The SVB collapse should have been a tale foretold
Trouble signs were apparent, most notably, the classic asset/liability mismatch that ultimately brought SVB down. It is said that crises happen slowly at first, and then all at once. Silicon Valley Bank’s shocking 48-hour collapse took many by surprise. Yet trouble signs were apparent, most notably, the classic asset/liability mismatch that ultimately led to a run on the bank and the Federal Deposit Insurance Corporation takeover.
This is a failure of both governance and supervision. SVB was highly concentrated on technology and life science companies, which tend to generate significant cash flows and require specialized banking services, as well as on their venture capitalist backers.
The boom in the tech sector, fuelled by interest rates close to zero, swelled the deposit base. Liabilities rose at a dizzying pace. The problem when liabilities go up that fast is that banks have trouble finding suitable assets to offset them. With no better alternatives to deploy the capital, the bank went big into the long government bond market.
The timing could not have been worse. SVB’s big investment in long-term, low-yield bonds exposed it to interest rate risk in a rising rate environment. As rates went up, the value of these bonds declined, resulting in mark-to-market losses for the bank. Many banks are being challenged in this environment, but the size of SVB’s problems were massive: adjusting for mark-to-market losses, the bank effectively had zero capital.
An earlier intervention by either the board or regulatory authorities was clearly warranted.
First, allowing an unhedged long-bond position to form at a time when inflation was high and central banks were jacking up rates was shockingly imprudent.
Second, someone should have noticed that SVB’s ratio of wholesale to retail deposits was a huge vulnerability. Retail depositors tend to be stickier in part because their accounts are small enough to be covered — up to $250,000 in most cases — by federal insurance. Wholesale depositors, on the other hand, hold a lot more so are not covered for that portion if a bank fails. Therefore, they tend to be quicker on the draw, removing their deposits if they get any whiff about the health of their banks. And SVB was very heavily weighted towards wholesale deposits.
The joint statement and decisive action of the Treasury, Federal Reserve Board and the FDIC might shore up confidence. Amongst other actions, the Fed on Sunday announced it will let almost everyone have their money. This is exactly the sort of lender of last resort action needed to calm fears.
However, it does raise other issues. In particular, what is the point of deposit insurance if all depositors are made whole? This is especially the case for non-systemically important banks. Clearly the plan is to use the FDIC’s fund to keep everyone whole and then likely jack up deposit insurance premiums for everyone else to replenish the fund. This keeps taxpayers off the hook, which is a positive, but punishes institutions that acted more prudently. Meanwhile, SVB was a well-known lobbyist for looser bank regulation and successfully hid mark-to-market losses by declaring its asset base held-to-maturity, permitting those losses to escape stress testing. Regulators have no small culpability here.
At the same time, the stress tests that were devised following the 2008 financial crisis were irrelevant in 2023. The failure here is not a lack of stress testing but gaming the stress tests through asset classification. It was a failure by the board and regulators to deal with the fact that after adjusting for mark-to-market the bank had almost no capital. That’s deeply problematic in a world where most of your deposits are a flight risk.
North of the border we are in better shape. The Office of the Superintendent of Financial Institutions has taken over the local SVB branch, a corporate lending branch that holds no deposits. More importantly, Canada has not seen the gaming of mark-to-market rules that may have loosened discipline at SVB.
It goes without saying that regulators should ensure financial institutions hold adequate capital to absorb potential losses in interest-sensitive assets during a rate increase cycle. This may include setting higher capital requirements for financial institutions with significant exposures. At the very least it involves more attention to the stress-testing process.
In Canada, banks typically hold a higher percentage of retail deposits and are far more diversified than SVB. Even so, our officials must increase their monitoring until matters settle, as OSFI has done with daily testing. They must also stand ready to step up as lenders of last resort for any domestically systemically important bank should the need arise — though it’s to be hoped the bail-in debt regime enacted after the 2008 financial crisis eliminates the need for bailouts and the moral hazard they induce.
SVB’s failure underscores both the critical need for prudent governance and supervision and what can go wrong without them. Boards and regulators must not only run stress tests but also ensure results are internalized and acted upon. In Canada, we must also resist any urge to loosen our market rules, which serve as a great prophylactic to poor decisions.