(Note: this is a draft of a section of my banking manuscript. It is in a very rough state; I want to get it out now, but I need to do some consulting work. It is obvious in retrospect that I should have started that manuscript before the inflation one, which I am still finishing off. The inflation manuscript work is somewhat scattered text changes, so I have not been able to publish as articles here.)
I am going to start off this primer with a discussion of generic bankruptcy procedures, at a high enough level that it captures the basic principles that hold across jurisdictions. I will then zero in on how American bank resolution differs from this basic procedure. There are a few reasons justifying this order of explanation.
Non-bank bankruptcies are more common, and so the discussion will start off in better alignment with most reader’s knowledge.
Banks liquidate a lot of firms and households in bankruptcy procedure. If you want to understand how banks operate, you need to understand bankruptcy.
Bank resolutions outside the United States might use procedures closer to the “generic bankruptcy procedure” that I describe here. The American banking system produces a lot of bank failures, and so they have developed a lot of specific bank windup procedures.
What is Bankruptcy?
A term that I often saw in the business press is “bankruptcy protection,” and one thing that I never thought about: what exactly is being protected? The answer was not obvious to me, but it is to protect the firm, and all of the associated interested parties.
Under normal circumstances, if a firm owes you money and does not pay, you drag them into court and get a judgement to force payment. In the worst case, you have a bailiff come and drag out property to make good your claim. However, if you have a firm that is circling the drain, you do not want it disappearing in a chaotic mess of uncoordinated legal judgements. Instead, the firm enters bankruptcy protection, which protects it against creditor claims. All the claimants — lenders, suppliers, workers, owners, the government — get paid out in a controlled process.
There are two ways in which the competing claims are settled. The first is that there is a negotiated solution, under the eyes of a bankruptcy judge. The second is that the firm is liquidated, and claims are paid out in order of priority, which are determined based on bankruptcy law (and the whims of bankruptcy judges, who have considerable leeway).
The optimal outcome for society is a negotiated resolution. This can allow the firm to continue operating, which can preserve at least some jobs and relationships with suppliers and customers. It also tends to preserve capital. In a liquidation, nobody in their right mind bids at “fair market value” for assets, as they know they face a “motivated seller” and will put a decent discount in their bid. This means that the recovery value can often be below the enterprise value (even with a discount on the enterprise value), other than in the case that the failed firm has assets that can be employed more usefully (mainly real estate assets).
In a negotiated solution, the credit claims on the original firm are typically converted into new debt and equity shares of the new firm. The amount of equity each class gets depends upon their negotiating power.
The liquidation procedure is important for a negotiated solution, as it determined the relative power of the negotiating parties. Creditors with seniority have considerable power — but the weaker parties can essentially threaten to drag out negotiations, forcing a somewhat conciliatory attitude. Furthermore, bankruptcy judges are expected to generate good outcomes for society, and can punish arrogant parties by inventing new precedents if they get sufficiently angered. This is an example of how finance is not a video game: prices and procedures are not handed down by impersonal algorithms, but rather result of negotiations within a society.
|Payment Priority (Draft!)|
If a bankrupt firm is liquidated, all of its assets are sold, often in some kind of auction or series of auctions. This might take considerable time, but for our purposes here, assume that it happens all at once. For now, assume that no assets have been used as collateral for loans.1
The diagram above shows what happens next. All creditors are organised into “classes” based on bankruptcy law. Each credit claim’s dollar amount acts as “share” within the class, and the class has a total amount due to it which the sum of all of the claims in that class.
The payment proceeds are then paid out to those classes, in order. Government taxes get paid first (of course). After that, payments are made to classes in order, until the funds raised in auction run out. In the diagram, the “Senior Creditor” gets paid out in full (100%). The “Next” class uses up the auction proceeds, but with a shortfall. Every claim in the class is paid out the same percentage (“X%”) of the claim amount. All the classes below that class get nothing.
Banks generally ensure that their loans are senior to all the creditors, including senior unsecured bonds. (People often omit the important “unsecured” qualifier when they discuss “senior bonds.”)
The original owners of the firm (“Equity”) is always at the bottom of the priority list, and got 0% in this example.
American Bank Resolution Procedures
The key difference between the generic bankruptcy process and American bank resolution is that the Federal Deposit Insurance Corporation (FDIC) administers the bankruptcy under the Federal Deposit Insurance Act. Different classes of creditors make claims, but do not negotiate a resolution — it is imposed by the FDIC. (Please note that various assertions in this section about the Act are based on the Bliss and Kaufman article. As always, nothing in this section represents legal or accounting advice.)
The Act creates two classes of super-senior creditors: insured deposits, and uninsured deposits. At the time of writing, the insurance limit on deposit accounts was $250,000, which meant that a depositor with a $1 million deposit would have two claims: a $250,000 insured deposit, and a $750,000 uninsured deposit.
I do not wish to discuss the mechanics of the resolution, but the usual procedure is that FDIC agents swoop in on Friday after the close of business, shutter the bank, and some new bank (possibly operated by the FDIC, or another bank assuming the failed bank) is open for business on Monday. (The Silicon Valley Bank failure on March 10, 2023 was unusual in that some geniuses organised a bank run over social media, forcing the FDIC to shut down the bank during business hours.) The objective is to minimise the disruption for depositors.
The FDIC allows insured depositors to withdraw cash from the new bank. This creates a drain on FDIC funds, which is matched by the FDIC assuming (subrogating) the original claim. (Subrogating is a cool word, and needs to be used more.) This means that the outflow does not affect the waterfall of claims on the bank assets, just moves the claim from the original depositor to the FDIC.
For uninsured deposits, they normally get a claim certificate that is a negotiable instrument. That is, they can sell the claim at a discount to raise cash. In the Silicon Valley Bank bankruptcy, the FDIC waived this condition and offered a full guarantee on uninsured deposits (which I will discuss below), so the two classes of deposits ended up merged.
Since the depositors (and subrogated FDIC) have priority, they get paid before all other classes of creditors. This means that the FDIC only takes a financial loss on a bank failure if the losses from the failed bank blow through every other layer of the bank’s capital structure — equity, preferred equity, and all the various seniorities of bonds. (The FDIC will incur administrative costs during the workout.) Since regulators are supposed to ensure that losses are less than the equity of the bank — never mind the other layers of capital — that requires a somewhat drastic regulatory failure for such losses to show up. Deposits are supposed to be safe, and all the other layers of the capital structure are sacrificed to make them so.
The Silicon Valley Bank Depositor Bailout
The Silicon Valley Bank (SVB hereafter) was an unusual institution, and had a large amount of uninsured deposits from firms. During the weekend after the failure (March 10-12 2023), some well-known personalities took to social media to demand a bailout. Given that said personalities had previously been agitating against other bailouts and were not particularly sympathetic, they managed to create an online community demanding that the uninsured depositors be thrown to the wolves.
Based on the available information at the time of writing, I think the FDIC’s decision is justifiable in that other options were worse. My assumption is that initial rumours/reports were true, which said that there were large firms interested in the assets of SVB, or taking over SVB itself, but those offers were seen as unacceptable to the FDIC (allegedly due to bank concentration concerns).
The options available to the FDIC were as follows.
Let one of those large firms buy out SVB on the weekend, probably at a highly distressed price due to the tight time constraint. SVB had an unusual portfolio of assets, and so it was much harder for acquirers to value the bank’s book than it would be for a mom and pop bank that makes mortgage loans. If one wants to beat one’s breast about the sanctity of market pricing, this was the market solution. The uninsured depositors would get paid out upon the market price of SVB’s assets. It would almost certainly hand SVB’s assets to an already powerful institution at a fire sale price — the rich get richer.
Reject whatever offers came in, and let the uninsured depositors twist in the wind while the FDIC tries to find better pricing for SVB’s assets. This would have the easily predictable effect of leaving loudmouths screaming about the risks of the banking system — and banking systems were a feature of the following week, even though the uninsured depositors were bailed out.
Do what the FDIC did: offer a guarantee to the uninsured depositors, and take time to find a buyer. The premise is that there is enough value to SVB’s franchise and/or assets that the losses will not be enough to blow through the subordinated parts of the balance sheet. And even if there are losses, they are chicken feed relative to the potential cost of a wider banking crisis.
I know my logic is not going to make the liquidationist wings of both the economic left and the John Birch Society right happy. Both sides want to see “undeserving” groups punished, and the existing capitalist order to fall apart. Although I am not a fan of this group of uninsured depositors, I do not see the attractiveness of handing a very large gift to the bottom feeders who would be the main beneficiary of a rapid liquidation.
Bank balance sheets are complicated, with a lot of different instruments being issued. This complexity reflects the desire to be able to calibrate potential losses to different investors with differing risk appetites. The objective is to find investors in the non-bank financial markets to absorb losses ahead of depositors, who do not want to spend time analysing their bank’s financial prospects.
If an asset is used as collateral, the lender has priority over the proceeds from that asset, which is treated independently of the rest of the failed firm’s asset pool. If the asset is worth more than the loan, the excess might end up disbursed to other creditors. If the collateral is insufficient to cover the loan, the shortfall ends up as an unsecured claim against the failed firm.