"I think inflation comes about when the government prints too much - by which I mean the central bank and broadly speaking the government spends too much."

- Kevin Warsh, during his Apr 2026 Senate confirmation hearing

Massive rounds of QE post-GFC did not raise prices. “But wait, Warsh is wrong” quips the sharp-eyed observer. “QE did not stoke inflation after the GFC.” While it is true that y/y US Core PCE averaged a paltry c.1.5% in the 2008 to 2014 period where the Fed was conducting QE, one must be cognisant of the limitations of monetary transmission in that era to understand why. QE created a regime of excess bank reserves, which increased the ability of banks to lend on credit – but not the willingness. The post-GFC era was one of low willingness to both lend and borrow these excess reserves, due to (a) stringent bank capital and leverage ratio requirements restricting the risk-appetite of lenders, while (b) borrowers were unwilling to incur more debt having experienced a near financial meltdown of the global economy and were conversely paying down debt. Therefore, much of these excess reserves remained unborrowed; the increase in base money supply via QE did not translate to broad money supply growth in the real economy.

2026 is not 2016. Ten years later, this situation has completely reversed. Borrowers have come out in force, or rather, in two large forces. The first is the US government. In 2025, the federal government spent c.USD7tn while collecting only c.USD5tn in revenues, leaving a budget deficit of c.USD2tn that will likely persist (or grow) in the years to come. The second is the hyperscalers. The AI Capex boom has seen debt capital raises to the tune of hundreds of billions of dollars with no respite. The similarities between these two borrowers are that they are relatively price insensitive; government budget outlays need to be spent and voter bases appeased with handouts, while stiff competition to achieve AI dominance has fuelled an extensive data centre buildout that seemingly justifies the gargantuan costs. Leaving bank reserve balances at c.USD3tn in an era when borrowers are much more willing to spend means that this supply of money can be quickly borrowed into existence – adding fuel to the inflationary fire. Kevin Warsh was not wrong about QE and inflation; he was simply late in being right.  

Size does matter. The fact that he now inherits a Fed balance sheet of close to USD7tn at the start of his term should then make one wonder if this is too uncomfortably large a monetary base for the new Fed chair to leave around. The duration constituents of the Fed balance sheet should also give one pause, being heavily weighted towards long-term assets. This was the result of “Operation Twist” in 2011 – selling USD400bn of short-term securities (<3Y) to purchase an equivalent amount of long-term securities (6-30Y), effectively flattening the curve to give downward pressure on long-term borrowing costs for businesses and consumers to stimulate the economy. Given his views, this should now go in reverse as the borrowing binge does not need more encouragement. It is no surprise that the Fed has discussed a “maturity-matching” strategy to shorten the duration of the balance sheet, reinvesting maturing treasuries into front-end bills to match the issuance preferences of the US Treasury – a strategy which could literally be called “Operation Untwist”.

Price versus supply of money. At this point, there is no dispute that Kevin Warsh will be tougher on inflation – markets rightly digested the message and priced in good odds for a rate hike for 2026. But let’s not forget, central banks have two policy tools that they can use to combat inflation, including (a) raising the policy rate (price of money) and (b) quantitative tightening (supply of money). Noting Warsh’s long-held views on the matter, markets might be right about the ends but not the means; in that Warsh may prefer adjustments to the supply of money to contain inflation and not the price. One could argue that this may even be more precise a tool, given that interest rates disproportionally affect the lower-income strata of society currently at risk of AI disruption, while money supply addresses directly the appetite of large borrowers like governments and hyperscalers, arguably the bigger risks to the inflation target today. 

Yield curve steepening the likely outcome. Putting Warsh’s narratives together, a steeper curve appears the most likely scenario. Let’s take stock of his beliefs:

1. AI is disinflationary. Warsh believes AI-driven productivity gains will serve as a structural disinflationary force, giving the Fed the leeway to cut interest rates.
2. No forward guidance. Warsh believes forward guidance traps the Fed into predefined paths and distorts market signals. This likely leads to higher term premiums.
3. Smaller balance sheet. Warsh may prefer tapering the balance sheet to combat inflation over raising the policy rate, increasing the net supply of long-term debt in the market.

Bond investors should therefore not be complacent around duration positioning, avoiding ultra-longs, and remain with a blended portfolio duration of 5-7 years. Should the market pricing of the hiking path presently be too hawkish, with the Fed ultimately hiking less than expected, we believe the biggest beneficiary would be short-term (2-3Y) IG credit – as represented by our Liquid+ Strategy – given that the 2Y/10Y spread is the flattest it has been in 15 months. “Playing the man, not the cards” may be the best strategy for fixed income, now that one faces a new Fed chair that is unwilling to show his hand. 

Figure 1: Hyperscalers are borrowing like never before

Source: Bloomberg, DBS


Download the PDF to read the full report.