All banking crises are repo and shadow banking blowups (like SVB and CS)

Every bank blowup in the last many decades has been in repo — the S&L Crisis, Bear Stearns and even Enron.

It is amazing that Bloomberg pundits, economists and investors still haven’t figured out, even after 2008, that bank blow-ups are about off balance sheet assets.

Some estimate that Long Term Capital Management (LTCM) had 200 times its capital in repo liabilities (creating the joke that they were not “long term”, had no “capital”, and had no “management” — as they used “models” like Black-Scholes).

Lehman had $800 billion in repo liabilities! AIG blew up by creating more liabilities than the company’s entire market cap with one derivatives trading group.

The idea that SVB — a risky tech lender — blew up because they were holding treasuries is absurd, ignorant as to how banking work and gives too much trust to balance sheets. Treasuries are the most common form of bank collateral and most big responsible banks hold it, but they don’t blow up. So what is actually happening?

Repo is happening

To be competitive in banking, your loans do not sit stale on balance sheets reaping yield, they are lent out on an overnight basis in repo.

Repo is both leveraged, as you only have to post collateral on margin, and it is overnight. This can be a dangerous combination.

What happened in the 80s with the S&L Crisis is that banks lent out their mortgage assets on an overnight basis assuming that the repo loans would be rolled over indefinitely and profitably, and they spent the cash on riskier investments. When their collateral declined in value, they didn’t have enough collateral or cash to meet margin calls.

This is exactly what happened in 2008 but with CDOs.

These loans on an overnight basis meant that they could blow up overnight. These off balance sheet loans mean that a bank could blow up with a good balance sheet. In fact, banks seem to always blow up with a AAA rating and a good balance sheet, because the risk is hidden in repo.

SVB had a great balance sheet (including treasury bill losses)

What seems to be overlooked about the SVB blowup is that they had good profits and a strong balance sheet.

They had profits every year:

And they had profits literally every quarter:

They had a $315 million profit just last quarter.

Here’s what’s really odd about the narrative that they blew up due to holding treasuries…

Looking at SVB’s 10K released February 4, they talk about “ten-year” treasury securities when they calculate their values. They reported a $1 billion (unrealized) loss on treasuries.

All securities are to be “marked to market”, meaning that their unrealized loss would still be accounted for in their profits, so their profit includes these losses. They also have a load of derivatives to hedge against security losses (TRS) and interest rate swaps which had gains.

They discuss that they are used for repo but they do not disclose where that money goes.

They had $200 billion in assets with $16 billion in positive equity.

Furthermore, the 10 year treasury bill yield peaked on Oct 24, 2022 and yields have been going down since. (Could someone who thinks the Fed “controls” rates and inflation explain how rates are going down when the Fed has been “raising” them during this time?) Yields went from 4.25 in October to 3.39 when they filed for bankruptcy on March 17.

In other words, SVB had a large gain on their treasury holdings between their last report and going bust.

So if SVB had a gain on these blamed treasuries since their last report, and their last report shows positive equity and enough liquid assets to more than cover all deposits… What really happened? Plus, they only lost $1 billion in treasuries, a tiny portion of their balance sheet. How in the world can this possibly explain a collapse?

Even if literally all of their deposits vanished they should have been fine, at least according to the balance sheet.

They were also fine according to the Fed’s liquidity stress testing metric — the Liquidity Coverage Ratio (LCR). That SVB had 150% high quality assets to cover liquidity strains like large withdrawals.

A $1 billion loss on treasuries in the prior tax year can’t possibly explain illiquidity from deposits. In fact, their balance sheet shows they were very profitable and full of “high quality assets” to cover all deposits.

The deposit narrative and treasury narrative can’t possibly explain SVB. They are narratives.

Maybe repo is the problem…

On the other hand, they had a ton of CDOs (which is exactly what caused the GFC) under various names being used in repo, if there was a problem with their collateral they would have been forced to sell treasuries.

Overleverage in repo using these mortgage assets, the riskiness of the type of mortgage derivatives used as collateral, and the type of risky investments made with this leveraged overnight money, all seem like way more likely candidates for a bank blow up. In fact, that’s the formula for all modern bank blowups.

Bank runs are a myth of the modern era

I double you’ll find a bank failure driven by a bank run, that is without prior malinvestment also present, at least in the last century.

Elmus Wicker showed using Federal Reserve deposit data that in the first three banking crises of the Great Depression deposits were only withdrawn after the bank failures started. People didn’t get spooked then take out their deposits and then banks failed — it happened exactly the other way around. And this was true all throughout the increasing mass bank failures of the 1920s. Banks blew up and then sometimes people started withdrawing money.

Bank malinvestment is the sole cause of bank failures. People don’t understand how fractional reserve banking works and just how much loss you would need to go under with a bank run. A bank failure would require large scale malinvestment.

Furthermore, we don’t have fractional reserve banking. Case in point, where are all these SVB losses? Not on their balance sheet.

Bank failures happen under lower rates, not rate increases

Right now the treasury is paying hundreds of billions to banks with no effort on their part. This is great news for banks.

Bank failures started to increase in the 80s after rates had been coming down. The S&L Crisis is considered to be from 1986 to 1995. A third of regional banks failed(!) under lower and lower rates.

I believe there were more bank failures in the year 1990 than the entire decade of the 1970s. Bank failures ballooned into the 2008 crisis, all under even lower rates.

The facts don’t add up with our narratives. The 70s wasn’t a period of bank failures and lower home buying, for example. Quite the opposite, as mortgage applications consistently increased. Our narratives are off because they are not based on observation. We don’t know what real monetary-caused inflation looks like (credit expansion) because we’re simply experiencing whipsaw effects (like the “inflation” and “labor shortage” of 2008).

Something else is going on.

Shadow Banking is the Cause

Banks use dollar denominated assets as collateral to clear trade and lend in repo. When there is a shortage of collateral, yields go down. When there is a shortage of collateral, lending and trade have to decrease.

The 2008 GFC was a dollar collateral shortage. Even the BIS report from 2009 was called “the US dollar shortage in global banking”. They are explicit that this dollar funding is secured with collateral. This is shadow banking.

This might seem contrary to those who watch Fed money metrics like “bank reserves”, which are simply when primary dealer banks deposit money with the Fed instead of elsewhere. They are paid for and it is an asset swap. QE is an asset swap and does not impact the notional supply of dollars in the systems, just changes the form in which they are held.

Shadow banking gets its name because the majority of banking is outside the sight and control of the Federal Reserve. Repo contracts aren’t even required to be disclosed by the Fed and this is why we don’t know about Lehman or Bear’s liabilities until after they blow up and need to be rescued.

If we can’t track lending, we also can’t track the money supply.

It is important to understand that “bank reserves” are an arbitrary designation of money that means deposits held specifically with the Fed, which is only available to a few select “primary dealer” banks. These banks can seemingly switch from storing their money from bank assets/collateral, to treasuries, to “bank reserves” and it would mean absolutely nothing but show up as a massive spike in the “money supply”. Just like you switching from bonds to cash. In fact, it could mean less collateral and therefore less repo lending.

Bank reserves tell us nothing about the state of the economy, nor does it even correlate with inflation.

We can only see the effects of the dollar collateral shortage in the price of dollar assets that can be used for collateral — like treasury and mortgage bonds — and spreads.

The Yield Curve:

The yield curve inversion has predicted every recession in the era of modern (collateral based) banking. There was even an inversion in 2019, before COVID and its recession (a topic for another day).

Without fail the yield curve predicts recessions.

Volatility in yields and lower yields

Related to the yield curve, isn’t it odd that 30 year treasury bonds are at a 3.5% yield? With inflation that loss should compound at an enormous rate to an enormous loss.

Is this the banks making a bold bet? Is it ignorance as to how powerful compounding is and how big their loss would be?

Bonds are required for collateral, collateral is required for trade and lending. Banks use dollar collateral — bank liabilities — for trade and lending, not dollars as in Federal Reserve liabilities.

Clearly banks see an arbitrage opportunity to lend long assets out in the repo markets on an overnight basis. Both are indicators of lower yields.

Since even before the creation of the Federal Reserve, yields go up before a recession (into the bubble) and come down during the crash and recession. This is the nature of commercial banking for the last centuries of financial crises.

Financial crises are monetary shortages. Banks create money when lending. Those obligations, and trade cleared on the backend by banks, are cleared with collateral, not dollar bills.

When these assets go bad the commercial banking system is halted by deleveraging caused by having overvalued (and therefore over leveraged) these assets in the increasing shortage. When assets decrease in value there is a nonlinear impact on deleveraging.

When collateral decreases in quantity and value there is not enough collateral to clear trade and pay back old bonds (causing defaults).

Yields have been going down for 4 decades:

Debt becomes increasingly overvalued in its usefulness to clear trade and pay off old debts, which are also used to clear trade. Lower yields are a result of collateral demand above collateral creation.

Assets are more over valued than ever before, possibly because we reached the pinnacle of this collateral shortage. That is, there is no new market to expand repo and dollar bond creation to. The global currency has finally run out of new dollar bond creation to pay off old dollar bonds.

What is most scary is that more institutions than ever before are reliant on this collateral to rollover overnight loans. They have long term obligations and short term money.

Loss of new collateral creation is showing up in metrics globally. For example, the loss of major real estate dollar bonds in China, like Evergrande, have caused a dollar shortage in Asia. They are having to liquidate reserves like never before to meet other dollar bond obligations because their sources of dollar collateral are blowing up. China is experiencing a tough choice: drain reserves or default.

Look at currencies blowing up all over the world — it’s because they are losing their dollar sources locally and because it costs them more and more collateral to secure dollars. The two aren’t unrelated, dollar lenders everywhere are losing dollar collateral sources.

Look at the yield curve. Look at yield curves globally.

What’s even more ominous is how the inversion has moved up sharply to the front end:

Conclusion

If the majority of bank activity is off balance sheet, why do pundits still jump to conclusions based on a single $1 billion loss?

Furthermore, there is far more leverage in repo than on balance sheet loans, without comparison.

Compared to 2008, there is far more leverage using far worse underlying assets with far riskier instruments. Worse, these dubious sources of collateral are more globally connected and systemic, underlying virtually all trade and lending.

CDOs, for example, were relatively new and niche in 2008. Today, it is the majority of SVB’s assets. More importantly, they are used to secure far more leverage in repo than 2008. A small bump in the road, like the tiny increase in mortgage default rates in 2007 that cause the global crisis, could cause a much larger liquidity crisis.

The Fed insuring more deposits won’t stop underlying assets from blowing up, which historically is the real cause of bank failures, (which in turn cause bank runs).

Lower rates won’t make a difference, not only because the Fed follows market rates, but because high rates are good for these bank assets. QE won’t make a difference because they can only swap one type of collateral for another, not create more. And they certainly can’t prevent defaults that would ruin the underlying assets/collateral.

This is a collateral shortage problem driven by: 1) an overvaluation of assets in the first place, 2) a decrease in new lending to create more collateral to pay off old bonds, 3) reliance on extreme leverage to clear all trade, and 4) reliance on dubious derivative instruments.

These risks multiply each other.

Repo is leverage, using leverage, in order to get leverage, in order to buy more leverage instruments. Volatility and price determine how much margin is given. Therefore, a further decrease in asset prices and/or an increase in volatility will create a wave of collateral calls. This will create a negative feedback of selling/liquidations, lower prices, higher volatility, more collateral calls, and exacerbate an already strained supply of dollar collateral.

It’s a good time to be in treasuries and good quality dollar denominated collateral (if there is any left) and a bad time to be in assets bought with this degrading collateral (like stocks).

Once again, as Buffett explains, treasuries are the best form of dollar denominated collateral. It is preferred over cash because it has no counterparty.

Collateral is valued solely based on the lender’s assessment of whether or not the market would buy the collateral should they have to take possession. If there is even a hint of defaults or volatility in mortgages or corporate bonds, treasuries could be required for collateral (like Bear Stearns with the start of the GFC in 2007). CDOs will be the first to go — and perhaps we’re already seeing that.