The Fed Chair Never Loses a Vote: Warsh Talks Against Guidance, but His Body Is Already Paving the Way for Hikes
Posen gives Warsh's Jackson Hole speech a B-, arguing that he talks against forward guidance while actually paving the way for hikes, and predicts the Fed will raise rates 75-100 basis points over the next six months.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Warsh never said he would hike, but the whole speech paved the way
Posen gives Warsh's Jackson Hole speech a B-, but says it could be treated as a B+ given the chaos of the preceding two months. The fourth section of the speech systematically lists reasons inflation could persist or even rise, yet never says ‘therefore we must tighten.’ Posen reads this as deliberate groundwork: if there are no hikes next, the market will suspect this is capitulation to pressure from Trump. Posen himself has argued for hikes for months, because inflation is real. Warsh also made clear a 2% core PCE target and refused to use wage inflation as a forecasting indicator — all mainstream moves.
— Adam PosenWithout defining the ‘right speed’, returning to target is an empty phrase
What worries Posen most is that at the end of the speech Warsh said inflation needs to fall in the ‘right direction’ at the ‘right speed’, without ever defining what the right speed is. Without a speed constraint, ‘returning to target’ is an empty phrase. From his confirmation hearing to two press conferences to the ECB Sintra speech, Warsh has consistently reserved the right to ‘decide at the last minute’, refusing to pre-commit to his own observation indicators. This is close to Greenspan's 1999 style, but Greenspan held up through personal ability and control of the committee; once he erred, the entire credibility collapsed with it.
— Adam PosenThe Fed chair never loses a vote
Posen explains the culture of the Fed committee: unlike the ECB, which seeks consensus, the Fed has historically given the chair more power. More critically, there is an unwritten rule — each meeting has only a few dissents, and the chair never loses a vote. Back then, when Volcker found he might lose a vote, he simply announced it would be his last meeting. This culture leaves the committee weak at correcting the chair, so Warsh's ‘reserve discretion’ style carries more risk than he imagines.
— Adam PosenBashing the central bank isn't dangerous; making it help sell Treasuries is
Posen distinguishes two kinds of political pressure: ordinary central bank bashing is part of the game and can even, when necessary, help the central bank ‘take the heat.’ What is truly dangerous is when a president or treasury secretary asks the central bank to help sell Treasuries — the central bank must then answer: ‘I am here to help you and your successor sell Treasuries’, and must never manipulate or cheat. Trump's threats to the Fed are far more serious than rhetoric — trying to strip regional Fed presidents of their votes, scrambling staggered terms, changing the mandate, touching the budget — these directly destroy functional independence.
— Adam PosenThe most radical reform may be making the central bank talk less
Posen thinks the communications committee led by Mervyn King is most likely to produce a radical proposal. King himself is a founder of inflation targeting and fan charts, but since retiring he has publicly doubted the central bank's forecasting ability and information supply: too much noise makes markets depend on official guidance, creating moral hazard and bubbles. Peter Fisher agrees, arguing that central bank ‘certainty’ makes markets less risk-sensitive. The two could push for less information disclosure rather than more. The ‘hall of mirrors’ metaphor in Warsh's speech already concedes that markets are not necessarily right.
— Adam PosenEconomists who join AI companies lose their independent voice
Posen argues that economists flocking to AI companies is not unprecedented — it happened in the internet era too — but this time it is more troubling. Companies like Anthropic hire economists partly to design AI transition policy, and these scholars genuinely believe they are doing good. The problem is that once inside, they enter a corporate hierarchy, lose their independent voice and freedom to choose topics, and their views get reasonably discounted by outsiders. Posen's hope is that scholars stay in research institutes, take low six-figure salaries, and still have public influence.
— Adam PosenAI displacement hasn't happened yet; programmer employment is still growing
Posen says labor market data does not yet show AI displacement — even employment growth for programmers and truck drivers continues. Brynjolfsson's J-curve theory holds that firms need time to reorganize their businesses, and people and AI will first have a period of collaboration. Garicano's messy jobs theory points out that the jobs most easily automated actually contain a lot of tacit knowledge and human relationships — a truck driver, for instance, does more than drive. Large-scale displacement may take about five years, and will show up first as reduced hiring of young people.
— Adam PosenIn six months rates will be 75-100 basis points higher than now
Posen reiterates that his earlier forecast of 4% year-end PCE inflation is coming true, with CPI already above 4% early in the year. The bigger problem is inflation inertia: although energy prices may pull the numbers down slightly over the next month or two, core services inflation is stubborn, with 3/6/12-month moving averages all rising. He expects the Fed to hike in September or December; if it hikes in September it will hike again in December, and in six months the federal funds rate will be 75-100 basis points higher than now, only then will inflation truly start to fall, spending the interim in a 3.5%-4.5% range.
— Adam PosenIn their own words · checked verbatim
You know, if you were grading it, it's a B-minus speech. Okay. It's a B- speech by normal standards. It's much more positive because of the situation we were in.
Adam Posen2:05
And right direction doesn't mean anything unless you're saying the target. I mean, without the speed, just to emphasize your point, without the speed, then we're back with what we had the last four years, which is inflation's above target.
Adam Posen7:13
Only so many people dissent on the committee at any one meeting. And the chair never loses a vote. Even Volcker, once it was clear he was going to lose a vote, He basically said, next meeting is my last meeting.
Adam Posen12:20
I think the communications fund is going to surprise people. Because both Mervyn and Peter, and I'm only referring to their public statements, have gotten pretty radically skeptical about central bank communications in recent years.
Adam Posen24:34
You're losing your independent voice. What you choose to work on, you may choose not to work on that, which you might have worked on because now you work for this company. So again, is it outright corruption of the sort we're seeing in the Trump administration? No. But it is unsettling.
Adam Posen34:51
I had no idea that the President of the United States would bomb Iran and ignore the fact that disrupting the Straits of Hormones might have some inflationary effects beyond whatever else you think of it. But I knew there was gonna be an inflation shock and we were primed to have more.
Adam Posen53:12
Figures
| Months of above-target inflation | 64-65 months | 13:22 |
| Core PCE target | 2% | 5:09 |
| Expected rise in the federal funds rate over the next six months | 75-100 basis points | 54:13 |
| Posen's earlier forecast for year-end PCE inflation | 4% | 51:11 |
| Possible understated GDP contribution from AI chips | 0.3 percentage points | 45:05 |
| How long the Fed chair's advantaged position has lasted | about 45 years (since Volcker) | 11:17 |
Glossary
- fiscal dominance
- The government forces the central bank to accommodate fiscal financing, stripping monetary policy of its independence.
- fan charts
- Charts that show forecast uncertainty using fan-shaped probability intervals.
- J-curve
- A trajectory in which a technology's initial effect declines, then rises significantly only after adjustment.
- messy jobs
- Jobs that look easy to automate but actually contain tacit skills and human relationships.
- staggered terms
- Regional Fed presidents' appointments are staggered in time to prevent a wholesale turnover.
How to listen
Macro traders watching the Fed's path, economists studying AI's impact on employment, and policy researchers concerned with central bank independence.
The last two minutes of subscription and Discord promotion can be skipped.