Weekly Monologue
This week's bevy of central bank meetings, particularly the Fed, provided a nuanced update on the outlook for global interest rates. In this week's note, we recap the meeting and cover the market reaction, offer some provocations on the future of Fed balance sheet operations, and take a peek at the runners and riders for the next Fed Chair. We had just enough space to add in some thoughts on global trade and US earnings momentum.
First things first. The Fed cut rates by 25 basis points on Wednesday, bringing the federal funds target range to 3.75% to 4.0%. No surprises there, but there was a lot of "news" in this meeting.
For one thing, the Committee was not united on the move; this was a rare meeting with dissents in both directions. Combined with some cautious language on the outlook from Chairman Powell, this has caused markets to rethink their dovish rate outlook.
Also, the FOMC officially called time on quantitative tightening (QT) and will stop shrinking the Fed's balance sheet on December 1.
We'll explore both of these notable developments later on in the note. Here, I want to focus on Powell's comments on AI-related investment, which were buried deep in the press conference and didn't get much media airplay. On a medium-term horizon (1-3 years), I believe this is emerging as the defining debate in monetary policy.
Tip of the cap to Chris Rugaber from the AP, who asked Powell whether his assessment that monetary policy is modestly tight holds any water, given the significant infrastructure investment underway. This is essentially the point we made in last week's Memo, in which we noted that broad financial conditions are easy and that there is no clear indication from the financial sector that more accommodative rates are needed.
While acknowledging the scale of the AI investment surge, Powell essentially said that it is entirely disconnected from interest rates because it's based on a long-term theme with strong cash flow projections. Therefore, the Fed will focus on the margins it can control.
In essence, Powell is setting up a dichotomy between the "AI economy" – driven by its own self-reinforcing dynamics – and the "real economy," which follows the old rules of macroeconomics.
Color me skeptical.
First, just because investment isn't financed using traditional loans and debt issuance doesn't mean it's disconnected from short-term rates. There are at least three mechanisms through which interest rates could influence energy and data center investment:
- Equity financing and collateral: High equity valuations, which are supported by lower interest rates, are used to raise capital or secure funding for projects.
- Discounting: The viability of some projects will still be affected by higher discount rates.
- Tobin's Q: Lower interest rates push up the value of existing capital relative to the costs of creating new capital, incentivizing new investment via balance sheet effects. This effect is less powerful in a situation of scarce replacement capacity, so it may be less relevant at present.
Second, and perhaps more importantly, it's the Fed's job to control, or at least lean against, "irrational exuberance" (to borrow a phrase from a predecessor whose record Powell would do well to study closely). This is where central banking is more art than science. A central bank governor has many tools at his disposal to calm potentially bubbly investment dynamics: communication, forward guidance, moral suasion, background political influence, and, yes, the level of interest rates.
To be generous to Powell, he may not believe he has any influence over the rabid AI frenzy. But even if the question is just "is the AI investment boom completely insulated from interest rate policy?" the answer is not as cut-and-dry as Powell makes out.
For now, there might be a "bit of both" situation at play. Some data centers and energy investments would very possibly survive much higher rates and/or an equity market downturn. But we shouldn't underestimate the influence of interest rates on the valuation effects and sentiment that powers investment. I suspect the true share of capital outlays truly beyond the scope of monetary policy is much lower than Powell has implied.
(Case in point: Meta is financing its latest surge of AI-related investment with a massive $30 billion bond, its biggest ever.)
We should not forget that similar arguments were made about fiber-optic cable in the 1990s, the majority of which is still "dark" in 2025. When Powell argues that the Dotcom bubble burst because the market was full of businesses without a model (as opposed to the "physical" AI buildout we're seeing now), he misunderstands the dynamics of innovation and investment.
All paradigm-shifting new technologies start with a colossal infrastructure build while giving rise to a large crop of bad businesses and a small group of winners (ask any VC principal) based on that infrastructure. It was actually the realization that the physical infrastructure build-out had gone too far that triggered the Dotcom crash. Cisco Systems was the Nvidia of the 1990s.
That's not to say I believe that will necessarily be the dynamic this time (in our recently published Quarterly Scenario Update, we gave this outcome a 35% probability); I just don't think we can or should shrug our shoulders and treat the AI economy as a standalone beast.
Reads of the Week
- How OpenAI Uses Complex and Circular Deals to Fuel Its Multibillion-Dollar Rise (NY Times). A nice visual resource for anyone trying to explain to friends and family how the AI buildout is being financed.
- The Role of Single-Family Rentals in the US Housing Market (St Louis Fed). A good summary of the research on investor-owned single-family housing.
- Maximum Sustainable Debt Across Countries: An Assessment using P-Theory (IMF). One for the nerds. The IMF estimates the maximum debt carrying capacity based on fiscal, macro, and market features. Key takeaway: the US can withstand a net debt of 160% of GDP.
Market Monitor
Two opposing dynamics were at play in markets this week. The first was the Fed meeting, where the hawkish dissent and Powell's cautious comments (more on that in the Macro Monitor) lifted and steepened the yield curve. Forward markets are now pricing ~60% odds that rates come down another 25 basis points in the 12 December meeting (from ~90% before the meeting) and are all over the map for January.
Remember—we'll likely get a large amount of data before that meeting as the government re-opens.
Equity markets looked set for a bad week until a strong series of earnings reports saved the day on Thursday and Friday. The Mag-7 rally hit new highs as Amazon, Apple, Google, Microsoft, and Meta all delivered knock-out earnings reports and doubled down on AI spend (though smaller companies whose AI investments are not yet paying off are getting a shorter leash). Other strong performers included Berkshire Hathaway and Visa (which also reported broad-based consumer spending strength).
Beneath the surface, market breadth deteriorated further. The Nasdaq closed the week up +3.2%, but the Russell 2000 small-cap index was down a tad. And although the S&P 500 rose +1.5%, seven of the eleven sectors fell (Consumer Staples, Real Estate, and Materials were hardest hit). The bull market is increasingly just an AI mega-cap story.
Other notable moves were Gold falling -3.2% on the week, and the DXY US Dollar index rising (+0.9%). That's what you'd expect when interest rate expectations increase, but it's nice to see the old rules being followed for once.
Macro Monitor
The key US and global data and events this past week:
- The Fed (Wednesday) lowered interest rates by 25 basis points to 3.75% - 4.0%
- The Bank of Canada (Wednesday) lowered interest rates by 25 basis points to 2.25%
- The Bank of Japan (Wednesday) kept rates unchanged at 0.5%
- The European Central Bank (Thursday) kept rates unchanged at 2.15%
Rate Path Reset
Let's dig a little deeper into the drivers of this week's US interest rate path reset. Powell was clearly on a mission to remove a little dovishness from markets ("a further reduction in the policy rate at the December meeting is not a foregone conclusion — far from it"). He got a lot of help from Kansas City Fed President Jeff Schmid, who dissented in favor of no cut.
Stephen Miran, President Trump's proxy on the Committee, predictably dissented in favor of a larger cut, setting up the first "symmetric" dissent since September 2019.
Schmid followed up on Friday with an explanation of his dissent. Here is his summary of the economic status quo:
"By my assessment, the labor market is largely in balance, the economy shows continued momentum, and inflation remains too high." — Jeffrey Schmid
He also referenced easy financial conditions. I might quibble with the "continued momentum" point based on what our Real Factor is telling us about September growth, but the rest is spot-on. The note is worth perusing in its entirety.
What remains to be seen is whether Schmid's side of the committee can swing enough members to pause easing. It might be challenging, given the number of committee members who are candidates for the top job and won't want to be seen as too hawkish before they get there.
Competitive co-dependence
A trade "deal" of sorts was struck between Presidents Trump and Xi on the sidelines of the ASEAN confab this week. Details are still trickling out from both sides, but the broad strokes are that China has backed off its latest rare-earth export controls and will resume US soybean purchases in exchange for access to top-of-the-line US AI chips and the removal of Fentanyl-related tariffs. President Trump sought to convey a positive message, saying that the US and China have "settled" their differences.
Our take:
- This will take trade and tariff-related uncertainty out of markets for a few weeks, clearing the way for the market rally to continue into the holiday season (after which we'll get the real read on tariffs).
- The "deal" only highlights the degree of co-dependence that still exists for the US and China. America's technological supremacy limits China's room for maneuver, while China's dominance in manufacturing and rare-earth minerals means the US can't leverage its advantage. This state of "competitive co-dependence" is not going anywhere soon.
The big winner here is the theory of comparative advantage.
What we'll be watching next week
- FedSpeak: Always worth watching the week after a meeting. We'll get Governor Cook on "The Economic Outlook and Monetary Policy" (Monday), as well as Vice Chair Jefferson on "AI and the Economy" and Miran on "Stablecoins and Monetary Policy" (both Friday)
- ISM PMIs: An important first look at October activity and price developments in the manufacturing (Monday) and services (Tuesday) industries.
Memo
Fed balance sheet operations could be a thing of the past
Bottom line: Opposition to QE is in the ascendancy
The Fed has announced at this week's meeting that it will end its balance sheet runoff policy (so-called "quantitative tightening" or "QT") on December 1. In technical terms, this means the Fed's Open Market Desk will now roll over its Treasury holdings. Agency debt and MBS holdings will continue to run off, with the principal reinvested in T-bills.
Those who were hoping that this might, on the margin, support longer-dated Treasuries will be disappointed. While keeping the (nominal) size of the balance sheet stable, the Fed will gradually "twist" its holdings toward shorter-dated securities.
QT has achieved its aims. As US debt issuance projections rise and rise, the Fed has returned to roughly its historical prominence in bond markets. Its balance sheet may have ballooned relative to GDP, but its public debt holdings as a share of total marketable US securities are not much different from the pre-QE era. That share is currently sitting at 21%, down from 36% in 2022 and closer to the 18% that prevailed before the GFC. What's changed is the absolute size of the public debt.
This week's FOMC meeting will be remembered as the one where the Fed ended QT. Could it also mark the end of balance sheet operations altogether (i.e., will we ever see quantitative easing again?).
This is not as outlandish a question as it may initially seem.
The five-candidate shortlist recently announced by Treasury Secretary Scott Bessent includes several names associated with skepticism or outright opposition to QE. Kevin Warsh is a longstanding critic. Michelle Bowman has led the charge to rationalize the Fed's balance sheet. Kevin Hassett, seen by many as the frontrunner, has previously supported QE in a limited context, but more recently expressed skepticism.
The intellectual leader of a campaign to limit "unorthodox" policy tools such as QE is John Cochrane of Stanford's Hoover Institution (where he is a colleague of Kevin Warsh), who this week published an essay arguing for a version of Fed independence with a narrow mandate and a more limited toolkit.
With the Fed's policy framework currently under review, this is something to keep a close eye on.
There is, of course, an irony at play. This movement toward a more constrained Fed could be read as an attempt to partially reduce the Fed's power and "independence" relative to Congress and the executive branch. At the same time, it's inherently hawkish in terms of constraining how the Fed can fight future downturns or ease the impact of public debt burdens. It would therefore run counter to the Trump administration's clear desire for lower interest rates and (this is a personal suspicion) its willingness to allow some back-door debt monetization if the need were to arise.