Who Will Buy US Debt? ​

Despite Jay Powell’s clear message on the FOMC’s reluctance to engage in pre-emptive cuts, the market has revised down its expectations for the Fed Funds’ trajectory, probably reacting to another round of direct comments from D. Trump on the central bank’s stance. Naming the next Fed President well ahead of the end of Powell’s term - materialising the much talked-about “shadow Fed” scenario – could be the harbinger of lasting volatility. Beyond the risk of contradictory signals in the months ahead, bringing Fed Funds into clear accommodative territory without the right dataflow to back it – assuming this would be the choice of D. Trump’s appointee - would likely be opposed by a majority of the FOMC. Central bank leaders can at times be put in a minority – this has happened at the Bank of England. But the Fed is not the BOE. Given its history, this would affect its credibility.

This is still a theoretical discussion, but all this noise is already having a tangible effect on the dollar, which continues to weaken. The expected rate differential with Europe is now narrowing faster, which adds to the generic discomfort with the US currency overseas triggered by the fiscal outlook and threats to the central bank’s independence. As the Senate is about to vote on the “Big Beautiful Budget Bill”, questions on the funding of the US fiscal deficits are getting more urgent. While the market commentariat is focusing on the very latest signs of international reallocation, a striking feature of the last 10 years has been the gradual decline in the share of foreign investors in the US treasury market: domestic players have taken up the slack, with US households, over the last 3 few years, bringing a decisive contribution. In the long run, maintaining this pattern would probably require a rise in the personal savings ratio, which for now has remained elusive. The quest for the next marginal buyer is now shifting to US banks. A proposition by US regulators last week would free up some significant space for US banks to participate more in the treasury market. This could offer some respite in the short run, but even if such changes in regulation can “nudge” towards treasuries, US banks have already been raising their sovereign exposure a lot these last 15 years. Beyond the technical changes, there could be fundamental limits to such approach.

Market getting gung-ho on Fed cuts

Despite Jay Powell’s clarity during the press conference on 18 June that the Fed was not in a hurry to resume cutting, reflected in a more hawkish Fed Funds forecast by the median FOMC member (see Exhibit 1), and the re-affirmation of this stance at his Congressional hearing last week, the market is now expecting more and faster cuts, with forward contracts roughly 50 bps below the FOMC’s median dot for the end of next year (see Exhibit 2). In the absence of any tangible turn in the dataflow these last few days, we suspect this new market positioning is driven by D. Trump’s renewed pressure on the central bank. His attacks on Jay Powell have heated up, and even if the Supreme Court has protected the current Chair from early termination, the market is probably starting to anticipate a “dovish tilt” by the FOMC even before Jay Powell’s formal replacement in May 2026.

This means that the gap between the Fed and the ECB is expected to narrow faster (by c.50 bps by December 2025 and by 100 bps by December 2026, see Exhibit 3). This is reflected in the 2-year spread between US and German government bonds, which has narrowed by more than 30 bps since late May. Expectations of a more dovish Fed add to the softness in the US dollar, but there remains a “risk premium” which cannot be explained by the difference in remuneration across the Atlantic: using the “old correlation” between the 2-year spread and the exchange rate, the euro should be at 1.08USD, still significantly below the actual 1.16 hit on Friday (see Exhibit 4). ​

If indeed it is the expectation of the nomination of a dovish successor to Jay Powell which is behind these market dynamics, then there is good case to consider that the “political risk premium” on the dollar should rise. D. Trump has made public his own “dot” for Fed Funds: he considers the central bank should immediately cut rates to 1%. This obviously goes far beyond what even Governor George Waller has publicly stated. He is ready to contemplate a cut as early as the next meeting but given where even the lowest “dot” is for the end of 2025 – 3.5% - it is unlikely that he would go as far as what the US President wants. But it not necessarily the Fed’s short-term trajectory which should be in focus. If the US administration manages to bring in a structurally more dovish Fed, then the risks that the inflation target would regularly be missed “from above”, combined with more permissive conditions for run-away fiscal policies and larger current account deficits should indeed take the dollar down.

Still, for the remainder of Jay Powell’s term, the market could be preoccupied with the risk that a “shadow Fed” emerges, i.e. that the President announces rapidly a successor, who would become vocal and constantly undermine the actual Chair’s communication and compete with him in steering investors. Such “cacophony” would obviously trigger volatility, and such muddled message, rather than a complete policy U-turn, could continue after the actual arrival of a dove at the chairmanship. Indeed, once formally appointed, the new Chair will need to deal with a stable FOMC (in principle, i.e. unless anyone resigns, there will be only one more Fed Governor leaving the board within Donald Trump’s mandate).

The Chair matters enormously, of course, but he or she probably could not steer the committee in a radical direction if there is no clear macroeconomic case for this. We explored Governor Waller’s dovish proposition last week, and it is well-grounded in a plausible economic scenario, but it would still need to be tested against the dataflow – in particular his view that the absence of labour market tension makes it unlikely the tariff shock proves persistent. By the time someone like Waller makes it to the chairmanship of the FOMC, this test may have already been performed. A question then would be “what next”? If Waller is right, the impact of the tariffs on consumption will be almost immediate, stopping second round effect on prices, but then the slowdown in the real economy should be short-lived as well. It is not entirely obvious if monetary policy will then need to move into accommodative territory: staying neutral (i.e. with rates at around 3%) could be enough. The discussion at the FOMC will likely be lively.

Any kind of dissent is rare at the Fed and the Chair of the FOMC has never been put in minority. But experience from other institutions suggests this is not unthinkable. At the Bank of England, both Mervyn King and Andrew Bailey have found themselves in the situation of being outvoted at the Monetary Policy Committee. But in the UK, the diversity of points of view at the MPC is actively sought, so that there has not been any major reaction to those occasions of dissent. Key to the UK case though is that dissent was not the product of an attempt by the government to influence monetary policy. In the Fed’s case, we suspect the market would be less indulgent since a “structurally dovish” Governor at odds with the majority of the FOMC would be seen as the reflection of the politicisation of the central bank, posing immediate questions on credibility.

In the meantime, there could be a counterproductive loop at play – from D. Trump’s point of view. Indeed, as long as the market reacts to these noises by taking the currency down, the FOMC in its current composition may be increasingly worried about the risks of imported inflation. While the recent spike in oil prices has been corrected – for now – US price dynamics could be increasingly affected by the depreciation in the currency. Claudia Sahm in a recent post has had the great idea of looking at import prices in the US for some key products (we now have the data until May). In “apparel” (basically clothing), it seems that exporters into the US have accepted to take a share of the tariffs in their margins. Indeed, import prices are measured before tariffs are paid. Exhibit 5 suggests that, as of May, Chinese exporters of apparel to the US cut their prices by 6% relative to the beginning of the year. This is significant, but should this move stop there, this would offset only one fifth of the tariff hike (should it stay at 30%). If on top of the tariffs, exporters need to deal with a further depreciation of the dollar, more pressure is likely to show on the US side (in wholesales and retailers’ margins and consumer prices).

True, it may well be that exporters to the US have not yet fully adjusted their margins, but prudence would call for taking the time to monitor the pass-through instead of jumping to pre-emptive cuts. In addition, while Jay Powell was probably right when he said two weeks ago that we are past “peak uncertainty” when it comes to tariffs, the very latest news flow suggests that it will take quite some time to get to the “landing zone” on these matters. Donald Trump’s threat on Canada, as a response to the enforcement of their Digital Service Tax, is a reminder that tariffs are still seen at the White House as a very handy “multi-purpose tool”.

Finding the next “marginal buyer” of Treasuries

In the same statement on his wish to see the Fed cut all the way down to 1%, Donald Trump indicated that the US Treasury should only issue short-term debt as they wait for a new monetary policy course. Although there is no sign as of now that the Treasury has revised its issuance approach (apart from the usual arrangements to delay the moment the “debt ceiling” is hit), we think this reflects some nervousness at the White House on the debt “snowballing” effects which are quite widely discussed within traditional Republican circles as the “Big Beautiful Budget Bill” is still discussed. ​

Given the recent dollar weakness, the market debate is focusing on the possibility that a structural re-allocation of international saving away from US assets – and treasuries in particular – would force a rise in US long-term interest rates. Now, when looking at the Treasury Department’s own estimates of ownership of its securities, the share of non-residents has been steadily falling since a peak in 2014. This is not a statistical artefact due to the growing involvement of the Fed in the US bond market: here we compute the share of residents in “debt held by the public” excluding the Fed’s holdings (see Exhibit 5). This does not mean that the US economy as a whole has become less dependent on flows of foreign savings: the current account deficit is still very much here, but those foreign savings can be invested in other US assets than treasuries.

If non-residents have been reducing their participation to the US bond market, which domestic investors have taken up the slack? The answer is: US banks (more on this in a moment) and US mutual funds…but the bulk came from “others”, with a particularly rapid increase over the last few years. This catch-all group comprises Hedge Funds and Private Equity Funds, which have raised their demand for risk-free assets as a basis for their collateralised operations, but within “others”, most of the progression came from households. As of Q4 2024, as per the Fed’s Flow of Funds data, American families directly own (i.e. excluding their indirect exposure when they invest in Mutual Funds) USD 2.8trn worth of treasuries, from almost zero 10 year ago (see Exhibit 8). The Fed made the point when questions were raised on the “replacement” of the Fed’s purchases of treasuries (see the paper here). Households have increasingly shifted away from time and savings deposits at their banks to Mutual Funds and purchased treasuries directly from the government (which is easily available in the US). Tax changes under “Trump 1.0” have probably pushed them in this direction. Indeed, what US individuals pay towards their municipal and state tax used to be exempt from federal tax – which gave an advantage to residents of tax-heavy states. This was capped in 2017 at 10K (we have recently commented on this issue which is one of the thorny matters being discussed between the House and the Senate). Since treasury securities are exempt from local tax, residents of tax-heavy states found a new interest in shifting their savings towards them. ​

Dynamically though, it is not obvious the US government can continue to count on American families to act as the “marginal buyer” of its debt as it balloons away. Beyond the obvious fact that without the rise in interest rates, households would probably never have increased as much their exposure, we fail to see how the tax treatment of direct holdings of treasuries by households could be made even more favourable. Over time, having US households as the structural marginal buyer would require a rise in their savings rate – which for now has not materialised. Besides, the attractiveness of treasuries over time deposits can become an issue in the long-term if this results in a compression of bank margins, especially for the smaller institutions which rely on traditional activities.

The focus is now squarely on banks to become the next “marginal buyer”, thanks to a regulatory reform. On 25 June, the Treasury Department, together with the Federal Reserve and the FDIC released a proposition – now open to consultation – to amend the “Enhanced Supplementary Leverage Ratio” (e-SLR) applied on the US Global Systemically Important Bank holding companies (GSIBs) and their depository subsidiaries. The e-SLR in its current form penalises banks for holding treasuries, despite the fact that, since they incur a zero capital charge, they normally do not affect their traditional, risk-weighted capital ratios. Indeed, the e-SLR imposes that capital amounts to at least 5% of total assets, including risk-free ones. In the agencies’ analysis (see link here, from page 45 onward), the e-SLR has counter-productive effects – from a financial stability point of view – since a low and high risk asset “consumes” the same quantum of capital, which would incentivise banks to skew their allocation toward riskier products – presumably advantageous from a return point of view. In addition, the e-SLR could exacerbate, rather than mitigate, financial crises when investors rush for cash. The beginning of the COVID market turmoil was a case in point: investors piled up deposits in large banks (seen as more solid) which the GSIB had trouble recycling into government bonds without raising capital at the same time so that they could comply with the SLR – which had to be suspended.

The proposal would change the rate used for the calculation of the e-SLR, lowering the capital requirement, but also requests opinions from market participants on some “reasonable alternatives”, including one under which treasuries would simply be excluded from the calculation of the total assets (without changing the rate itself). In the agencies’ estimates, the baseline proposition or the alternative excluding treasuries from the calculation would free up USD 1.1 to 1.4trn for additional treasuries held for investment purpose at Category 1 GSIBs (the biggest ones), and USD 2.1 to 2.5trn for their trading portfolio. As an order of magnitude, overall, “depository institutions” (mostly banks, in the Fed’s nomenclature) at the end of 2024 held USD 1.7trn in treasuries.

These numbers are impressive, and Scott Bessent stated several months ago that the “carve out” solution could lower treasury yields by 30 to 70 bps. Yet, a usual issue with regulatory reform is that, as much it can “nudge” banks in a certain direction, the decisions will ultimately be made on a sound investment case. One factor to consider is that, since the Great Financial Crisis, American banks have already significantly raised their exposure to treasuries (see Exhibit 9), which now stand for more than 6% of their total assets. As much as reforming the e-SLR makes sense to avoid the replication of liquidity shortfalls on the US bond market, such as the one of 2020, which alone would probably reduce the risk premium on treasuries somewhat, we maintain what is probably a fairly typical macro-economist view of these issues: financial engineering can bring short-term solutions, but ultimately, if overseas investors “go on strike”, the fate of the treasury market will depend on US fiscal policy and on the saving behaviour of the US private agents. Either the deficits must fall, or private savings must rise. This is hard to do without engineering at the same time a transitory slowdown in aggregate demand. ​