Countries outside the US/UK/Europe bloc are leaving the dollar-based trade community at a rate which has begun to alarm US geopolitical strategists, and neocons in particular. Announcements of large trade deals moving away from the dollar arrive more or less daily, and are typically measured in multi-billions of dollar value. De-dollarisation is in progress, but how much of a threat is it to US geopolitical hegemony?

The trend, which was quietly in train before the Ukraine War but picked up speed when the US froze Russia’s non-resident currency reserves, is now being openly acknowledged as a strategic threat by senior people in the US Administration.

Last week US Treasury Secretary Janet Yellen, interviewed on CNN, conceded that using dollar-connected sanctions could undermine “the hegemony of the dollar”. But she was not worried, going on to say that the depth and liquidity of the US Treasuries market serves to defend its status, because the world has no practical alternative to dollar reserves and dollar trade settlements. Is her confidence justified?

Out of Washington’s $31,000bn Treasury debt balance only $5,000bn is owed to foreign holders. These are a mix of states, sovereign wealth funds, banks, corporates and private investors. $1,100bn of that debt is owned in Japan, $1,100bn by investors whose holdings are buried in various major tax havens, $800bn by the government of China, and $645bn by UK holders. Those account for 70% of the total, and after them come $250bn holdings by Canadian and Brazilian holders. Below the $250bn mark a miscellany of holders in almost every economy on the planet own the balance.

It can be seen that most US foreign-owned government debt is owned by members of the Western Bloc plus Japan, and in this bloc there are no obvious signs of dollar flight. So far so boring. Less boring is the picture of flows into and out of US Treasuries.

The US runs a large trade deficit, nudging just under $1,000bn in 2022. 2022’s outturn is the most recent number in an unbroken line of trade deficits that began in 1976 (the US’s last recorded trade surplus was in 1975, at $12.4bn). The sum of all trade deficits from 1976 to 2022 comes out at $15,000bn.

Only one third of that has been paid for with foreign-owned Treasuries. Some has been exchanged for net purchases of US-based assets (property, farmland, equities and businesses). The rest has been exchanged for dollars supplied by the US Treasury to provide the circulating capital for world trade. We’ll look at how much in a moment, but this is the supply that a past French President described (accurately) as an “exorbitant privilege”, and which falls into the exotic economic category of “seignorage”.

Most world trade (84% last year) is priced and settled in dollars. The WTO reported that global trade in 2021 totalled $22,000bn, so that’s $18,000bn of dollar settlements, but rather less than $18,000bn of circulating dollar cash. Trade payments cycle within much less than a year. The duration of that cycle defines the frequency with which a given offshore dollar is traded, and in turn that is the “velocity” of trade dollar flows. Like other “velocity of money” data, no-one that I can find reports a hard number, probably because the actual velocity number is unknowable given the scale and diversity of global trade, but we can make a reasonable estimate.

The international trade cycle of “order, deliver, clear customs, settlement, re-order” is rarely shorter than 90 days and can be more than a year. Between those extremes the likely average is probably 120 days. If that is the right number then the velocity of those trade dollar balances would be a fraction over “3” (transactions per year). That in turn would imply that the quantity of trade dollars feeding the global trade’s circulating capital would mathematically be equal to one third of the total value of all dollar trade settlements, which would put it at around $6,000bn.

If we add those two numbers together – $5,000bn for reserves and $6,000bn for circulating capital, we are left with a $4,000bn gap in the source of offshore dollars. That is probably accounted for by, say, $2,000bn for US asset sales, $1,000bn of static cash in foreign-owned bank deposits and finally the $1,000bn in US dollar banknotes that the Fed estimates is sitting in vaults and under mattresses around the world.

A central plank of the global economy’s deal with the dollar is that stocks of offshore dollars must remain safe from political interference. A second plank is that the dollar’s value must be largely preserved, forcing at least some level of fiscal prudence on the US government.

Up until now the second condition has been met fully – the dollar index, first set at 100 in 1973 after the US defaulted on the dollar’s convertibility into gold, sits today at 102.4. In relative currency terms, at least, the dollar’s value has been preserved almost perfectly.

The first condition was well served for almost two generations, but came unstuck in 2022, when Washington collaborated with Brussels and London to freeze some $360bn of Russian sovereign financial assets. While the actual sum is small compared with global dollar balances (only 2.5% of that total of $15,000bn), the breach of principle is large. For the first time foreign holders of dollar balances saw clearly that the risk profile of their dollar balances had materially changed.

In consequence, some 8% of dollar reserve balances were quickly exchanged for other currencies (and gold). There was no visible impact on dollar values of that flight, and Ms Yellen would (accurately) say that the lack of adverse consequences proves just how deep and liquid the foreign dollar market is, and how essential the dollar has become to the global trade and reserves system.

At one level she would be right, but when we look at the nature of flows into and out of foreign-owned dollar balances, is there evidence that she is less right? Is there in fact a structural threat to the offshore dollar economy?

While the US continues to run a large trade deficit a calm continuation of the dollar system requires that each year a new $1,000bn of trade-deficit dollars be soaked into one of financial reserves, circulating capital for international trade, static cash balances or cash stashed under those beds.

Within those categories only one, circulating trade capital, has an intrinsic growth vector, which is caused by a linear connection between the value of world trade and the value of the circulating capital that finances it. If trade values grow then trade needs more circulating capital.

Reserves have a slight connection with GDP growth, so long as economies grow reserves at the same speed as their GDPs. Slight, because it is conceivable that reserves might move out of sync with GDP in some circumstances, in either direction.

Global trade grows more or less in line with GDP, and runs at approximately 22% of GDP. Therefore when global GDP grows 2% (adding $1,800bn), world trade may be expected to expand by $390bn. With a money velocity of 3x that would create the need for an additional $130bn of circulating cash.

If the dollar kept its 84% share of trade values that would in turn create a demand for 84% of $130bn in new foreign dollar holdings – or $109bn. So if the dollar keeps it share of global trade, and if global GDP grows, the US can hope to pay for just one tenth of its trade deficit by exporting dollars to global traders.

Those “ifs” are looking a bit soft. Even before the Ukraine war the dollar’s share of global trade was slipping, and it now looks like it might slip some more. The balance is sensitive. If the dollar share of global trade slipped to 80% that would reduce dollar use by $240bn (4% of $6,000bn of circulating capital), wiping out demand up-steps from trade growth.

A similar problem lurks in reserve flows. At present global foreign currency reserves excluding the US sit at around $10,000bn, with the dollar providing about half. Those reserves amount to 15% of non-US global GDP ($70,000bn in 2021). Non-US global GDP growth of 2% per year should add GDP of $1,400bn and, ceteris paribus, generate new global reserves of 15% of that, or $210bn. In the days when the dollar accounted for 60% of reserves the US could rely on a demand for 60% of that $210bn, covering another tenth of its deficit. But the dollar’s share of reserves is falling, and its share of new reserves is probably falling even faster.

So where does that leave Ms Yellen, trying to finance a $1,000bn trade deficit using that exorbitant privilege? Somewhat uncomfortable, it seems to me.

While the dollar appears to have shrugged off threats to its hegemony this year, future demand for reserve and trade dollars looks worryingly weak from Washington’s perspective. With the risk landscape enhanced by sanctions, demand weakening and US trade deficits likely to grow, is this the moment at which net flows into the dollar fade to a critical level?

That is certainly what China, Russia, Saudi, Iran and now Brazil appear to want. 2023 has seen a spate of announcements of bilateral trade deals to be priced and settled in yuan, Saudi riyals and rupees. Ms Yellen can take comfort that the sums involved are so far small by comparison with total world trade. Even the largest component, the China/Russia energy trade based on the ruble/yuan pair, is only $1.5bn per day, with a net working capital requirement of just $180bn. Other non-dollar currency pairs are harder to track but might take another $50bn away from the requirement for dollar-based working capital. Together those remove just 4% of demand for trade dollars – unwelcome but not intrinsically critical when seen from Washington’s perspective.

Ms Yellen’s argument is that for the rest of the world the original reasons for using the dollar, both for trade capital and for reserves, are as compelling as they have been for two generations – where else can an economy find the liquidity and scale of the dollar market? Some look to the euro, and more will do so in future, but the euro comes with the same “weaponisation” risk as the dollar, to which must be added higher regulation, a shallower market pool and weak foundations in a federation of states who are far from the best of friends. At least the dollar has one mind and one purpose.

Beyond the euro the pool of candidates thins dramatically. In theory the yuan could step up as a liquid source of reserves and working capital, but probably not any time soon with existing capital controls and Beijing’s unpleasant experience of capital flight the last time it eased those controls.

If we look at world currencies outside the Western/Nato/Japan bloc, and exclude China (for capital controls), and Russia (for high political risk) we find only around $20,000bn of GDP and $5,000bn of world trade, in economies which frequently have some form of capital controls, not inconsiderable political risks attached, and which collectively run a net trade surplus and have no need to export their currencies anyway. That landscape may explain why the core BRICS states have included discussion of a new currency – some sort of Shanghai Cooperation Currency Unit – at their forthcoming meeting in South Africa. We have been here before, when the European Commission invented the European Currency Unit (the ecu) as a precursor to the euro.

Sadly for the gold bugs, gold is not the answer, at least at any gold price this side of the stratosphere. We saw earlier that new global reserves accumulate at around $1,400bn per year. At current gold prices ($2,000 per ounce, or $60 per gramme), the world would need 15bn grammes (15,000 tonnes) of gold per year if it wanted to replace dollar reserves with gold.

But global new gold production in 2022 was just 3bn grammes, and only one third of that was left for investment after jewellery and the tech manufacturing markets had taken their shares. There isn’t enough annual gold production to meet even one twentieth of net new reserves formation, at least at current gold prices.

Which doesn’t mean that cash-holders won’t try. Russia and China are both stacking gold as fast as they can buy it, and gold bugs would argue that one resolution to the stresses building up might be a substantial increase in the price of gold, as reserve holders chase physical metal in place of dollar assets.

Which leaves the world with the dollar, still, and a justification of Ms Yellen’s confidence that Washington’s 2022 “abuse” of its exorbitant privilege against Russia will not have any systemic adverse effects beyond the one-off 8-point drop in dollar share of reserves that we saw in 2022. But annual demand for new offshore dollars is threatened by quite small moves away from the dollar in both trade finance and reserve accumulation, and the US depends on a constant new flow of demand for offshore dollars to cover its trade deficit.

With the structure at risk a new material risk might enter play – that a negative feedback loop establishes itself, following the path: “dollar demand falls enough to undercut US trade deficit financing > US dollar exchange value falls > dollar holders fear that future dollar value will fall even more > dollar holders begin slow retreat from dollar reserves and trade finance > dollar squeeze tightens > rinse and repeat”.

This possible negative feedback loop would not exist in a vacuum. Weaponisation of the dollar in 2022 turned Russia into a firm enemy of the dollar, and India, Iran and Brazil into cautiously hostile frenemies. Washington’s evident enthusiasm for a war with China over Taiwan has had a similar effect on China’s view of the dollar. It is possible that Beijing might choose this moment to weaponise its own holding of US Treasuries, by dumping them on the market to undercut global confidence in the dollar index. The cost of a dollar attack – perhaps 30% of China’s $800bn holding – is not small, but might be less than the cost of a hot war with the US, and would certainly do less physical damage to Taiwan and to the young men and women of the People’s Liberation Army and Navy.

A look at history shows that we have been here before. In 1956 Britain, in the twilight of its own hegemonic role, tried to enforce its will over a strategic piece of territory in Egypt by force of arms. The military campaign was a textbook success, but the attempt foundered swiftly when President Eisenhower threatened to crater the pound by dumping US holdings of British Treasuries on the market. 1956 marked the end point of British power and the birth of US hegemonic power. It is possible that China might have its eye on a similar play. We do know that both Russian and Chinese strategists are keen students of history.

Setting aside apocalypse now, or apocalypse in the near future, where does that leave the US? Its confidence (for now anyway) seems well founded, but perhaps Ms Yellen might have a view to the delicate balances that exist between the competing flows of the US trade deficit and the world’s likely future appetite for new stocks of offshore dollars, before completely discounting the risk of a future dollar crisis.