Warsh's double dilemma: when the Middle East rewrites the Fed's playbook

• The energy price shock will delay the Fed’s single rate cut in 2026. Rising energy prices should push US inflation towards +3.6% y/y in April-May, up from +2.8% expected before the Middle East conflict, under the assumption that oil prices remain around 90 USD/bbl on average in Q2. Despite a still weak labor market, the Fed will have to keep rates on hold at least through the summer amid risks of de-anchoring inflation expectations. We continue to expect only one 25bps rate cut this year, but pushed out to September.
• Fed chair nominee Kevin Warsh combines a dovish view on interest rates with a hawkish approach to the Fed’s balance sheet, which could drain liquidity from a highly leveraged financial system. Increasing frictions in the US money market would have a global impact on financial markets mainly though the channels of FX and interest-rate volatility. Warsh opposed additional rounds of quantitative easing (QE) in the 2010s, warning that prolonged asset purchases fueled financial excess and risked inflation. With hindsight, his concerns about loose financial conditions and asset valuations appear partly validated. But reducing the Fed balance sheet drastically or even returning to a scarce-reserves regime would ignore the intermediation mechanisms on which the US funding market has been founded since the GFC. The critical point is not the amount of reserves but the dealers’ balance sheets available to warehouse Treasuries and intermediate repo flows. Dealers face regulatory limits that are capping the size of their balance sheets. They cannot do both: absorb heavy Treasury issuance and provide smooth funding for the private sector. It would drive up prices for hedging instruments (swap spreads and cross-currency basis) but in a worst-case scenario such mismatch could trigger a deleveraging spiral affecting valuations of risky assets and increasing default risk for highly leveraged entities (e.g. hedge funds). The Treasury and the Fed have options to mitigate this, but it would require a massive easing of leverage regulation and/or a substantial decrease in Treasury issuance, the first being more likely than the latter.
• We expect the Fed to move cautiously towards a modestly smaller balance sheet from Q4 2026 at the earliest, keeping Treasury holdings steady while continuing the run-down of the USD2trn MBS holdings. This would limit money-market volatility but still risk higher mortgage rates. Wars typically involve monetizing deficits, i.e. restarting QE and departing from Warsh’s pledges for rapid quantitative tightening (QT). The Fed will likely adopt a hybrid model: a leaner balance sheet combined with standing repo facilities as permanent liquidity backstops, preserving rate control while mitigating systemic funding risks. Regulatory moves on bank liquidity regulations could also support the transition toward lower bank’s reserves. We would expect banks’ reserves balances to drop gently from 9.5% of GDP currently to reach the level of the 2019 money-market meltdown (below 7% GDP) only by Q4 2028, though uncertainties are large and episodes of money market volatility will likely become more frequent. The MBS reduction would however blunt the administration’s attempts to lower mortgage rates. The conflict in the Middle East could also drastically change the course of monetary policy in the US. In a downside scenario, quantitative easing may be needed to help finance the effort of the war. 
  

The Fed is likely to keep interest rates on hold at its next meeting on 18 March amid upward pressure on inflation, which we expect to peak at +3.6% by April/May. The conflict in the Middle East poses a difficult dilemma for the Fed, with inflation set to pick up amid higher energy prices while the labor market still looks fragile. Non-farm payrolls declined by -92K in February, and though part of the weakness was due to strikes in the healthcare sector, the private sector excluding healthcare continued to shed jobs. But with inflation now stuck above the 2% target for five years, the Fed clearly does not have the luxury to focus on labor market risks. Inflation expectations are especially sensitive to energy (and food) prices and could risk de-anchoring if the Fed is perceived as too complacent with inflation risks. We expect the Strait of Hormuz to open again by the end of April. In the meantime, partial re-retouring, additional production (Russia, US) and tapping strategic reserves should cover over 50% of the gap in oil and gas, capping oil prices at 90 USD/bbl on average in Q2 2026. We estimate that US energy CPI will rise from +0.4% year-on year in February to peak at around +15% in April. Headline inflation would rise from +2.4% to a peak of +3.6% in April-May (Figure 1, see pdf).

We continue to expect one 25bps rate cut this year, but pushed out to September. We have long argued that the Fed will be limited in its ability to cut interest rates in 2026, and we were expecting only rate cut in June prior to the conflict in the Middle East – a much more hawkish view than what is currently priced in financial markets. We now expect the Fed to wait until September, when inflation should start to decline though still remain above target. The labor market is likely to remain weak, or even weaker further, through the summer as higher inflation eats into households’ incomes and some corporates’ profit margins. If the Fed stays on hold in the near term, the risks of second-round effects on prices, which would compound the initial energy price shock, may be smaller than in 2022. 

Kevin Warsh, President Trump’s nominee for the next Fed Chair, has assuaged financial markets, thanks to his strong credentials and perceived prudent approach on monetary policy. Warsh, who is expected to take over in May, served on the Fed Board from 2006 to 2011 during the height of the Global Financial Crisis (GFC). At the time, he supported an aggressive monetary response to stabilize markets. However, once the recovery gained traction, he opposed the second and third rounds of quantitative easing (QE) and stepped down in 2011, reportedly amid policy disagreements over continued asset purchases. Importantly, his concerns about QE were centered on the blurring of monetary and fiscal policy. He has warned about financial stability risks, asset bubbles and inflation pressures stemming from an expanded balance sheet. With the benefit of hindsight, many of Warsh’s views proved to be well founded: Financial conditions have been very loose since the aftermath of the GFC, equity markets have been buoyant, financial risks have built up and inflation picked up – though at a much later stage. Other critics of the Fed’s balance sheet policy include Treasury Secretary Bessent, whose line of attack is on the net operating losses that the Fed makes by paying interest on banks’ reserves – although Bessent is not pushing for a sharp shrinkage of the Fed’s balance sheet. More recently, Warsh argued in op-ed in 2025 for a significant reduction in the Fed’s holdings.  

However, Warsh’s hawkish views on the balance sheet sit alongside more dovish views on policy rates. Warsh argues that money is too tight for small corporates and calls for lower policy rates. He believes strong GDP growth is being driven by a positive supply shock – notably AI – which should generate disinflationary forces, which will in turn be reinforced by the Administration’s deregulatory agenda. Besides, in 2025, Warsh wrote that the Fed “should abandon the dogma that inflation is caused when the economy grows too much, and workers get paid too much,” signaling limited concern about demand-driven inflation pressures. He also suggested that balance sheet “largesse can be redeployed in the form of lower interest rates to support households and small and medium-size businesses.” While the operational implications remain unclear, this could imply a preference for shifting from asset purchases toward more direct credit-support mechanisms.

Warsh’s criticism of the Fed’s bloated balance sheet could find support among FOMC members. The Fed has already shrunk its balance sheet significantly from its 2022 high of 35% of GDP to less than 20% through a reduction of asset holdings (Treasuries and MBS). All Treasuries maturities were reduced except long-term ones (Figure 2, left). The pre-GFC average of the Fed’s balance were close to 5% of GDP, when the Fed operated under a ‘’scare’’ reserve system rather than an ‘’excess’’ reserves one: banks’ reserve balances were essentially zero prior the GFC (Figure 1, right). In end-2025, following tensions in the repo market, the Fed decided to launch Reserve Management Purchases (RMPs), i.e. Treasury purchases at short maturities (T-bills) to increase reserve balances again. The T-bills purchases are partly funded by the reinvestments of MBS and longer-dated Treasuries paydowns. However, there was some disagreement among FOMC members around the extent to which the Fed should increase its purchases, with the minutes noting “a couple of participants added that effective standing repo operations may allow for a smaller balance sheet ”. 

The transition towards a reduced balance sheet would entail risks for financial stability as the monetary policy system has shifted from reserve fine-tuning to private-sector intermediation. The current ample reserve balance sheet policy is the result of the structural breaks of the GFC. Before 2007, the Fed ran a scarce‑reserves corridor system in a funding market structured by large scale unsecured lending, small dealer balance sheets, loose leverage constraints (pre-Basel III), off balance-sheet repo and a limited role of non-banks (money market funds, hedge funds etc.) in cash supply. After the liquidity collapse during the GFC, the Fed expanded its balance sheet by large‑scale asset purchases – creating such excess reserves that the corridor system could no longer be handled with fine‑tuned open‑market operations – and endorsed the role of market maker of last resort. The US funding market structure shifted from one where banks arbitraged small reserve shortages via interbank lending to one where administered rates and central‑bank facilities anchor pricing and balance‑sheet‑constrained dealers intermediate between non-banks and the Fed. Consequently, the critical point of the system shifted away from the amount of reserves to the amount of dealer balance sheet available to warehouse Treasuries and intermediate repo flows. If balance-sheet capacity is constrained, market liquidity can be scarce despite ample reserves. Going back to a pre‑GFC scarce reserve system would not be “doing what worked for decades” but simply ignoring the intermediation nature of the USD funding market. The current ample‑reserves floor system with Interest on Reserve Balance (IORB) and standing facilities (Overnight Reverse Repo ON RRP, Standing repo facility SRF) has been a crucial pillar of central banks’ volatility compression in the post-GFC era. It has immunized the monetary policy system against shifts in the reserve demand. Changes in reserves today only have a minimal impact on movements in money-market spreads. Under the previous “scarce reserves” regime, a 1% change in reserves could trigger a response of up to 3–4 bps in the money-market spreads (Figure 3). A return to scarce reserves with daily open market operations would create volatility and money-market rate dislocations as rather than simple management of reserves, intermediation capacity is the key point of friction.

A removal of the Interest on Reserve Balance (IORB) would have more negative effects than positive ones – making it unlikely. Alternatively to the Fed balance sheet discussion, there is an ongoing debate on removing the Interest On Reserve Balances. (IORB). The IORB is intended to steer banks’ behavior by guaranteeing an administered return on their reserves. The costs and benefits of operating this system are politically challenged. Critics see the IORB as a subsidy for the US banking sector. The Fed sees the IORB as the price to sustain bank credit flow. Eliminating the IORB “would be extraordinarily disruptive” as it would impair the Fed’s control of short-term rates, reduce the volume of bank-intermediated transactions, and thereby diminish private-sector liquidity. In addition it would generate massive money market volatility and downward pressure on short-term rates as banks would substitute reserves with interest-bearing alternatives such as newly issued Treasuries or the Fed’s reverse repo facility (RRP, which is the lower bound of the policy rate corridor). Replacing reserves with Treasuries would increase banks’ liquidity risk and raise Basel III capital charges, boosting banks’ demand for central bank liquidity and therefore counteracting efforts to shrink the Fed’s balance sheet. Shifting to the RRP would also compress banks’ net interest margins, reduce their lending capacity and restraining credit provision to the economy. Given these adverse consequences, the costs of a IORB removal outweigh fiscal benefits. 

The Fed is facing a funding paradox where an ample balance sheet coincides with scarce intermediation capacity. While the discussion is focused on the scarcity of reserves, we find that the Fed is already operating in a framework of scarcity but from the intermediation side. Intermediation in funding markets is the willingness and ability of balance‑sheet‑constrained dealers and banks (subject to regulatory leverage and liquidity limits) to act as a broker between cash‑rich non-banks (money market funds, pension funds etc.) and collateral‑rich issuers (US Treasury, corporates etc.). During the QE era, balance‑sheet constraints were rarely the binding margin. With QT, excess liquidity declined and dealers’ balance-sheet capacity decreased as did their debt-absorption ability. This situation prevails today. With Treasury (notably T-Bills) issuance taking over an increasing part of the scarce dealers’ balance sheet, funding becomes more expensive for other investors, especially those running leveraged risky strategies (i.e. basis trade, asset swap spread). In Figure 4, we can see the effect on balance sheet scarcity: the more Treasuries issuance increases, the more dealers must increase their holdings on their books. Their cost of carry is reflected in the risk premium, which is the asset swap spread, and in the increased demand for cash reflected by the SOFR-Fed Funds spread (SOFR can be understood as the price for balance-sheet capacity) as well as volatility in short term rates. 

The recent decision of the Fed to resume buying short-term US Treasuries (T-Bills) must be seen as a reaction to this scarcity, acting as cash support to absorb heavy US Treasury issuance. If intermediation scarcity prevails we will see further Fed interventions mostly though the SRF (Temporary Open Market Operations) and further asset purchases. The probability of Fed interventions is linked to the ratio of bank reserves on bank total assets. Above 10% reserves are “abundant”, below 10% reserves are “ample” while 8% marks reserves as “scarce”. As shown in Figure 5, when the market reaches the level of absorption stress the Fed generally intervenes and restores market liquidity. These interventions are often accompanied by announcements  of new facilities (SRF in September 2019) or programs (Bank Term Funding Program following Silicon Valley Bank Crisis, March 2023).

The Fed is moving from being a structural provider of liquidity to a relief valve while the balance sheet is playing a limited role. The problem is that non-banks (esp. leveraged hedge funds) are not “plumbed” to the Fed’s valves. Elevated and volatile repo rates with tighter haircuts can force them to unwind risky leveraged strategies (such as basis trades) and sell Treasuries into less liquid market, amplifying price moves and volatility. Such a  destabilizing deleveraging spiral was seen in March 2020, for instance. The Treasury and the Fed have option to mitigate these risks. A substantial easing of leverage regulation (i.e. SLR, ISLR) could alleviate some balance-sheet constrains and free up cash that could ease the liquidity pressure. A substantial slowdown in Treasury issuance could also free balance-sheet capacity, but remains unlikely. Both would anyway only be temporary. 

US money markets are the primary providers of liquidity and leverage across the global financial system. When US money-market spreads widen and volatility rises due to pressure on SOFR, the FX swap market and interest-rate swaps become the main transmission channels. For FX swaps this occurs because the same intermediaries use their constrained balance sheets to price FX derivatives. Therefore intermediaries can pass on the cost of scarce balance sheets through higher risk premia onto the end users, resulting in higher FX volatility and widening the cross-currency basis, appreciating the USD, which tightens global financial conditions. The transmission channel through interest-rate swaps operates via Treasury issuance dynamics. Swap spreads reflect liquidity at specific maturities as well as the banking system’s absorption capacity. If newly issued debt is not enough subscribed by the private sector, primary dealers (banks) must by law keep it on their books, affecting their funding capacity of the private sector. Negative swap spreads are a sign that issuance absorption is tight. This reduces the demand for duration, increases bid-ask spreads, tightens trading volumes and steepens the curve. This effect is reinforced by huge positions of hedge funds in bond basis trades (long cash bond, short future) funded in US money markets. Official numbers from the CFTC (Commodity Futures Trading Commission) indicate a total volume of USD1,300bn with a USD1,000bn net short position in the US basis trade (Figure 6).

Scarce money market liquidity therefore creates a risk of fixed income market selloff and increases the risk of a massive deleveraging spiral. The US Treasury market is now more exposed to this risk as hedge funds have become the marginal buyer of US Treasuries since the of the US curve (on a FX hedged basis) has become relatively unattractive for foreigners. But hedge funds are not buy-and-hold investors; they use these US Treasuries as collateral to borrow cash in the repo market to fund leveraged positions such as yield curve arbitrage, swap spreads, credit spreads, securitized products on bonds, crypto or stock indices. This demand from leveraged investors compresses implied volatility on many asset classes and is a reason for the unusually low global volatility. But this mechanism only works as long as there is subdued volatility on money-market rates. This is why any change in the monetary policy system by the Fed would be very risky (Figure 7).

The transmission of rising money-market rates volatility runs via higher risk premia that would hurt leverage, affect risky-asset valuations and increase default risk within the banking and non-bank system. It would unfold the following way: First, a margin call due to rates volatility would trigger emergency refunding demand and so a selloff in leveraged assets. Fringe assets would be the first in line to correct (leveraged crypto, highly leveraged ETFs). Second, default risk among leveraged investors (hedge funds) would rise. Since hedge funds usually fund themselves on the sponsored repo market, the sponsoring banks would finally have to step in and bail out the clearing house if hedge funds fail. This procyclicality is the result of the “Active Treasury Issuance” (ATI) strategy. Its objective is to massively issue T-bills to release excess liquidity created by QE and parked into the RRP (reverse repo program) to counter restrictive effects from Quantitative Tightening (QT). But this aggressive supply of T-Bills also increased the potential for leveraged trades for non-banks (hedge funds). ATI therefore contributed to fuel asset prices and compresses asset volatility. So far we do not see stress on sponsored repo, but outstanding volumes indicate a slight deleveraging trend since the beginning of the year. The deleverage risk from the non-bank sector is reinforced by additional layers of leverage: collateral transformation and collateral re-use. Collateral transformation refers to the practice where counterparties swap lower‑quality assets, such as investment‑grade corporate bonds, for high-quality assets like US Treasuries in uncleared bilateral markets. These transactions occur off-balance sheet, their  outstanding amount is hard to quantify. Collateral re‑use (rehypothecation) occurs when dealers re‑use received collateral forming collateral chains in which the same asset runs through multiple transactions. This is a widely used practice as around 85% of collateral received by the biggest primary dealers is said to be re-used. In a massive sell-off scenario, collateral re-use might trigger a chain reaction across markets that the Fed liquidity back stop could only partially contain as it is occurs outside its sphere of influence in the non-bank and/or uncleared segment of USD money markets (Eurodollar system).

Given these market fragilities, we expect the Fed to embark cautiously on a balance-sheet reduction path, from Q4 2026 at the earliest. Treasury holdings could be maintained at flat levels, while the Fed could continue the run-down of the USD2trn MBS holdings. The Fed would then instead be relying more heavily on standing repo facilities as permanent liquidity backstops, preserving rate control while mitigating systemic funding risks while reserves decline. Regulatory moves on bank liquidity regulations, such as a lowering of liquidity requirements for banks, which include banks’ reserves, could support the transition towards a smaller balance sheet and lower reserves. This would maintain financial market stability but risk higher mortgage rates. We would expect it would take a bit of time for Warsh to convince a majority of the FOMC to start reducing the Fed’s balance sheet. Most likely, a compromise could be found on flattening total Treasuries holdings, rather than shrinking them, but to let the MBS run-down continue at the current pace of around -USD17bn per month. In this baseline scenario, we would expect the Fed to start shrinking its balance sheet in Q4 2026. Banks’ reserves would drop from 9.5% GDP currently to less than 7% in Q4-2028 – a level close to the 2019 money-market meltdown. Put otherwise, under prudent balance-sheet reduction, it would take a bit of time before the Fed potentially faces money-market volatility on the back of scarcer reserves. However, forecasting is fraught with large uncertainties. In particular, should nominal GDP grow faster than we expect, we could see a faster reduction of reserves as a share of GDP.