A successful tariff wall is set to slow growth, but the downturn begins with a dollar liquidity crisis in China and ends with a crash in the US as funding pressures ricochet throughout the global banking system. Because they've historically run a large trade surplus with American consumers, Chinese banks have accumulated a large base of Eurodollars, with trade serving as their chief source of dollar funding. The tariff wall appears to target China, which will throttle the flow of Eurodollar deposits to the mainland, worsening offshore dollar funding stresses.

The payment chain is the supply chain in reverse. Since the trade war began just before Covid, China has successfully evaded US duties and kept the dollars flowing through entrepot trade via Mexico, ASEAN, and other countries, continuing to funnel goods to ultimate US consumers. Deploying a tariff fortress across all US imports, however, is an attempt to make that impossible.

Hardly a blind spot, China is of course aware of this vulnerability, having been carefully preparing for it with multiple points of leverage that go far beyond retaliatory tariffs and entrepot trade.

Chinese banks will transfer Eurodollar deposits to their New York branches - keeping them onshore and within the Fed's system - or park them into onshore nostro accounts with big US banks. That appears to have been status quo until sometime in 2023. Since then, Chinese banks have largely made Eurodollar deposits into nostro accounts based in Hong Kong - offshore and outside the Fed's system. Citibank, in particular, reports a uniquely massive and uninsured foreign deposit base, a large portion of which appear to be Chinese claims. An unprecedented 79% of Citi's domestic deposits are uninsured, a lurking solvency issue which would be realized in the event that Chinese banks need to raise dollars.

There are also problems with engineering dollar shortages, namely that if borrowing rates are too punitive, Chinese banks and non-banks would force liquidate dollar-denominated assets - equities, Treasuries and credit - at fire sale prices, an outcome that would not only have adverse consequences for the solvency of Chinese lenders, but for US lenders and the stability of the global financial system.

In any case, the US sees an opportunity for asymmetric pain to develop - hurting China and forcing their banks into a role that is more subservient to the dollar - but likely overestimates its own leverage. This sets the stage for a battle in the ongoing war of Great Powers, and there will be blood.

1. Recession? Crisis? "Detox"?
2. How Foreign Banks Obtain Liquidity
3. 'Chinadollars'
4. Dollar Shortages
5. Shoring Up USD & Hong Kong
6. Gold, Tariffs and the 'Three Strategies'

- DeathTaxesandQE

Recession? Crisis? "Detox"?

'Recession' has recently been touted as a mainstream idea. I said in my article from February 1st that tariffs were more likely to slow the economy than stoke it with higher inflation, creating recession risks that the market had not yet priced. Had a broad tariff wall on imports been implemented in 2021, it would've proven much more inflationary in my opinion, because disposable & real incomes were higher, so consumers would be more willing to buy goods at higher prices. At this stage in the cycle, however, a price hike from tariffs would more likely force consumers to pull back and businesses to retrench.

But let's totally ignore the debate about "what tariffs may or may not do" for a moment: the fiscal impulse and immigration were both were significant tailwinds for GDP during the Biden administration, so as each of these peak and level out, expectations of US growth will reset lower. The growth scare should thus not be a surprise. Incoming labor and GDP data will be weaker than it was over the last couple of years to reflect what Scott Bessent calls a "detox period." That of course also means lower stock prices and higher volatility.

Does that mean a recession (an "official" one, anyway) is looming? Perhaps. When people think about recession, there's at least one common denominator we all know of: an equity selloff. One that vaporizes trillions of dollars in household wealth with every 2% downswing in the S&P500.

While that's the most exciting, it's hardly the only leg of a recession...

... there are multiple things to characterize a real shock.

First, when all is said and done, you definitely do not improve the deficit or make the debt trajectory less unsustainable. During a recession, especially in modern America, you simply cannot get anything even resembling a balanced budget. Not even close. If asset prices don't post returns (stocks don't go up) and/or businesses aren't making money (real GDP stalls or declines), tax receipts will slow, which then blows out the deficit.

Another feature of US recessions is that they are global. A hallmark of today's monetary regime, especially the post-GFC paradigm of ample reserves (ample 'bank cash'), is that dollars are more accessible at home and onshore than they are abroad and offshore. But the dollar needs of foreign banks & non-banks is no less significant. That means downturns - anything from a tiny bank failure to a worldwide crisis - will tend to begin overseas, where dollars can easily grow scarce but dollar liabilities total over $13 trillion, before creeping back home as those funding pressures inevitably ricochet throughout the entire banking system.

BIS

Finally, you don't get recessions and economic slowdowns, even those that are papered over and thus go unnamed, without a systemic credit event, which means a lender going belly-up. This could be the failure of a commercial bank lending to the public, or the collapse of a hedge fund (a shadow bank) "lending" to the Treasury via its basis trades. Credit doing poorly is a calling card of recession. The failure of Silicon Valley Bank and Credit Suisse in March 2023 was as much a recession feature as were the two, back-to-back quarters of negative real GDP and bear market in stocks that preceded it.

The Fed had been injecting reserves - cash that can only be used by banks to settle payments between one another - via QE in the two years following March 2022. But even reserves count towards a bank's leverage, which their balance sheet can only hold so much of. What doesn't count towards a bank's leverage are overnight loans to the Fed - at the reverse repo facility - so that's where they deployed the excess reserves.

Reserves are the core of bank liquidity - they're actually used for day-to-day bank operations, and their in-house dealer businesses call on reserves to settle trades, like buying Treasuries. Reserves are "bank cash", and have basically remained within the same range since "balance sheet taper" began in June 2022. There's a red line beneath which reserves are considered scarce, called the Lowest Comfortable Level of Reserves or LCLoR. Frankly, it's anyone's guess where exactly that red line is. But based on the 2022-2023 banking stress, a good estimate is probably in the neighborhood of $3 trillion or slightly more. Though it's come close, levels of bank cash haven't breached that number since March 2023.

removing reserves via QT on June 15 2022, but they have remained within the same, $600 billion range since then.

If cash shortages induce bank failures, and bank failures mark recession, then it follows that cash shortages are a feature of recession. "Cash", "reserves", "dollars" - these should all be thought of to mean the same thing, just like how I use "shortage", "scarcity" and "illiquidity" interchangeably.

It's hard to get a recession if lenders are flush with a surplus of dollars. But that "surplus" looks to be peaking, and especially offshore.

How Foreign Banks Obtain Liquidity

The recent and notorious case study of a crash prompted by global dollar shortages was Covid. The shock that lockdown measures brought to manufacturing and services activity in the first quarter of 2020 led to missed payments, forcing more and more firms to become 'deficit agents' as they relied on borrowing to manage their cash flows. As the crisis evolved, banking systems around the world quickly became deficit agents themselves - foreign banks suddenly needed cash to lend to their constituent firms.

Local central banks can deal with missed payments in their local currency easily. The Bank of Japan can print yen-denominated reserves, the People's Bank of China can print renminbi-denominated reserves, and so on. Foreign banks in need of foreign currency funding is thus not a global problem. The main concern during Covid was about missed payments of dollars, because everybody seemed to have dollar funding needs, but only the Fed can print dollar-denominated reserves and provide dollar liquidity when other spigots run dry. As Zoltan Pozsar commented in 2020, "Dollar funding is always the orphaned child of crises as the regions where the pressures flare up have no control over it and the Fed, uncomfortable with the reality of it being the de facto central bank of the world, takes time to add liquidity."

In "allied" jurisdictions, and during normal times, foreign banks should not have a problem obtaining dollar liquidity. Many hold Treasuries and are set up to borrow dollars in repo, but, if they're not, or if they don't have the balance sheet space for repo, they can bid for dollars in FX swaps with a G-SIB (a major US bank). If the price is right, Citi, to use an example, will lend it's dollar reserves to Deutsche Bank.

In repo, Treasuries serve as collateral to secure the dollar loan. In FX swaps, reserves in foreign currency serve as collateral to secure the dollar loan:

FX Swap, based on 3-month term repo rates.

The spot rate is today's exchange rate - where is EUR/USD trading today? That's the rate at which the currencies are first exchanged. The forward rate is the rate that each counterparty (Citi and Deutsche Bank) agreed to re-exchange the currencies for at the end of the swap. In that sense, the difference between the forward rate and the spot rate is the cost of the FX swap.

Now let's say I'm a European-based investor earning 2.4% by lending euros within Europe. What's stopping me from converting into dollars, lending those dollars to earn a superior 4.3%, and then converting back into euros to walk away with a superior return?

Absolutely nothing is stopping me, but doing that comes with new risks. A rise in EUR/USD during the trade would cut into and could even wipe out my profits when it's time to convert back into euros. I can eliminate that exchange risk by entering into an FX swap, but that means paying the forward rate, which increases as the difference in interest rates grow. If repo rates (the price of money; the benchmark interest rate that central banks cut and hike) in Europe were the same as they are in the US, the forward rate and spot rate should be equal.

The price of the forward rate takes interest rate differences into account. Notice in the example from above how, in normal times, it costs Deutsche Bank $0.51 to borrow $108.10 for three months. He'd earn $0.51 more if he lent those dollars for three months in the US versus in Europe, so the opportunity is priced in.

Alas, like with many things in markets, there is no such thing as a free lunch.

"Supply and demand" is the closest thing there is to a universal law in finance. What if everyone wants to own an asset? All else equal, its price increases. What if lenders are in short supply? All else equal, the price of obtaining money - the interest rate - increases. In other words, when there is a market, supply and demand dominate.

What if everyone wants dollars, and/or dollar lending has stalled? FX swaps price supply and demand by including an FX basis (a basis) in the forward rate, where a more negative basis signals more need for dollars.

The basis is special to each currency pair: sometimes, European banks are in need of dollars more urgently than Japanese banks, in which case, the EUR/USD basis would turn more negative than the JPY/USD basis. Some systems, like Australia, don't invest in a lot of US assets, keeping the AUD/USD basis rather positive. Overall, this makes Australian banks net providers of dollars. In March 2020, everybody was scrambling for dollars, but the panic was more like a nightmare for Korean banks, who paid eye-watering rates as the Korean won basis collapsed.

source: Conks

But if private FX swaps prove too expensive because the liquidity is not there and the FX basis is too wide, as often happens when US banks are hoarding reserves, many can tap the Fed's FX swap line by bidding for dollars through their local central bank, who enters into an FX swap with the Fed.

For example, let's say Deutsche Bank needs dollars but, for whatever reason, the marginal US lenders are pulling back from EUR/USD FX swaps. Without any Fed rate cuts, the FX basis widens and forward rates spike such that the FX swap implied yield (the US rate, r_q) rises from 4.31% to 6.31%. Most borrowers would say that market-based FX swap costs are way too high.

Deutsche Bank doesn't have to go through Citi - it can bid at dollar auctions held by the European Central Bank. Key to the Fed's FX swap line is that the foreign borrower, Deutsche Bank, is essentially just borrowing dollars from his home central bank, the ECB. Thus, he doesn't need to post Treasuries; he can post local collateral, like German government bunds, but receive dollars, in a special type of repo with the ECB:

repo rates.

For example, if 7-day repo rates are 4.3%, a 7-day central bank FX swap would cost 4.55%. Which only makes sense when private dollar liquidity dwindles. The swap lines are mostly used once in a blue moon, but when they are, it usually signals offshore dollar banking stress in those nations.

How far do the Fed's FX swap lines reach? There are five, standing FX swap lines that serve as a fixture of the global dollar system and won't go anywhere anytime soon (yes, you should ignore politicized reports that suggest otherwise, as swap lines have recently been politicized). Unsurprisingly, it's in these economies where US banks, and the dollar, penetrate most deeply:

1. Bank of Japan
2. Bank of England
3. European Central Bank
4. Swiss National Bank
5. Bank of Canada

On March 19, 2020, the Fed backstopped dollar liquidity by adding nine, temporary emergency swap line arrangements. Amidst the panic, private US lenders were hoarding reserves, but each of the banking systems of each of these countries had (and still have) "dollar books" and thus needed dollars:

1. Reserve Bank of Australia
2. Banco Central do Brasil
3. Danmarks Nationalbank (Denmark)
4. Bank of Korea
5. Banco de Mexico
6. Reserve Bank of New Zealand
7. Norges Bank (Norway)
8. Monetary Authority of Singapore
9. Sveriges Riksbank (Sweden)

Japanese banks have an indisputably large dollar book, as you can tell from their persistently negative FX basis and outsized swap line usage in 2020. Since the GFC, Europe is not nearly as dependent on dollar liquidity as Japan or Korea, but that doesn't mean their banking system wouldn't run into solvency issues if the standing swap lines were actually pulled.

source: NY Fed

But not all jurisdictions are allies and so not all have access to the Fed's FX swap lines. And many of those jurisdictions, like each of the "90+ countries" where Citi operates, still need dollars. If a country runs a trade surplus with the US (like ASEAN - who suffered a dollar crisis in 1997 - does now), that's a source of dollars. And the largest foreign banks have offices in New York, so, in times when reserves held at New York G-SIBs are ample and funding costs are low, will first borrow dollars in private venues. But what if the demand for dollar funding is not met with an adequate supply of funding? Those market-based costs can rise prohibitively high, or even spike. With no central bank swap lines, what to tap?

In March 2020, a new facility emerged that supplied dollars to banking systems falling within this category. Called the Foreign and International Monetary Authorities (FIMA) Repo Facility, it gave foreign central banks access to Fed repos, and so, unlike FX swaps, it required the central bank to post Treasury collateral, instead of their local currency:

It might seem like only a handful of international systems fall into this category - large dollar books but not an "ally" - and yet, 30 central banks (in addition to the 19 with FX swap lines) signed up for and used the FIMA repo facility during Covid. It's also known as a China Repo Facility. The dollar funding needs of Chinese banks are thus backstopped by the PBOC's large Treasury holdings - $760 billion as of today. Rather than devaluing the renminbi or burning through (selling) its war chest of dollars, in a bind, the PBOC can repo (borrow dollars against) its Treasuries.

FIMA repos allow for those foreign central banks to monetize their Treasuries without outright, acute selling.

This is important, because China is a special case: it does have a lot of dollar funding needs, but no access to swap lines. US banks are less willing to accept Chinese renminbi as collateral in FX swaps, or at least at favorable rates, so their dollar funding in the money markets (repo, FX swaps) is thin. But thin as it may be, China has consistently relied on obtaining funding in the most classic way - through its trade surplus.

'Chinadollars'

Even without Fed swap line access and fewer dollar lenders willing to be counterparty in FX swaps, Chinese banks have largely avoided any problems in managing their dollars because they consistently run a trade surplus with the US, which proves a reliable, steady source. So long as Chinese merchants continue to accept dollars from American customers and deposit them into their local, onshore and usually international bank, dollar funding pressures in China only flare up during a crisis.

BOC, ICBC, AB of China and CCB.

Bank deposits - the money you, me, and anyone all over the world use day to day - are booked as the bank's liability: it's our money and the bank owes it to us. A Chinese merchant who collects $100 during a transaction with an American will put those bank deposits into his bank account, which, in this example, he holds with the Bank of China (BoC):

These dollars, deposited anywhere outside of the US, are called Eurodollars (a misnomer as they don't need to be in Europe). For the sake of argument, these might as well be called "Chinadollars."

One of the most important points to understand here is that, in the Eurodollar system, bank deposits held outside of the US can function as a form of short-term funding and substitute for bank reserves. That’s because, for any dollar deposit anywhere, there must be a corresponding claim on dollars—a matching asset that the bank can 'spend,' including to meet withdrawals. In the US, that claim usually takes the form of bank reserves held at the Fed. Outside the US, it often means a claim on dollars held with a correspondent bank, typically in New York.

The Eurodollar system is how dollar credit has propagated throughout China, penetrating everything from local government bonds to cross-border construction loans for the Belt & Road Initiative.

What BoC decides to do with that $100 in Eurodollar deposits is important, here. And they have a few options.

Of course, BoC could always just to go to the Chinese central bank, the People's Bank of China (PBOC), and convert those dollars into renminbi. That's one option. But BoC is a massive institution - one of the largest banks in China - constantly doing business not just with Americans but in dollars, so would be unlikely to do this. A small, more rural and domestically-focused commercial bank without much need for dollars might, but BoC probably wants to keep its dollars as dollars.

Like most major Chinese banks, Bank of China is international, with a New York branch that has a reserve "master account" at the Fed. The US branches of foreign banks can accept Eurodollars in offshore branches and then transfer the funds to an onshore branch. The $100 Eurodollar deposit could be transferred from Shanghai to New York, for example. This is basically an accounting trick, as the New York branch "owes" the Shanghai branch $100:

The foreign bank does not have to transfer its deposits to the New York branch because, if they do business in dollars, they will also have a nostro account. A nostro account, also called a correspondent account, is a foreign bank’s dollar deposit account held with a US bank, giving it direct access to the dollar system for payments and settlements. Many foreign banks (like "national banks" in China) will deal with dollars but do not have New York offices and, thus, do not have direct access to Fedwire. For them, a nostro account is their spot to park reserves on the Fed's balance sheet, albeit by going through their correspondent bank.

In this example, BoC transfers Eurodollar deposits into its nostro account that it holds with the New York branch of its correspondent bank, Citi:

nostro account, usually at a G-SIB, where it can deposit and settle Eurodollars.

The nostro account does not have to be onshore. The Bank of China can hold claims on Eurodollar deposits in a nostro account at a bank’s offshore branch, such as the Hong Kong branch of Citi. What matters is that the nostro account is held with a correspondent bank that provides access to the dollar payment system, not necessarily that it is located in the US:

It's this last example, including Hong Kong specifically, that's probably most relevant to think about, because this is the shift that would allow China more autonomy over, and greater leverage against, the US banking system. It also happens to be the shift that Chinese commercial banks have quietly made.

Dollar Shortages

For all the push-and-pull we've faced on the topic so far, where harsh duties are applied on certain nations only for them to be walked back days if not hours later after "successful negotiations", the broad tariff wall set to take effect tomorrow is without a doubt going to throttle the international flow of dollars. It will appear nascent at first, because the understaffed USTR faces a herculean challenge implementing Section 301 tariffs across the board, but will evolve over time. The market still refuses to realize this.

The tariff wall appears to be focused on China, targeting not just most industrial imports but selecting nations that China has used or could use to bypass tariffs, like Mexico and Canada, respectively. To pinpoint how this shortfall of dollars through the entire system, we'd have to think about exactly which banking systems are most reliant on dollars, and to what extent.

Do we have a precedent where a dollar shortfall emerged that hit some systems harder than others?

Yes, we do.

During Covid, dollar funding stresses emerged all over the world. Lockdowns disturbed global trade and resulted in firms missing their payments on dollar debts, payments that foreign banks came to rely on. We can pinpoint which specific banking systems were exposed as the most vulnerable - at the time - by looking at their FX basis. Remember as I said earlier, when a foreign borrower needs dollars, they're willing to bid for them by offering higher rates in the money markets. In FX swaps, the more negative the basis, the higher the forward rate and the more a foreign bank is willing to pay for dollars.

MIT

Australian and New Zealand banks are the outliers, as they are allies who also invest in fewer dollar assets. See how the basis for most currencies collapsed in early March, before spiking higher the day the Fed announced uncapped FX swap lines with 19 countries. That March 19 announcement calmed private lenders in FX swaps, so countries without swap lines also faced easier funding rates.

But after the March 19th swap lines, the bases for renminbi - onshore CNY and offshore CNH - were still left far below their 2019 means, lower than that of any other major foreign currency. This is clear evidence that,  without access to swap lines and even at "normal rates", Chinese banks struggled to obtain dollar funding.

MIT

Covid is a special case, and a lot has changed in the five years since. Not only based on the recorded trade balance with the US, but China's import bill has shifted away from dollars. It's incredible to think that by early 2020, China paid for nearly all commodity imports in dollars, including its imports from Russia, a statistic that has obviously adapted since.

The Chinese trade surplus with the US - net trade with the US - remains their primary source of dollar funding. A balance of payments is also supposed to include capital flows (like foreign direct investment or FDI), but I omitted this as the data was not clear:

source

The dip in 2016 just reflects August 2015, when Chinese authorities devalued the renminbi. Effectively, it cheapened the costs of imports in US dollar terms - there wasn't less value flowing, it was just that the value of those dollars increased relative to Chinese renminbi.

Note however that a sizeable portion of the trade surplus includes American companies manufacturing in China's "bonded zones", whose profits never truly entered the Chinese system in the first place. In fact, that reshoring explains most of the contraction since 2018 - not that fewer dollars were entering the mainland than before.

This has become challenging to measure since the trade war began in late 2018 because, just by looking at the BoP data from the census (above), it would seem like our trade deficit with China peaked in 2018 and never returned. But as the New York Fed noticed over a year ago, "while the U.S.’s reported trade deficit with China from the beginning of 2018 to 2024 declined from $375 billion to $295 billion, China’s reported trade surplus with the U.S. increased from $278 billion to $360 billion."

That's (at least) $65 billion in 'missing imports':

Clearly, China has skirted trade barriers, mostly via entrepot trade where it first exports product destined for the US to "conduit nations", like Mexico and Vietnam, who will ultimately re-export those same products to the US. In the end, Chinese merchants are earning dollars, even if that's not apparent in the accounting.

And this is where I will admit that there may be some faults in the exact way the data was collected. To make an example, in estimating how much aluminum imported from Mexico was originally from China, I referenced historical data from Harmonized System (HS) codes which, for most aluminum products, is shipped under HS codes 7601-7616. I made the assumption that US end demand for aluminum kept in line with real GDP. So if, before the trade war, the US was importing $100 in aluminum from China and $20 of aluminum from Mexico every year, then suddenly $70 from China and $60 from Mexico, I would count $38 of that Mexican trade as originally from China.

So, if anything, this is a moderate overestimate.

It's not perfect, but even if off by $15 billion here and $25 billion there, the trend is clear. That is, not only the unsurprising fact that we've been feeding China a constant stream of dollars that increases each year for over three decades (with the exception of global downturns and 2019), but that the trade war has been largely unsuccessful in decoupling from China's supply chain so far.

That's why a tariff fortress (tariffs on all imports regardless of the exporter's origin) is so much different from the targeted tariffs (tariffs on Chinese imports) from 2019. If it's not only Chinese aluminum slapped with a 25% tariff but all aluminum, the old strategy of reexporting through another country won't work like it has since the trade war began.

Chinese banks would have to raise dollars by another means. China hawks in the US seem to think that, without swap lines, this is not feasible. That's not the case, and Chinese commercial banks are well-prepared to weather a funding shortfall until the US buckles.

Shoring Up USD & Hong Kong

A point I was sure to include at the beginning of this article is that China is of course very aware of and has been carefully paying attention to America's turn towards deglobalization, especially when it comes to trade, patching up vulnerabilities along the way. Although the reality is that China can't unshackle itself of reliance on dollars over just a few years, there have been substantial changes made in this regard, including buying a neutral reserve asset - gold - for trade settlement, but also manipulating the balance sheet of their commercial banks.

Below is the total cross-border "dollar book" of Chinese commercial banks: assets (such as claims on nostro accounts) and liabilities (like dollars borrowed from New York) that are positioned outside of China. We can clearly see a measured decline; and yes, the peak coincides with March 2022, at around Zoltan Pozsar famously called for the birth of a multipolar financial system:

war chest of Treasuries. source: BIS

Looking at the breakdown of cross-border exposure of Chinese commercial banks to dollars, we can see that, though they have also de-dollarized their balance sheets overall, many of their cross-border dollar claims have simply shifted, presumably out of New York and London, over to the “international financial center bringing east and west together” - Hong Kong:

BIS data; Bloomberg

The data showing dollar claims shift to Hong Kong is confirmed by two things. CHATS, a large-value payment system (where interbank transactions are settled) based in Hong Kong and used by at least 30 Chinese financial institutions, saw its dollar volumes spontaneously double in 2023, corresponding to the surge:

CHATS, one of the largest dollar settlement systems outside of the US. Over 30 Chinese accounts participate in CHATS. source: Hong Kong Interbank Clearing Limited

Could it just be that China wants to avoid watchful eyes in the US, and as such, transferred its claims from New York to Hong Kong? That's what it looks like.

One might think "Not so fast! Those are claims on gold." That's probably true as Chinese banks were buying gold furiously towards mid- to late-2023 (in order to sell to PBOC), and CHATS volumes would reflect that. But looking at the individual, publicly-available balance sheets of G-SIBs' Hong Kong branches, like Citibank Hong Kong, we also see a rise in "placements from and dues to foreign banks" liabilities corresponding to 2023. This data doesn't account for all of the dollar claims in Hong Kong, only the biggest four US banks, but the trend is clear and the correlation is not spurious:

In hindsight, China clearly made some big moves to safeguard their financial system in 2023. Not only did they accumulate gold on the PBOC's behalf (presumably that's a large portion of the $1 trillion increase in 2023's CHATS volume), but, rather than incur a risky, violent asset sale, reorganized their deposits such that they would be held outside of the US.

"Yeah. And?" you might think. "What's the big deal?" Not a ton, unless one of those Hong Kong-based US G-SIBs is poorly capitalized, in which case it's a huge deal!

In April 2023, in the wake of the regional banking panic where everyone seemed to be focused on which bank was next to die, a team of four academics at four separate universities presented evidence showing that one of the four US banks with “assets above $1 trillion” was effectively insolvent.

Stanford Business

There are only four US banks with “assets above $1 trillion”: JPMorgan, Wells Fargo, Bank of America and Citi. In mid-2023, a small hedge fund team and I scoured through the call reports of each, marking the value of its assets to market by hand, and then compared that value with the share of uninsured deposits, which means any deposit not insured by the FDIC. A bank is considered insolvent if the market value of its assets – after paying all uninsured depositors – is insufficient to repay all insured deposits. That fact disincentivizes banks to hold huge amounts of foreign deposits, all of which are uninsured. The only trillion-dollar bank who clearly, or even closely, matched the study's results was Citibank.

An email with one of the study's authors refused to name which trillion-dollar US bank was technically insolvent in 2023, but confirmed the technique used that implied the results reached by the Stanford team matched our own.

Looking at credit spreads, bank health would seem to have improved in the two years since SBV collapsed. But Citi's balance sheet tells a different story, it's assets anemic and its deposit liabilities only appearing to be more precarious. Their most recent call report from December lists deposits across all branches totaling $1.324 trillion, but the breakdown is what's concerning.

88% of its deposits are uninsured!

December 2024 FFIEC 031

Contrast that to JPMorgan (or Bank of America and Wells Fargo), whose deposit base is mostly funded by insured, domestic retail money. In the event that Chinese banks need to raise dollars, they will first tap the stash in Hong Kong, most of which is at the Citi branch! That's, in effect, not only a run on the fourth largest US bank, even if slowly, but on one who is recklessly capitalized as is.

Just over 15 years ago, Citi received the largest bailouts in US banking history during and after the financial crisis. Yet its stock traded at 99 cents in early 2009. Sheila Bair, who at the time was the Chair of the Federal Deposit Insurance Corporation (FDIC), reflected on the bank's litany of issues in her 2012 book, Bull by the Horns:

Therein lies the final piece to the 'path of tariff pain': facing a dollar shortage on the mainland due to tariffs reducing the dollar deposits of firms, Chinese banks would draw down their nostro accounts in Hong Kong to maintain stable funding conditions. Large withdrawals would tighten liquidity and expose leverage risks within the offshore dollar system - rippling back to their global correspondent banks, at least one of which looks to be ill-prepared for offshore withdrawals.

As Hong Kong branches face stress, they would tap interbank markets for liquidity, driving up FX swap costs, repo and other funding rates. If this proves insufficient, they would turn to their parent bank’s New York branch for emergency funding, potentially forcing fire sales of bank assets. When this occurs at Citi, the balance sheet strain could trigger broader funding pressures - raising the risk of a confidence shock and deposit flight in the US.

A bank run not just at any G-SIB, but one where 78% of domestic deposits are uninsured, has got to be the nightmare scenario for the Fed and for any regulator.

"When all these pressures flare up" is really a matter of how severely the US trade balance with China is kneecapped due to tariffs. The trade war during the first Trump administration appeared as a $72 billion shortfall in trade surplus-funding to China by end-2019, but it was almost immediately concealed by the funding pressures of a global pandemic. If Trump keeps his word, an arbitrary estimate would be at least $120 billion by end-2025, and more in end-2026.

The cost of reshoring supply chains away from China may not just be a recession, but a financial crisis, having untold effects downstream. Something I'll think about and look into more closely in the future.

Gold, Tariffs and the 'Three Strategies'

Chinese culture is, for lack of a better word, rich—its written history dates back to the time of ancient Egypt and possibly earlier. In On China, the late diplomat Henry Kissinger describes the East Asian empire as a singular nation due to its longstanding system of government based upon the idea that China is the center of the world and the epitome of civilization:

Societies and nations tend to think of themselves as eternal. They also cherish a tale of their origin. A special feature of Chinese civilization is that it seems to have no beginning. It appears in history less as a conventional nation-state than a permanent natural phenomenon. In the tale of the Yellow Emperor, revered by many Chinese as the legendary founding ruler, China seems already to exist.

​Chinese political leaders frequently invoke their shared, ancient history to underscore this cultural continuity and national identity. President Xi Jinping often links his Party’s achievements to China’s broader struggle for national rejuvenation. Some of these narratives have been accepted and are widely known: for example, the Chinese concept of "long memory", where goals are not about next year or next election, but truly about the future. Kissinger compares this theme with the Soviet's penchant for chess: "If chess is about the decisive battle, wei qi (an ancient Chinese board game) is about the protracted campaign."

In The Three Strategies, one of the few, great ancient Chinese treatises on military strategy from the 3rd century BC, Huang Shigong—a legendary strategist said to have advised Zhang Liang, a key figure in the founding of the Han Dynasty—draws upon The Art of War by Sun Tzu to outline a three-pronged approach that a state should follow to secure victory:

• Moral Legitimacy (道德权威): Rulers should not always resort to bruteness, and must have the people's support. Virtue is essential for lasting power.

• Strategic Deception (诈): Like Sun Tzu stressed in The Art of War, deception and adaptability are crucial. It advises to make "the fighting ground more difficult for the enemy".

• Military Force (武力 兵势): Ultimately, force is necessary, but it should be used efficiently, decisively, and most importantly “when the moment is ripe.”

China is making headway in moral legitimacy. We can observe that progress by seeing the price of gold & gold reserves at PBOC rise and watching China negotiate with vulnerable countries on bilateral trade by offering to settle in a neutral medium - gold.

What's the strategic deception angle to their strategy? It could be what I described in this article: China is leveraging offshore financial flows to complicate US efforts at deglobalization. The tariff fortress will backfire against American G-SIBs more severely than it would've just a few years ago.

The third prong, of military force, is left to the reader's interpretation of China's goals.

Kissinger would likely understand this strategic thought framework, but Wall Street remains blinded by its own hubris, confident in tricks that worked some time ago, but no longer.

- DeathTaxesandQE