Intro
Austan Goolsbee, president of the Federal Reserve Bank of Chicago, sits down at Jackson Hole to work through whether policy is restrictive, what the AI build-out is doing to prices in his district, and why he dissented against cutting rates at the last meeting of the year. Joe Weisenthal and Tracy Alloway press him on the neutral rate, the bond market, the dot plot, and what the FOMC feels like under a new chair.
Guest: Austan Goolsbee, president of the Federal Reserve Bank of Chicago
Hosts: Joe Weisenthal and Tracy Alloway (Bloomberg)
Published: 28 August 2026 on Odd Lots · recorded 27 August 2026
Listen on Omny | 46 min
Key Takeaways
Goolsbee is nervous about inflation, not about growth
"I'm OK with waiting to see, but I'm a little nervous that the inflation side has, over the last six months, not been looking great"
Services inflation is the part he cannot blame on tariffs or oil
He dissented against front-loading rate cuts and says he wants evidence first
"The reason I dissented in the last meeting of the year when they cut rates was I'm not comfortable front loading the rate cuts, counting on this inflation to be transitory and go away"
The data center build-out only matters for policy once it leaves its own lane
Chicago Fed district businesses tell him land, construction and HVAC workers are all being bid away
If the AI build-out is driving aggregate inflation, rates are too low
"If that starts happening in the aggregate, we're not restrictive enough"
His long-run landing spot is a 3% policy rate, 2% inflation, 1% real
R-star exists as a concept and is useless to him as a guide
"I always called R-star our Sasquatch"
Productivity has not yet done what the AI optimists promised
"We've now gotten six months in a row of pretty crummy productivity growth"
He splits the post-COVID inflation run-up roughly two-thirds supply, one-third demand
The credit he gives the Fed is for holding the inflation anchor, which TIPS showed never broke
He supports less forward guidance and rejects the idea that the bond market should set policy
The Federal Reserve Act names employment and prices, not the stock or bond market
The SEP's multi-year forecasts cost the Fed credibility when they miss
Dissenting cost him nothing inside the committee and plenty outside it
The FOMC feels very different under the new chair, down to the snacks
Jackson Hole, a Black Bear, and Whether Policy Is Restrictive
Weisenthal opens by noting the conversation is being taped on 27 August, ahead of Chairman Warsh's speech, and says the whole reason for being in Jackson Hole is to talk to as many people as possible about monetary policy and central banking
Alloway introduces Goolsbee as one of the show's favorite Fed presidents, then immediately walks it back to insist the show does not play favorites
Weisenthal reports seeing a black bear with some of the show's producers the day before, which Alloway says she has never managed at Jackson Hole; Weisenthal's verdict: "Hopefully not indicative of anything in the broader economy."
Asked flatly whether policy is restrictive right now, Goolsbee refuses the question as posed: the answer depends entirely on what you believe underlying inflation is, because what matters is the real rate — the policy rate minus expected or actual inflation over some period
His long-run landing spot: "I loosely think 3% rates with 2% inflation and 1% real is kind of an eventual landing spot"
The arithmetic cuts both ways, he says: if inflation is running 3% or more, the real rate is a lot lower than it would be with inflation heading back to target
Where he actually lands: "I'm OK with waiting to see, but I'm a little nervous that the inflation side has, over the last six months, not been looking great"
R-Star as Sasquatch, and What AI Might Do to It
Alloway frames the harder version: restrictiveness is measured against R-star, R-star is unobservable at the best of times, and AI is a huge structural change sitting on top of that
Goolsbee, who says he was an academic for 30 years, has a standing joke for it: "I always called R-star our Sasquatch" — you can never see it until after it has left, and here was a footprint
The concept exists, he says, but he does not find it helpful for deciding what the next policy move should be, because it is not observable even in the best of times
Over a longer horizon he does think faster productivity growth raises R-star — faster growth requires a higher steady-state interest rate
The distinction he says matters is whether the productivity gain is expected or unexpected, a point he made in a speech at the last Hoover conference
Unexpected and landing on you: inflation falls, behavior has not adjusted, rates can come down
Widely hyped in advance, with everyone spending against a coming bounty: the economy can overheat in the short run and rates have to go up
He puts the mid-to-late 1990s in the second category, as the case the US has already lived through
Decomposing the Inflation: Tariffs, Oil, and the Services Problem
Weisenthal asks him to set the theory aside and just decompose why inflation is still warm and drifting the wrong way
Goolsbee says he warned before Liberation Day that a big enough supply shock does not behave like a one-off: COVID taught that a supply shock large enough can take far longer to pass through than initially forecast
The textbook case for tariffs is a one-time increase to the price level and a temporary inflation shock, and he says the evidence that this was actually true was thin
Stacking a war-driven, one-time oil price shock on top of the tariff shock before the first one cleared is what he calls a dicier proposition
Services inflation is the piece that worries him, because it cannot be attributed to tariffs or to oil prices, which makes it a deeper level of concern
He is hopeful the bulk of it was temporary — and openly uncomfortable saying so: "I'm giving myself hives even using that word" about transitory
The test he sets is that the temporary part has to actually show up as temporary: it cannot be that each quarter the Fed says it is about to go away, just not yet, three months from now
That is why he was comfortable waiting a meeting or two on one or two moderating inflation readings
Why More Supply Shocks Should Change How Central Banks Think
Alloway asks whether the central bank should be building persistent uncertainty into its mandate, given shock after shock in recent years; Goolsbee agrees, then corrects the framing to its thinking rather than its mandate
His read is that the mix of shocks has changed: the traditional demand-driven business cycle, where unemployment falls as inflation rises, is not the main thing happening, and commodity prices are not the driver
When inflation and unemployment move together, as they do under supply shocks, there is no automatic playbook for what to do
He says the Fed and central banks everywhere should at least be contemplating what they do if shocks keep arriving at this frequency
On the labor market, he blames the uncertainty itself: tariffs and war have piled uncertainty on top of itself, and that is what has produced the low-hiring, low-firing combination, which he calls not a normal combination
When the Data Center Build-Out Stops Being One Sector's Story
Weisenthal asks whether the AI build-out shows up in the district as tighter supply chains and competition for real resources — labor, materials, commodities, parts
Goolsbee's threshold test is whether the pressure escapes the sector: AI investment going up while another industry competes for electricians and construction workers and complains bitterly about it is not the same thing as the economy overheating. It has to get out of its lane and drive up wages and prices outside that lane
His aggregate check remains the national unemployment rate and the national GDP growth rate, on the logic that if one sector rising causes another to fall, the aggregate is unchanged
What he hears on the ground says the distance is not large. In Cedar Rapids, Iowa, he asked what the biggest problem was: "the data centers are buying up all the land. They're driving up the prices. Nobody can construct, can do any construction. You can't get an HVAC person."
The policy conclusion he draws is explicit: if that turns into old-fashioned excess demand with output above potential, driving inflation, then "If that starts happening in the aggregate, we're not restrictive enough"
Alloway points out people have complained about not being able to get a contractor forever, and now the reason is just the data center
Goolsbee agrees, and takes it to the boogeyman stage: "I half expect people to be like, I need to go to the dentist. They can't see me for three weeks. Thanks a lot, data centers."
Consumption-Led Growth Versus Investment-Led Growth
Alloway asks how Fed policy should treat an economy growing on investment versus one growing on consumption
In the short run, Goolsbee says, it is not that different — overheating has come from housing construction running ahead of what the economy could handle, from consumer spending with the savings rate at zero, and now from business investment
He is blunt about the instrument: the Fed has a very unsophisticated tool, and it can raise the interest rate or lower it
What makes that work is that the most cyclical industries tend to be the most interest-rate-sensitive ones, which he says is why it makes sense for the Fed to be the tip of the spear
Long run, the answer changes, because investment is supposed to raise potential output five or ten years out — and if productivity growth stays high, he thinks rates could be higher for the best possible reasons, calling productivity growth manna from heaven
Weisenthal says growing 3% a year with incomes rising and no inflation is the normalization everyone wanted
Goolsbee's caution is that it has not shown up in the data yet: "We've now gotten six months in a row of pretty crummy productivity growth. So let's not all conclude before it's actually manifested that we've had a change of era."
What Actually Brought Inflation Down
Weisenthal asks for his theory of the case on the last few years: inflation fell without unemployment taking off, but housing has cooled significantly, and the old rule that the housing cycle is the business cycle has stopped holding
Goolsbee's split: supply was a major driver of the run-up and the healing of supply a major component of the fall — "I think it was loosely two-thirds supply and one-third demand then. I wasn't there when it went up, so you can't blame me."
His complaint about the debate is that people want it both ways: blaming the run-up on fiscal and monetary stimulus, then denying policy any credit for the decline by attributing that entirely to supply chain healing
The shoe that did not drop was the inflation anchor, and he points listeners to the TIPS market as the evidence: "Even as CPI was pushing close to 10% inflation, if you looked at inflation compensation in TIPS, it remained steadily at 2.3% CPI."
He notes that CPI of 2.3% is conventionally taken to correspond to PCE of 2.0%, which is exactly the target
To him that is proof the 2.0% target did what its advocates said it would, and losing the anchor would have made the inflation far harder to get rid of
His counterfactual for what conventional wisdom would have prescribed: an AI trained on all of history, told inflation was double or triple target, "would have said, jack the interest rate up to 20% and have a huge recession because that's the only way you ever get rid of inflation"
He gives the Fed credit for recognizing that a component was not permanent and that the much maligned or mocked immaculate disinflation was in fact possible
He concedes the other side of the ledger without hedging: on getting started, they were slow out of the gate, and you cannot look back and say they weren't
Reading the Bond Market Without Letting It Set Policy
Alloway lays out the picture: short-term rates steady, longer-term yields rising, TIPS not pricing much inflation, and asks what that combination is telling him
His first instruction to himself is not to overreact — a central banker should not get into every blip in Twitter of the bond market, and needs a bit of time to sort out what is driving it
Three candidate drivers, and he thinks it is probably a combination of all of them: expectations of higher inflation, expectations that the Fed's rate path has to be higher, or simply more competition in issuance and more bonds being put out
He puts little weight on the US creditworthiness story: "if you really think that a country is going to experience default, the rates aren't whatever, 5%, five and a quarter percent. That's just a historically pretty normal rate."
He flags the reflection problem in reading any market reaction — part of what the market is expressing is what it thinks the Fed will do
The line he closes on comes from Paul Volcker: "Paul Volcker used to tell me, our job is to act and the market's job is to react and let's not get the order mixed up."
Less Forward Guidance, but Not a Market That Referees
Alloway puts Warsh's on-the-record position to him: less forward guidance from the Fed, with the market playing more ball than referee, and the bond market sending a useful signal through yields
Goolsbee declines to characterize anyone else's message and answers only for himself, adding that it is healthy for any organization to refresh what it does and have a rethink
On the substance he agrees with less forward guidance, meaning the practice of saying where rates are going over the next six months and pre-committing to vote a certain way if X happens
His objection is mechanical: that kind of guidance adds to volatility and threatens a tying of hands he does not think is healthy
He draws a hard line at the next step. Gathering information from markets he is totally for; treating that as an instruction on what to do he is not
His authority for it is the statute: "The Federal Reserve Act says by law what we're supposed to look at: maximizing employment, stabilizing prices. Doesn't say anything about stock market, doesn't say anything about bond market."
Press Conferences and Dot Plots Were Built for a Problem That Went Away
Weisenthal's own explainer: press conferences and dot plots are recent innovations, brought in because the Fed had a specific problem in 2008 and 2009, and possibly solved it at the time — so it is fair to ask whether they, or the current number of meetings, still make sense
Goolsbee says everything should be on the table, with a joke about the FOMC's furniture: it is the biggest table he has ever seen in his life, with room for plenty of stuff on it
He points to the outside task force headed by outside people now working through the communications questions, and says the rethink is healthy
He fills in the history Weisenthal raised: the policy rate was at zero, the formulas said it should be around negative 6%, and the Fed faced the unprecedented problem of what to do when the rate is already at zero
Guidance at the zero lower bound — that rates were not going up for years, not until unemployment came down, the Evans rule — were creative solutions to the specific problem of that moment, and the environment now is totally different
A tangent between the hosts: Alloway asks whether she and Weisenthal were both at Bloomberg when the dot plot started, and he corrects her that he was and she was not; she remembers an all-hands committee convened to work out how to display the dot plot graphically, which he confirms
Goolsbee's punchline is that this is a generational memory question — his mother's generation remembers where they were for John F. Kennedy, and these two remember the day the dot plot came out
Alloway's own version: she remembers where she was when the Bank of England retired its fan charts, and says she was really sad about it
What He Would Change About the SEP
Alloway asks what problem the comms task force is actually solving now, since the zero-bound problem is gone, beyond a new chair wanting to lead the staff
Goolsbee agrees the logic of the old tools deserves revisiting, because the Fed is nowhere near the zero lower bound
He has been on public record for years about the Summary of Economic Projections, and his objection is a credibility one: when members write down what gets read as predictions that do not come true, the Fed pays a price
The cost, in his telling, is people going back to the record: "wait, aren't you the bozos who said that by now, whatever, inflation would be 2%, the unemployment rate would be something?"
His starting point is asking whether the Fed really needs projections one, two and three years out plus the long run
On the dot plot itself he sees a purpose that is not being served. He thinks it matters that the world can understand the committee's reaction function and worldview, and the dot plot could do that job
The reason it does not is that the assumptions are not tied to the dots — the median inflation forecast and the median rate are not necessarily the same person, so it does not actually serve as a reaction function
He thinks there is a lot that could be done with the SEP
Reaction Function Is Not Forward Guidance
Weisenthal's distinction, which he says gets conflated constantly: people who say they do not want forward guidance mean they do not want the Fed pre-committing or holding the market's hand, which is fine — but it is still valuable to understand how the central bank thinks about its tools in relation to the data
Goolsbee endorses the split. Forward guidance to him means literally stating what he intends to do with rates and what he thinks the appropriate move is at the next meeting; that is the tying of hands. Saying how he sees the economy and what he is watching is a different thing
Weisenthal gives his own worked examples: in the 2010s the Fed needed to communicate that it would not overreact to a hot inflation print because the priority was employment; by 2022 and 2023 it needed to communicate that it would tolerate labor market softness to get inflation down. That, he argues, is reaction function
Goolsbee agrees, and offers a microcosm from his own first months at the Fed in early 2023. The public argument then was that inflation could not fall until wage growth fell, since wages are the lion's share of services costs
His answer was that the argument gets the dynamics backwards: wages are stickier than prices, so when shocks hit, prices go up first, then wages, then prices come down, then wages
That is why high wage growth does not make him nervous about whether inflation can fall — and he offers it as an example of what a reaction function actually is
The Goolsbee Reaction Function, and the Dissent Behind It
Weisenthal says Warsh's press conferences have been different from his predecessors', that forward guidance is evidently ending, and that he still cannot get a handle on the new chair's reaction function
Goolsbee refuses, cheerfully: he tells Weisenthal he constantly tries to get him in trouble, says he is not allowed to talk about somebody else's reaction function, and directs him to ask the chairman
Alloway backs the question with the state of her inbox: every sell-side analyst note for the past five or six weeks has been about the need for clarity on the Fed's reaction function
What he will give is his own, and it is weighted to one side: he is especially tuned to the inflation side
The dissent, in his own words: "The reason I dissented in the last meeting of the year when they cut rates was I'm not comfortable front loading the rate cuts, counting on this inflation to be transitory and go away" — he notes he is not a voter this year and was last year
What would satisfy him is evidence, specifically evidence that inflation is heading back to 2%; if he gets it, he says he is totally fine, and the path back to the 3, 2, 1 he described earlier is on
His account of the recent path: substantial progress on inflation, then a stall, then movement the wrong way, and now a bit of easing that is still iffy
What makes him nervous is services: if the evidence comes back high and going the wrong way there, with no progress, he says he is going to be nervous
What It Feels Like to Dissent, and Whether Silent Dissents Exist
Alloway asks the personal question, and he warns her he might not answer it before answering it
His fear going in was cinematic: "I was afraid, like, uh-oh, are they going to come, like the goons are going to come punch me? They didn't really."
He was not alone in dissenting, found colleagues respectable about it, and had laid out his criteria in advance, saying repeatedly that he was uncomfortable front-loading too many cuts
The circumstance mattered to him: that was the meeting where the government was shut down and the committee did not have the data, so his position was to get some data before acting rather than keep cutting
No mean or nasty calls came from the chairman or other members. Plenty came from the public, in his summary of it: you jerk
He puts the current run of dissents in historical perspective: more lately than in the immediately preceding period, but by historical standards still far fewer than in the old days
Alloway raises silent dissents, from a Goldman Sachs analyst note observing that there were three official dissents at the last meeting and asking what the unobservable count is; she finds it funny that economists are out there trying to count something that is not happening
Goolsbee says one version is real and one is not. Not everyone votes at every meeting, so there certainly are people at the table who would dissent if they had a vote and whose only outlet is saying publicly what they think about the economy
The other version — people voting differently from what they actually think — has not been his experience
He notes the minutes show basically what everyone said, and the word-for-word transcripts follow. Weisenthal says five years; Goolsbee corrects him to four years, nine months
The FOMC Under a New Chair
Weisenthal returns to a question he asked Goolsbee at Jackson Hole a year earlier, about why dissents are generally rare and whether that is genuine agreement or a chair skilled at corralling the committee. Goolsbee's answer then was both, and he credited Chair Powell as quite good at it
Asked whether it feels different now, Goolsbee does not hedge: it feels very different, personally and institutionally, though he notes he only served under one chair before this
He points to the unusual fact that the former chair has moved his seat over a couple of spaces and is still in the room
The detail that says the most: "There's no snacks. The snacks are outside with the phones." Members are not supposed to bring them in
The new chair has a different worldview, visible in press conferences and in his talks, and has publicly said he wants to stir up more debate
Goolsbee's response to that is open but pointed: he thinks there was debate before, and if the answer is more of it, or changing the format of the meetings, he is open to thinking it through
Weisenthal floats a two-popes tension with Powell still in the room; Goolsbee declines that one too, saying Weisenthal is trying to get him in trouble on a different road, and that his read is simply that there is a new person and new leadership bringing the job into his own
What He Is Listening For in the Chairman's Speech
Alloway asks, with the interview taped the day before, what would make him sit up in his seat during Warsh's Jackson Hole speech
He says he will be listening for reaction function, broadly defined — how the chairman sees the economy
The specific question he is hyper focused on is whether inflation is going to be transitory or persistent, because persistent inflation forces action from the Fed
If it is going away on its own, or if the committee can even believe it might, he says that puts them in a very different circumstance
Joe and Tracy After the Interview
Weisenthal thanks Goolsbee for playing ball and for tolerating the attempts to get him into trouble; Alloway's verdict is that he plays along
Alloway's joke about the speech, recorded before it happened: it would be funny, but not really, if Warsh spent 40 minutes on financial innovation and payments — stablecoins and FedNow — while everyone waited for something on the economy
Weisenthal's takeaway is the forward guidance and reaction function distinction, which he says the discourse handles sloppily
His historical point: go back to Greenspan and there were no dots, no press conferences and much shorter statements, so the Fed is plainly capable of operating with far less literal communication and hand-holding
What he thinks is still missing from the chairman: the reaction function. Goolsbee gave his — the burden is on the inflation side to improve, and otherwise that would probably call for higher rates — and as of the 27th the chairman has not
Weisenthal picks up the negative 6% Taylor rule figure Goolsbee cited as the reason all the extra communication tools were layered on in the first place, and says maybe they did not need to exist as long as they did
Alloway remembers people declaring the death of forward guidance when the Fed deviated from it under Powell, and notes everyone is declaring it again now
She adds that the speech could carry early findings from the task forces, which is why anything from payments to communications is on the table
Goolsbee's bottom line is that he will not cut again on faith: he wants evidence that this inflation shock is not persistent, and if the AI build-out starts pushing prices up across the whole economy rather than one sector of it, he thinks policy is not restrictive enough as it stands.
Products, Companies & Tools Mentioned
Data centers and the AI build-out (The investment boom his district reports is bidding up land, construction and HVAC labor, and the thing he is watching to see whether it leaves its own lane)
TIPS (His evidence that the inflation anchor held — inflation compensation steady at 2.3% CPI even with CPI near 10%)
The Summary of Economic Projections and the dot plot (The Fed publications he has criticized for years, on credibility and for failing to convey a reaction function)
The Fed's communications task force (Outside-led review of press conferences, guidance and projections that he says should put everything on the table)
ChatGPT (His stand-in for conventional wisdom: an AI trained on all of history would have prescribed a 20% policy rate and a huge recession)
Goldman Sachs (Source of the analyst note on silent dissents that Alloway raises)
Bank of England fan charts (Alloway's own example of a retired central bank communication tool, and one she was sad to lose)
FedNow and stablecoins (Alloway's joke about the speech nobody wants: 40 minutes on payments while the economy question goes unanswered)
Books & Resources Mentioned
Goolsbee's speech at the last Hoover conference (Where he worked through what expected versus unexpected AI-driven productivity would mean for R-star and for policy now)
FOMC minutes and the word-for-word transcripts (His answer on how much of the committee's disagreement eventually becomes public, the transcripts on a four-year, nine-month lag)
A Goldman Sachs analyst note on silent dissents (Counts three official dissents at the last meeting and asks what the unrecorded number is)
The Evans rule (Cited as one of the creative zero-lower-bound solutions built for a problem the Fed no longer has)
Get the latest market chatter and takes as they happen:
X | Threads | Instagram | YouTube | TikTok | Facebook