s more or less a central bank-backed savings account. The required operations for such a bank are massively reduced, compared to other banks.

Paying interest on the deposits and the ability to use the bank account directly for payroll or service payments isn’t required immediately, as other accounts could be used for this. Founders, for the time being, could simply push money from the deposit account to these fully-featured accounts once a month or so as needed, to make payroll and payments.

You might say - but if deposits aren’t available to be lent out, then this will cause banking to cease to function. This would be true at a huge scale, but startups really have a drop in the ocean in terms of overall deposits. If startup deposits were ring-fenced in this way, there would be no measurable reduction in liquidity for lending in the market.

Companies that are not venture-backed generally break even or make a profit, so they don’t need much runway stored in a bank for them to operate and the risk of their deposits being lost isn’t so acute. These companies typically can operate from revenue, and if there is surplus they often look to invest it in operations to enhance performance (like marketing).

An implementation of this kind of simplified, safer value store probably wouldn't be enacted by giving companies accounts with a central bank. Central banks deal with other banks and not companies or individuals. Just as I write this, our banking provider, Mercury, has announced that they now have FDIC Insurance for up to $3m and are working on $5m, through their new feature called “Vault” (it sure sounds secure 🙂). It certainly makes sense for banking providers for startups to provide this security as a service.

Up to $5m of cover provides for what would be pre-seed and seed stage startups. At pre-seed stage, like Delphi, it’s possible that startups are slightly less vulnerable, because ultimately, the founders could just go without pay and cover their own service expenses for a while - not comfortable, but actually many founders do this if they bootstrap to start with. At seed this becomes more difficult, as your team is larger. However, they may be willing to go without pay for a while, as they have relatively high equity.

The most vulnerable companies are probably Series A and B stage startups: they have the highest operating costs, without much in the way of enhanced operating leverage (meaning being closer to living off revenue) that usually happens at around Series C onwards. By the time you’re at Series A and B, there are many more people employed who won’t be willing to work without pay, and operating expenses are too high to be covered by personal funds. It may not ever be possible to have $25m to $50m of FDIC insurance. This is why I still think the simple value store I mention above is a good idea, if possible.

Art Levy @Alevy1511

1. Up to $2.25M FDIC coverage across 9 partner banks, including JPMorgan Chase, minimizing your exposure to any one bank, and even to Brex itself.

Art Levy @Alevy1511

2. 100% of funds in Brex business accounts are always available for withdrawal. Brex doesn’t lend your deposits away, so it cannot experience a bank run.

This is a further step in the right direction from Brex. However, the flaw in the argument is that, while Brex does not lend your money and therefore your balance with them is 100% available, the 9 partner banks most likely won’t ring-fence funds.

If there was systemic liquidity risk³, it could still be the case that your funds aren’t available for withdrawal, as Brex are not a bank and cannot hold your funds themselves. The risk of this happening, if managed well across 9 banks, is probably fairly low, but this is why I believe there needs to be a bank that does not lend or really do anything except hold deposits.

This is definitely my naïve and highly simplistic idea for a solution - I ran it past someone who understands economics better than me, and they pointed me in the direction of someone who really knows:

The Fed raises three main objections. The first is macroeconomic: The Fed worries that narrow banks could mess with the implementation of monetary policy, because if they succeed they will keep a lot of money at the Fed, increasing the size of its balance sheet...Second, it worries that narrow banks will take funding away from regular banks, making it harder for those banks to trade stocks and bonds (a business largely funded by repo), and maybe even making it harder to make loans. Third, the Fed worries that having too safe a bank would be bad for financial stability: In times of stress, everyone will flee from the regular banks to the super-safe narrow banks, which will have the effect of bringing down the regular banks.

So it turns out my idea, of course, has been thought of before. It has been not been allowed.

Now that I’ve been schooled, some of my assumptions above are wrong or problematic:

1. Not just startups would want this narrow bank - pretty much anyone would want it during times of turbulence. Just as we’re seeing deposits from smaller banks move to banks too big to fail at the moment, if the narrow bank existed, deposits from banks too big to fail would move to narrow banks… then the banks too big to fail would need bailing out…

2. If a good interest rate could be had from a narrow bank, it would encourage normal banks to take on higher risk to provide an attractive (higher) interest rate. Then the banks too big to fail would need bailing out…

3. If we were to say only startups should have this facility, this is suggesting regulation, which is not a good bedfellow for startups… if we regulate this then why not how they operate 😬.

The truth is the banks that are too big to fail are the narrow banks. It doesn’t matter that they lend, do some crazy stuff on their trading floors or fail to manage interest rate risk. Fundamentally, government has shown that while the West exists, these banks won’t become insolvent. So your money is safe in them while money is worth anything at all = civilisation as we know it continues (oh, I’m beginning to get this whole crypto thing now).

And yet despite all these new learnings, I still yearn for some way to safely store value that doesn’t rely on some form of insurance (actual and government bail out). It just feels fair… call me a dreamer.

1. https://amp.theguardian.com/business/2023/mar/13/silicon-valley-bank-why-did-it-collapse-and-is-this-the-start-of-a-banking-crisis

2. 20 bps less than IOER would be fine - https://www.newyorkfed.org/markets/domestic-market-operations/monetary-policy-implementation/repo-reverse-repo-agreements

3. I never thought I’d see the day when Credit Suisse would need bailing out