Why Silicon Valley Bank failed: a primer

SVB’s overindulgence in its HTM portfolio and excessive uninsured deposits led to its swift collapse. It is, however, not a failure of the banking system.

Banks have one of the simplest business models on paper. They borrow money from one party promising to return that money at a time or on demand, with interest. They then lend that money to another party willing to pay an even higher interest. The difference between these two deals is the profit that the bank pockets.

In practice money management is considerably more complicated. I worked briefly in the treasury of a large bank and am intimately familiar with how banks manage their surplus funds to turn a profit. Among the many monetary instruments available to us are long-term bonds. These could come from governments—provincial, federal or international—or they could come from corporations (we lend to companies by letting them ‘raise money’ through bonds, which is really just a fancy manner of giving them a loan). But not all bonds are made equal.

The Silicon Valley Bank crash early last month led to fears of an industry-wide collapse. Banks in America and elsewhere, though, remained intact and for good reason: SVB’s collapse was a result of its own bad decisions, not a reflection of the banking system as a whole. Specifically, it had to do with their over-indulgence in long-term bonds which, as we shall see presently, meant they bet on interest rates being low. When reality struck, a rise in interest rates saw their asset value sink like a wrecking ball tossed into a lake. People panicked and the rest is history.

Understanding bonds

A bond, like a loan, is the somewhat risky prospect of lending money (buying a bond) in return for which the entity issuing the bond (taking out a loan) agrees to pay a specific sum, at a specific rate, every year (coupon rate) until such a time as when the lender sells the bond (at market value) or until maturity (held to yield, paid at par). We normally make the call to hold to maturity (HTM) or hold for trading (HFT) at the time of purchase and there is some red tape when it comes to moving these instruments across portfolios, which is not our concern for now. I will also not be using the parenthetical jargon, preferring simpler words instead, but they are there should you be interested in exploring these ideas further.

When banks buy bonds we normally strike a balance between trading and holding till maturity as both approaches offer different benefits. The reason we are required to choose our portfolios at the time of purchase is to prevent banks (who usually procure these bonds in millions of dollars or equivalent) from reacting to the market too wildly, which would induce a constant sense of unease in the economy.

To understand just how a bond value fluctuates, you need to understand two ideas: every bond has a duration which tells you by when the issuer of the bond—that is, the bloke to whom you lent money—will repay you, the possessor of the bond; second, there is a ‘price’ for the bond that varies inversely as the interest rate. So if you have a five-year bond and the interest rates went up by 2%, your bonds just dropped by something like (2 x 5) = 10%. A ten-year bond you had bought at the same time and under similar terms would go down by a solid 20%.

Note that this is back-of-the-envelope mathematics. Your bond may not drop by 20% exactly; it may only drop by 18%, depending on other terms (stated on the bond). Your takeaway here is the nature of the variation of bond value against interest rates.

Let us try to understand the same ideas with no mathematics at all. Let’s say there is a company looking to raise $1,000 so they issue a bond for $950 and a duration of one year. (I said ‘no mathematics’, I did not say ‘no numbers’.) You do some calculations and realise that if you ‘invested’ $950 today, i.e. gave this company you money, and they gave you $1,000 after a year, they effectively gave you $50 more, which is a return of around 5.3%, way better than your savings account. So you go ahead with it.

Now a few weeks after you buy this bond things get better economically and another company comes up looking to raise $2,000 through a bond of $1,800 for a year’s duration. Now you calculate again as you did before and you realise that this second company is giving you $100 for the same investment as the first company or a return of around 11%. Suddenly, for all prospective bond buyers (i.e. the ‘market’) your first bond making 5% returns does not look like a great purchase when they can get 11% returns for the same investment with the new company.

If you want to sell your bond (even hypothetically) your bond must now compete with the second company’s offerring, giving at least 11% return. In other words, you would have to sell your $1,000 value bond at $900. For you, who paid $950 for it, that’s a loss of $50 or 5%. Look at the final, big picture now: the coming of a second bond offerring 5% more than your first bond (a 5% rise in interest) saw the value of your bond fall by 5%. This is how the inverse nature between bond price and interest rate is born.

Now if the bonds are government-issue, when the federal/reserve bank raises interests they are effectively raising the interest rate of government-backed bonds. Replace our companies from the above examples with the government and you get the idea.

Playing with interest rates

Now why would central banks raise interest rates? Think of these rates as tools used to control liquidity in an economy. There are multiple rates that come under this ‘interest rate’ umbrella but most apply to banks and not the people directly, so think of all of these as one single rate indirectly affecting you, the individual.

If there is a tonne of cash (high supply) circulating in a country the value of one unit of currency falls (low demand). This could go on to have lots of other side effects like reducing the value of your properties because they are after all measured in this unit of currency. To prevent this, the central bank raises interest rates. Think of this as making money costlier. This would encourage banks to park their funds in the central reserve because the public would not borrow cash at higher rates and it would in turn encourage the public to spend their money more cautiosly.

When a bank buys a bond and the interest rate shoots up, we saw already that the bond prices falls. That is, the bonds a bank already possesses lose value (we call this getting ‘discounted’). Do the banks actually go on a loss, though? Not really.

Remember I said previously that some of these bonds are classified into our HTM portfolio, meaning we do not sell them but rather hold them right until they mature. What banks are really doing here is betting on the possibility that somewhere in the future, within the duration of these bonds, the interest rates will fall again and return these bonds to a premium. That way the banks actually stand to profit.

But what if the bonds reach maturity and the interest rates never fall, or at least hold strong after a net rise? Then the bonds reach maturity, the banks have to sell and the banks take a loss. Generally, there we keep a nice mixture in our HTM and HFT portfolios and, despite taking some losses, make a comfortable net profit.

This is generally the case. Sometimes things go bad.

Unrealised losses AKA depositor panic

As of now American banks have about $620 billion worth of unrealised losses. They hold bonds whose prices have fallen to such a point that if all these banks sold all such bonds right now, they would end up making a combined loss of $620 billion.

The FDIC chairman said as much in his speech days before SVB collapsed:

First, as a result of the higher interest rates, longer term maturity assets acquired by banks when interest rates were lower are now worth less than their face values. The result is that most banks have some amount of unrealized losses on securities. The total of these unrealized losses, including securities that are available for sale or held to maturity, was about $620 billion at yearend 2022. Unrealized losses on securities have meaningfully reduced the reported equity capital of the banking industry.

That last sentence is important. Banks are required to provide for their unrealised losses reflecting it as a proportionate reduction in realised profits. This prevents us from showing off an inflated balance sheet that hides a chasm of losses that are just round the corner. It also gives investors an idea whether their money is in safe hands or not. Banks work this way across the world.

The financial prescription that rescues banks when such risks fail is either investing in more secure instruments, loaning out traditionally, spreading assets across portfolios etc. This is known as hedging. Entire companies are built on this idea (hedge funds) but nothing except frail regulatory guidelines enforce this on banks and keep them in line, which means it is only a matter of time before someone’s unchecked appetite for risk or manically miscalculated investment drive sees them hoarding low-return, long-term bonds hoping that interests rates never go up. This is what happened with Silicon Valley Bank.

The treasury group at Silicon Valley Bank bet against any rises in interest rates. Their approach was that since they did not think the federal bank would raise interest rates, they could afford to load up on long-term bonds that offerred very little return. As soon as their bet failed and the fed raised its rates, SVB’s HTM portfolio saw its bonds fall in value all the way to becoming potential losses.

Depositors in SVB, comprising mostly rich tech bros and young companies like Roku (who notably had $500 million in SVB) saw their money getting eroded in real time. Federal banks in most countries have a deposit insurance scheme known by various names. In the US, deposits in a bank up to $250,000 are guaranteed; in India this number is around ₹5,00,000; in Germany this is €100,000. This is one reason why depositors spread their money across multiple banks—these guarantees are per depositor, per bank.

One might argue that had SVB been given time they may have learnt their lesson and recouped their losses. But the same people that helped drive its expansion—techies drowning in their smartphones—also led to its collapse. As news hit their ears that their bank had irresponsibly invested their money they took to their smartphones in panic to swipe their money out of SVB and into other accounts. The Bank soon found itself suffocated of cash and unable to pay its depositors. Too much of its money was tied up in loss-making long-term bonds to satisfy its surprising, short-term needs.

Feel no pity for SVB

Arguing that SVB should have been given time to recover their losses is like saying you lent money to a drunkard and are now counting on him to get a windfall so your money will definitely come back to you… some day. Sure, you could have given to your severely alcoholic friend but only if you had put a lot more of your money somewhere else too—like in the market, in real estate, in the hands of a considerably more reliable pal—just to cover any losses the drunk guy might leave you with. In other words, you could have hedged your risk. But you did not and you alone are to blame for it. So why should things be different for SVB?

The thing about classifying assets into portfolios is that bank shareholders are appraised of this, usually every quarter if not more frequently. The people on the board of a bank are especially aware of it as well. And senior executives even more so: treasury operations are a card banks play close to their chest. Bankers at the highest level of decision making at SVB knew exactly what they were doing when they bet unhedged against the interest rate. Their downfall is textbook in nature and deserves no lifeline.

In cases such as this, depositors are protected by a pool of money to which all other banks contribute and which is monitored by the central bank of the country. This is the same pool from which banks borrow from the central bank for the short term or to cover exigencies. Like SVB, other banks following suit should have been allowed to crash with only depositors offered a safety net.

However, the federal government chose instead to offer banks a lifeline in the form of loans backed by these same long-term bonds.

Banks are a necessity, human greed is not

Society does not often recognise the merits in the banking system. Most only see it as a one-sided evil. Why should I park my money somewhere when I can invest it myself? Why do banks get to invest my money and profit from it?

Banks have short-term liabilities that they balance with long-term assets because without this imbalance we would end up living in a financially precarious world. Think about it: if bank assets were short-term or on-demand, like your savings account, you would owe the bank your loan dues whenever they asked you for it. Dining out one Friday night? You might get an SMS from the bank asking you to repay your car loan in full by Monday.

Banks exist to ease the economy as well as to keep it healthy and afloat. They make money on the way because there are an umpteen number of people running this delicate ship who need to get paid, but banks as an industry are highly regulated. This regulation is why most banks failing is the result of either sidestepping rules (due to greed or simply, consciously bad decision making) or because a whole new class of maturity instruments were cooked up (think credit default swaps from the 2007 financial crisis in the US).

All this is also why banks should not repeatedly be cast lifelines when their failure is not due to the system but of their own making—as was the case with the over-ambitious Silicon Valley Bank.

Cover image: The monument sign in front of the parking lot of an SVB Financial Group office at 3005 West Tasman Drive, Santa Clara, California, next door to the headquarters. Picture by Minh Nguyen..