Source: Reuters

The problem is that the behavior of the Treasury curve simply does not support these conclusions.

Looking only at the long bond while ignoring the rest of the curve produces an incomplete interpretation of what markets are actually pricing. The Treasury market does not trade as thirty independent securities. It trades as an interconnected term structure where each maturity reflects different expectations regarding monetary policy, economic growth, but also the evolution of the policy rate. The recent move therefore should be interpreted as another phase of yield curve normalization following a reduction in the probability of additional Federal Reserve tightening rather than evidence of an inflation panic or a buyers’ strike against US government debt.

The distinction matters because the implications are fundamentally different.

At the July FOMC meeting, the Federal Reserve kept the federal funds target unchanged despite three dissents favoring an immediate increase (9-3). Chairman Kevin Warsh maintained an unmistakably hawkish tone, emphasizing that inflation risks remained significant and that policymakers should not become complacent. Nevertheless, markets interpreted the meeting as reducing the probability of an imminent rate hike rather than increasing it. The two-year Treasury yield declined materially immediately after the meeting, precisely because it is the maturity most sensitive to expected Fed policy.

Source: Reuters

The 10-year yield changed comparatively little, while the 30-year yield drifted higher. But no matter what, that is not the profile of an inflation scare. If anything, at least to me, this is a classic example of a bull steepening process in which the front end rallies more aggressively than the long end while the curve gradually returns toward a more normal upward slope.

Source: Federal Reserve Bank of St. Louis

Understanding this distinction requires moving beyond the simplistic idea that every increase in long term yields must reflect deteriorating confidence.

Historically, yield curve inversions emerge when monetary policy becomes substantially tighter than what markets consider consistent with long run equilibrium. The inversion is therefore an artificial configuration generated by policy rates remaining above longer-term expectations for nominal growth. When investors begin to assign a lower probability to further tightening, the first reaction almost always appears at the front of the curve because those maturities directly reflect expected policy rates. The long end behaves differently (especially since they are out of the control of a central bank). Rather than collapsing alongside the front end, it often adjusts more slowly, producing a steepening of the curve.

If investors genuinely believed inflation were spiraling out of control, they would demand higher compensation across the entire intermediate and long end of the curve. Inflation expectations would push five-year, seven-year and ten-year yields sharply higher alongside the thirty-year bond. Instead, much of the recent adjustment has been concentrated at the very long end while policy-sensitive maturities have declined. The market is therefore differentiating between near-term monetary policy and long-term equilibrium. They are not pricing a generalized inflation shock.

This also explains why interpreting the move as a fiscal event is problematic. The fiscal narrative implicitly assumes that investors have suddenly become unwilling to finance US government deficits. Yet Treasury auctions continue to clear smoothly, bid-to-cover ratios remain historically healthy, and foreign participation has not collapsed. For instance, the last 30y auction drew a bid-to-cover ratio of 2.44x, in line with recent sales. Indirect bidders, mostly foreign investors, took nearly 78% of the supply. Moreover, the Treasury market remains the deepest and most liquid sovereign bond market in the world, as we discussed so many times on this blog.

The notion of bond vigilantism also struggles to explain the relative behavior of different maturities. A genuine buyers’ strike would normally produce a broad repricing of duration risk, not an isolated adjustment concentrated toward the longest maturity. Even more importantly, such an episode would likely coincide with widening inflation breakevens and substantial increases in Treasury term premia across maturities. That is not what current market indicators suggest. Not even remotely close, see the image below.

Source: Federal Reserve Bank of St. Louis

This distinction becomes especially important when considering recent developments in energy markets. Oil prices undoubtedly increased following geopolitical disruptions in the Middle East. Conventional commentary immediately concluded that higher oil prices necessarily imply higher inflation and therefore higher Treasury yields. Yet history demonstrates that oil shocks are considerably more complex than this simple relationship suggests.

An increase in oil prices simultaneously raises measured inflation while reducing real disposable income. Higher gasoline prices function as a tax on consumption. Households devote a larger share of their income to energy expenditures, leaving fewer resources available for discretionary consumption. Firms experience higher input costs while demand simultaneously weakens. Investment slows. Consumption slows. Credit demand deteriorates. About these things I discussed here or here.

Consequently, long-term Treasury yields need not rise because investors expect permanently higher inflation. They may instead reflect uncertainty regarding future monetary policy responses. This distinction is essential. Markets increasingly appear to believe that central banks could eventually tighten policy not because underlying demand is overheating but because policymakers continue interpreting temporary energy price increases through conventional inflation-targeting frameworks. In other words, markets are pricing the probability of future policy mistakes rather than persistent inflation itself.

Europe offers perhaps the clearest historical parallel. During 2008 and again in 2011, the European Central Bank tightened monetary policy largely in response to commodity-driven inflation despite rapidly deteriorating economic conditions. Those decisions are now widely regarded as policy errors because the inflation shock originated from supply disruptions while demand remained weak.

The current Treasury curve resembles markets assigning a non-trivial probability that central banks could once again overreact to temporary commodity-price movements. This interpretation also aligns remarkably well with recent macroeconomic data. Second-quarter GDP figures across advanced economies appeared superficially encouraging. Consumer spending improved modestly in the United States. Several European economies avoided outright contraction. And even so, the strength appears driven by temporary factors including front-loaded purchases and AI-related capital expenditure concentrated within a relatively narrow set of firms.

This combination hardly resembles the environment associated with sustained inflationary expansions. Instead, it resembles what we have repeatedly observed since late 2023: short-lived mini-cycles generated by temporary shocks rather than durable recoveries. Each time activity stabilizes, financial markets quickly extrapolate “resilience”. Each time, subsequent data reveal that much of the improvement reflected temporary distortions rather than genuine cyclical acceleration.

The Treasury market appears considerably less enthusiastic than headline commentators. Perhaps the most overlooked aspect of the recent move concerns what happened to the 2s10s spread. The spread had narrowed significantly during periods when markets assigned elevated probabilities to additional Federal Reserve tightening. Following the latest FOMC meeting, however, the spread widened materially as two-year yields declined faster than ten-year yields. Historically, this pattern is consistent with the gradual removal of policy-induced distortions.

Source: CNBC

Another frequently overlooked consideration concerns the role of Treasury collateral within the global financial system. Treasuries are not simply government liabilities. They constitute the primary collateral asset underpinning global repo markets, derivatives margining and/or wholesale dollar funding. Their pricing also depends on collateral demand and, of course, dealer balance sheet constraints. Consequently, modest increases in long-end yields need not imply deteriorating sovereign creditworthiness. They may instead reflect the (evolving) microstructure of Treasury intermediation.

Ultimately, the rise of the 30-year Treasury yield above 5.2% appears far less dramatic than headlines suggest. Rather than signaling the collapse of confidence in US public finances, the recent move fits a much more coherent narrative: policy-sensitive maturities have begun adjusting downward as markets reduce the probability of further near-term tightening, while longer maturities continue participating in the gradual normalization of a yield curve that has remained deeply distorted for several years. Bull steepening is an evidence that markets are attempting to return toward a more normal term structure after an extended period in which monetary policy held the front end artificially elevated.

If anything, the recent behavior of the Treasury curve confirms that markets remain primarily focused on monetary policy uncertainty rather than fiscal sustainability. Inflation expectations remain comparatively well anchored. Growth continues exhibiting characteristics of temporary mini-cycles rather than durable expansion. Energy shocks continue functioning more as taxes on consumption than engines of persistent inflation. Under those conditions, interpreting the rise in the thirty-year yield as evidence of Treasury rejection risks confusing the shape of the curve with the message it is actually transmitting.

However, the current market dynamics suggest that the rise in yields is not primarily the result of a traditional bond vigilante reaction. The market is not necessarily questioning Japan’s ability to service its debt or suddenly pricing a sovereign solvency crisis. Instead, the recent move appears to reflect an adjustment in the JGB market, where the traditional buyers of Japanese government bonds have become reluctant to absorb additional duration risk at a time when the Bank of Japan (BOJ)’s future path of interest rates is uncertain.

This distinction is important because the mechanics behind rising yields are fundamentally different. A bond vigilante episode occurs when investors actively withdraw confidence from a sovereign issuer and demand a higher risk premium because fiscal sustainability has deteriorated (maybe the case of Greece or Ireland during the Eurocrisis). The current JGB market appears closer to a buyers’ strike, where investors are not necessarily rejecting Japanese government debt but are temporarily unwilling to increase their holdings because the risk-reward balance has changed. The issue is mainly about the availability and willingness of marginal buyers.

The latest pressure on Japanese government bonds emerged after the government announced a long-term economic strategy designed to mobilize more than ¥370 trillion ($2.29 trillion) in combined public and private investment through fiscal 2040. The plan focuses on strengthening strategic industries, increasing investment in technology, but also on reducing vulnerabilities exposed by recent global supply chain disruptions.

From a growth perspective, the strategy reflects Japan’s attempt to address long-standing structural weaknesses, including weak productivity growth and insufficient domestic investment. However, bond markets evaluate fiscal announcements not only through their potential economic benefits but also through their implications for future government financing requirements. The possibility of additional fiscal commitments raises questions about future JGB issuance at precisely the moment when some of the largest domestic holders of Japanese government bonds are reassessing their exposure.

As we know, domestic investors, particularly financial institutions, pension funds, and insurers, have historically represented the core demand base for government bonds. This domestic ownership structure allowed Japan to maintain extremely low borrowing costs despite its elevated debt-to-GDP ratio because the majority of government debt was absorbed internally.

That equilibrium depended heavily on a low-interest-rate environment. For decades, Japanese banks and institutional investors could accumulate JGBs because the opportunity cost of holding government bonds was limited and the BOJ provided a powerful stabilizing presence through quantitative easing (QE) and, later, yield curve control (YCC).

That framework is now changing. The BOJ’s decision to gradually unwind its ultra-loose monetary policy, including the abandonment of negative interest rates and the adjustment of its YCC framework, has introduced a new source of uncertainty for bond investors. The BOJ has emphasized that future policy decisions will remain dependent on inflation and wage developments, but the market is attempting to determine where the eventual equilibrium interest rate will settle.

Moreover, one of the most important factors behind the recent JGB volatility is that higher yields do not automatically generate stronger demand when investors are uncertain about the direction of rates. In theory, rising yields make bonds more attractive because they offer higher returns. However, for large financial institutions managing substantial fixed income portfolios, the decision is more complex.

Japanese banks, insurers, and pension funds do not simply evaluate bonds based on yield levels. They must consider duration exposure and potential mark-to-market losses. A rapid increase in long-term yields can significantly reduce the market value of existing bond holdings, particularly for institutions that accumulated large portfolios during decades of declining or stable interest rates. This is why major Japanese banks reported ¥7.05 trillion ($45 billion) in unrealised losses on JJGB holdings in the final three months of 2025, with Japan Post Bank and Mitsubishi UFJ Financial Group (MUFG) accounting for the largest increases.

This creates a situation where investors may prefer to wait for greater clarity rather than immediately increase allocations. A bank buying a 10-year JGB at a 2.8% yield faces the possibility that yields could continue moving higher if inflation remains persistent or if the BOJ signals additional tightening. In that scenario, the institution would experience valuation losses shortly after purchasing the bonds.

The rational response is often to delay purchases. This behavior reduces liquidity and weakens the traditional stabilizing mechanism of the JGB market. When major institutional investors step back simultaneously, even moderate increases in government bond supply can create disproportionate market moves because the marginal buyer becomes less predictable. As such, the issue is that they are uncertain about the correct price for duration in a market transitioning away from decades of monetary intervention.

The behavior of Japan’s largest financial institutions demonstrates this shift. Mizuho Financial Group has significantly reduced its holdings of Japanese government bonds, lowering its portfolio from approximately ¥25.1 trillion several years ago to around ¥8.3 trillion. The institution has indicated that it prefers to wait for greater confidence that the BOJ’s rate hiking cycle is approaching its conclusion before rebuilding a larger bond position.

The importance of this development extends beyond one institution. Japan’s financial system was built around the assumption that JGBs were effectively risk-free assets from a price volatility perspective. While the probability of default remains negligible, the market risk associated with holding government bonds has increased substantially because yields are no longer anchored by central-bank intervention.

The result is a paradox: higher yields should theoretically attract buyers, but rising volatility can prevent buyers from entering the market until yields have moved even higher.

Bond markets are particularly vulnerable to self-reinforcing dynamics. Initial selling pressure pushes yields upward. Higher yields increase uncertainty about future valuations. Increased uncertainty discourages institutional investors from purchasing bonds. Reduced demand then requires yields to rise further in order to attract buyers. This mechanism can create significant market stress without requiring any deterioration in fiscal fundamentals.

The situation resembles a liquidity adjustment. Japan does not have the same characteristics as emerging market sovereign crises, where foreign investors suddenly question debt sustainability and currency stability. The majority of Japanese government debt remains denominated in yen, and the country maintains a large domestic savings base. The challenge is that the marginal buyer is changing. For years, the BOJ itself represented a dominant force in the JGB market. Through its asset purchase programs, the central bank absorbed a significant share of government bond issuance and suppressed volatility. This is why the BoJ said at its meeting last month that it would stop reducing its monthly bond purchases next year, instead levelling off at around ¥2tn ($12.5bn) a month.

Nevertheless, with the BOJ gradually retreating from that role, private sector investors must become the primary mechanism for determining prices. That transition is likely to remain volatile because markets are effectively rediscovering the appropriate yield structure after decades of artificial suppression.

The Japanese currency adds another layer of complexity. Traditional monetary theory suggests that a country moving toward higher interest rates should experience currency appreciation as investors seek higher returns. Yet the yen has remained weak despite the BOJ’s tightening cycle. This divergence indicates that domestic interest rate differentials alone cannot explain currency movements, as I explained in this post.

The yen’s performance continues to be influenced by global liquidity conditions and the structure of the global dollar system. Japan has historically been one of the largest sources of funding liquidity because investors borrowed yen at extremely low rates to finance investments in higher-yielding assets elsewhere, creating the well-known yen carry trade.

The unwinding of these positions depends not only on Japanese monetary policy but also on global financial conditions, particularly the availability of dollar liquidity through the offshore dollar system (the Eurodollar system). Research from institutions such as the Bank for International Settlements has repeatedly confirmed the importance of global dollar funding conditions in shaping international financial markets. A stronger yen requires not only higher Japanese rates but also changes in the global demand for dollar-based liquidity. This explains why the yen can remain weak even as Japanese yields rise.

In conclusion, the current JGB market adjustment represents a transition away from the financial repression that characterized Japan’s economy for decades. Extremely low interest rates allowed the government to maintain high levels of debt while domestic institutions accumulated government bonds as stable, low-risk assets. That model is gradually disappearing. The question facing Japan is whether the financial system still has the capacity and willingness to absorb additional government borrowing without significantly higher yields.

If Japan’s fiscal expansion succeeds in generating stronger productivity growth and even higher nominal GDP, higher yields could eventually become manageable. However, if government borrowing expands faster than private sector demand for bonds, financing costs could rise at precisely the moment when institutional investors remain cautious. As such, the recent increase in JGB yields is not a classic bond vigilante event. If anything, it is a repricing caused by the withdrawal of traditional buyers from the market and the uncertainty surrounding Japan’s transition toward a post-zero-interest-rate environment.

Of course, as usual, the Bank of England’s latest proposal to limit the amount of leverage that hedge funds can employ when financing UK government bonds has largely been interpreted as another attempt to reduce financial stability risks created by non-bank financial institutions. The immediate narrative is straightforward: highly leveraged basis trades amplify volatility during periods of market stress, and regulators therefore seek to constrain excessive borrowing before the next crisis emerges (about leveraged basis trades I discussed here).

There is certainly some truth to that interpretation. The liability-driven investment (LDI) crisis of September 2022 demonstrated how leveraged positions can rapidly transform what initially appears to be an orderly repricing into a disorderly liquidity event. Yet focusing exclusively on hedge fund leverage risks missing the institutional transformation that has taken place over the past fifteen years.

The Bank of England has repeatedly argued that leverage among non-bank financial institutions can create destabilizing feedback loops during periods of stress because margin calls force investors to liquidate government bonds precisely when market liquidity is already deteriorating. What deserves greater attention, however, is why those forced sales become so destabilizing today compared with earlier decades.

The answer lies in dealer intermediation. Before the GFC, primary dealers functioned as genuine shock absorbers. When large institutional investors sold government bonds, dealers expanded their inventories, financed those securities through repo markets, and gradually redistributed the risk across the financial system. Temporary imbalances between buyers and sellers rarely translated into severe market dysfunction because dealer balance sheets possessed considerable elasticity.

That world no longer exists. Post-crisis regulatory reforms fundamentally altered the economics of market making, we already went through a process of hedge-ification, as I discussed here. Supplementary leverage ratios, Basel III capital requirements, liquidity coverage ratios, G-SIB surcharges, stress-testing frameworks, you name it, have substantially increased the balance sheet cost of intermediating sovereign securities. Or, from a more technical point of view, Basel III rules and national regulations imposed leverage ratios, which cap banks’ balance sheets regardless of riskiness of assets, liquidity coverage and net stable funding ratios, which increase the cost of holding inventory and of course stricter risk limits and internal capital charges. As the chart below shows, hedge funds have become major players in these markets, especially in the UK:

Source: Reuters

The implications are profound. Market liquidity is no longer determined primarily by the willingness of investors to hold government debt. Instead, it depends on the willingness and ability of regulated intermediaries to temporarily warehouse that debt.

This distinction fundamentally changes how we should interpret hedge fund leverage. Much of the recent discussion focuses on relative-value trades, particularly basis trades that exploit pricing differences between cash government bonds and futures. These strategies typically employ significant leverage obtained through repo financing. Under normal market conditions, the trade appears relatively low risk because the spread between the two instruments eventually converges. However, the stability of the strategy on uninterrupted access to funding. As documented by the Financial Stability Board’s work on non-bank financial intermediation, leverage transforms relatively small price movements into large funding shocks because lenders demand additional collateral whenever volatility rises.

Source: Financial Times

Once margin calls begin, leveraged investors must either obtain additional cash or liquidate positions. The resulting sales increase market volatility, which generates further margin calls, producing a classic liquidity spiral. The LDI episode provided perhaps the clearest recent example of this mechanism. The Bank of England itself concluded that the crisis was driven by forced collateral sales associated with leveraged liability-driven investment strategies.

Importantly, government bond yields rose not because investors suddenly questioned the creditworthiness of the United Kingdom but because balance sheet constraints prevented dealers from absorbing the extraordinary volume of sales. This distinction remains underappreciated. Traditional finance often interprets sovereign yields through expectations of inflation, fiscal sustainability, or even (or especially) future monetary policy. But there is a variable that deserves equal attention: intermediation capacity.

When dealer balance sheets become constrained, sovereign bonds can temporarily trade at prices that reflect funding scarcity (and not even related macroeconomic fundamentals). The same lesson emerged during the US Treasury market turmoil of March 2020. Investors around the world attempted to sell US Treasuries during one of the largest flights to liquidity in modern history because they needed USD funding liquidity. The resulting dysfunction forced the Federal Reserve to intervene aggressively, purchasing Treasuries not because government solvency was in question but because dealer balance sheets had reached their effective limits.

The common element linking March 2020 and September 2022 was therefore not excessive sovereign borrowing. Not even close. If anything, it was all about the insufficient balance sheet elasticity. Seen through this lens, the Bank of England’s proposal appears considerably more logical. Constraining hedge fund leverage effectively reduces the size of potential forced liquidations that dealers may eventually be required to intermediate. Rather than expanding dealer balance sheet capacity, a politically and regulatorily difficult objective, the authorities are attempting to reduce the magnitude of future balance sheet demands. But it represents a demand-side solution to a supply-side problem.

One potential concern is that leverage restrictions merely redistribute activity rather than reduce systemic leverage. Financial history repeatedly demonstrates that tighter regulation of one segment often pushes leverage toward less regulated institutions. Futures contracts that are used today by non-banking financial institutions to engage in US Treasury markets, as I discussed in my latest research paper (you can read it here). The growth of private credit following tighter bank regulation is another example of this phenomenon.

The same may occur within sovereign bond markets. If leverage becomes more expensive for UK-based hedge funds, trading activity could migrate toward offshore entities, alternative financing structures or synthetic exposures that remain outside the immediate regulatory perimeter. The Financial Stability Board has repeatedly emphasized that monitoring leverage within non-bank financial institutions remains considerably more difficult than within the banking sector because reporting standards remain fragmented internationally.

Another issue concerns market liquidity itself. And that’s because leverage is not inherently destabilizing. It is also used to finance the entire market-making activity. For example, relative-value arbitrage contributes to pricing efficiency by narrowing discrepancies between closely related securities. If leverage restrictions become excessively stringent, bid-ask spreads could widen while market depth deteriorates, particularly during normal market conditions.

As such, reducing leverage decreases systemic fragility but may simultaneously reduce secondary market liquidity. And an equally important implication concerns repo markets. Modern sovereign bond markets cannot be understood independently from secured funding markets because repo financing effectively transforms government securities into monetary instruments. A government bond is no longer simply a duration asset. It simultaneously functions as collateral supporting cash creation throughout the financial system. Consequently, restricting leveraged repo activity inevitably affects monetary transmission. This observation becomes particularly relevant during periods when central banks are simultaneously shrinking their balance sheets through quantitative tightening.

As reserves decline, repo markets assume an important role in reallocating liquidity across financial institutions. If leverage restrictions reduce repo intermediation at precisely the same moment reserves become scarcer, authorities may inadvertently tighten financial conditions more than intended. This interaction shows why monetary policy cannot be analysed independently from financial market structure.

Another dimension concerns the growing importance of non-bank financial institutions. Over the past decade, pension funds, insurers, asset managers and hedge funds have become progressively larger participants in sovereign bond markets, partly because banks have reduced balance-sheet-intensive activities following post-crisis regulation, as argued above. Ironically, regulation designed to reduce banking sector risk has increased the systemic importance of institutions that central banks traditionally supervise less directly.

The Bank of England’s proposal therefore reflects an institutional reality. Central banks find themselves responsible for stabilising markets whose dominant participants fall largely outside conventional banking supervision. This evolution raises questions about the future architecture of sovereign debt markets. Should authorities continue regulating leverage among end-investors while leaving dealer balance sheet constraints largely unchanged? Or should policy instead focus on expanding the capacity of regulated intermediaries to absorb temporary market imbalances?

Recent debates surrounding the SLR in the US demonstrate that these questions extend well beyond the United Kingdom. Numerous market participants have argued that relaxing selected balance sheet constraints during periods of stress could improve Treasury market resilience without materially weakening bank safety. Whether similar discussions eventually emerge in the United Kingdom remains uncertain. Nevertheless, the underlying issue appears remarkably similar.

Ultimately, interpreting the Bank of England’s proposal solely as an attempt to control hedge fund leverage risks overlooking the more significant transformation taking place across advanced financial systems. The central problem today is the interaction between leverage, collateral markets and, of course, the inelastic dealer balance sheets. The modern sovereign bond market functions as both a capital market and a monetary infrastructure. Government securities are collateral instruments and funding vehicles simultaneously. Once these multiple functions are recognised, episodes such as the Treasury market dysfunction of 2020, the UK gilt crisis of 2022 and the recurring episodes of repo market stress appear considerably less mysterious.

They all reflect the same reality, namely that financial stability depends on who possesses the balance sheet capacity to intermediate it when everyone else wants to sell at the same time. From that perspective, the Bank of England is attempting to compensate for a financial architecture in which the private sector’s ability to absorb sovereign collateral has become structurally less elastic than the markets it is expected to support. And that is because of the post-GFC regulations, whether we like it or not.

If nominal rate spreads alone explained exchange rate behavior, recent developments should have produced at least some appreciation pressure on the yen. The Bank of Japan has progressively exited the exceptional monetary accommodation regime that defined the post-1999 period. In plain English, policy rates are no longer negative and yield curve control has been substantially relaxed. Yet the yen continued to weaken. That divergence immediately suggests that the mechanism cannot be reduced to textbook carry incentives.

The more convincing explanation begins with Japan’s structural position within the global dollar system. Japan is not merely a low yield jurisdiction. It remains one of the largest exporters of capital globally and continues to maintain one of the world’s largest net international investment positions. Japanese banks, insurers, pension funds, institutional investors, you name it, collectively hold trillions of dollars in foreign assets, with a significant concentration in USD-denominated securities - including US Treasuries, agency debt, corporate bonds and/or private market exposures.

This distinction matters because those portfolios generate a permanent demand for dollar funding that exists independently of current policy rates (shocking, right?). Traditional carry trade narratives assume that investors borrow cheaply in yen and purchase higher-yielding foreign assets. While such trades still exist, a growing share of international allocation now occurs through synthetic balance sheet structures. The Bank for International Settlements has repeatedly documented that FX swaps and forwards have become one of the largest channels through which institutions obtain synthetic dollar funding, with total outstanding positions now exceeding $80 trillion notionally and the US dollar appearing in approximately 90% of contracts.

Source: Bank for International Settlements

This changes the interpretation of exchange rate dynamics. For instance, a Japanese institution purchasing Treasuries often does not physically obtain dollars in the traditional sense. Instead, it may hedge currency exposure through rolling FX swaps, effectively borrowing dollars synthetically while maintaining yen funding on the liability side. Under normal conditions this structure appears frictionless because dealer banks absorb the mismatch between cash and synthetic funding demand.

But this intermediation process depends critically on dealer balance sheet capacity. Once dealer constraints tighten, obtaining dollars becomes more expensive independently of policy rates. This mechanism has become more important since the post-2008 regulatory transition. Research from the BIS and multiple central banks has shown that covered interest parity (CIP), historically treated as one of the most stable relationships in international finance, has persistently broken down in the modern era because arbitrage itself consumes scarce balance sheet capacity. I have already written about CIP here, in case you are curious to explore the topic further.

Source: Bank for International Settlements

In more detail, under CIP, interest rate differentials should be offset through forward exchange markets, eliminating arbitrage opportunities. In practice, persistent deviations now emerge because dealers require compensation to intermediate funding flows. The result is the cross-currency basis. The cross-currency basis effectively measures the premium foreign institutions must pay to obtain dollars synthetically. When demand for dollars increases while intermediation capacity remains constrained, the basis becomes (increasingly) negative. At that point exchange rates begin to reflect balance sheet scarcity (more than a simple portfolio allocation).

This loop creates conditions under which currency depreciation becomes self-reinforcing. A stronger dollar raises effective funding costs and higher funding costs reduce dealer willingness to intermediate. As a result, reduced intermediation increases synthetic dollar scarcity. Dollar demand rises further and exchange rates adjust accordingly. From this perspective, the yen’s depreciation becomes more a reflection of persistent structural demand for dollar liquidity.

The March 2020 episode provides a useful empirical example. During the pandemic shock, institutions globally rushed to secure dollar liquidity. Japanese institutions were among the largest users of dollar funding markets. FX swap bases widened dramatically. The three-month basis moved to approximately –140 basis points against the dollar as funding markets became impaired.

Source: Bank for International Settlements

Importantly, this was not caused by changing interest rate expectations. It reflected a shortage of intermediation capacity. And, of course, the Fed responded not (only) by adjusting policy rates but by massively expanding central bank swap lines and introducing facilities such as FIMA repo to prevent forced sales of Treasuries.

That intervention demonstrated something fundamental. Dollar shortages emerge not because the world lacks dollars mechanically (hard to believe we could ever reach this stage), but because the institutions responsible for distributing those dollars become constrained. This distinction becomes relevant in the current environment. Why? Because the Bank of Japan can raise rates and still fail to alter the structural demand for offshore dollar liquidity.

Meanwhile, higher oil prices create an additional source of pressure. Japan remains structurally dependent on imported energy. Because global crude markets remain overwhelmingly denominated in dollars, higher oil prices mechanically increase external demand for USD liquidity. This relationship aligns with the dominant currency paradigm developed in recent international macroeconomic literature, which shows that trade invoicing itself generates persistent demand for dollars independently of exchange rate movements.

This point is frequently misunderstood. Oil shocks are often interpreted as challenges to dollar dominance (something I argued against a couple of weeks ago, you can read here). If anything, they frequently strengthen short term dollar demand because importers must secure additional funding to maintain trade volumes. And this was what happened during the US-Iran conflict. But for Japan this creates a particularly difficult combination. As we already know, energy imports require more dollars. At the same time, existing foreign portfolios require more hedging. And as the story goes, dealer capacity remains constrained, synthetic funding becomes more expensive, the exchange rate absorbs the adjustment and intervention therefore becomes limited as a policy tool.

Japanese authorities can sell reserves and purchase yen, but intervention addresses price rather than structure. Even today, there were renewed official concern regarding exchange rate volatility and preparedness for intervention if necessary. Historically, however, intervention has rarely reversed sustained depreciation when funding conditions remain unchanged. This is ultimately why the recent episode should not be interpreted as evidence that Japanese monetary normalization has failed, nor is it evidence that markets irrationally prefer dollars. Rather, it reflects the continued dominance of a global financial architecture organized around offshore dollar creation and synthetic funding.

And, as a conclusion, the important variable is not what rate the Bank of Japan sets. It is whether global balance sheets still require dollars. And so far, there is little evidence that they require fewer. If anything, recent developments suggest the opposite. That is why the yen’s depreciation should no longer be interpreted primarily as a carry trade story. And, more importantly, that is why the yen at 161.5 may therefore tell us less about Japan than about the continued resilience of the Eurodollar system itself.

Kevin Warsh’s first meeting as Chair of the Federal Reserve generated a predictable initial reaction. Financial media quickly classified the outcome as hawkish: rates were left unchanged at 3.50–3.75%, the updated projections shifted upward and, as a result, and markets materially reduced expectations of near-term easing. Several commentators immediately interpreted the meeting as the beginning of a new monetary regime. I am not convinced that this is the correct interpretation.

The June 17 decision matters less because of the level of the policy rate and more because of what it revealed about how the institution may operate under new leadership. The Federal Reserve kept rates unchanged in a unanimous vote and maintained its ample reserves operating framework while acknowledging continued “inflation” pressures (I will get to this point in a minute). At the same time, however, the meeting introduced meaningful changes to communication practices and to how uncertainty itself may be transmitted into markets.

Most commentary after the meeting focused on the updated projections. Market participants emphasized that a growing number of FOMC members now anticipate at least one additional tightening move later this year and interpreted this as evidence that the committee has become more restrictive. That reading is understandable but incomplete. The upward shift in projections should not automatically be interpreted as an intention to tighten in the short term. To me, it appears to reflect a reassessment of the persistence of supply-side inflation disturbances, particularly energy-related ones, and the probability that these disturbances remain politically and institutionally difficult to ignore. This distinction matters because the current macro environment does not resemble the inflation episodes that historically justified aggressive tightening cycles (we discussed about this here or here).

Recent inflation pressures emerge from sectors affected by supply disruptions and energy developments (not some form of broad-based overheating). It is enough to consider the annual PPI, which indeed rose to 6.5%. However, the increase was largely driven by energy prices, with gasoline surging 23.4% during the month and crude petroleum prices climbing nearly 12%. Moreover, let’s take a look at the 5y/5y forward inflation expectations. Nobody believes we are going through an inflationary cycle.

Source: Federal Reserve Bank of St. Louis

The Fed’s own communication explicitly referenced supply-side pressures and uncertainty linked to geopolitical developments. Yet despite acknowledging this context, Warsh avoided providing a clear policy path or a timetable for reaction.

That decision is arguably more important than the rate hold itself. One of the most interesting developments during the meeting was the deliberate reduction of forward guidance. Reports following the decision highlighted that the official statement was substantially shortened compared with previous meetings and that Warsh deliberately moved away from detailed signaling about future policy intentions. Kevin Warsh even refrained from submitting his own rate projection in the Summary of Economic Projections, signaling skepticism toward the modern practice of central banks attempting to shape expectations through increasingly “precise” communication.

If this interpretation is correct, then what happened on June 17 was not a hawkish pivot in the conventional sense. I see it as an institutional repositioning -ish. Since 2008, monetary policy in the United States gradually evolved beyond overnight rate setting. The Federal Reserve relied on communication itself as an operational instrument. Forward guidance became part of policy transmission. Markets became conditioned not simply to react to rates but to interpret future commitments embedded in projections and speeches.

Warsh’s first meeting suggests discomfort with this model. Importantly, there was no attempt to dismantle the post-crisis operational framework. The committee reaffirmed its commitment to maintaining ample reserves and did not announce accelerated balance sheet normalization or additional tightening of quantitative measures. I also criticized this potential normalization in this post. I am glad that, so far, no major changes have emerged in this direction. Despite Warsh’s historical criticism of balance sheet expansion and extraordinary monetary accommodation, operational continuity remained largely intact.

And that continuity deserves more attention than it received. Because if Warsh intended to signal a genuinely hawkish regime change, there were several obvious mechanisms available: accelerating quantitative tightening, lowering reserve abundance, maybe introducing a more restrictive reaction function. And luckily (at least for me), none of these happened. Instead, what changed was the willingness of the institution to reduce the amount of certainty it supplies to markets. Of course, could be argued that this creates a different type of tightening. Why? Because traditional tightening raises funding costs directly, but institutional tightening increases uncertainty around future funding costs.

That distinction may explain why markets reacted most aggressively at the front end of the Treasury curve. Following the meeting, short-dated yields (including 2s) rose more meaningfully than long-end yields. That move is difficult to reconcile with expectations of imminent inflation acceleration or immediate tightening. A more plausible interpretation is that markets began pricing a wider distribution of future policy outcomes. In other words, investors became less certain that the Fed would intervene quickly if growth deteriorates and less certain that rate cuts remain the default response to softer conditions.

This interpretation becomes even more relevant once we consider the macro context. Over recent months markets confronted conflicting signals, such as slower labor market momentum, rising concerns in private credit and, of course still-elevated headline inflation. Under these conditions the Fed faces a more complicated trade-off than traditional Taylor-rule frameworks imply. If energy shocks ultimately act more as a tax on consumption and investment than as a source of persistent inflation, as I argued so many times before on this blog, then aggressive tightening becomes difficult to justify. Raising rates into deteriorating demand conditions may stabilize inflation expectations but at the cost of amplifying weakness elsewhere in the financial system.

Warsh’s first meeting did not resolve this tension. His repeated emphasis on uncertainty and his reluctance to offer explicit guidance suggest that the Fed may allow markets themselves to absorb the burden of interpreting incoming data rather than attempting to pre-commit policy responses. This approach has important implications. For years markets benefited from an implicit assumption that the Federal Reserve would continuously compress uncertainty through communication and increasingly transparent reaction functions. That assumption may no longer hold. It seems like Warsh appears more interested in preserving policy optionality. But the thing is, this should not automatically be interpreted as hawkishness. If anything, optionality means refusing to guarantee accommodation. The consequence is that market volatility may increase even without changes in the policy rate itself.

As such, this meeting was about redefining the role of communication in monetary policy and that may ultimately prove more consequential than any 25-basis-point move. And this lead me to another (important) observation. The market interpreted Warsh’s debut as hawkish because expectations became less dovish. But those are not identical things. A central bank that withdraws guidance without withdrawing liquidity is not necessarily becoming more restrictive. And so, that distinction could become one of the defining features of the next phase of US monetary policy.

To be continued!

In the figure below we can see the Bund response to USTs supply shock (in bps):

Source: European Stability Mechanism

The empirical finding is both interesting and plausible. Financial markets remain highly integrated, and sovereign bond markets exhibit substantial co-movement. What deserves closer scrutiny, however, is the interpretation of the mechanism. The ESM framework implicitly treats Treasuries and Bunds primarily as competing safe assets within a global portfolio allocation problem. Yet modern sovereign bond markets operate through collateral and funding channels that are difficult to reconcile with a pure portfolio balance framework. Moreover, the notion that Treasuries have become “less special” immediately raises a deeper question: special in what sense?

The literature on convenience yields traditionally interprets specialness as investors’ willingness to accept lower yields on safe assets because they provide liquidity services. The classic work of Ricardo Caballero and Emmanuel Farhi links these premia to global shortages of safe assets and elevated demand for liquidity. Similarly, research from the Federal Reserve Board emphasizes that Treasury convenience yields reflect their role as exceptionally liquid instruments that can satisfy precautionary demand during periods of uncertainty.

However, the post-2008 financial system fundamentally altered the meaning of liquidity. A Treasury security is no longer merely a safe asset. It is also repo collateral, margin collateral, maybe a high-quality liquid asset (HQLA) under Basel regulations. From this perspective, convenience yields and collateral values are not identical concepts. The ESM paper largely interprets declining convenience yields as evidence that Treasuries are losing some of their unique attractiveness relative to Bunds. Yet an alternative interpretation is that the observed decline reflects changes in dealer balance sheet economics.

This distinction matters because the Treasury market has undergone profound structural transformations since the GFC of 2007-9. We heard about so many (complex) regulations, Supplementary leverage ratios, GSIB surcharges, liquidity coverage requirements, enhanced stress-testing frameworks, just to name a few. And these regulations have fundamentally altered how dealers absorb sovereign issuance. Research from the Bank Policy Institute (BPI) consistently highlights that post-crisis regulation has reduced dealer balance sheet elasticity, increasing the sensitivity of Treasury market functioning to large issuance cycles.

Source: Bank Policy Institute

Under such conditions, Treasury supply shocks are no longer transmitted exclusively through investor portfolios. They also (or mostly) operate through dealer balance sheets. The September 2019 repo disruption provides a useful example. At the time, aggregate reserves remained historically elevated relative to pre-crisis norms. Yet funding markets experienced severe stress because dealer balance sheets could not expand sufficiently to absorb collateral flows. The resulting spike in repo rates had little to do with investor perceptions of Treasury safety and much more to do with intermediation constraints. In previous posts (for example here or here) I’ve shown that balance sheet constraints were central drivers of the episode.

The same lesson emerged during March 2020. The Treasury market experienced one of the most severe liquidity disruptions in its modern history despite a global flight toward safety. Investors were attempting to sell Treasuries not because they doubted Treasury creditworthiness but because they needed liquidity. Dealer balance sheets rapidly became constrained, producing severe market dysfunction. Research from the BIS demonstrates that Treasury market stress during the pandemic reflected balance sheet capacity limitations as much as changes in investor demand.

These episodes complicate the ESM interpretation. If Treasury supply spillovers increasingly depend on dealer intermediation capacity, then a reduction in observed spillovers may not indicate that Treasuries are becoming less special. Instead, it may indicate that the transmission mechanism itself has shifted from portfolio allocation toward balance sheet management. The issue becomes even more complicated once reserves enter the discussion. The ESM analysis treats Treasury supply largely as an independent variable affecting sovereign yields. Yet from a monetary perspective, Treasury issuance cannot be separated from reserve dynamics. Treasury issuance alters the composition of private sector balance sheets and affects money market conditions through its interaction with the Federal Reserve’s balance sheet.

Recent research from the Federal Reserve Bank of New York demonstrates that Treasury issuance interacts directly with reserve balances and the overnight reverse repo facility, creating effects that extend far beyond conventional duration channels. This observation becomes especially important when considering the post-2022 period of quantitative tightening. A traditional portfolio-balance framework would predict substantial upward pressure on yields from the combination of increased Treasury issuance and Federal Reserve balance sheet reduction. Yet much of the adjustment occurred through declines in ON RRP balances rather than through large increases in term premiums.

Source: Federal Reserve Bank of St. Louis

Consequently, interpreting Treasury supply purely through convenience yields risks missing the monetary architecture in which these securities operate.

A second limitation of the ESM framework concerns international transmission. The paper implicitly assumes that Treasury spillovers reach Bund markets through relative pricing channels. Yet sovereign bond markets today are linked through common funding conditions as much as through investor substitution. For example, research from the ECB documents how dollar funding conditions increasingly influence euro-area bond markets through global collateral and repo channels. This implies that Bund-Treasury co-movement may reflect common funding constraints rather than merely substitution between safe assets.

Indeed, the distinction becomes particularly relevant when examining periods of elevated Treasury issuance. If additional Treasury supply expands the stock of collateral available to global funding markets, the resulting effects on Bund yields may arise through changes in repo conditions and dealer balance sheet utilization. Not through changes in investor demand for duration. This interpretation aligns more closely with the modern monetary system, where sovereign securities serve dual roles as investment assets and monetary instruments. The implication is that the concept of “specialness” itself may require reconsideration.

The ESM paper correctly identifies a changing relationship between Treasury supply and Bund yields. Yet the conclusion that Treasuries are becoming less special risks conflating changes in convenience premia with changes in monetary functionality. Treasuries remain the dominant collateral asset in global repo markets. They remain the primary reserve asset for central banks. They remain central to margining, derivatives clearing, and dollar funding. None of these functions has diminished materially. What has changed is the institutional environment in which Treasuries operate.

The distinction matters because markets are not currently pricing a 1970s-style inflation regime. Rather, they are pricing the possibility that central banks, confronted with several months of elevated headline CPI driven by energy costs, may feel compelled to maintain tighter policy for longer than macroeconomic conditions would otherwise justify. This is a fundamentally different mechanism than the standard “debt crisis” interpretation that often dominates public commentary.

The structure of the WTI futures curve continues to point toward an immediate and potentially prolonged supply disruption. Front-month prices remain heavily elevated, while persistent backwardation suggests markets continue to price acute near-term scarcity linked to the Strait of Hormuz disruption. However, the relatively anchored deferred end of the curve also indicates that markets do not expect structurally stronger long-term demand, but rather anticipate that weakening global growth will eventually suppress consumption. And this is also proven by the International Energy Association (IEA), which noted that “(W)orld oil demand is forecast to contract by 420 kb/d y-o-y in 2026, to 104 mb/d, 1.3 mb/d less than our pre-war forecast. The biggest decline is in 2Q26, down by 2.45 mb/d, of which the OECD accounts for 930 kb/d and the non-OECD for 1.5 mb/d”. Lower demand everywhere…

Source: Author’s representation

This distinction is important. Higher oil prices generated by supply disruptions do not function in the same way as demand-driven overheating. They operate more like a tax on consumption and production, compressing real incomes. Historically, sustained energy shocks have tended to weaken growth over time rather than generate persistent inflationary spirals unless accompanied by strong wage dynamics (which is not the case here).

This is precisely where current market pricing becomes revealing. If investors genuinely believed that the recent rise in yields reflected the emergence of broad-based inflation, Treasury Inflation-Protected Securities (TIPS) markets would be behaving very differently. Longer-term breakeven rates would be steepening aggressively and forward inflation expectations would be repricing materially higher.

Source: Federal Reserve Bank of St. Louis

Instead, the evidence shows something far narrower. As recent market analysis correctly emphasizes, near-term breakevens have risen largely because markets expect several months of elevated gasoline-related CPI prints, not because investors foresee a structurally inflationary environment. The five-year breakeven rate has moved higher, but not remotely to levels historically associated with genuine inflation scares. During the 2022 supply shock, comparable breakevens moved dramatically higher because inflation pressures were broad-based across goods, services, wages, supply chains, basically everything. Today, the move is overwhelmingly concentrated in energy-sensitive pricing.

Even more importantly, longer-term inflation expectations remain comparatively stable. The five-year/five-year forward inflation measure has not exhibited the type of repricing consistent with a market expecting runaway inflation (same for 10-year/10-year). Instead, the curve remains relatively anchored, suggesting that investors continue to view the current shock as transitory in macroeconomic terms, even if persistent in energy markets.

Source: Federal Reserve Bank of St. Louis

This is where the Treasury curve itself becomes especially important. The most significant move has occurred in policy-sensitive maturities, particularly the 2-year yield. Under normal conditions, sharp moves in the 2-year segment are interpreted as expectations of imminent Federal Reserve tightening. Yet short-term funding markets tell a more complicated story.

SOFR futures and Treasury bill pricing have not behaved as though markets expect an immediate sequence of aggressive rate hikes. Instead, they suggest that traders see the Federal Reserve trapped in a “strange” position: inflation data may remain elevated because of energy prices, but growth conditions continue to deteriorate. This distinction helps explain the recent shape of the curve. Markets are not pricing strong growth and inflation. They are pricing the increasing probability that central banks may remain hawkish longer than fundamentals justify because headline CPI remains politically and institutionally difficult to ignore. However, markets also expect the Fed to be forced into becoming more dovish as economic problems intensify, and that is what the three-month SOFR curve is beginning to show.

Source: Federal Reserve Bank of St. Louis

This mechanism has become increasingly visible over the past several years. Central banks, particularly after the post-pandemic inflation episode, have become highly sensitive to energy-driven headline inflation. Even where disinflationary forces dominate, elevated oil prices can alter policy expectations simply because policymakers fear the reputational consequences of appearing complacent.

The market understands this dynamic well. The rise in intermediate and longer-term yields therefore reflects not a sudden reassessment of sovereign creditworthiness, but the expectation that central banks may maintain restrictive policy stances even as growth weakens.

Importantly, the relationship between yields and inflation expectations has itself become more unstable. In recent episodes, real yields and breakevens have occasionally moved in opposite directions, reflecting uncertainty about whether central banks will respond aggressively enough - or too aggressively - to energy shocks. This reinforces the idea that markets are increasingly focused on policy reaction functions rather than inflation itself, what I already argued on this post written last month.

At the same time, there is little evidence that investors are abandoning Treasuries as a safe asset. Foreign demand remains structurally strong, even if official reserve managers occasionally liquidate holdings to obtain dollar liquidity during stress episodes. The recent decline in yields during periods of intensified geopolitical tension further undermines the “Treasury rejection” narrative. If markets genuinely feared fiscal collapse or inflationary monetization, yields would not decline when geopolitical risk intensifies.

Instead, the opposite has often occurred. As growth fears rise, investors continue to seek duration exposure despite elevated oil prices. This is why a comparison between March 2025 and March 2026 shows that Major Foreign Holders of Treasury Securities increased from $9.05tn to $9.34tn. This is because markets increasingly interpret energy shocks through a disinflationary lens over the medium term. Higher oil prices compress consumption, weaken margins, but also reduce investment capacity. In other words, they eventually become growth-destructive.

Recent labor market data reinforce this interpretation. For instance, hires are now roughly at pandemic-crisis levels, pointing to a fragmented labor market.

Source: Federal Reserve Bank of St. Louis

This matters because persistent inflation typically requires not only supply-side pressure but also resilient aggregate demand. Without strong wage growth and hiring, energy-driven price increases tend to dissipate into weaker growth rather than broader inflationary persistence.

The yield curve is therefore reflecting a highly asymmetric policy environment. Central banks face pressure to respond to elevated headline inflation, yet macroeconomic conditions increasingly argue for caution. Markets are pricing the risk that policymakers err on the side of hawkishness because institutional credibility considerations push them toward maintaining tighter conditions.

This distinction is crucial because it changes how we interpret curve steepening and yield volatility. The current environment does not resemble a classic sovereign debt crisis. And the Treasury curve, SOFR markets, and TIPS show that markets are not pricing an inflationary boom. They are pricing the possibility that central banks remain too hawkish for too long in response to energy-driven CPI pressures. That is a profoundly different narrative than “bond vigilantes” rejecting government debt.

Indeed, the behavior of longer-term inflation expectations strongly suggests that investors still believe disinflationary forces dominate over the medium term, especially as it will lead to lower consumption. Oil shocks may temporarily elevate headline inflation, but without strong second-order effects they are unlikely to generate a sustained inflationary regime.

The two biggest stories in private credit right now are actually the same story, and the FSB’s 6 May report is the first official document to (sort of) admit it. On one side, private credit funds have spent five years lending heavily to PE-owned software companies whose moats agentic AI is now busy eroding. On the other hand, those same funds (and their insurance affiliates) are bankrolling the AI data centre buildout through SPVs that keep the debt off everyone’s balance sheets. Same asset class, same blind spots, same retail investors holding the bag through semi-liquid BDCs.

A caveat. We’re not predicting a 2008-style blow-up. Howard Marks made the point well in his April memo: the volatility now visible in semi-liquid vehicles is mostly flow- and sentiment-driven rather than underlying credit deterioration. The deterioration is coming, but on a longer fuse than the panic suggests. What we are saying: the structural fragilities the FSB flagged are real, the data gaps are worse than the headline numbers suggest, and the AI angle adds two distinct (and partially correlated) sources of stress the asset class hasn’t priced.

What the FSB said

The FSB’s Report on Vulnerabilities in Private Credit dropped on 6 May. The numbers aren’t subtle. Private credit globally stands at $1.5–2.0 trillion at end-2024, the US alone at roughly $1 trillion. Banks have extended approximately $220 billion in credit lines to private credit funds, but commercial estimates put the real figure at $270–500 billion. The precise number is unknown to the regulator itself. Retail investors now account for roughly 13% of the asset class, up from near zero a decade ago. The sector, in the FSB’s own words, “has not been tested during a severe economic downturn”.

Source: Financial Stability Board

Two things deserve attention. First, the FSB flags the rise of payment-in-kind (PIK) interest, where borrowers capitalise interest rather than paying cash, keeping reported defaults low while distress builds beneath the surface (a point covered in an earlier post on selective defaults). Second, the report takes the gloves off on data gaps: most FSB authorities cannot separately identify private credit funds in their reporting frameworks, lack loan-level data, and/or face legal impediments to sharing what little they do have. Regulators are flagging systemic risk in an asset class they cannot see in any granular sense. From the lender side, it’s worse. Much of what one would want to monitor sits inside fund-of-funds, feeder vehicles, and insurance general accounts with no public disclosure.

AI risk #1: the SaaS loan book

There are two AI fault lines, and the industry consistently muddies them. The BIS estimates that outstanding private credit loans to SaaS firms rose from roughly $8 billion in 2015 to over $500 billion by the end of 2025.

Source: Bank of International Settlements

Morgan Stanley puts software exposure in direct lending at around 26%; UBS puts 25–35% of portfolios at risk from AI disruption. Apollo quietly halved its software exposure in 2025, from around 20% to 10%. And when Apollo de-risks that aggressively, well… Most of the time it is a signal.

SaaS was the perfect borrower profile (recurring revenue, sticky customers, asset-light), and PE sponsors paid 25–30x EBITDA on its strength. Then, agentic AI started reducing customer headcount, AI-native competitors entered defended markets, and per-seat pricing began giving way to usage-based models. The S&P North American software index fell 15% in January 2026, its worst month since October 2008; Ares fell more than 12% in a week, Blue Owl 8%, and KKR almost 10%. Roughly 15% of SaaS borrowers are already struggling to cover interest expense and distressed PIK deferrals reached approximately 6.4% of total private debt volume in Q1. And the 26% headline exposure is almost certainly understated. Many software-driven businesses are classified as “business services,” healthcare IT, or fintech. Half the “business services” book in a typical mid-market PE fund is really workflow software with a services wrapper.

AI risk #2: financing the buildout

The second fault line runs the other way: private credit is bankrolling the AI infrastructure buildout through structures designed to be hard to see. McKinsey estimates global data centre spending could reach $7 trillion by 2030. The interesting financing is off-balance sheet.

The template is the Meta–Blue Owl Hyperion deal in October 2025: a $27 billion A+-rated SPV anchored by PIMCO, with Meta holding 20% of the JV and leasing back the completed campus under a long-term operating lease (for a detailed overview of how AI centre buildout is funded today, you can check my Substack post written back in December). None of the debt touches Meta’s balance sheet. Equity is roughly 8.5% of total financing. An RVG from the operating tenant gets the structure to investment grade. The template is being replicated across Microsoft, Oracle, and others. Three BIS officials called these structures “shadow borrowing”: obligations economically akin to debt but largely outside corporate balance sheets, with banks supporting them via funding lines creating “new shock transmission channels”.

The underwriting concern is sharper than the regulator commentary suggests. NAV lenders apply a 30% LTV ceiling against diversified portfolios of mature PE assets, with cash sweeps and concentration limits. Hyperion effectively runs at ~91% LTV against an asset that doesn’t yet exist, financed against future cash flows from a tenant whose business model depends on AI customers whose monetisation curves are not yet known. Or, in the wise words of Matthew McConaughey in The Wolf of Wall Street, “It is a fugazi. It’s a whazy. It’s a woozie. It’s fairy dust”.

The RVG and the long-term lease are real mitigants, but they’re correlated: if Meta’s AI strategy doesn’t generate the revenue to support the lease, both become more questionable at the same moment. The investment-grade rating is genuine but conditional, as it depends on the tenant continuing to honour a lease that is, for accounting purposes, off its balance sheet and, for economic purposes, very much on it.

The systemic question isn’t “will Hyperion default?” It almost certainly won’t. It’s what happens when twenty of these structures exist, written against tenants of varying credit quality, when the AI capex cycle hits its first downturn. Nobody has stress-tested that at a system level. The BIS hints, the FSB alludes; neither has modelled it.

What to do about it

Four things matter. Ask managers to break software exposure into infrastructure (AI-resistant), vertical SaaS with deep workflow embedding (variable), and horizontal application software (highest displacement risk). If they can’t or won’t, that’s the finding. Ask what share of the portfolio has activated PIK toggles in the last two quarters and what share has coverage below 1.5x. For data centre exposures, look through the SPV: the RVG, the hardware obsolescence schedule, and the credit quality of the offtake. And read the disclosure that isn’t there: if a fund doesn’t separately identify AI-displacement risk, PIK activations, or off-balance-sheet exposures, the absence is informative. The data gap is the position.

Both the software loan book and the AI infrastructure buildout are funded from the same capital pool, by the same managers, through the same banks, and are marked by the same methodologies. When stress materialises, it won’t be separated into “software defaults” and “infrastructure write-downs”. It’ll move through the same plumbing simultaneously. That’s the question the FSB raised but didn’t answer. Marks’s “healthy balance between belief and doubt” applies. So does a blunter line from a previous comment thread: the documentation has become toilet paper, and recoveries when the next crisis hits will look very different from the last. Whether that view is too dark is exactly what we’re about to find out.

At a mechanical level, swap lines are straightforward. The Federal Reserve enters into an agreement with a foreign central bank to exchange currencies at a fixed rate, with a commitment to reverse the transaction at a later date. The foreign central bank can then lend those dollars into its domestic financial system. The Fed explicitly states that these arrangements are meant to “enhance the provision of U.S. dollar liquidity” and support global funding markets.

Historically, these facilities have been activated in periods of systemic stress. Their use during the Global Financial Crisis and again during the COVID-19 shock is well documented. Swap lines were expanded significantly in 2008 and 2020 to alleviate global dollar shortages and stabilize financial markets.

The key issue these facilities address is structural: the global financial system is deeply dollarized (I know I’ve been repeating this a lot lately). Research from the Bank for International Settlements (BIS) shows that a large share of cross-border lending and liabilities is denominated in USD, even outside the United States.

Source: Bank for International Settlements

This creates a persistent demand for dollar liquidity that cannot be met domestically in foreign jurisdictions. In normal conditions, this demand is intermediated through private markets such as FX swaps and repo. In stress scenarios, however, these channels become impaired. Disruptions in FX swap markets can lead to sharp increases in the cost of obtaining dollars, reflecting a global dollar shortage. This is basically what happened, most recently, during the pandemic crisis. In more detail, the role of FX swaps is central to understanding why dollar swap lines are not discretionary policy tools. A large share of global dollar demand is not generated by trade flows per se, but by institutional portfolio allocation and hedging practices. FX swaps and forwards account for tens of trillions of dollars in outstanding positions, with the USD on one side of nearly 90% of all contracts. Under normal conditions, this system functions smoothly. Dealer banks intermediate these flows, providing balance sheet capacity to absorb the mismatch between cash and synthetic dollar funding. However, this intermediation is not frictionless.

This is where the FX swap basis becomes critical. The cross-currency basis measures the deviation from covered interest parity (CIP), which in theory should eliminate arbitrage opportunities between currencies. In practice, persistent deviations exist because arbitrage is balance-sheet intensive. When demand for dollar funding increases and dealer capacity is constrained, the basis widens, typically becoming more negative for foreign currencies against the dollar.

During the COVID-19 shock in March 2020, this mechanism was fully exposed. The demand for dollars surged globally as institutions sought liquidity, hedged positions, and reduced risk. At the same time, dealer banks faced binding balance sheet constraints and reduced their intermediation capacity. The result was a sharp widening of the FX swap basis across major currencies, as documented by the BIS: the three-month basis reached approximately –144 basis points for the yen and –85 basis points for the euro.

Source: Bank for International Settlements

When this happens, institutions must liquidate assets to obtain cash. Crucially, those assets are often US Treasuries. This dynamic was visible during the March 2020 market turmoil, when foreign official institutions sold Treasuries to raise dollar liquidity, but also during the oil shock of February-April this year.

Source: Fleming & Ruela (2020)

In response, the Federal Reserve not only expanded swap lines but also introduced the FIMA repo facility, allowing foreign central banks to temporarily exchange Treasuries for dollars without selling them outright. This is the core point, swap lines are not about funding foreign economies. They are about preventing forced asset sales that would destabilize the US Treasury market itself (in times of distress).

This logic becomes even clearer when considering countries with strong external positions, such as the UAE. These economies do not face solvency constraints, but they are still embedded in the global dollar system. In stress scenarios, they may need dollar liquidity to support domestic financial institutions or stabilize exchange rates. Without access to dollar funding, the adjustment mechanism is straightforward: sell liquid assets. Most of the time, countries under currency pressure often draw down reserves by selling liquid dollar assets, including Treasuries, because these are the most readily convertible instruments available, as we discussed in this post written a couple of weeks ago.

Swap lines provide an alternative adjustment mechanism. Instead of selling Treasuries, foreign central banks can obtain dollars directly from the Federal Reserve, using their own currency as collateral. This reduces pressure on asset markets and stabilizes both sides of the system. In this sense, swap lines represent a shift in how global liquidity is managed. Traditionally, countries accumulated large reserves as a form of self-insurance. These reserves were held primarily in US Treasuries, reflecting their liquidity and safety. However, reserve accumulation is (or could be costly) and sometimes inefficient. Swap lines offer a more flexible alternative: access to liquidity when needed, rather than holding large stocks of low-yielding assets.

This changes the logic of the system. The presence of credible swap lines can reduce the need for precautionary reserve accumulation. At the same time, access to swap lines is selective. The Federal Reserve maintains standing arrangements with the European Central Bank, Bank of Japan, Bank of England, Bank of Canada, and Swiss National Bank. This selectivity introduces a geopolitical dimension, but not in the conventional sense of coercion. Instead, swap lines function as a mechanism of institutional integration. They embed participating countries more deeply into the dollar system by providing privileged access to its core liquidity infrastructure.

The role of the US Treasury complements this framework. Through the Exchange Stabilization Fund (ESF), the Treasury can provide financial support in ways that resemble swap arrangements, though these are less formalized. Unlike Federal Reserve facilities, ESF operations carry a stronger political signal, reflecting maybe strategic commitments.

This distinction matters in the current context. Extending swap lines or similar arrangements to countries such as the UAE would not primarily address funding shortages. It would signal a deeper integration into the dollar system and a commitment to mutual financial stability. Making it even more complicated for China, for instance, to advance its petroyuan. We already know how actively China, particularly through its yuan-denominated futures contracts, has sought to advance financial and commodities cooperation with the UAE. And importantly, this is not even a bailout. Swap lines are structured as temporary, collateralized exchanges between central banks. They do not involve fiscal transfers or permanent resource allocation. These arrangements are designed to address liquidity pressures, not solvency issues.

The political optics, however, can be challenging. In periods of domestic economic stress, such measures may be perceived as support for foreign economies at the expense of domestic priorities. Yet this interpretation overlooks the feedback loop between global and domestic financial stability. Disruptions in global dollar funding markets could impact US markets. Preventing these disruptions is therefore not an act of external support, but also of domestic stabilization.

Ultimately, swap lines should be understood as part of the infrastructure that sustains the global dollar system. They are integral components of a broader framework that ensures liquidity and market functioning. Their expansion or strategic deployment reflects a shift from a system based on passive reserve accumulation to one based on active liquidity provision. This shift reinforces the central role of the United States in global finance, not through coercion, but through the provision of essential financial infrastructure. In that sense, swap lines are a mechanism through which the United States preserves the stability of its own financial system while simultaneously anchoring the global role of the dollar. This is much needed in this global context.

Source: DeVere Group

It happens so because oil markets remain deeply embedded in a USD-based financial architecture. Pricing, invoicing, hedging, even collateral practices are overwhelmingly denominated in USD, and that is because of liquidity and institutional depth. As highlighted by the Bank for International Settlements, the dollar’s dominance in trade invoicing reflects network effects and the availability of deep financial markets that no alternative currency currently replicates (hence the wave-like pattern discussed in that research report from BIS). This becomes especially relevant during supply shocks, when the need for hedging and financing rises sharply.

The closure of Hormuz has led to a significant increase in crude oil prices, which directly raises the dollar funding needs of importing economies. Countries in Asia, which rely heavily on imported energy, must now secure larger volumes of dollars to finance higher-priced oil shipments. This dynamic has been widely documented in commodity cycles: higher oil prices do not reduce reliance on the dollar. If anything, they intensify it by increasing the nominal size of transactions and the associated demand for dollar liquidity. That is why commodity price shocks are so tightly intertwined with tighter global dollar funding conditions, particularly across emerging markets. This dynamic helps explain why the DXY index reached its peak in March. Although the index is a weighted geometric mean heavily dominated by the euro (accounting for more than 57%) it nonetheless confirms that in periods of crisis, especially those tied to energy markets, global liquidity tends to consolidate around the dollar. As can be seen in the image below, since the outbreak of the conflict involving Iran, the index has moved higher, reaching on March 30 levels last observed in May–June 2025. Only in recent weeks has this upward pressure begun to ease, amid tentative expectations of a swift resolution to tensions in the Middle East…

The idea that higher oil prices could catalyze de-dollarization rests on a misunderstanding of how currency substitution works in commodity markets. Switching invoicing currency is not a marginal (or magical) decision. At the end of the day, it requires the development of parallel financial infrastructure, including liquid derivatives markets and/or reserve management practices. While China has made efforts to promote yuan-denominated oil contracts, such as those traded on the Shanghai International Energy Exchange (INE), these markets remain relatively small and lack the depth and global participation of dollar-based benchmarks like Brent and WTI. In other words, the persistence of the USD in global trade is closely tied to financial market liquidity and the availability of safe dollar assets.

Moreover, the closure of Hormuz introduces an additional layer of dollar demand through financial channels. Rising oil prices increase margin requirements in futures markets, expand the notional value of hedging positions, and elevate the need for USD-denominated collateral. Commodity traders and energy firms must secure additional dollar liquidity to maintain positions and manage risk. This creates a feedback loop in which higher prices generate higher funding needs, further entrenching the role of the dollar. This is precisely why many currencies, from the Indian rupee (see figure below) to the South Korean won, have reached historic lows in recent months.

This pressure is not theoretical. It is already visible in the behavior of central banks and sovereign institutions. Although this shift has been widely interpreted as further evidence of de-dollarization (for a contrarian position, you can read here what I wrote two weeks ago), I have argued that the decline in foreign official holdings of US Treasuries held in custody at the Federal Reserve Bank of New York is better understood as a liquidity response (not as a structural retreat from the USD). Central banks sell Treasuries (or gold, if we discuss Turkey) to obtain dollars needed for foreign exchange intervention or to finance higher import costs. As noted by the Council on Foreign Relations, countries facing currency pressure often liquidate dollar assets precisely because those assets are the most liquid and readily convertible into cash.

A particularly telling development in this context is the request by the United Arab Emirates for access to dollar liquidity through swap arrangements. While details of such requests are often not fully disclosed, the pattern is consistent with past episodes of dollar scarcity, namely central banks seek access to USD funding lines to stabilize domestic financial conditions and ensure the smooth functioning of trade and financial markets. The Federal Reserve’s swap line network has historically served as a critical backstop during periods of global stress, including the 2008 crisis and the 2020 pandemic shock. The fact that even major energy exporters such as the UAE seek dollar liquidity confirms a key point, that participation in the global oil market often increases the USD demand.

This is also why, despite ongoing tensions in the United States and the conflict involving Iran, there is little evidence of a so-called “Sell America” dynamic. For instance, the number of days in which US equities, the USD, and bonds have all declined simultaneously has fallen to just nine in 2026, the lowest annual reading in eleven years.

Source: The Kobeissi Letter

The role of the Strait of Hormuz itself must also be properly understood. It is a critical physical chokepoint for oil flows, but it does not determine the currency in which those flows are priced or settled. Even if transit becomes more costly or restricted, these costs are incorporated into USD-denominated prices rather than triggering a shift in the currency regime. The financial infrastructure surrounding oil markets remains unchanged, and market participants continue to rely on USD-based systems for pricing, hedging, or settlement.

In fact, the current environment confirms once again the asymmetry between oil exporters and importers. While exporters benefit from higher revenues, these revenues are still largely denominated in USD and often recycled into dollar assets (that is why we cannot have a “Sell America” in the context of the oil crisis). Importers, on the other hand, face increased financing needs and must secure additional dollar liquidity. This asymmetry ensures that the marginal impact of a price shock is an increase in global dollar demand rather than a diversification away from it.

The behavior of global markets reinforces this interpretation. Periods of geopolitical stress and commodity price volatility are typically associated with a strengthening of the USD (see the DXY image above again). Even when central banks sell Treasuries to raise liquidity, the demand for USD-denominated assets remains robust. This reflects the central role of the dollar in global financial markets, where US Treasuries serve as the primary form of high-quality collateral.

It is therefore misleading to interpret current developments as evidence of declining trust in the dollar. The observed adjustments in reserve holdings and funding strategies are responses to liquidity needs within a USD-centric system. They do not signal a structural shift in the global monetary order. On the contrary, they demonstrate the continued reliance on the dollar as the ultimate source of liquidity and stability in times of stress.

In conclusion, the closure of the Strait of Hormuz does not undermine the petrodollar system (not even close). By increasing oil prices and tightening global liquidity conditions, the disruption amplifies the demand for dollars across the global economy. The narrative of de-dollarization overlooks the structural features of financial markets that sustain dollar dominance. Rather than accelerating a transition to alternative currencies, the current shock highlights the resilience and indispensability of the dollar in global trade and finance.

It is true that Crude Oil WTI (NYM $/bbl) Front Month (CL1) reacted immediately, with the price reaching a bit over $93 yesterday. Of course, it didn’t hold - eventually pushing to $95, as speculation grew that Israel had already violated the agreement through strikes in Lebanon.

Source: MarketWatch

BUT (and this is an important but), oil markets are forward-looking and deeply financialized, meaning that prices incorporate not only current supply conditions (which, on paper, eased after the ceasefire announcement) but also expectations about future constraints and, of course, risk premia. I say that supply conditions eased on paper is because a temporary easing in geopolitical tensions does not immediately unwind existing supply uncertainties, nor does it reverse precautionary behavior by producers or refiners who adjust contracts and hedging strategies in response to volatility. As a result, even if diplomatic progress reduces the probability of extreme tail events, the residual pricing of risk tends to remain embedded in futures curves and physical delivery premia. For example, Fatih Birol himself, the ​head of the International Energy Agency (IEA), said that the current oil and gas crisis is worse than 1973, 1979, 2022 together.

And it will likely remain that way for a few more months. Why? Because the conflict has already removed approximately 11 million barrels per day of crude production from the market, while export flows from the Middle East Gulf have collapsed from roughly 15 million barrels per day to about 7 million barrels per day, which is both a persistent and substantial physical shortfall. When combined with refinery run cuts (reducing product supply by an additional ~3 million barrels per day) the global oil market is operating under a cumulative supply deficit on the order of ~6 million barrels per day, a magnitude that is only partially reflected in forward pricing.

Source: Kpler

Crucially, the disruption is a reconfiguration of trade flows. With the Strait of Hormuz effectively closed (until yesterday), traditional export routes have been severely constrained. Limited rerouting capacity, such as Saudi Arabia’s east–west pipeline to Yanbu and UAE exports via Fujairah, has absorbed only a fraction of displaced volumes. Significant quantities of crude, particularly from Iraq, have been stranded due to infrastructure constraints, with Iraqi exports falling from around 4 million barrels per day to under 1 million barrels per day. All of this shows that crude oil transportation will no longer be the same starting today. It may take months for the market to become even somewhat (though not fully) functional again. That said, there are two sides to this, one good, and one less so.

Let’s start with the good one. Even if these disruptions are sustained, the oil market do not operate in a one-directional inflationary manner, as we discussed in the past few days here or here. The standard partial equilibrium intuition that higher energy prices translate directly into higher consumer price indices overlooks the demand-side contraction that accompanies such shocks. Energy functions as a pervasive input into both production and consumption. When its price rises, it effectively reduces real disposable income for households and compresses margins for firms, leading to adjustments in consumption and hiring decisions. These adjustments introduce a deflationary (or at least disinflationary) impulse that can offset or even dominate the initial price increase. And this is where the less encouraging news comes in. Both consumer credit and perceptions of household finances are deteriorating as we speak, with or without an oil price shock.

For instance, the recent data on US consumer credit provides the first “useful” empirical anchor for this mechanism. Total US consumer credit increased by $9.48 billion in February 2026, following a downwardly revised $7.67 billion gain in January, and falling short of expectations for a $10 billion increase, according to the Federal Reserve’s G.19 release.

This corresponds to an annualized growth rate of approximately 2.2 percent, a pace that remains modest relative to prior expansionary periods. Within this aggregate, revolving credit increased by only $0.71 billion (just 0.6%) after a $2.57 billion gain in January, while nonrevolving credit rose by $8.79 billion following a $5.10 billion increase in the previous month.

Source: Federal Reserve’s G.19 release

The distinction between revolving and nonrevolving credit is important because it reflects different drivers of household behavior. Revolving credit, primarily credit cards, is closely tied to short term liquidity needs and, of course, consumer confidence regarding income stability. It tends to expand when labor market conditions are strong and households expect stable or rising incomes. And it contracts or slows when uncertainty increases. Nonrevolving credit, by contrast, is more closely linked to structural borrowing decisions (such as vehicle purchases and so on), and is less sensitive to immediate cyclical fluctuations (even though uncertainty matters a great deal here too). This “increase” in revolving credit suggests that households remain cautious in their use of unsecured borrowing. This caution is consistent with an environment in which labor market uncertainty is perceived to be rising. In such conditions, households tend to reduce their reliance on high-cost credit.

And the labor market is not looking good at all. The perceived mean probability that US unemployment rate will be higher one year from now increased.

Source: Federal Reserve Bank of New York

Moreover, as shown by the US Bureau of Labor Statistics, both job openings and even hires are at low levels. This is prompting households to maintain precautionary savings and reduce their reliance on expensive credit, as discussed. More importantly, this cautious behavior cannot, in itself, generate inflationary pressures, regardless of whether the oil shock persists or a permanent agreement is reached between the United States, Israel, and Iran.

Source: US Bureau of Labor Statistics

Household financial perceptions corroborate this cautious stance. Reports from the same New York Fed survey indicate that a larger share of households perceive their current financial situation as worse compared to a year ago, and expectations for future financial conditions have also deteriorated, with sentiment reaching weaker levels relative to prior months. These sentiment indicators influence consumption decisions, particularly for durable goods and discretionary spending. So, hard to believe that an inflationary spike could come anytime soon.

Source: Federal Reserve Bank of New York

Survey-based measures of inflation expectations/household sentiment further confirm this interpretation. Median one-year-ahead inflation expectations increased to 3.4 percent, while three- and five-year-ahead expectations remained comparatively stable at 3.1 percent and 3.0 percent respectively, based on the New York Fed’s Survey of Consumer Expectations. At the same time, median expectations for commodity price changes showed a notable increase, with gas price expectations rising quite a bit. And exactly as we discussed in this post a few days ago, households are distinguishing between short-term price pressures (particularly in energy) and longer-term inflation expectations, which remain anchored. In the image below you can see that although the median one-year-ahead expectation has edged up slightly, it is still at the level seen at the end of 2025, when oil prices were just above $60 per barrel, while the three- and five-year expectations have remained basically flat.

Source: Federal Reserve Bank of New York

This distinction is critical. If inflation expectations at longer horizons remain stable, then wage-setting behavior and investment planning are less likely to embed persistent inflationary dynamics. Short-term spikes in energy prices may influence headline inflation, but without a corresponding shift in expectations, the second-round effects that typically sustain inflationary regimes are limited.

In this context, rather than fueling a generalized increase in demand, energy price increases are likely to act as a drag on real income, thereby reducing consumption growth. This mechanism is consistent with historical episodes in which oil shocks have coincided with slower growth rather than sustained inflationary regimes, particularly when monetary policy does not accommodate the initial price increase.

It is also important to consider the role of credit conditions in amplifying or dampening these effects. In an environment where revolving credit usage is lower and households are not aggressively expanding balance sheets, the multiplier effect of income shocks is reduced. Without leverage-driven consumption expansion, higher energy costs translate more directly into reduced discretionary spending.

At the same time, firms facing higher input costs may respond by adjusting production plans rather than passing costs through fully to consumers, particularly if demand conditions are weakening. Margin compression can lead to reduced investment and hiring slowdowns (and this could be seen in the BLS report), all of which contribute to a deceleration in economic activity. These responses further reinforce the demand-side channel through which oil shocks can exert a disinflationary influence over time.

In such an environment, the primary risk associated with higher oil prices is not a runaway inflation spiral, but rather a combination of slower growth and weaker consumption (among others). Central banks are likely to interpret these developments through the lens of growth risks rather than purely inflationary pressures, particularly if longer-term inflation expectations remain anchored. This is why I view the oil price shock and the broader tensions in the Middle East not as a factor that would take us back to 1970s-style inflation, but rather as something that could further weaken economic growth, which is already visibly slowing due to issues in the private credit market and the labor market.

Source: Financial Times

This interpretation misreads both the reserve management and the structural role of Treasuries in the global financial system. What appears as “selling” is (or could) be better understood as liquidity mobilization under stress. And the key to understanding these flows lies in the interaction between global dollar funding needs and the pricing of energy. Oil is overwhelmingly invoiced in US dollars, meaning that a sharp increase in oil prices mechanically raises the demand for dollar liquidity across importing economies. The recent geopolitical shock involving Iran and the disruption risks around the Strait of Hormuz have pushed energy prices higher, forcing oil-importing countries to secure additional dollar funding precisely when external balances are deteriorating. That is why countries such as Turkey, India, and Thailand have likely sold reserve assets (including Treasuries) to finance higher import bills and stabilize their currencies, with Turkey alone selling approximately $22 billion in foreign securities over a short period. For example, as early as a week ago it was already known that Turkey had sold both part of its gold reserves and $16 billion worth of US Treasuries, but with the stated objective of supporting the lira, not due to any loss of confidence in US debt (or the USD).

In such an environment, foreign central banks are not “losing trust” in Treasuries (whatever that means). USTs remain the most liquid and readily deployable reserve asset (close to gold). When central banks need dollars, either to finance higher import bills or to stabilize their domestic currencies, Treasuries are the first asset to be mobilized. This is a just demonstration of its core function.

This process operates through two closely linked channels. First, higher oil prices worsen trade balances for importing economies, increasing the need for dollar payments. Second, currency depreciation amplifies the domestic cost of those imports, creating incentives for central banks to intervene in FX markets. Intervention, in turn, requires the sale of FX reserves, often composed of Treasuries or dollar deposits, to support the local currency. The result is a visible decline in official holdings, but the driver is a surge in dollar demand, not a shift away from dollar assets.

And this can be demonstrated by looking at the US Dollar Index (DXY). Although this index is a weighted geometric mean heavily dominated by the euro, accounting for over 57%, it still confirms that, in times of crisis, especially energy-related ones, markets tend to gravitate toward the dollar. Since the outbreak of the conflict involving Iran, the index has risen, reaching on March 30 (so just two days ago) a level last seen in May/June 2025. Only in the past two days has it begun to decline, amid hopes for a swift resolution to the Middle East conflict, particularly following statements suggesting that the military campaign in Iran would continue for only another two or three weeks, with or without a deal. A genuine loss of confidence in the dollar would have manifested very differently…

Getting back to USTs, I should make it clear once again that these instruments are uniquely suited to this role because of their liquidity and depth. Unlike other reserve assets, they can be sold quickly, in large quantity, and with minimal market impact under normal conditions. This is precisely why they are held in the first place. In times of stress, they become the primary buffer through which central banks absorb external shocks. That is why their use reinforces their status as the global system’s ultimate liquidity instrument.

At the same time, the behavior of bond markets does not support the narrative of “eroding” trust. If foreign selling reflected a structural shift away from US assets, one would expect a sustained rise in yields driven by a shrinking buyer base and increasing risk premia (across all or nearly all maturities). If you recall, in last week’s post I showed that, while bond yields did rise at the long end of the curve (the 30-year segment), the same pattern was observed across jurisdictions. This pointed primarily to uncertainty regarding the policy reaction function, rather than any outright loss of confidence. Today, however, we are observing the opposite dynamic. The yield on the 10-year US Treasury note has declined toward approximately 4.28% in recent sessions (reflecting strong demand for safe assets).

This is consistent with the standard flight-to-quality (or safety) dynamic, where Treasuries benefit from risk aversion even as some official holders liquidate positions for liquidity purposes. Moreover, this confirms what I argued in that post: there is no genuine inflationary risk. Instead, the Federal Reserve (along with other central banks, including the Bank of Canada in the North American context) views second-order risks as more pressing, particularly those related to weak economic growth and labor market problems (just yesterday, JOLTS confirmed that the US labor market is not looking good at all). These concerns are being prioritized over the so-called inflationary risk, which markets had already dismissed from the outset, as can be seen, once again, in the 5-Year, 5-Year Forward Inflation Expectation Rate chart below.

Source: Federal Reserve Bank of St. Louis

As such, the role of oil prices in this adjustment is critical but most of the time misunderstood. While higher energy prices are typically labeled as inflationary, their macroeconomic impact is more complex. For oil-importing economies, they act as a tax on consumption and production, reducing real incomes and compressing corporate margins. This dynamic tends to weaken aggregate demand, leading to lower investment and slower growth over time. In that sense, the shock is fundamentally disinflationary in the medium term, as demand destruction takes hold.

This helps explain why long-term yields have declined in the past couple of days even as oil prices have risen. Markets are increasingly interpreting the shock not as a persistent inflation driver but as a drag on economic activity. And finally, the same logic applies to monetary policy expectations. While central banks may initially respond to higher headline inflation with a hawkish bias, the deterioration in growth conditions limits their ability to sustain restrictive policies. As a result, the expected path of short term rates shifts downward, pulling long term yields lower as well. There is finally clarity when it comes to policy reaction function…

Against this backdrop, the observed decline in foreign custody holdings should be interpreted with caution. First, not all changes in custody data reflect outright sales. Some positions may be shifted to other custodians, including private institutions or domestic accounts, without any change in exposure. Second, even when sales occur, they represent a reallocation within the dollar system rather than an exit from it. Dollars obtained through reserve sales are typically redeployed within the same ecosystem, whether through trade payments, FX intervention, and/or short-term liquidity management.

Moreover, the global demand for dollar assets remains structurally strong. The dollar continues to dominate international trade invoicing, financial contracts, reserve composition, you name it. This structural demand creates a persistent bid for Treasuries that is not easily displaced by cyclical flows. When external shocks increase the demand for dollars, central banks do not abandon Treasuries (they just convert them into cash). And it is precisely what it allows for rapid adjustment to changing conditions without requiring large shifts in portfolio allocation.

It is also important to distinguish between liquidity-driven flows and portfolio rebalancing driven by expectations of returns or risk. The current episode clearly falls into the former category. There is no evidence of a broad-based reallocation away from dollar assets toward alternative reserve currencies or asset classes. Instead, the observed flows are concentrated among countries facing immediate external pressures, particularly those with high energy import dependence. In fact, one could argue that the episode confirms the enduring centrality of the dollar system. The fact that countries respond to higher oil prices by mobilizing dollar reserves shows the lack of viable alternatives. Even as geopolitical tensions rise, the operational logic of the system remains intact: dollar liabilities must be met with dollar assets, and Treasuries remain the most efficient bridge between the two.

In conclusion, foreign central bank sales of USTs should not be interpreted as a signal of declining confidence in the dollar. They reflect the opposite: the continued reliance on Treasuries as the primary source of global dollar liquidity in times of stress. Rising oil prices increase the demand for dollars, forcing central banks to deploy their most liquid assets. At the same time, bond markets are signaling concerns about growth rather than inflation, with yields declining across major economies. The apparent contradiction between official selling and falling yields is therefore just a reflection of different layers of the same system: one driven by liquidity needs, the other by macroeconomic expectations.

Source: Robin J. Brooks

The reality is more technical. Long-term yields are rising primarily because of changes in term premia and central bank reaction functions, not because markets are suddenly “punishing” governments for higher debt. Understanding this distinction is essential.

At its core, a long-term yield can be decomposed into two components: the expected path of short-term interest rates and the term premium. The first reflects where markets believe policy rates will be over time (it is easily influenced by the central bank behavior). The second reflects compensation for holding duration risk, including several “uncertainties”, we can think of inflation and policy, but also balance sheet risks. As multiple studies and central bank analyses highlight, recent increases in long-term yields have been driven disproportionately by higher term premia rather than changes in expected policy rates.

This distinction matters because it immediately challenges the “debt-driven yield spike” narrative. If yields were rising purely because of fiscal deterioration, we would expect a clearer, more direct relationship between debt levels and borrowing costs. Instead, what we observe is a broad, synchronized increase across jurisdictions with very different fiscal profiles, from the US to Germany to Japan. That pattern suggests a global repricing of risk, not a country-specific fiscal crisis.

As aforementioned, the real driver is uncertainty, specifically uncertainty about inflation, policy, or even the future structure of financial markets. As the IMF and other institutions have emphasized, increases in long-term real yields are often associated with rising real risk premia, reflecting uncertainty about economic and policy trajectories.

This is precisely what we are observing today. The post-2020 macro environment has fundamentally altered how investors price long-term bonds. Inflation is no longer anchored in the same way it was during the pre-pandemic era. Central banks are no longer operating under a credible “lower for longer” regime. And fiscal policy, while relevant, interacts with these dynamics indirectly, by increasing uncertainty about future inflation and policy responses (rather than mechanically pushing yields higher).

In fact, even when central banks are cutting rates or signaling easing, long-term yields can rise. This apparent paradox has been visible over the past year. Short-term rates are influenced directly by central banks, but long-term yields respond to broader expectations about growth or inflation. As a result, it is entirely possible, and increasingly common, for yield curves to steepen even as policy rates decline.

This is why framing the current move as a “debt problem” misses the financial mechanism entirely. Debt may contribute to uncertainty, but it does not mechanically determine yields. The key variable is how investors price risk over long horizons. And, more importantly, how effectively a central bank can communicate. Take the latest Federal Open Market Committee session as an example. As we can see in the figure below, based on the economic projections of the Federal Reserve Board, inflation risks appear to rank relatively low on the list of concerns among board members.

And this is not unusual at all. If anything, it simply confirms what markets have been signaling for weeks, that there is no meaningful inflationary risk, as reflected in the 5-Year, 5-Year Forward Inflation Expectation Rate (see figure below). This is precisely where central bank communication becomes critical. Even though the forward inflation expectation rate is trending downward, and even though board members do not perceive an inflationary risk stemming from the Israel–Iran conflict, Jerome Powell still delivers a statement that leans toward headline-making, namely, that inflation is not under control, or at least “not as much as we had hoped”. That kind of messaging clearly introduces a layer of uncertainty, one that can quickly be reflected in long-term yields.

Source: Federal Reserve Bank of St. Louis

Another critical factor is the re-emergence of the term premium after a decade of suppression. During the quantitative easing era, central banks compressed term premia by purchasing long-duration assets and signaling prolonged policy accommodation. That regime has now shifted. As central banks reduce their balance sheets or simply step back from large-scale interventions, private investors must absorb more duration risk. Naturally, they demand higher compensation for doing so.

This dynamic is already visible in market commentary, where rising long-end yields are increasingly attributed to term premium normalization rather than policy rate expectations. In other words, the market is no longer pricing a world of permanently low volatility and central bank backstops.

Importantly, this process is global. The repricing of term premia is occurring simultaneously across developed markets, reflecting shared “beliefs”. And now we have to go back to the inflation/deflation issue. At first glance, rising oil prices and geopolitical risk appear to explain the increase in yields. Indeed, financial media often frames the current bond sell-off as a reaction to war-driven inflation fears. But this interpretation is, again, incomplete.

From a macroeconomic perspective, oil shocks are not unambiguously inflationary. While they increase headline inflation in the short term, they also act as a tax on consumption, reducing disposable income and dampening demand. And this is also discussed in a really interesting CEPR paper, from Lutz Kilian and Xiaoqing Zhou (2021). Using advanced vector autoregressive models, the authors show that a one-off, unexpected spike in gasoline prices leads to a jump in US headline CPI, but this effect fades quickly, disappearing after about two months and becoming statistically negligible. And this is why the forward inflation expectation rate is telling us that we have nothing to worry about inflation-wise. Historically, long-term, sustained increases in oil prices have often been associated with demand destruction, weaker growth, and ultimately disinflationary pressures.

Even recent reporting acknowledges that higher energy prices risk slowing global growth and could trigger recessionary dynamics. This creates a paradox: oil shocks can raise inflation temporarily while simultaneously weakening the economy. And this could have deflationary (or at least disinflationary) implications. In that sense, the current geopolitical environment should not automatically lead to higher long-term yields. If anything, a purely macro-driven interpretation might suggest downward pressure on yields as growth expectations deteriorate.

So why are yields rising instead? The answer lies in central bank behavior and policy expectations. Markets are not reacting to the direct macroeconomic effects of oil shocks, but to how central banks might respond to them. If central banks interpret higher energy prices as persistent inflationary pressure, they may delay rate cuts or even consider tightening policy. This expectation feeds directly into long-term yields. Recent developments support this view. Central banks are increasingly signaling concern about inflation risks linked to energy prices, even as growth outlooks weaken (the case of the ECB).

This is the key mechanism. Yields are rising not because oil is inflationary per se, but because markets expect central banks to treat it as such. In other words, the bond market is pricing a policy reaction function, not a direct macroeconomic outcome. And this distinction is critical. It basically shows that long-term yields are fundamentally financial variables, shaped by expectations and institutional behavior.

Moreover, the interaction between central banks and financial markets introduces an additional layer of complexity. When central banks signal tolerance for higher inflation (or uncertainty about their reaction function) term premia tend to increase. Investors demand compensation for the risk that inflation will remain above target or that policy responses will be delayed. Or even what the response will be. As long as you don’t know how a central bank will interpret a clearly defined event, it becomes extremely difficult to prepare your own response.

This dynamic has been evident in recent months, where long-term yields have risen even as economic data softens. The market is effectively pricing a scenario in which central banks face a trade-off: tightening policy to contain inflation at the risk of exacerbating weaknesses in labor markets and private credit, or maintaining a looser stance and risking inflation persistence. That trade-off is particularly relevant today. The global economy is showing signs of fragility, especially in segments such as private credit and labor markets. Yet central banks remain cautious about declaring victory over inflation. This tension creates uncertainty and uncertainty, in turn, raises term premia.

From a financial perspective, this is the core story behind rising long-term yields. It is not about debt levels. It is not primarily about war. And it is not even directly about inflation in a mechanical sense. At the end of the day, it is about how uncertainty is priced in long-duration assets.

Even arguments that emphasize fiscal deficits ultimately feed into this same mechanism. Deficits matter insofar as they increase uncertainty about future policy, and bond supply, all of which could contribute to higher term premia. But they do not operate through a simple, direct channel. And this becomes even more important in the United States, where we discussed that both the “maturity wall” and the deficits are primarily driven by bills rather than bonds, allowing for easier reinvestment, as these are, first and foremost, financial instruments (the post on the maturity wall can be read here).

In fact, focusing too heavily on debt risks obscuring the more important changes. The end of the low-volatility, central-bank-dominated regime means that investors must once again price risk explicitly. That process is inherently volatile and leads to higher yields. But the drivers of these changes are fundamentally financial, not purely macroeconomic.

In conclusion, debt levels play a secondary role, mediated through their impact on uncertainty rather than through a direct mechanical channel. Geopolitical shocks, such as the Israel–Iran conflict, influence yields primarily through their effect on policy expectations, not through their immediate macroeconomic impact. The result is a global repricing of duration risk in a world where inflation is less predictable and central banks are less dominant (or clear). Understanding this distinction is essential because it shows that the current environment is part of a structural transition in how long-term interest rates are determined.

As you probably know, the current debate around private credit in 2026 is increasingly framed in familiar, almost reflexive terms, such as opacity, weak underwriting and, more importantly, the migration of risk outside the regulated banking sector. And, of course, the parallels with 2007–2008 are tempting and not entirely unfounded. Yet the analytical shortcut, namely equating stress in private credit with the preconditions of the global financial crisis, obscures more than it reveals. The structure and balance sheet interconnections that defined 2008 are materially different from those shaping private credit today. The result is a system that can generate losses, even sharp ones, without necessarily producing systemic collapse. It is true that banks are (still) interconnected with private credit, but their exposure is not even remotely close to what have been in 2008, as we can see in the figure below.

Source: Reuters

To begin with, scale matters. And not just as a trivial metric, but as a constraint on propagation. The pre-crisis mortgage securitization complex was both enormous and deeply embedded across financial institutions. At its peak, the US mortgage-backed securities market was approximately $7.2 trillion, representing a significant share of global securities and sitting directly on bank balance sheets (as an AAA-rated tranches). By contrast, the private credit market today is roughly $2 trillion globally and less than 1% of total securities outstanding. This difference is important. Mortgage securitization in 2008 was a core funding and collateral mechanism for the banking system. We recall how Bear Stearns engaged in repo transactions backed by MBS, among other instruments, and how these assets had become critical to the functioning of funding markets (you can also read more about this here)1. Private credit, by comparison, is largely a parallel lending channel, not a (foundational) layer of financial intermediation.

This distinction becomes clearer when examining where the risks actually sit. In 2008, the critical vulnerability was the embedding of structured credit products, such as Collateralized Debt Obligations (CDOs), Residential MBS (RMBS), and related derivatives, within the balance sheets of systemically important banks (or as part of bank-sponsored SPVs). These instruments were financed through short-term wholesale funding markets, including repo and asset-backed commercial paper. The result was a classic maturity mismatch: long-dated, illiquid assets funded by runnable liabilities. When confidence eroded, funding evaporated, forcing fire sales and generating a feedback loop between asset prices and balance sheet solvency.

Private credit does not replicate this configuration. By design, it is largely funded through locked-up capital structures. Institutional investors, you can think of pension funds, insurers, endowments, commit capital to private credit funds that deploy it into loans, typically held to maturity. This liability structure is fundamentally different from the runnable wholesale funding that characterized pre-crisis shadow banking. Even where liquidity mismatches exist, such as in interval funds or vehicles offering periodic redemptions, they are generally bounded and subject to gating mechanisms, as recent episodes of redemption limits have demonstrated.

Of course, this does not imply that private credit is immune to stress. On the contrary, several indicators suggest rising fragility within the asset class (again, please take a look at the articles that I mentioned at beginning of this post). Default rates in certain segments, particularly software and technology-linked borrowers, are expected to increase meaningfully, potentially reaching levels not seen since the pandemic. The growing use of payment-in-kind (PIK) structures, where borrowers capitalize interest rather than pay it in cash, has also raised concerns about the true health of credit exposures. The share of such structures has more than doubled in recent years, with a notable increase in “distressed PIK” arrangements that signal borrower weakness (you can read about this here).

Source: Bloomberg

However, these developments show a deterioration in credit quality, not to a systemic liquidity crisis. The distinction is crucial. Credit losses can be absorbed over time, particularly in vehicles designed for illiquidity and long-term capital deployment. Liquidity crises, by contrast, are nonlinear events driven by the inability to roll over short-term funding. The architecture of private credit limits the latter, even if it does not prevent the former.

A second structural difference lies in the distribution of risk across the financial system. In 2008, risk was both concentrated and obscured. Banks held large inventories of structured products, often financed through leverage and off-balance-sheet vehicles. The interconnectedness of these exposures, through derivatives, guarantees, and/or funding markets, created a dense network of contingent liabilities. When losses materialized, they propagated rapidly across (trans-Atlantic) institutions.

Today’s private credit exposures are more dispersed and, importantly, less levered within the core banking system. Banks do have exposure, but it is typically indirect, through credit lines to funds, securitized risk transfers, or asset management subsidiaries, rather than through large-scale direct holdings. This matters because it limits the immediate transmission of losses into the banking system’s capital base (see the first figure again).

Moreover, post-crisis regulation has materially altered the capacity of banks to warehouse risk. Capital and liquidity requirements, including leverage ratios and stress testing frameworks, have constrained balance sheet expansion and reduced the incentive to hold illiquid credit exposures. While this has contributed to the growth of private credit (effectively shifting lending activity outside traditional bank), it has also reduced the likelihood that losses in that sector will directly impair systemically important institutions.

The role of leverage further differentiates the current environment from that of 2008. Leverage in the pre-crisis system was both high and procyclical, embedded in structured products and amplified through repo financing. When asset prices declined, margin calls forced deleveraging, triggering fire sales and reinforcing price declines.

Private credit, in contrast, operates with lower embedded leverage at the asset level. Loans are typically senior secured and directly negotiated, with covenants and structures tailored to specific borrowers. While leverage does exist, particularly at the fund level or through subscription credit facilities, it is generally less dynamic and less tied to daily market pricing. This reduces the likelihood of rapid, forced deleveraging cycles.

Even in more adverse scenarios, the macroeconomic impact of private credit stress appears contained. Estimates suggest that losses under severe conditions could amount to around 0.4% of US GDP, material, but far from systemic. This is not a trivial number, but it is orders of magnitude smaller than the losses associated with the global financial crisis, where the impairment of bank balance sheets triggered a collapse in credit creation and economic activity.

Another important distinction lies in valuation dynamics. One of the defining features of the 2008 crisis was the mark-to-market nature of structured credit (because repo markets played a crucial role). As prices fell, institutions were forced to recognize losses immediately, often based on illiquid or distressed market conditions. This accelerated the downward spiral, as accounting losses translated into capital constraints and forced asset sales.

Private credit, by contrast, is largely marked to -model rather than to -market. This has drawn criticism, particularly regarding transparency and the potential for delayed recognition of losses. Yet this same feature can act as a stabilizing mechanism in the short term. The absence of daily market pricing reduces the likelihood of sudden valuation shocks triggering margin calls or forced deleveraging. Losses are still realized, but they emerge more gradually, through restructurings and eventual (selective) defaults.

Again, this is not to suggest that opacity is benign. Limited transparency complicates risk assessment and may obscure the build-up of vulnerabilities. But opacity alone does not generate systemic crises. It must interact with leverage and liquidity mismatches. In the absence of these amplifiers, its effects are more likely to be localized.

The growing involvement of non-bank financial institutions, particularly insurers, adds another layer of complexity. Insurance companies have significantly increased their exposure to private credit, with allocations rising rapidly in recent years (as you can read here). These institutions are structurally suited to hold illiquid assets, given their long-duration liabilities. However, concerns have emerged valuation practices and capital treatment, particularly where ratings may not fully reflect underlying risk. Even here, the systemic implications are nuanced. Insurers do not rely on short-term funding in the same way banks do, and their liabilities are generally less prone to sudden withdrawal. While losses could affect capital and profitability, they are less likely to trigger immediate liquidity crises.

What, then, are the actual transmission channels through which private credit stress could propagate? The most plausible mechanism is a tightening of credit conditions for middle-market firms. Private credit has become a significant source of financing for these borrowers, particularly as banks have retrenched from certain lending segments. A pullback in private credit, whether due to rising defaults, investor caution, or fund-level constraints, could reduce the availability of financing, leading to slower investment and hiring.

But this is a macroeconomic channel, not a financial stability one. It operates through the real economy rather than through the plumbing of the financial system. The distinction is again critical. The global financial crisis was a crisis of the financial system itself, with cascading failures across banks and funding markets. A private credit downturn, by contrast, would more likely resemble a cyclical credit contraction, with localized stress and broader economic spillovers.

This interpretation is consistent with recent market behavior. Despite rising concerns and isolated defaults, broader financial indicators do not point to systemic stress, as there is little evidence of strain in core funding markets (and when we had something like that, back in December, it was not [only] about the private credit, we discussed this topic here). This does not preclude future deterioration, but it suggests that the system’s core remains functional as of now.

None of this should be read as complacency. The private credit market is evolving rapidly, and its scale and interconnectedness are increasing. The rise of retail participation and the growing links to other parts of the financial system all warrant careful monitoring. There is also the unresolved question of how these markets would behave under a severe macroeconomic shock, particularly if accompanied by rising interest rates and declining earnings.

Yet the analytical framework must remain grounded in structure. The conditions that made 2008 a systemic event, such as maturity transformation or concentrated exposures, are not replicated in today’s private credit market. We could arrive at such an outcome, for instance, if the growing importance of CLOs persists, as we discussed in the case of insurers. But we are not there yet, even though these instruments are clearly on an upward trajectory, as you can see in the figure below. Until then, the risks are real but they are configured differently.

Source: AXA Investment Managers

The more appropriate comparison may not be to 2008, but to other episodes of credit cycle adjustment, where losses are absorbed within specific segments without destabilizing the entire system. Private credit may experience rising defaults (I mean, everybody expects that). But absent a mechanism that transmits these shocks into the core of the financial system, the probability of a systemic crisis remains limited.

To evaluate these propositions, it is essential to unpack not only what tiering does mechanically, but also how it interacts with the operational architecture of money markets, and why the US system differs fundamentally from that of the European Central Bank, where tiered remuneration on reserves was an established feature. According to Central Banking, tiered reserve remuneration is used by some central banks to manage excess liquidity and interbank dynamics beyond simple policy rate adjustments.

At its core, tiered interest on reserves means that different tranches of reserve balances earn different interest rates, typically with a lower rate on the marginal tranche or an exempt tier that earns zero or less than the policy rate, while only a limited band earns the full administered rate. The idea is to reduce the effective interest paid to banks on excess reserves (or on portions thereof) while maintaining the central bank’s ability to control the benchmark rate for money markets. Theoretically, tiering can reduce the opportunity cost of central bank liabilities and compress the central bank’s interest expense. But also to tailor incentives for liquidity management. Between us, all of this would matter at most from a political perspective, it would barely change the narratives about the supposedly enormous cost borne by the American taxpayer. I know, it doesn’t make much sense. But that was the reasoning behind those who, last summer, were seriously discussing the need to eliminate interest on reserve balances (IORB) (I wrote about this topic here, or even here, in the case of the UK).

Source: Federal Reserve Notes

However, the analytical environment in which tiering operates matters enormously. In the euro area, tiering was introduced against a backdrop of persistent excess liquidity resulting from asset purchase programmes and a money market structure in which interbank markets behave very differently from the US context. Because yes, the European financial system is far more bank-centered than the US system. The European Central Bank adopted tiered remuneration while operating large-scale liquidity support programmes, aiming to mitigate the impact of negative rates on bank profitability without undermining policy transmission. In more detail, the European Central Bank introduced tiered remuneration on reserves in October 2019 primarily to mitigate the side effects of its negative interest rate policy. When the ECB pushed its deposit facility rate below zero, excess reserves held by banks were automatically subject to a penalty rate. Given the scale of liquidity created through asset purchase programmes, this significantly increased the cost burden on euro-area banks.

As such, the ECB formally adopted a two-tier system under which a multiple of banks’ required reserves would be exempt from the negative deposit facility rate, while balances above that threshold would continue to be remunerated at the policy rate (which was negative at the time). The ECB explicitly stated that the objective was to support the bank-based transmission of monetary policy and preserve the effectiveness of negative rates while limiting their impact on bank profitability. So until 2022 it was just an one-off solution, not something permanent.

In contrast, the Fed already operates in a floor system anchored by interest on reserves, a regime where reserve abundance (or ampleness?) is a feature of modern monetary operations designed to support rate control across a range of conditions. The shift from a scarce‑reserves regime to an ample‑reserves framework, driven by large QE programmes, is well documented in Federal Reserve research. Tiering in that context would alter the equilibrium dynamics of liquidity demand and dealer intermediation, before anything else.

To understand why this matters, it helps to revisit how the current system works. In the post‑crisis era, the Federal Reserve abandoned the scarce reserves operating regime and transitioned to an ample reserves or floor system. Under this framework, the Fed pays IORB to help establish a floor on short‑term interbank rates: banks would not lend reserves on the market at a rate lower than what they can earn risk‑free from the central bank (but sometimes, s*it happens). This mechanism has allowed the Fed to control short‑term rates even when reserves have been very large, as they were following quantitative easing (QE) and remain elevated compared to pre‑crisis levels (even though the balance sheet is on a downward trend at the moment, the so-called quantitative tightening or QT).

Source: Federal Reserve Notes

Under a simple, uniform IORB regime, all reserve balances held by banks at the Fed earn the same administrative rate. This uniformity creates a direct link between the policy rate and banks’ incentive to supply liquidity in money markets. Banks treat reserves as a risk‑free asset that also helps satisfy regulatory liquidity needs, while non‑bank intermediaries, money market funds, for example, interact with bank balance sheets primarily through repo and other secured funding channels.

Tiering, by contrast, would complicate this linkage. If only a portion of reserves earns the full administered rate, and the rest earns a lower rate, banks face a different incentive structure. In theory, this could encourage banks to hold fewer reserves and instead seek other interest‑bearing assets. In practice, the effects would reverberate through funding markets and the very plumbing that makes modern rate control possible.

First, tiering alters the marginal value of reserves and therefore banks’ liquidity preferences. In a uniform IORB regime, reserves provide two broad services: they settle interbank payments and they provide a safe, liquid cushion that helps banks meet regulatory requirements and manage intraday payment flows. These functions do not depend on reserves earning interest. Mainly, they depend on reserves being available at moments of stress. If tiering reduces the remuneration on a significant portion of reserves, banks could view those balances as somewhat less liquid in economic terms, even if they remain legally equivalent. Reduced perceived liquidity can make banks more cautious, prompting them to hoard other high‑quality assets, to limit intraday exposures, or to tighten lending standards, none of which improves market functioning.

Second, tiering has implications for dealer balance sheet elasticity, which is already constrained by post‑crisis regulation. Supplementary leverage ratios, G-SIB surcharges, liquidity coverage requirements, all that means that banks, and especially dealers with large balance sheets, face a trade‑off between holding reserves and other safe, liquid instruments even if both are risk‑free in economic terms. If tiering effectively reduces the return on a portion of reserves, it exacerbates this trade‑off. Dealers can become more reluctant to intermediate Treasury flows or to warehouse inventory, because holding reserves that earn less interest relative to alternatives raises their effective cost of capital. This is grounded in observable money market behavior. Episodes like the September 2019 repo stress show that funding markets can become unstable when dealer balance sheets are already stretched and when intermediation capacity is limited. The Federal Reserve Bank of New York’s analysis of the 2019 repo event pointed to intermediation capacity and regulatory constraints as central stress factors.

This regulatory lens helps explain why movements in short‑term funding rates, including the Secured Overnight Financing Rate (SOFR), cannot be interpreted purely as a function of reserve supply (as I discussed in detail here). SOFR is a secured funding rate that reflects the cost of financing high‑quality collateral (primarily US Treasuries) in the repo market. In normal conditions, abundant reserves and elastic dealer capacity help keep SOFR close to the policy target. in stressed conditions, or when balance sheet constraints bind, SOFR can trade above the policy rate even if reserves are ample.

A tiered regime might also impact the functioning of the overnight reverse repo (ON RRP) facility, which has played a complementary role to IORB in anchoring short‑term rates, particularly when dealer capacity is constrained or when non‑bank liquidity demand is high. The ON RRP allows a broader set of counterparties to place cash at the Fed at a known rate, establishing a floor on market rates. If tiering reduces banks’ incentive to hold reserves because only a portion earns the full rate, more funds may migrate into the ON RRP, reducing intermediation through banks and increasing the relative role of direct access to the central bank’s liabilities. While this may look attractive in some respects, overreliance on ON RRP can have unintended consequences. It can also weaken the traditional bank‑centric intermediation channel and further elevate the sensitivity of short‑term rates to non‑bank flows and technical demand. This is one reason why the Fed has been cautious in managing both the ON RRP and IORB settings, recognizing that changes in one instrument affect the equilibrium across money markets.

By contrast, the European Central Bank’s use of tiering reflected a very different structural context. The ECB introduced tiered remuneration years ago in response to prolonged periods of negative policy rates and large excess liquidity resulting from asset purchase programmes. Banks in the euro area hold excess liquidity not because of reserve requirements but because deposits typically exceed the operational needs of the money market. Tiering was designed to mitigate the cost of negative rates on banks’ balance sheets without undermining the transmission of policy to broader financial conditions. The pre‑existing dynamics of the euro money market, where unsecured interbank activity is less central, and where the deposit base functions differently, means that the distortionary effects of tiering are not identical to those that would arise in the US system.

Ultimately, the debate over tiered reserve remuneration should not be framed as a narrow question about reducing the Federal Reserve’s balance sheet or limiting interest expense. It is a question about institutional design. The Fed operates in a market-based financial system where dealer balance sheets, repo intermediation and/or regulatory capital constraints are deeply intertwined with the level and remuneration of reserves. Altering the marginal return on reserves reshapes (or could reshape) incentives across the liquidity ecosystem. In a bank-centered system operating under negative rates, as in the euro area, tiering was a defensive measure to protect transmission. In the United States, where reserves anchor a floor system and market-based finance plays a dominant role, tiering would instead redistribute pressures across funding markets and balance sheets in ways that may reduce elasticity precisely where it is most needed. Before embracing tiering as a cost-saving reform, policymakers would need to demonstrate not only that it lowers interest expense, but that it preserves (or improves) the resilience of the monetary transmission mechanism itself. Without that assurance, tiering risks becoming a solution in search of a problem.

There is no debate, Kevin Warsh is way more experienced than I am, but I feel compelled to say that t(his) prescription misunderstands (to some extent) both the function of central bank liabilities in a modern (mostly non-banking) monetary system and the mechanics of money markets that have evolved over the past decade. Shrinking the balance sheet aggressively, without addressing the operational and structural context, risks destabilizing the very systems that underpin efficient monetary policy implementation.

To evaluate the balance sheet question, we must first go beyond simple narratives that equate balance sheet size with inflation risk or fiscal (or maybe monetary?) dominance. These narratives implicitly presume a household-budget analogy, that a bigger balance sheet is necessarily bad, which fails to capture how contemporary monetary operations actually work.

In the post-crisis world, the Fed shifted from the old scarce-reserves operating regime to a floor system anchored by interest on reserve balances (IORB). Reserves are no longer scarce. They are created to be abundant by design, and paying interest on reserves allows the Fed to control short-term rates even with very large reserves. This framework is fundamentally different from pre-2008 open market operations. For a clear explanation of this floor system and IORB, you can also read this.

Source: Oxford Economics

In the old regime, open market operations added or drained reserves to nudge the federal funds rate. But when the reserves stock grew dramatically after QE, this mechanism lost traction, in the sense that reserves were so abundant that small changes no longer moved rates. Hence, the Fed introduced IORR/IOER, now unified as IORB, to pay banks interest on their reserves and create a policy rate floor. This allows the Fed to set a target rate even when reserves are “ample” and supply vastly exceeds demand.

Source: Liberty Street Economics

The result is that reserves are now part of the plumbing of money markets, namely liquid balances banks use to meet payment obligations and regulatory requirements, not discretionary holdings that automatically drive lending or inflation. Excess reserves do not inherently cause credit growth. They just sit on bank balance sheets because banks choose liquidity and safety, especially when the yield curve and IORB environment make holding reserves or safe assets rational. As research shows, abundant reserves alone do not generate runaway credit or inflation when central banks control the interbank rate via IORB and other tools. Thus, a shift toward fewer reserves, which would necessarily occur if the Fed were to shrink its balance sheet significantly, would not simply “normalize” policy. It would risk returning to a regime where interest rates are harder to control and where scarcity of reserves itself creates (or could create) volatility in money markets.

Of course, I should make it clear that one of the clearest operational signs of stress in modern money markets is not simply the level of reserves, but the interaction between reserve supply and balance sheet constrained intermediation. Aggregate reserves can remain ample while funding markets nevertheless destabilize, because post-crisis regulation has made liquidity increasingly immobile. Supplementary leverage ratios, GSIB surcharges, but also liquidity requirements limit dealers’ ability to expand repo books precisely when marginal intermediation is most needed.

This is why movements in the SOFR–Fed Funds spread cannot be interpreted as a pure “quantity of reserves” story. As I have argued elsewhere, the spread widens when imbalances collide with dealer balance sheet constraints. In more detail, problems start when Treasury issuance increases collateral supply → money market funds shift cash placement → deposits migrate → leveraged strategies expand → but primary dealers are unable (or unwilling?) to intermediate these flows due to regulatory capital costs. When reserves are plentiful and balance sheets are elastic, banks can absorb short-term funding pressures. When reserves fall or balance sheets are constrained, rates spike unpredictably.

The September 2019 repo episode shows this dynamic clearly. Reserve balances had declined under QT, but the decisive factor was that dealer balance sheets could not expand to accommodate a sudden increase in repo demand. Overnight rates jumped far above the policy target, forcing the Fed to intervene. As documented by the Federal Reserve, the stress also reflected a failure of intermediation capacity under regulatory constraints (not merely reserve scarcity). The same thing happened (to some extent) in December 2025, as you can read here.

But this raises another important observation, namely that shrinking the balance sheet isn’t the same as tightening monetary policy. One common misconception is equating QT with monetary tightening. In reality, policy rate decisions and balance sheet size are orthogonal tools. The Fed can raise or lower policy rates irrespective of asset holdings, as balance sheet operations affect liquidity structure, not the stance per se. A tighter balance sheet would not automatically lower long-term yields or cool asset markets. On the contrary, because the private sector becomes the marginal holder of safe assets as the Fed sells down its portfolio, scarcity of collateral can raise term premia and widen spreads. A smaller balance sheet in a world awash with high demand for safe, liquid collateral is thus not necessarily contractionary in the desired way. It can instead render markets less predictable and more volatile.

This dynamic shows why central banks are cautious about QT today. A reduction in reserves rearranges who holds liquidity. This affects repo rates but also the term structure pricing. These channels do not operate in a vacuum, and the feedbacks can be destabilizing.

Additionally, the emergence (and later near collapse) of the overnight reverse repo (ON RRP) facility demonstrates how balance sheet liabilities interact with nonbank liquidity. RRPs became a crucial tool to absorb excess liquidity and help anchor short-term rates. When the reverse repo facility dwindled to negligible levels, further reserve removal would compress liquidity in funding markets and make rates more sensitive to cash flows, not a desirable outcome for a monetary authority seeking predictability.

This shows that ample reserves remain foundational to how the Fed targets policy rates and stabilises money markets. Reducing them without ensuring alternative liquidity anchors will inevitably push rates toward scarcity dynamics, undermining the Fed’s ability to control policy.

Shrinking the balance sheet in a world of structurally higher Treasury issuance and complex funding needs would amplify money market pressure rather than relieve it. Treasury bills, repo markets, (but also SOFR-linked instruments) form a dense web of money-like liabilities that require deep (liquid) clearing mechanisms. Reserves serve as the ultimate safe asset for banks and are critical for payment settlement and regulatory compliance.

As such, the debate over the Federal Reserve’s balance sheet often misses the deeper institutional issue, namely regulation shapes how money markets function, not merely how large central bank assets appear on a ledger. The expanded Supplementary Leverage Ratio (eSLR) reform shows this (you can read more about this reform in this post). Under the current SLR, banks are penalised for holding reserves and central bank securities, the safest assets, because these counts increase the leverage exposure denominator. This creates a paradox. Holding the very liquidity that underpins settlement and market functioning raises regulatory costs, constraining dealers’ balance sheets precisely when elasticity is needed.

As a result, dealer intermediation capacity diminishes. Moving forward, repo market liquidity weakens and rates like SOFR become more volatile, especially around reporting dates. The eSLR reform directly addresses this incentive distortion by excluding reserves and central bank securities from the denominator, not as a reward for size, but to realign banks’ incentives with systemic needs for safe asset holdings and intermediation capacity. In this light, eSLR is the reform that matters far beyond headline balance sheet figures.

Thus, shrinking the balance sheet without a corresponding redesign of money markets would distort short-term rates and, by extension, risk premia across the yield curve. The consequence is not better policy transmission but less stable markets. That’s why I argue that eSLR reform could do better things for money markets than simply reducing the balance sheet (just for the sake of it).

As a last observation, the notion that a larger balance sheet inevitably leads to fiscal dominance is speculative at best. The danger of fiscal dominance arises when monetary policy becomes subordinated to financing government deficits. But this is a function of policy coordination, not balance sheet size per se. As long as the Fed retains operational independence and targets inflation via traditional tools, the size of the balance sheet is secondary to its function in monetary implementation. And no, Trump Administration is not controlling the Federal Reserve as we speak, not even close.

So, for me, the real question is not whether the Fed should have a smaller balance sheet but what level of reserves and liquidity is appropriate given the current market structure and policy regime. Also, what kind of reforms we should promote to diminish the money markets volatility. The ample reserves framework was just a response to regulatory changes and the post-crisis evolution of money markets.

As such, the argument for shrinking the Fed’s balance sheet appeals to an intuitive desire for “normal” monetary conditions. But intuition divorced from operational context misreads how modern money markets function. The size of the balance sheet is not a distortion that needs elimination but (an important) part of the plumbing that allows policy rates to be set and liquidity to be distributed.

Today’s discussion focuses on the alleged return of gold-backed money, particularly in the current environment. If you subscribe to the Austrian School, the idea probably sounds appealing. If you are post-Keynesian, it likely feels unbearable. If you are like me, you try to approach the topic as apolitically as possible.

So why does this subject matter? Because discussion of a return to gold-backed money, whether through tokenized gold, state-level legal tender laws, reactions to US fiscal dynamics, you name it, ultimately conflates asset prices and symbolic politics with how money actually works in a highly interconnected global financial system. Advocates of gold as money often start from fears about US federal debt and, as a consequences, diminishing confidence in fiat currencies. At face value, long-term debt growth and elevated gold prices might seem to lead to (structural) shifts. Yet these observations misunderstand both the role of sovereign debt markets and the operational mechanisms of modern fiat money.

The premise that public debt alone undermines monetary credibility assumes a direct linkage between government liabilities and currency viability. This view treats sovereign finances as if they function like household balance sheets, an accounting analogy that does not translate into monetary reality. The United States issues debt denominated in its own fiat currency. It has never failed to settle obligations in dollars. More importantly, fiscal aggregates like gross debt or deficit figures do not in themselves trigger inflation or loss of confidence. Inflation is not simply caused by debt levels on a balance sheet (there are other already well-known reasons).

In more detail, central to the gold standard narrative is the idea that rising US debt signals degradation of the dollar’s value. But real-world reserve and capital flow data do not support an abrupt de-anchoring of confidence in the dollar (as I’ve argued so many times in the past few weeks, see here or here). According to the latest IMF Currency Composition of Official Foreign Exchange Reserves (COFER) data, the share of the US dollar in allocated global reserves remained high in the third quarter of 2025, at about 56.9 %, with only a modest decline from prior readings, and with total global reserves still expanding. This pattern reflects continuity in dollar demand rather than abandonment of dollar reserves in favor of gold or other currencies.

Source: International Monetary Fund

Treasury International Capital (TIC) data provide further empirical grounding against panicked “de-dollarization” (or gold-ification?) scenarios. In November 2025, global investors recorded a net inflow of $212 billion into US long-term and short-term securities, including strong net purchases of long-term Treasuries by both private and official foreign investors. Far from dumping dollar assets en masse, foreign portfolios showed continued accumulation. The raw level of foreign holdings of US Treasuries likewise reached a record high of approximately $9.355 trillion in November 2025, confirming the sustained global demand for these assets.

And, frankly speaking, even a reduction would not be problematic per se. It would have to be situated within a much broader discussion of the eurodollar system. There are circumstances in which a sovereign (or a private institution) may urgently need USD liquidity, especially when its domestic banking system is highly dollarized and it lacks access to Federal Reserve swap lines. In such cases, Treasuries can be mobilized (and sold) to obtain dollar funding, either for domestic stabilization purposes or, as in China’s case, potentially to provide liquidity onward to other institutions.

This distinction matters. While it is true that certain actors, most notably China, have reduced their Treasury holdings from earlier peaks and diversified into other assets, including gold, this behavior reflects liquidity management and reserve diversification strategies rather than any coordinated attempt to abandon the dollar system. News reports note China’s recent sell-downs and increased gold accumulation as part of strategic reserve management. But even if certain large holders tilt away from longer-dated Treasuries for portfolio reasons, the broader pattern is not one of systemic flight from dollar assets. Over many years, balance adjustments by one holder have been offset by increased holdings from others, including European economies and private institutional investors. More importantly, these shifts do not correlate with an exodus from the dollar system. They instead reflect normal reserve and portfolio diversification within a highly interconnected and liquid market framework. This is also why, over the past two-three days, despite all the noise around China, US Treasury yields have remained within normal ranges, with no abnormal reaction (see the chart below for the 10-year UST). That’s because everyone understands this is not an economically driven decision. If investors genuinely believed the narrative that China is “changing course”, we would have seen far greater volatility - especially given that China is, after all, a large holder.

Modern money does not depend on commodity backing. It is (mostly) defined by institutional function, as a unit of account used in contracts, or a medium of exchange accepted for taxes and obligations, not to mention as a settlement asset integrated into clearing and payment systems. Central banks, commercial banks, markets (as a whole) coordinate through sophisticated infrastructures that enable intraday liquidity, risk transfer, and/or balance-sheet netting. Commodity money, historically tied to gold, cannot replicate these functions. Gold itself does not generate the liquidity necessary for modern financial intermediation, nor does it adjust elastically to credit demand in the way that fiat and credit systems can. The operational reality is that money today is largely created through private bank credit and reconciled through central bank liabilities, not through physical convertibility to a fixed supply commodity.

The recent enthusiasm around blockchain tokenization of gold does not change these fundamentals. Representing gold ownership via digital tokens does not make gold into money, it makes it into a more easily traded claim on an asset that still sits outside the monetary hierarchy. Tokenized gold remains a claim on a store of value, not a settlement asset integrated into payment and clearing systems. Even stablecoins pegged to fiat currencies, themselves far more directly tied to monetary systems, remain anchored to traditional sovereign money and rely on banking rails and/or central bank facilities for settlement. Research on tokenized and stablecoin holdings suggests that, while they influence market pricing and liquidity conditions, they do not replace the clearing role of sovereign money.

State-level legal tender experiments, such as statutes in Utah, Florida, or proposed bills in Georgia, are often cited as evidence of grassroots demand for commodity-based money. In practice, however, these initiatives are largely symbolic and do not displace the dollar as the unit of account. Even where states have enacted legal tender laws for gold and silver, taxes, budgets, and official accounting continue to be denominated in dollars. Vendors and recipients ultimately settle in dollars through existing banking infrastructure. This reflects the constitutional reality that states may designate gold and silver coins as tender, but they cannot create a parallel monetary system that displaces federal currency or central bank operations. The choice of medium at the point of payment does not alter the overarching monetary structure.

Drawing macroeconomic inference from rising gold prices is also analytically weak. Commodities, including gold, respond to a range of factors. Price moves alone do not indicate that markets are preparing for a return to commodity money. They reflect risk premiums and expectations about inflation, monetary policy, but also cyclical conditions. For instance, part of the recent rise in gold prices has been driven by inflows into exchange-traded funds (ETFs). In more detail, in January alone, gold ETFs attracted US$19bn, which is the strongest month on record. January’s net buying, combined with a 14% surge in the gold price, pushed global gold ETF assets under management (AUM) to a new record of US$669bn, a 20% increase on the month. I find it difficult to see a clear connection between ETFs and gold-backed money, unless I’m missing something.

Source: Gold.org

The thesis that gold must return as money because fiat currency is inherently unsound also neglects the fact that fiat money’s role has evolved. Unlike earlier monetary regimes, modern fiat systems support credit creation and liquidity provision. Gold’s fixed supply, which was once lauded as a cure for monetary instability, becomes a liability in a world where liquidity must expand and contract procyclically to support economic activity. Attempts to revive a gold standard would impose rigid constraints on money supply and credit, potentially exacerbating economic downturns rather than stabilizing them.

Critics sometimes argue that central banks are quietly reducing dollar holdings or holding assets outside official statistics, and that gold accumulation by emerging markets supports a latent de-dollarization trend. But the empirical record shows otherwise. In more detail, all these commentaries are overlooking another crucial technical detail, namely valuation effects. At first glance, the raw data suggest that the US dollar’s share of allocated global reserves fell from 57.79 percent at the end of Q1 to 56.32 percent at the end of Q2, a decline of 1.47 percentage points. Taken at face value, this looks like evidence that central banks actively reduced dollar holdings.

Source: International Monetary Fund

But once reserves are adjusted for exchange-rate movements, the picture changes dramatically. Holding exchange rates constant, the dollar’s share would have slipped only marginally, to 57.67 percent. In other words, roughly 92 percent of the reported decline in the dollar’s share over the quarter was driven by currency valuation effects, not by active selling. As non-dollar currencies such as the euro and yen appreciated against the dollar during this period, their reserve values mechanically increased, lowering the dollar’s percentage share even though central banks largely maintained their underlying allocations.

This matters because central banks manage reserves primarily for liquidity and operational needs, not for geopolitical signaling. Their portfolios are typically adjusted gradually and conservatively, and large discretionary reallocations away from the dollar would show up clearly in constant-exchange-rate data.

In sum, the argument that the world is returning to a gold standard because of US debt or rising gold prices conflates price movements and fiscal anxieties with the institutional foundations of money. Money is defined by its function, as a unit of account, medium of exchange, and settlement asset, as I noted before, and those functions today rest on fiat currencies backed by central bank authority and integrated financial infrastructure. Gold may retain value as an asset or hedge, but it cannot re-establish itself as money in the modern sense without displacing the institutions and credit systems that define contemporary financial intermediation.

In more detail, the marketing of this “capital war” framing suggests that nations could actively use capital flows, asset holdings and/or reserve positions to coerce rivals or protect themselves, potentially causing dramatic shifts in global investment patterns. Dalio’s commentary points to historical precedents in which economic and financial coercion preceded traditional military conflict, and to contemporary market responses, including surging gold prices and repositioning by some institutional investors, as early evidence of a shift in sentiment.

But despite the appeal of this narrative, the underlying reasoning is deeply flawed in its analytical foundations. The idea of a capital war abstracts the mechanics of finance into a geopolitical tool without sufficiently confronting the structural realities of the global monetary system. It merely conflates correlation with causality, and it underestimates the robustness of existing financial infrastructure, particularly the centrality and resilience of the US dollar system, and the incentives faced by sovereign and private holders of capital. A careful examination shows that what Dalio calls a capital war is mostly a repackaging of well-known trends, such as diversification of reserve portfolios, responses to sanctions, geopolitical risk premiums in specific commodities and so on, not necessarily evidence of an inevitable system-breaking conflict.

To understand why this matters, one must first recognize that warnings about capital wars typically start from the premise that capital flows and reserve holdings are easily weaponized. In Dalio’s framing, when geopolitical relations deteriorate, the logic goes, investors (including sovereign holders) will reduce their exposure to perceived adversaries’ assets, particularly sovereign debt, thereby inflicting economic pain. It is asserted that even allies would become reluctant to hold each other’s debt in times of conflict, preferring “hard currency” or other assets outside governmental control to traditional sovereign bonds (think about gold and silver again).

But this premise ignores both the incentives and institutional constraints that govern how capital actually moves. First, what is often described as a shift away from US debt (because this is what it is all about…) in response to geopolitical stress tends to be incremental and cyclical, not sudden and structural. Sovereign and private investors adjust portfolios based on risk-return trade-offs, maybe liquidity needs, not solely geopolitics. There is little evidence of coordinated or systemic sell-offs of US Treasury securities by central banks in recent history, even during periods of meaningful geopolitical strain (this is something I also discussed in this post). In fact, when risk aversion rises, demand for US Treasuries typically increases because they remain the deepest, most liquid safe asset globally. This pattern contradicts the idea that geopolitical tension leads directly to asset abandonment. Rather, it reflects flight-to-quality behavior that benefits the US sovereign market in times of stress.

Source: European Parliament

Second, the concept of capital wars assumes a level of fungibility and policy control over capital flows that simply does not exist at scale. Countries and institutions cannot readily “dump” assets like Treasuries without incurring significant market impact and self-inflicted losses (I’ve discussed the reasons here). Financial markets are not static repositories of capital to be deployed at will. If you want, they are dynamic systems in which massive, uncoordinated sales would depress prices and raise yields, hurting the seller’s own balance sheet before inflicting meaningful strategic pain on targeted economies. Even if a government wanted to weaponize $2 trillion in Treasuries, there is no alternative market of equivalent depth and liquidity ready to absorb the flows, a point confirmed by analyses of Treasury holdings and global safe asset markets. The systemic role of Treasuries as collateral and liquidity anchors means that aggressive selling would create self-harm rather than leverage.

Third, warnings about capital wars often fail to recognize the institutional mechanisms that mitigate systemic stress. Central banks and international authorities have a range of tools, from swap lines to coordinated liquidity facilities, that stabilize liquidity and funding markets even under duress. The Federal Reserve’s standing swap arrangements with other major central banks, for example, were activated during global crises precisely to ensure the continued flow of US dollar liquidity, not to shield US authorities from foreign coercion. These mechanisms have dampened the very feedback loops that “capital war” scenarios assume would manifest abruptly and violently.

A second layer of critique comes from a deeper understanding of what money and capital actually are in the modern system, an understanding that is often obscured by geopolitical metaphors like “war”. Ok, I am not suggesting that Ray Dalio is unaware of these dynamics. Rather, the issue may be an overreliance on textbook econometric models that abstract away from how the global financial system actually functions in practice.

But in technical (but also practical) terms, a global monetary economy based on fiat money and extensive banking intermediation is not primarily about static piles of sovereign bonds that can be switched on and off like strategic resources. Rather, it is about the continuous creation and destruction of credit and money, much of which occurs outside formal sovereign balance sheets through bank lending and offshore dollar markets such as the eurodollar system. This system, which allows USD-denominated liabilities to proliferate internationally through banking and shadow banking networks, means that the monetary base is far larger and more elastic than simple reserve holdings would suggest (there are so many articles and books on this system, maybe you can start with this one). The creation of credit through private banking means that markets are governed by endogenous liquidity conditions rather than exogenously fixed stocks of capital.

From this perspective, the real “war on fiat” that some analysts warn about is not a geopolitical confrontation between sovereign reserves, but a contest over the perception and function of money itself. This is the argument advanced by commentators who view the eurodollar system as central to understanding monetary dynamics: modern money is not a passive asset owned by states, but a set of promises and credits continuously negotiated across banks, firms, and households (on different hierarchical layers, thanks Perry Mehrling). Risk in this system is endogenous, driven by shifts in leverage or liquidity conditions, not by geopolitics alone. That analytical lens casts Dalio’s capital war narrative in a different light. For example, what appears to be geopolitical conflict over sovereign assets may actually be structural adjustments in how money is created (and used) in markets.

This is why the US dollar has not experienced significant disruptions. For instance, just last month it was reported that the USD global transactions use jumps to new high. Likewise, the Treasury Department’s November 2025 report showed that foreign holdings of US Treasuries reached an all-time high of $9.355 trillion in November, ‌up from October's $9.243 trillion. Compared with ​a year earlier, Treasuries owned by foreigners were up 7.2% in November. More than that, overall net foreign acquisitions of US long-term and short-term securities and banking flows amounted to a net inflow of $212.0 billion. Of that total, net foreign purchases of long-term US securities were $221.8 billion, and foreign official institutions alone were net buyers of $64.0 billion in long-term US securities. Although the net change in short-term holdings (Treasury bills) was smaller, this data confirms that foreign entities are still investing in US assets rather than abandoning them. More than that, data from the IMF’s COFER database show that the share of the US dollar in allocated global foreign exchange reserves remained near historical norms in 2025, with around 56–57 % of disclosed reserves still held in dollars. A stable dollar share in official reserves, even amid geopolitical stress, argues against wholesale de-dollarization. And even if it may not seem directly related, all of these dynamics ultimately tie back to the way the eurodollar system governs global financial relationships. A country cannot wage a true “capital war” against another that provides the very infrastructure on which global finance relies, no matter how many political narratives are invoked.

It is precisely this eurodollar framework that brings me to the point concerning the historical examples often cited to justify fears of a capital war. Dalio has pointed to episodes such as European pre-World War I financial restrictions and modern sanctions against Russia, Iran, and North Korea as historical precedents where capital was used coercively. While such episodes involved financial pressure (or weaponized interdependence, how it was called in the literature), they occurred in specific contexts that cannot be generalized mechanically to the US Treasury market today. Those cases often involved relatively closed financial systems and targeted sanctions against smaller economies. In contrast, the US Treasury market is the largest, most liquid, and most globally embedded sovereign market in existence, and its role as the cornerstone of international finance remains intact. There is no comparison.

Critically, the very attributes that Dalio identifies as vulnerabilities, the USD dominance, high foreign holdings of Treasuries, the centrality of US financial markets and so on, are also the reasons why systemic rupture is unlikely. Deep markets with broad global participation are inherently harder to weaponize because many actors have conflicting incentives and diversified portfolios. Sovereign holders such as central banks do not rebalance in isolation. Most of the time, they operate within a web of financial and political objectives that favor gradual, risk-managed adjustments rather than abrupt disinvestment. Let’s take the ECB as an example. Does anyone seriously believe that a central bank of this kind, regardless of whether public or political pressure were ever to intensify, could sever all forms of engagement with the US dollar, whether in the form of asset holdings or swap line arrangements? Such an outcome is both a logical and a financial impossibility, particularly given the profound degree of dollarization embedded in the European banking system.

Another (or the last) flaw in the capital war narrative is the implicit assumption that sovereign debt holdings are levers of political influence in a zero-sum game. This assumption overstates the political control sovereigns have over private international capital and underestimates the role of private sector incentives and market liquidity. Foreign buyers of Treasuries, including pension funds, mutual funds, insurance companies, private investors etc., are driven by risk management and portfolio diversification, not by alignment with sovereign geopolitical strategy. Their behavior cannot be reduced to simple political alignment or opposition.

The capital war framing also obscures a more important point. The stability of the global monetary order ultimately depends more on confidence in the underlying institutions and network effects of money. Currencies gain and retain international roles because they are useful in trade, payments, invoicing, and as a store of value, not because one country holds another’s debt. Network effects and liquidity advantages mean that once a currency like the dollar is entrenched, it is extremely hard to dislodge except through prolonged systemic breakdowns, which are themselves rare.

In short, the idea that the world is on the verge of a capital war, where geopolitical tension transmutes into financial conflict of existential proportions, is a dramatic narrative that obscures more than it clarifies. It takes structural phenomena that are already well understood within monetary economics and reframes them in militaristic terms that appeal to intuition but lack analytical precision.

This is not to dismiss legitimate questions about reserve diversification or even the long-term sustainability of fiat systems, but it is to caution against conflating these issues with an imminent capital war that markets and policymakers are powerless to prevent. A more grounded approach recognizes that capital flows are shaped by a complex set of incentives, regulations, and market structures where no single actor can unilaterally weaponize money at scale without significant self-impact.

But let’s set aside the preliminaries and go straight to the substance. The argument of a maturity wall is typically presented as self-evident: a large volume of US government debt is scheduled to mature over this year, refinancing needs are historically high and higher yields imply greater fiscal strain. In this framing, rollover itself becomes a risk factor, implicitly treated as a liquidity (but also solvency) constraint.

But this interpretation is analytically flawed. It rests on a fundamental misunderstanding of the composition of US Treasury debt. Second, it overlooks the role of short‑term instruments in the global dollar system and their function as liquidity and collateral anchors for financial markets worldwide. Finally, and perhaps most importantly, it avoids a detailed discussion of the institutional mechanisms that absorb Treasury refinancing flows and ensure smooth rollover operations. Once these elements are examined in detail, the notion of an impending maturity wall loses much of its force.

According to the US Treasury’s own reporting and the Joint Economic Committee’s Monthly Debt Update, about one-third of publicly held marketable US Treasury debt will come due within the next 12 months. The Treasury Borrowing Advisory Committee (TBAC) and Treasury data show that this pattern is consistent over time given the high share of short-term Treasury bills and notes in the outstanding stock.

In more detail, as of late 2025, Treasury statistics indicate that Treasury bills, instruments with maturities of one year or less, make up roughly 21–22 percent of all marketable US Treasury debt outstanding, with Treasury notes comprising about 50–51 percent and Treasury bonds around 16–17 percent. Because bills naturally mature within a year by definition, and shorter-dated notes also come due, the share of marketable debt scheduled to mature within 12 months is approximately 33 percent. In absolute terms, if total publicly held marketable Treasury debt is roughly $30.5 trillion (this is the debt held by the public, we also have the intragovernmental debt at $7.62 trillion, but this debt doesn’t concern us), then about $10 trillion of that is on a maturity path in the next 12 months. It is precisely this arithmetic reality that is often misinterpreted as a funding cliff. But in reality, it reflects the high turnover inherent in a system that relies heavily on short-term instruments for cash management and liquidity provision.

Source: Joint Economic Committee

And now we can discuss why I don’t think the so-called “maturity wall” represents any problem. Treasury bills are not designed to be “paid down” in the conventional sense. They are designed to roll. Their economic role is closer to that of reserve-like instruments than to that of long-term financing vehicles. Bank for International Settlements (BIS) research has repeatedly emphasized that short-term Treasuries function as core collateral and near-money assets in the global financial system, supporting repo markets, margining and, ultimately, dollar liquidity provision.

This distinction between rollover volume and rollover risk is critical. A $6 trillion bill stock that turns over four times per year will generate $24 trillion in annual maturities without increasing the outstanding debt stock at all. Treating this as a refinancing burden conflates flow with stock. What matters for market stability is whether there is sustained demand to hold that stock at prevailing yields. On this metric, evidence points in the opposite direction of crisis.

Demand for Treasury bills has expanded structurally over the past decade, driven by regulatory and institutional factors. Post-crisis liquidity regulations have increased demand for high-quality liquid assets, while the growth of money market funds has created a large investor base explicitly oriented toward short-duration government securities. As of early 2025, US money market fund assets exceed $6 trillion, with a significant share invested in Treasury bills and overnight repos backed by Treasuries.

Foreign official institutions also play a central role. Reserve managers prefer bills precisely because they offer liquidity without duration risk. BIS data show that a substantial fraction of foreign official Treasury holdings are concentrated at the front end of the curve, reflecting the use of Treasuries as instruments of liquidity management rather than long-term investment. Let’s take the Treasury International Capital (TIC) report from February 2025 (but issued in April 2025, specifically during the “dash for cash” event). That report showed that official foreign institutions sold a net $19.6 billion in US Treasury bonds and notes, securities with maturities of over one year. But the same foreign buyers added $61.6 billion in Treasury bills, which reflects a growing preference for short-term assets that mature in less than a year.

Concerns about a maturity wall also frequently invoke the fiscal implications of refinancing at higher rates. While it is true that interest costs have risen as older low-coupon debt rolls off, the composition of maturities matters here as well. Bills reprice quickly, but they also reflect current policy rates. When the Federal Reserve eventually eases, bill interest costs will adjust downward faster than those on longer-term debt. This asymmetry gives the Treasury flexibility. Locking in long-duration issuance at peak yields would embed higher interest costs for years. Maintaining a higher bill share allows fiscal costs to adjust more dynamically with the rate environment. The trade-off is higher rollover frequency, but in a system with deep structural demand for short-term safe assets, that trade-off is manageable.

Another misconception embedded in maturity wall narratives is the idea that rollover volumes crowd out private borrowers or strain bank balance sheets. This argument assumes that banks are the marginal holders of Treasury bills. In reality, banks’ holdings of bills are modest relative to those of money market funds and foreign institutions. Moreover, banks’ demand for Treasuries is shaped by regulatory liquidity requirements, not by maturity profiles per se. When banks do hold bills, they do so precisely because of their liquidity properties. There is little evidence that bill issuance has displaced private credit or generated balance sheet stress in recent years (I’m sorry if I missed such events).

The repo market provides further evidence against the maturity wall thesis. Treasury bills are among the most widely accepted forms of collateral in repo transactions, often trading “special” during periods of high demand. Far from being a liability, the abundance of bills supports repo market functioning by providing high-quality collateral that can be reused and rehypothecated. A reduction in bill supply, not an increase, would be more likely to generate stress.

Internationally, the maturity wall narrative often intersects with concerns about foreign selling. But this too reflects a misunderstanding of incentives. Foreign holders of bills value their liquidity and dollar-denominated nature. Selling bills en masse would undermine the very liquidity buffers they are designed to provide. Moreover, any attempt to exit short-term Treasuries would require reinvestment into alternative dollar assets, many of which ultimately depend on the same Treasury market for pricing and collateralization. The idea that rollover itself creates a vulnerability to foreign pressure confuses gross exposure with net dependence.

None of this implies that US fiscal dynamics are irrelevant. Persistent primary deficits and political polarization around budgeting are real issues. Over time, they can interact with inflation expectations and monetary credibility (we are not there yet despite what headlines say). But these are medium- to long-term concerns. They do not arise mechanically from a calendar-year concentration of maturities, particularly when those maturities are concentrated in instruments explicitly designed to roll.

The maturity wall narrative ultimately reflects an accounting view of debt that abstracts from institutional context. It treats sovereign finance as if it were household finance, where refinancing risk is binding and liquidity is scarce. The US Treasury market operates under very different conditions. Its liabilities serve as monetary instruments and its investor base is structurally geared toward rollover.

Seen through this lens, the absence of a Treasury maturity wall this year is not surprising. High bill maturities are the predictable outcome of a system that uses short-term debt to supply global dollar liquidity. The relevant risks lie elsewhere, maybe in inflation control and the long-run credibility of fiscal institutions. Confusing those risks with short-term rollover mechanics obscures more than it reveals.

Source: European Central Bank

This deep embedding of euro area actors in USD‑denominated markets makes any notion of strategic leverage through Treasury holdings both technically and economically implausible. At the most basic level, US Treasuries serve as the primary safe asset of global finance. They are the largest sovereign debt market in the world, with a size approaching USD 30 trillion, more than double its size since 2018 and far exceeding the scale of European sovereign debt markets, both individually and collectively, which remains at just 40% of the size of the US counterpart.

Source: European Central Bank

The sheer depth and liquidity of this market underpin why it is central to global financing and safe asset demand. This is even accounting for recent market frictions following policy shifts and tariff announcements in the United States.

For euro‑area financial institutions, this matters because Treasuries are essential collateral for funding operations. European banks rely on US dollar funding extensively through wholesale markets such as commercial paper, repo transactions and FX swaps. In the ECB’s own Financial Stability Review, euro‑area banks’ funding demonstrates that nearly 23% of total funding is denominated in foreign currencies, with the US dollar accounting for the largest share (about 17%). These positions are predominately obtained through wholesale instruments, including unsecured funding (e.g., commercial paper) and repos, which by themselves account for roughly three‑quarters of total dollar funding. These instruments are inherently short‑term and highly dependent on market confidence and liquidity conditions.

And, more importantly, this reliance on US dollar funding is not a fringe phenomenon. A dedicated ECB analysis of euro‑area banks’ US dollar activities reveals that dollar assets and liabilities are concentrated in capital markets business, with wholesale activities dominating. Holdings of US dollar debt securities, including Treasuries and agency mortgage‑backed securities, are a major component of high‑quality liquid assets for euro area banks, and these are used to secure short‑term funding and meet regulatory liquidity requirements. Not to mention that Euro area banks’ dollar-denominated lending is estimated to be close to €700 billion at the least (close to 10% of the total loan book).

Source: European Central Bank

This deep integration creates a tension at the heart of the argument about “weaponization”. If the EU attempted to use its Treasury holdings strategically, for example, by threatening coordinated sales to exert pressure on the US government, the most immediate effects would not be felt primarily in US fiscal metrics or targeted political outcomes. Instead, such actions would disrupt global funding markets and collateral chains on which euro area banks and non‑bank institutions rely. Large sales of Treasuries would be likely to push yields higher and depress prices, tightening dollar liquidity conditions globally. A stronger dollar, the likely response to higher US yields, would exacerbate funding pressures for European banks that hold dollar‑denominated liabilities, precisely at the moment they are exposed to rollover risks and liquidity mismatches.

Otherwise put, any yield impact from coordinated foreign sales would be diffuse and global rather than fiscally targeted. Treasury yields are global benchmark rates. If yields rise due to selling pressure, the effect is transmitted immediately across global asset prices, which ultimately leads to higher discount rates and tighter financial conditions. These effects do not map cleanly onto US fiscal metrics such as debt sustainability ratios or deficit financing capacity. Instead, they tighten conditions for leveraged actors worldwide, including European banks and funds that rely on Treasuries as collateral. And more than that, higher Treasury yields do not automatically translate into political pressure on the US government because they often strengthen, rather than weaken, US monetary and financial power. Rising yields tend to attract capital inflows, reinforce dollar demand and tighten dollar liquidity abroad. From a systemic perspective, this can actually increase foreign dependence on US dollar funding channels, including Fed swap lines. As a result, the adjustment burden falls asymmetrically on non-US actors. The US state, which issues debt in its own currency and controls the ultimate liquidity backstop, remains largely shielded from the immediate consequences.

Part of this dynamic is captured in global safe asset substitution patterns documented by the Bank for International Settlements (BIS). BIS research highlights that while US Treasuries and German Bunds are both considered safe assets, they play qualitatively different roles in international portfolios. Investors around the world treat Treasuries as a global safe asset and their substitution patterns have spillover effects that extend far beyond any single bilateral creditor relationship. In essence, US Treasuries are substituted with global bonds and other risk assets during portfolio adjustments, but Bunds remain primarily regionally substitutable. This confirms why Treasuries have global reach and systemic implications.

Moreover, official reserve managers have shown some willingness to diversify away from dollar assets in recent years, but these adjustments have been gradual and limited in scope, reflecting the difficulty of displacing the dollar’s central role in reserve portfolios. Even as some European central banks increase their euro‑denominated holdings at the margin, US Treasuries still comprise a significant share of official reserves globally. One survey of European central bank reserve allocations found that Treasuries still make up the largest share of foreign reserves among respondents, despite a year‑over‑year decline in that share as allocations to euro area sovereign debt increased.

The European banking sector’s dependence on US dollar funding has also drawn attention from European supervisory authorities, especially European Banking Authority (I discussed this in detail here). Reports have highlighted that a meaningful portion of European banks remain exposed to dollar funding risk, and that currency mismatches on banks’ balance sheets are growing, with USD‑denominated funding increasing alongside total exposures. This has been flagged as a potential vulnerability, especially in periods of market stress when dollar funding could become constrained or more expensive.

Even from a policy coordination perspective, the notion of weaponizing Treasuries would encounter fundamental institutional constraints. The EU and the euro area lack a centralized fiscal authority with unified control over all Treasury holdings. Private institutional investors, sovereign wealth funds and a multitude of national central banks hold these securities under varying mandates and legal regimes. There is no clear mechanism by which the EU could credibly direct all of these actors to act in concert for geopolitical purposes without upending market norms and violating legal or procedural constraints. In contrast to centralized reserve managers such as the People’s Bank of China, which can adjust holdings strategically (albeit gradually), the EU’s fragmented reserve ownership complicates any coordinated strategic action.

The asymmetry in financial statecraft is further confirmed by the role of the Federal Reserve and dollar liquidity channels. During times of stress, the Fed’s swap lines with other central banks, including the ECB, serve as backstops for dollar liquidity, helping to stabilize funding pressures and prevent disorderly market conditions. Even European officials have contemplated pooling dollar reserves to reduce reliance on these swap lines, implicitly acknowledging how central these arrangements are to European banking stability.

The broader geopolitical context makes this dynamic even more acute. The dollar’s predominance in foreign exchange markets remains overwhelming relative to the euro. While the euro plays a significant role in global trade invoicing and payments, the dollar continues to command the largest share of transactions and reserve holdings. This entrenched position reinforces the dollar’s attractiveness as a funding currency and an investment anchor, complicating efforts to pivot away from it strategically.

Indeed, the very nature of Treasuries as collateral highlights why they are rather a public good in global finance (not a simple lever of coercion). They underpin the pricing and risk management of derivatives and act as the baseline for monetary policy transmission and interbank credit. Disturbing this infrastructure without a credible alternative would risk damaging the institutional foundations of European capital markets, whose stability is predicated on access to deep and liquid dollar markets, among other things.

In practice, European policymakers seeking greater strategic autonomy must focus not on coercive tactics involving Treasury holdings but on building resilient financial infrastructure and expanding the international role of the euro. This means deepening European capital markets, enhancing the liquidity of euro‑denominated assets but also developing payment and settlement systems that reduce reliance on dollar‑centric channels. It also involves fostering regulatory and institutional reforms that promote financial integration and reduce fragmentation within the euro area itself. Only through such structural efforts can Europe begin to mitigate vulnerabilities associated with dollar dependence.

By way of a brief conclusion, I consider myself a convinced European, born and raised here. But at times it is worth resisting the temptation to promote ideas that may ultimately return against us like a boomerang. The idea that the EU could weaponize its holdings of US Treasuries misunderstands both the nature of these assets and the dynamics of the global financial system. Any attempt to manipulate these holdings for geopolitical ends would risk destabilizing the very institutions and markets that depend on them, tightening dollar funding conditions and ultimately harming European financial stability more than US fiscal policy. The path to strategic autonomy lies not through coercive use of Treasury holdings but through long‑term structural reforms that address the euro area’s dollar dependence and bolster the international role of the euro.