Via the Rational Walk, a handy (technical, but easy to read) account of some of what went wrong with Silicon Valley Bank (SVB). It’s well worth reading in full, but, as the author (Ravi Nagarajan) sets out, the problem has been there in plain sight:
What seems to be missing from much of the coverage of SVB’s failure is that the bank’s financial situation should have come as no surprise to anyone. Indeed, the company’s 2022 10-K, filed with the SEC just two weeks ago, made it quite clear that the bank was actually undercapitalized and vulnerable to a bank run.
[A] quote at the beginning of this article, taken from SVB’s 10-K, indicates that the bank had $16 billion of stockholders’ equity as of December 31, 2022. This is accurate from a GAAP accounting standpoint. There is no reason to believe that there was any fraud or other nefarious activity at SVB. However, it must be emphasized that the bank did not actually have stockholders’ equity of anywhere near $16 billion if we consider the market value of the securities portfolio.
But under GAAP (accounting) rules, not all those securities had to be valued at market. To put it simply, securities held by a bank that are “available for sale” (AFS) have to be shown at their market value (“marked to market,” to use the jargon), when calculating what they are worth, a calculation that is bundled into the overall calculation of stockholders’ equity (the bank’s net capital). However, another set of securities in the bank’s portfolio did not have to be marked down or marked up in this way. Those are the securities that a bank plans on holding to maturity (HTM).
The thinking behind this (it’s something I alluded to in my earlier post) is that if a bank plans on holding bonds (and SVB appears to have been holding high-quality bonds) until they mature, the bank will be fully repaid. It doesn’t matter how the bond trades in the interim. In the end the bank (assuming it bought the bonds at par, call it 100) will receive 100. In the event that the bank bought the bond in the market at a premium or discount to par, the difference between the purchase price and 100 will (I assume) be amortized over the holding period, but the underlying assumption continues to be that the bank will receive 100 when the bond matures.
All this makes sense until it doesn’t.
The Rational Walk:
We can see that stockholders’ equity is reported as $16 billion. However, this figure depends on HTM securities being carried at amortized cost rather than at market value. If the HTM securities were instead accounted for as AFS securities, SVB’s stockholders’ equity would drop to under $1 billion leaving the bank severely undercapitalized. It is important to emphasize that there is no evidence that the bank’s management did anything that was not permitted by accounting and regulatory standards. The situation was disclosed in the 10-K published just two weeks ago!
SVB’s investments were predominantly high quality mortgage-backed securities — that is, high quality from a credit standpoint. However, like all fixed-income securities, these investments declined precipitously due to rising interest rates. At the end of 2022, the HTM portfolio had large unrealized losses . . .
Again, those unrealized losses wouldn’t matter so long as SVB could hang onto the securities where those unrealized losses currently lurk. When the bonds matured, SVB would, it was (reasonably) expected would be paid in full. That doesn’t stop the market price of those bonds fluctuating in the interim. As referred to in my earlier post, it will vary with interest rates. If a bond with a nominal value of 100 carries an interest rate of 3 percent, and interest rates rise to a number above 3 percent, the bond’s market price will fall below 100, even though the buyer or holder who holds the bond until maturity will receive 100 at that time. The further away from maturity that a bond is dated the greater the price will fall (or rise, if interest rates fall).
So, why did SVB load up its portfolio with long-dated securities?
According to Nagarajan, one key reason was that, in order to earn a return on its deposits, SVB had to invest in longer-dated bonds as, in an ultra-low interest rate environment, short-term bonds yielded next to nothing. But the moment SVB invested in longer-term bonds it was increasing the interest rate risk it was taking. I don’t know what, if anything, it did to hedge that risk. And when rates rose dramatically (albeit from a ridiculously low basis), the bank’s AFS portfolio took a hit, but the unrealized losses on the HTM part of its portfolio began to pile up, too.
When confidence began to ebb in SVB, it had to sell HTM securities to meet departing depositors’ demands for cash. The unrealized losses thus became realized, losses that ate away at SVB’s capital base and undermined confidence still further.
It gets worse.
The Rational Walk:
When management is forced to liquidate securities classified as HTM, the unrealized losses become realized. In addition, the status of the remaining HTM portfolio is called into question. Rather than being held to maturity, those securities suddenly must be regarded as available for sale. As soon as this happens, they must be marked to market on the balance sheet and the bank’s stockholders’ equity plummets.
And that’s a recipe for a rout.
Back in December, I wrote about the HTM/AFS issue, and noted this from the Financial Times:
To keep capital ratios stable, many banks have shifted assets away from AFS toward HTM. More than half of the $690bn in unrealised losses are from the HTM basket.
For now, US banks remains awash in liquidity and are suffering no obvious financial stress. But rising deposit outflows and the increase in unrealised losses could become problematic if they need to sell investments to meet unexpected liquidity needs.
My conclusion:
The appearance of “problematic” and “banks” in the same sentence rarely bodes well for what is to come.
Like me, and many others, Nagarajan sees a lot of the blame lying with ultra-low interest rates, and quotes this from Thomas Hoenig, a former president of the Kansas City Fed. They date from a few years ago:
“An entire economic system. Around a zero rate. Not only in the U.S. but globally. It’s massive. Now, think of the adjustment process to a new equilibrium at a higher rate. Do you think it’s costless? Do you think that no one will suffer? Do you think there won’t be winners and losers? No way. You have taken your economy and your economic system, and you’ve moved it to an artificially low zero rate. You’ve had people making investments on that basis, people not making investments on that basis, people speculating in new activities, people speculating on derivatives around that, and now you’re going to adjust it back? Well, good luck. It isn’t going to be costless.”
They are words that deserve to be written in fire.
What next?
My view continues to be that we are not looking at 2008 part deux. That doesn’t mean that the economic unwinding ahead will be easy. I highlighted one or two areas of concern back in November, here.
Bank runs are, as I observed in my comment earlier today, a mix of the rational and the irrational. That makes them intrinsically hard to predict, and, given the nature of fractional reserve banking (our system, in other words), potentially very dangerous.
The Rational Walk:
SVB is the latest example of the pain [arising out of the retreat from ultra-low interest rates] manifesting itself in the real world. The simple analysis of SVB’s financial position shown in this article can be repeated with any bank. The presence of unrealized losses on HTM portfolios is not unique to SVB.
The fear in the financial markets this weekend stems from the fact that investors are scrutinizing other financial institutions for similar vulnerabilities. Fractional reserve banking is a system that requires confidence. Other banks in a position similar to SVB could face bank runs next week which is why market participants fear contagion.
The failure of SVB itself is not a systemic risk to the economy, but systemic risks could arise if the bank run on SVB becomes a stampede next week.
To stress, that doesn’t mean that there will be a “stampede,” far from it, but it’s an excellent summary of some of the darker fears now running through the financial markets.