Liquidity dilemma of non-USD stablecoins: systemic failure in rule design

Regarding why non-USD stablecoins have experienced stagnation, the common market explanation is insufficient demand. But this judgment is wildly incorrect.

The demand for cross-border trade is enormous—historical data shows that daily foreign exchange transactions involving non-major reserve currencies exceed $320 billion. Demand is not the problem; the core issue lies in systemic paralysis on the supply side: a series of international banking regulations, especially Basel III and subsequent supervisory frameworks post-2008 financial crisis, have imposed increasingly stringent constraints on bank capital structures. These rules invisibly destroy banks’ motivation to provide liquidity for non-USD channels.

The essence of the liquidity problem: a supply-side crisis

On the surface, this appears to be a financial market issue, but at its core, it is a systemic failure of regulatory incentives. Basel III’s higher capital requirements, thicker capital buffers, and reduced leverage directly suppress banks’ return on equity (ROE). Under this pressure, many large global banks are gradually withdrawing from emerging market channels—not because these markets are unprofitable, but because regulatory costs make any profit negligible.

This has caused a structural failure in the global foreign exchange markets outside the G7, creating an invisible “liquidity vacuum.” In non-USD cross-border settlements, almost no direct bilateral liquidity exists between currencies. All transactions are forced to flow through the USD as the central hub—this is not market choice but the result of regulatory mandates.

First dilemma: the hidden tax of the LCR liquidity coverage ratio

One of Basel III’s key tools is the Liquidity Coverage Ratio (LCR). The logic is simple: banks must hold enough “High-Quality Liquid Assets” (HQLA) to survive a 30-day stress scenario. It sounds prudent, but in practice, it becomes a deadly noose for non-USD market-making.

The problem lies in the definition of HQLA. The scope of Level 1 HQLA is extremely strict: cash, central bank reserves, and qualifying sovereign or central bank debt. Most critically, these assets must be traded in large, deep, and active markets.

As a result, only reserve currencies (USD, EUR, JPY) with stable, deep, and liquid markets can reliably meet this definition. They have global repo and cash markets that remain resilient under stress. Non-USD assets, especially emerging market currencies, are essentially “substandard assets” in the eyes of the LCR framework.

This means: if banks want to trade in Brazilian real or Mexican peso, they must hold inventories of these currencies. But under LCR, these inventories are liabilities, forcing banks to hold additional USD assets to “compensate” for this risk. In other words, the regulatory system directly penalizes non-USD inventories, effectively imposing a clear “capital tax” that makes maintaining these channels extremely costly.

Even more severe is the “trapped assets” problem. Suppose a global bank holds large amounts of cash in a Brazilian entity, but due to capital controls, asset segregation rules, or operational barriers, it cannot immediately transfer these funds to London. Under the LCR framework, this cash cannot be counted toward the group’s liquidity buffer. The bank must “double finance”—hold trapped liquidity locally and hold redundant liquidity centrally. This makes supporting non-USD channels structurally more expensive than USD-centric systems.

Second dilemma: the liquidity-term tax of FRTB market risk capital

Even if banks ignore the inventory constraints of LCR, they face a second set of equally strict restrictions: market risk capital requirements.

Under the Basel trading book review framework (FRTB), banks must hold capital buffers for potential losses from market volatility. A key parameter here is “liquidity horizon”—the assumed time needed to unwind a position without causing a price collapse.

For major currency pairs (like USD/EUR, USD/JPY), the assumed horizon is only 10 days. For “other currency pairs” (including BRL/MXN and other emerging market pairs), the horizon doubles to 20 days.

What does this mean? When banks calculate risk for non-designated currency pairs, the model assumes they need 20 days to exit during a crisis, while major pairs only require 10 days. The direct result is that the capital buffer required for BRL/MXN risk is much higher than for USD/EUR risk—an explicit capital surcharge on non-USD inventories.

Worse, 20 days is only the minimum. FRTB explicitly allows regulators to increase the liquidity horizon based on the specific trading desk conditions. If regulators deem a channel high-risk or illiquid, they can arbitrarily extend it to 40, 60 days, or even longer. This introduces enormous regulatory uncertainty—banks cannot accurately predict capital costs. As a result, banks prefer to only participate in “designated currency pairs,” whose horizons are capped at 10 days.

Another mechanism is the “Non-Modelled Risk Factors” (NMRF), which systematically kills non-USD channels. To use internal models (which can save capital), banks must prove the market is “genuinely observable.” The requirement is at least 24 observable prices per year (roughly twice a month).

If trading is sparse and cannot meet this standard, the channel is classified as NMRF. Once designated as NMRF, banks cannot use capital-efficient models and must rely on stress scenarios for capital calculation. This is a catastrophic penalty.

This creates a vicious cycle: low trading volume → inability to meet requirements → classified as NMRF → capital requirements skyrocket → trading becomes unprofitable → banks exit → trading volume declines further. Banks are unwilling to market-make for a channel unless they already have sufficient liquidity—an endless deadlock.

Third dilemma: the complexity penalty of the G-SIB scoring system

Global Systemically Important Banks (G-SIBs) face additional capital costs based on their systemic impact if they fail. This should incentivize prudence, but instead encourages withdrawal.

The G-SIB methodology considers not just size but a five-factor model (size, cross-jurisdictional activity, interconnectedness, substitutability, and complexity), each weighted 20%. A direct consequence is that a non-USD channel can significantly increase the scores for cross-jurisdictional activity, substitutability, and complexity.

For example, a US bank trading USD/BRL only needs to book in New York, but to reliably provide BRL/MXN, it often needs local balances in Brazil and Mexico for settlement. Supporting each additional channel requires local financing, local accounts, and local liquidity buffers, raising the “cross-jurisdictional” score. Banks are thus incentivized to revert entirely to their home jurisdiction and trade everything in USD.

Similarly, the “substitutability” indicator is ironic: normally, being the sole provider is a competitive advantage, but in the G-SIB system, it’s a liability. If a bank is the only liquidity provider for BRL-MXN, it becomes a “critical infrastructure.” Regulators assign higher scores, leading to higher capital costs. A bank might think: “We dominate this niche channel, but being the sole provider pushes our G-SIB score higher, raising capital requirements. Better to shut it down and only trade USD.”

These scores are deadly because they directly translate into CET1 capital requirements. Moving from one G-SIB bucket to the next can increase the required capital by 0.5 percentage points. For a bank with $1 trillion in risk-weighted assets, that’s an additional $5 billion in capital. Even if a trading desk earns $50 million annually, if its activities push the bank into a higher G-SIB bucket, it could trigger billions in capital costs. The opportunity cost is clear—any smart bank would abandon such channels.

Why traditional FX market solutions are doomed to fail

The global FX market outside the G7 has already become structurally dysfunctional. Direct bilateral liquidity between emerging market currencies has almost disappeared, creating a “liquidity vacuum” in non-USD cross-border settlement.

This is not market failure but an inevitable result of rule design. The root cause of liquidity issues has never been insufficient trade demand but the prohibitive infrastructure costs taxed by regulatory frameworks. For decades, we relied on correspondent banking networks, but under the heavy G-SIB scoring pressure, this network is collapsing.

As long as global trade depends on balance-sheet-constrained institutions holding inventories, non-USD channels will remain fragmented, costly, and inefficient. Mimicking the old system with on-chain solutions is doomed from the start. Any attempt to replicate traditional FX market logic on-chain will inherit the same systemic failures.

The future of non-USD stablecoins must rely on DeFi-native innovations to bootstrap liquidity. Only by completely abandoning dependence on traditional financial infrastructure can we break the regulatory-imposed liquidity vacuum.