A brief but serious recap of the whole saga at Silicon Valley Bank (SVB)

2023-03-10 11:30

Read this article in 13 Minutes

How this plays out depends on a number of factors, such as whether SVBS suffer a more severe run or even fail

Originally written by @Degg_GlobalMacroFin

1. Unprecedented technological cattle

In the second half of 2020, after the pandemic fears subsided and the Federal Reserve remained committed to keeping interest rates at zero for a long time, quantitative easing continued and inflation disappeared, the world ushered in a funding boom for tech companies. The rapid growth of startups loans and venture capital lines enabled tech startups to accumulate large amounts of cash and deposits. These deposits have largely flowed into Silicon Valley Bank (SVB), the most important bank in Silicon Valley and one of the top 20 banks in the United States. During the year and a half between June 2020 and December 2021, deposits in SVBS nearly tripled from $76 billion to more than $190 billion (Figure 1).

2. "Buy with your eyes closed"

In the face of a large inflow of funds from the debt side, the investable funds from the asset side of SVB also increased rapidly. The Fed hasn't started raising interest rates in 2020-2021, and if you put your money in the Fed's reserve account, it will earn a measly 0.1% annualized. The SVB option is to buy a lot of Treasuries and MBS. From its 10-Q, SVB increased its holdings by $12 billion between mid-2020 and the end of 2021, increasing its holdings from $4 billion to $16 billion (Figure 2). More importantly, SVBS increased their holdings of MBS by about $80 billion, increasing their holdings from over $20 billion to $100 billion (Figure 3). What is this concept? With total assets of about $200 billion, SVBS has half of its assets in MBS, or about 70% of the more than $110 billion of new deposits that will flow into the bank in 2020-2021. This is almost unbelievable, even absurd, for a commercial bank whose main business is loans.

3. "Cash is Junk"

SVBS 'cash and cash equivalents (including reserves, repurchases, and short-term debt) did not grow significantly compared with the rampant increase in their holdings, rising only from $14 billion to $22 billion between mid-2020 and mid-2021, and even falling to $13 billion by the end of 2021. Not even close to the level in mid-2020 (Figure 4). This reflects the fact that SVBS have aggressively allocated long-term assets without setting aside an equal proportion of cash to cope with deposit outflows.

4. "Robust" accounting

We know that commercial banks' purchase and fixed collection products are mostly accounted for by available-for-sale (AFS) and hold-to-maturity (HTM). SVBS are no exception. Its $16 billion of Treasuries are measured entirely in AFS, while $100 billion of MBS are mostly measured in HTM (Chart 5). The benefit of AFS and HTM is that fluctuations in the market value of assets (mtm) are not directly reflected in profit and loss, at most affecting unrealised gains and losses under other comprehensive income (OCI), and can be rolled back. The downside, however, is that once AFS and HTM are forced to be sold, a profit or loss needs to be recognized in the current period.

5. Fed rate hikes and unrealized losses

The average yield on the AFS and HTM assets is very low as SVBS 'asset purchases are concentrated during the low interest rate period 2020-2021. From the perspective of 10-K, the average yield of AFS is only 1.49%, and that of HTM is only 1.91% (Figure 6). With the Fed's rapid rate hikes in 2022, these low interest rate purchases of AFS resulted in an unrealized loss of more than $2.5 billion for SVBS in 2022 (Figure 7), If the unrealized losses of 100 billion MBS measured in HTM are taken into account, the total unrealized loss is up to $17.5 billion (HTM unrealized loss is about $15 billion, Figure 8).

6. Deposit outflow

These unrealized losses don't become losses as long as you don't sell them, so they are often considered "floating losses not losses." The problem is that the Fed's rapid rate hike in 2022 has made life difficult for tech start-ups around the world. They can't get financing, their stock prices have been falling, but their research and development has to continue, so they can only continue to eat into their deposits in SVBS. Combined with other factors such as the Fed's balance-sheet reduction, deposits in SVBS have been flowing out since peaking in March 2022. In 2022, total deposits fell by 16 billion yuan, accounting for about 10% of total deposits. In particular, non-interest-bearing deposits in demand deposits plummeted from 126 billion yuan to 81 billion yuan, greatly increasing the pressure on interest payments on the debt side (Figure 9).

7. Negative convexity of MBS

In particular, when the interest rate rises, residents are willing to change loans slowly rather than repay loans in advance, so the duration of MBS will be extended. As a result, the duration of a large amount of MBS held by SVBS will be longer and longer, and it will be more and more difficult to cope with the continuous capital outflow from the debt end. So since the end of last year, SVBS have faced a situation in which they have large losses on asset-side MBS that will not mature for a while and their cash reserves are not abundant. Debt deposits have been flowing out and the cost of debt continues to rise.

8. A strong man breaks his wrist?

SVBS management actually had a few other options, like going to the lending market to borrow repo, or going to FHLBs to borrow advance, or issuing debt to meet deposit outflow pressure. But the problem is twofold. First, the current interest rate curve is seriously inverted, the borrowing cost of the short end is much higher than that of the long end, rather than borrow the short end to maintain the long end to maturity, it is better to directly cut the loss. Second, start-up deposits are less likely to flow out, so it is better to cut leverage directly than to use short-term borrowings -- which may cause stock prices to tumble in the short term, but may be the safest thing to do in the long run. The short pain of a strong man's broken wrist may be optimal in this environment.

9. Panic

When SVBS announced that it had sold $21 billion of AFS assets and incurred a loss of $1.8 billion, the market's panic was manifested in several ways. First, will the unrealized loss of 15 billion won against 100 billion won of HTM assets that have not yet been sold turn into a real loss? After all, the total market value of SVBS is less than $20 billion. Second, issuing a large number of shares will dilute the rights and interests of the original shareholders, which itself is negative. Third, most of SVBS 'customers are technology companies, so they are not covered by deposit insurance and are prone to bank runs. In the first 12 hours, many technology company executives said they would withdraw all their money from SVBS. Fourth, it is unclear whether other banks heavily exposed to technology companies will suffer runs and whether the crisis will spread.

10. Look ahead

How this plays out depends on a number of factors, such as whether SVBS will suffer a more severe run or even fail. There are at least two ways to watch the crisis unfold in the coming days. One is whether the interbank market and repo market will be concerned about the overall financial health of small and medium-sized banks. Will there be a local squeeze on liquidity? Watch to see if the EFFR and SOFR 99% levels rise significantly over the next few days. The other is to see how the market perceives the risk of tech related loans/assets. For example, will banks with high tech exposure face a more severe run? 99% of bank panics are FUDs (blind panics), but the remaining 1% of real panics often turn into crippling financial crises. Let the bullet fly a little longer.