Summary of Howell's Monetary Matters interview, November 2026. Section numbers (§) in the Logic Flow diagram refer to the sections below.
Logic Flow diagram in five steps: new debt demand, liquidity squeeze, quiet policy response, turbulence phase, next upswing, with a dashed secondary channel showing Chinese liquidity driving gold.
1. The big picture: liquidity is past its peak
Howell's Global Liquidity Index measures the growth of money flowing into financial markets, not its level. That growth peaked in late 2025 after a roughly three-year upswing and is now falling, even though the level is still fairly high.
He divides the cycle into four phases: Rebound, Calm, Speculation and Turbulence. Markets have been in Speculation for almost a year. That is a late-cycle phase where money can still be made but volatility is high and returns are poor quality. The signs are visible: the Russell 2000 looks like it is in a bear market, defensives such as healthcare are outperforming, and energy has been strong all year. He thinks markets are now moving into Turbulence, where liquidity tightens further and making money gets harder. The data point to defensive positioning, especially longer-duration government bonds, which runs against the recent headlines.
2. AI absorbs liquidity rather than creating it
The AI data-centre build-out consumes capital. Howell argues that transformative technology booms usually end in recessions or financial crises because every capital-intensive industry is cyclical. Investors treat AI as a long-term trend rather than a cycle. His example is fibre optic cable about 20 years ago, when several leading companies were in Chapter 11 within five years.
Financing has also changed:
12 to 18 months ago, the assumption was that 80% of AI capex would come from cash flow and 20% from markets.
That has flipped to roughly 20% cash flow and 80% external financing.
Hyperscaler debt is approaching $1 trillion a year this year and next, with equity issuance close behind.
3. The debt maturity wall
On top of new borrowing, there is a refinancing wave. During COVID, near-zero rates encouraged borrowers to borrow more and to term out existing debt. That debt now returns to be refinanced from 2026 to 2030, at much higher rates.
Howell's key point is that financial markets are no longer mainly capital-raising systems. They are refinancing systems for rolling over existing debt. The GFC and COVID were heavy rollover years, and both times policymakers had to add liquidity. The 2026 to 2030 wall offsets the negative years of 2021 and 2022 and the low years of 2023 and 2024, and it comes alongside about $2 trillion a year of new federal debt plus the hyperscaler borrowing.
4. The debt-liquidity nexus
This is, in his words, the essence of the problem:
Debt needs liquidity in order to be rolled over.
Liquidity needs debt, because modern lending is collateral-based. The World Bank says 77% of global lending is collateralised, mostly with existing government debt, especially US Treasuries.
The system can spin out of control in two ways:
Collateral side: stress in repo markets, visible in the MOVE index or SOFR spreads.
Refinancing side: credit spreads widening and term premia shifting.
He sees cracks starting to appear.
His long-run chart of debt-to-liquidity for advanced economies shows equilibrium around 2x. Well above that, refinancing crises follow, and he argues that essentially all recent financial crises were refinancing crises. Well below it, surplus liquidity creates asset bubbles. Fitting a sine wave suggests roughly a 16-year cycle, and he thinks we may be swinging from the "everything bubble" towards an "everything bust". The ratio is still on the abundant side, though, and the full bust might not come until the early 2030s, possibly after one more ordinary cycle.
A core line: liquidity is cyclical, debt is exponential, so refinancing tension is always present.
5. The Fed: managing liquidity
Howell thinks the Fed and Treasury do not want to tighten. With this much debt outstanding, aggressive tightening is too dangerous. He cites his former Salomon Brothers boss Henry Kaufman, who once called it a Fed that doesn't want to tighten.
The late-2025 repo scare: In autumn 2025, while the Fed talked of shrinking its balance sheet, bank reserves fell about $400 billion below the roughly $3 trillion target. The SOFR minus IORB spread blew out past the danger zone, trade fails rose, and repo rates spiked. The Fed responded with Reserve Management Purchases (RMPs), which Howell calls QE by another name.
Since then, the Fed has targeted reserves very closely. Over four years the average deviation from the reserves target was about $300 billion. In the last nine months it has fallen below $90 billion. The small recent uptick in global liquidity mostly reflects Fed action last month. RMPs fell from $40 billion a month towards zero, mostly for seasonal reasons, and have picked up a little since.
He sees a shift from an abundant-reserve regime (old-style QE) to an ample-reserve framework: enough liquidity for stability and not much more.
6. The Treasury: buybacks and bill issuance
The Treasury's focus is on bond-market volatility (MOVE index) rather than yield levels, because it must keep refinancing about $40 trillion of debt. It uses two tools:
More Treasury bills, which starves the market of longer coupon issuance.
Buybacks, which Howell thinks aim to reduce volatility, not lower yields. A $6 billion programme cannot move a $40 trillion market. Some off-the-run bonds have been retired at about 60 cents on the dollar, which he calls a win for taxpayers. He sees some correlation between MOVE spikes and larger buyback announcements.
Volatility matters because it affects the collateral multiplier: volatile collateral cannot be leveraged as efficiently, which drains liquidity.
He describes a three-legged policy stool: set policy rates, target money-market liquidity, and suppress bond volatility. He reads Warsh's comments at the meeting as hinting at this, Greenspan-style.
7. "Treasury QE" and the strong real economy
Howell defines liquidity as roughly assets divided by average duration. When the Treasury issues more debt at shorter maturities, private-sector liquidity rises. He calls this Treasury QE.
Bills are about 23% of federal debt, and he expects a push to about 30%. He suggests Bessent's "3-3-3" slogan should be "3-3-3-30".
Each 5-point rise in the bill share is worth roughly $3 trillion of net liquidity stimulus.
Fed QE is becoming minimal while Treasury QE dominates. Bessent criticised Janet Yellen for doing this, but now copies it, and other governments are following.
The catch is that Treasury QE flows mainly into the real economy, not financial markets. Howell treats money as one pool split between the two: when the real economy is strong, it crowds money out of markets. His charts show the global business cycle and financial liquidity moving almost exactly out of step. That fits the late-cycle pattern: a hot economy alongside falling market liquidity.
8. Bond markets are behaving normally
He pushes back on the media view that rising yields reflect fiscal fear:
US domestic liquidity leads the slope of the yield curve by about nine months. Falling liquidity implies flattening.
The consensus at the start of 2026 expected steepening. Curves have instead flattened in the US and Japan.
Term premia are flat to falling in the US, Germany and China, and even in France they are going sideways. That signals growing demand for safe assets.
Yields are being driven by rate expectations tied to strong nominal GDP, which is growing at its fastest pace since the mid-1980s. AI capex accounts for about a third of recent growth.
With his estimate of a roughly -50bp US term premium, fair value for the 10-year is about 5.5% to 6%, close to current levels.
In the long view (UK consol data back to 1700 spliced with US data), the 2021 low near 0.5% was a historical anomaly. Yields are now just above the long-term average, so bonds deserve a portfolio place for the first time in years. He expects the curve to move from the current bear flattening towards bull flattening in Turbulence, with some easing in long yields but no COVID-style collapse.
9. Why growth should slow in 2027
He builds a composite z-score of three headwinds:
A stronger dollar (DXY trending up).
Higher oil, mostly through product prices. Implied crude from diesel is over $160 a barrel.
Higher bond volatility (MOVE), which tends to lead the ISM.
This points to strong growth now and a sharp downturn in 2027, made worse by fading fiscal stimulus and AI capex possibly being squeezed by higher rates. He does not forecast a recession outright, but the probability has risen. A slowdown would let Warsh turn dovish. The Fed may still need to raise rates if momentum continues, but its pace is much gentler than Powell's tightening in 2021 and 2022.
10. Warsh and the balance sheet
Howell sees Warsh as Greenspan-like and expects him to keep rates near current levels rather than cut. The Fed and Treasury are working as a pair to keep the Treasury market stable, which Howell regards as the real top priority alongside the dual mandate. He points to the UK's Liz Truss episode, when the Bank of England switched from QT to QE overnight, and expects the ECB to use its TPI tool if French spreads blow out.
Despite Warsh's past hawkishness on the balance sheet, Howell thinks shrinking it is unrealistic. A growing debt stock needs a growing central bank balance sheet. He expects the Fed's task forces to sanction a larger balance sheet by 2030, effectively QE under a different acronym. This fits his "Capital Wars" framing: in a US-China contest, policymakers cannot afford financial disruption.
11. Asset allocation
Positioning for Turbulence
Long-duration government bonds: the main green light. He is shifting from short to longer duration.
Equities: reduced. P/E multiples are falling as expected. What is unusual is late, strong earnings growth holding indexes up, and that may be cracking.
Credit: a net seller. CCC versus B spreads are widening steadily, and CoreWeave's spreads blew out in October. Credit is where stress shows first.
Commodities: strong in 2026, but now being trimmed. He remains bullish long term on real assets, but not for the next six months. Commodities versus TLT is a useful gauge.
Defensive sectors: healthcare and consumer staples.
Bitcoin and gold
Bitcoin is the most liquidity-sensitive asset. It peaked just after the liquidity cycle did and is now stabilising. He expects a retest of about $70,000. If that holds, it could signal a base and a lift-off ahead. Its elasticity is roughly 45 to 50% per 10% move in liquidity, though that is declining.
Gold is driven mainly by Chinese (PBOC) liquidity, this is an unconventional position, not Western debasement or Russian sanctions. PBOC liquidity has nearly doubled since 2022. Gold's peaks this year coincided with the PBOC balance sheet flatlining from around February or March. Gold is holding near $4,000, consistent with current Chinese liquidity. China must eventually inflate away its debt, using a "dual circulation" system with capital controls to separate internal and external yuan values. That is very bullish for gold over time but neutral for now. Gold's elasticity is about 15% per 10% move in liquidity.
He is more bullish on Bitcoin and gold than on stocks, holding both weighted by volatility, because they lead in the early cycle.
12. Timing and signals
The sine wave points to a liquidity trough in the second half of 2027, possibly earlier. Expect choppy mid-cycle moves first.
He would want about three data points showing a clear upturn, using three- or six-month changes. Liquidity leads most assets by about six months.
Confirming signs: a change in direction of yield curves, Bitcoin holding $70,000, and the two-year Treasury, which leads SOFR and Fed funds.
Equities do best in Calm, the longest phase. Rebound and Speculation each last about 9 to 12 months, and Turbulence is usually shorter.
He agrees that a big policy pivot may need real pain first. Not every downswing brings a crisis, but the refinancing burden means policymakers will eventually have to "print money in large size".
Bottom line
Liquidity is falling as a hot real economy, huge refinancing needs and AI financing crowd out market money. The Fed and Treasury are quietly managing funding stability rather than tightening hard. Howell favours defensives and long bonds now, with Bitcoin and gold as the assets to watch for the next upswing, probably in 2027.
