The Economy That Refused to Break: Richmond Fed's Barkin on Stubborn Resilience

Once you have crossed the vast sea, no other water is worth the name.

— Yuan Zhen, “Thoughts of Parting IV” (Tang dynasty, c. 9th century; author’s translation)

What This Episode Is About

This Odd Lots episode was recorded at Jackson Hole on August 31, 2026. Tracy Alloway and Joe Weisenthal sat down with Richmond Fed President Tom Barkin — his third consecutive year on the show from the same location.

The framing was refreshingly concrete. Beyond the theoretical questions floating around Jackson Hole — the neutral rate, the term premium, AI’s effect on productivity — Joe wanted something more grounded: what are central bankers actually hearing from real businesses? Barkin is arguably the Fed official who travels the most, and his working method is to accumulate anecdotes from chambers of commerce, town halls and factory floors, then synthesize them into something that can sit on a policy table.

The spine of the conversation is a puzzle: with gas prices up, prices still high, and uncertainty everywhere, why do American consumers keep spending? Barkin’s answer is far more textured than “rich people have money.”

The Main Points

1. Consumers aren’t rich. They’re creative.

Barkin contrasts this cycle with the Great Recession. Back then, people lost jobs, houses, cars; savings were destroyed and retirement had to be rebuilt. The result was five, six, seven years of secular stagnation where nobody spent. COVID went the other way: money in pockets, stimulus, equity values up, home values up — and a mindset he describes as “bound and determined to spend.”

The interesting part is the lower-income end. What he hears about is “a very creative consumer figuring out ways to find money.” Concretely: the growth of private label, the migration to Walmart and dollar stores, people dropping insurance, and more people living at home. He flags the second reading of that last one: when the press writes about young adults moving back in with parents, the emotional register is “oh no, my kids are back.” The economic register is different — those are people not paying rent, spending that money on something else.

2. Two remarkable details about borrowing from the future.

These are the most alive moments in the episode. Auto lenders tell him people are 60 days delinquent rather than 120 days — because they need to find money somewhere, but they don’t want to lose the car. Gas providers tell him people skip the bill in summer, because nobody has a problem with gas in summer. Winter is what matters.

This isn’t “the consumer is broken,” and it isn’t “the consumer is fine.” It’s calculated, prioritized deferral. Money is pulled forward from the future along a clear hierarchy: keep the car, deal with heat later. Barkin’s conclusion: as long as markets are healthy and people have jobs, they’ll keep finding a way to spend.

3. AI’s biggest impact on the economy so far is political, not productive.

This is the most counterintuitive line in the episode. Every chamber of commerce meeting, every town hall — Barkin says he gets asked about jobs, water and data centers. “You can see the issues on people’s minds by the questions they ask.”

Productivity? Outside a few areas where you can genuinely substitute an agent into an existing structure — call centers, programming, heavy documentation and compliance — he doesn’t think AI is producing a massive productivity effect yet. The productivity boom we are seeing, which is significant, he attributes to 2022: firms were short of workers, so they invested in automation, new staffing models, different operating processes. They’re reaping those benefits today.

4. Data centers have no political base.

This is the sharpest observation of the hour, and its logic isn’t economic at all. If you have a manufacturing plant in your hometown, Barkin says, then the kids of the people who work there are on the baseball team, the football team, the hockey team. Data centers employ very few people. So once construction ends, nobody in town has a friend who works at the data center.

The standoff follows: economic developers talk about the tax base; citizens talk about water, about how it looks, about not trusting AI. It brings tax dollars but not enough workers — hence no constituency. He jokes that towns should name an elementary school after Microsoft or Google and tell residents “here’s who built you this school.” But that’s not how people think about it.

5. Crowding out is a chicken-and-egg problem.

Asked whether the AI buildout is squeezing other industries, Barkin concedes that switchgears, transformers and electricians are very hot and in very short supply. But he describes something more specific than “construction has stalled”: construction has migrated from one sector to another. Office buildings aren’t being built. Multifamily starts are way down. Industrial is coming back. Home building is okay, not great.

His pushback on multifamily developers is excellent. When they complain projects don’t pencil and want to talk about interest rates, he says: is it really rates? We had the same rates in 2004 and 2005 and you were building plenty. They then acknowledge construction costs are up, labor is up, they have to put more equity in. But you could also argue data center construction is what made the other construction expensive — and that’s the chicken-and-egg.

6. 65 months versus 47-plus-18: two ways to slice the same numbers.

Warsh’s speech noted inflation has been above target for something like 65 months, and said that sits squarely with the Fed. Tracy asked Barkin — who was in the room for much of it — how he reacts.

His answer is the most methodologically valuable thing in the episode. You can take a 65-month view, he says, or you can take a 47-month view followed by an 18-month view. The second reads: inflation happened, maybe we were a little slow, we raised rates, inflation came down. By March 2025 you have 2.3%, 2.4%, everything heading the right way — the plane is going to land. Then came a series of external shocks (AI, tariffs, oil), pushing inflation back up.

The crucial move is how he marks his own position: “which I’m not arguing for, I’m just saying it’s a perfectly defensible way to do it.” He explicitly grants the other reading too — don’t give me your excuses, it’s been 65 months, and maybe rates aren’t that restrictive. That, he says, is the argument they’re going to have.

7. Why tariffs vanished from the conversation.

A mechanical reason nobody talks about. The headline tariff numbers went up, Barkin says, but collections were never as large as the headline numbers, and then the Supreme Court ruled and refunds started flowing. So for the past three or four months, people have been collecting refunds rather than paying tariffs — and when something works in your favor, you don’t talk about it as much.

Pass-through splits cleanly. B2B firms, “to a person,” are convinced they’ve passed costs through successfully. B2C is harder: easier if you serve wealthy customers, brutal if you don’t — and those selling into big box retailers say they’re having “a devil of a time,” because the retailer’s mindset is “I need to find some price to give to my customers.” Joe’s line: big box retailers are our strongest soldiers in the fight against inflation.

Going Further

”The data looks great — why does it feel wrong?”

This is probably the most common dissonance of the past two years. GDP is fine, consumer spending is fine, unemployment is low, and yet everyone you know is cutting back. You start to suspect the statistics are lying.

Barkin’s episode offers a better answer: the statistics aren’t lying; aggregation is compressing two opposite things into one number. Wealthy households spend because of the wealth effect. Less wealthy households also spend — because they found ways. Drop the insurance. Skip the summer gas bill. Move back home. Go 60 days delinquent instead of 120.

In aggregate data these look identical: consumer spending growth. Their durability is not remotely identical. The first rests on asset prices. The second rests on future slack — and future slack runs out.

So when a macro number feels wrong, the first question is: how many distinct mechanisms are being summed into this figure? Aggregation erases mechanism, and mechanism is what tells you about durability. Barkin’s own condition is refreshingly explicit: as long as markets are healthy and people have jobs, this keeps running. Which means the failure condition for this structure is employment, not consumption. If you want to know whether spending breaks, don’t watch spending. Watch jobs.

”All the good news is out — why isn’t my corner moving?”

Suppose you follow a sector that should benefit from a huge theme. The theme is hot, capital is pouring in, and your particular corner is flat. You start doubting your read.

Barkin’s pushback on multifamily developers is a diagnostic template worth stealing. They say it doesn’t pencil and want to talk rates. He replies: same rates in 2004 and 2005, and you were building plenty. That question blocks the convenient, ready-made explanation and forces the real answer out — costs up, labor up, more equity required.

Learn the move, because it treats a very common laziness: we attribute difficulty to whichever variable is most topical and easiest to say out loud. Rates are high. Tariffs. The macro environment. What these share is that they’re all true and none of them is the actual binding constraint.

The real diagnostic is to find a historical point where that variable was identical but the outcome was different. If the same rates produced buildings in 2004 and don’t now, then rates aren’t the decisive thing — or at least not the only one. Whatever else differs is where your answer lives.

And Barkin’s follow-on matters more: you could argue data centers made the other construction expensive, which is why it isn’t happening. That moves you from “it isn’t rates” to “it might be crowding out” — the same dollars, the same electricians, the same transformers, bought by whoever can pay more.

So the sequence when your corner won’t move: block the glibbest explanation, find a historical comparison point, then ask whether something else is bidding for the same inputs. A hot theme doesn’t only lift things. It also removes supply.

”Which story am I supposed to believe?”

The most useful thing in this episode isn’t any economic judgment. It’s how Barkin handles the 65-month question.

Same history, two slices: 65 continuous months above target, or 47 months plus 18 months of “it was landing and then got hit.” Same numbers, wildly different attribution of blame. Barkin doesn’t pick. He does three things: states both readings fully, explicitly says he is not arguing for the second, and grants that the first stands up.

The investing analogue is direct. You hold something underwater. You have a story where it’s a temporary external shock and the structure is intact, and a story where it’s structural and you just won’t admit it. Most people pick the comfortable one and then collect only confirming evidence.

Barkin models the alternative: write both out fully, then mark which side you’re on and where that position is weak. Note that he wrote the strongest version of the side he doesn’t use (“don’t give me your excuses”).

One step further: where do the two stories diverge? If the external-shock reading holds, indicators return to trend once the shock recedes. If the structural reading holds, they don’t come back even after the shock is gone. That divergence point is your failure condition. It’s worth more than either narrative, because narratives are for convincing yourself and divergence points are for checking your answer.

Barkin’s read on the inflation environment has the same shape. Ten years ago the normal included fracking suppressing energy prices, demographics keeping wage costs in check, e-commerce pushing goods prices down, globalization pushing everything down. It’s not ridiculous, he says, to imagine that a decade later you have a bundle of factors pushing the other way. Then the metaphor: you sail differently with the wind behind you than into it — you can still get there, you just have to tighten your sheet.

Worth Looking Into

• The episode itself (Aug 31, 2026, from Jackson Hole). If you listen to only two minutes, listen to the auto-loan delinquency and summer gas bill anecdotes.
• Warsh’s Jackson Hole speech, which this conversation references constantly — particularly his treatment of inflation breadth and one-off shocks.
• The 2024 Odd Lots road trip with Barkin to Mount Airy, North Carolina. The hosts spend the closing minutes debating a 2027 return, and the reason is instructive: in 2024 they didn’t ask a single AI-related question. Same place, same people, three years apart, is a genuinely good observational method.
• If the forward-guidance debate interests you, Barkin holds two opposing thoughts openly: transparency builds credibility and lets markets do some work for you (the Bernanke theory), and too much forward guidance boxes you in (the 2021–22 lesson).

The One Thing to Take With You

One idea: write the side you disagree with in its strongest form.

The best thing about Barkin’s answer on the 65 months isn’t which side he took. It’s that he articulated the most forceful version of the reading he doesn’t use — “don’t give me your excuses, it’s been 65 months and it’s been over” — before saying how he actually thinks. Most people can’t do this, not because they can’t imagine the argument, but because writing a strong case for the other side is uncomfortable. You discover they aren’t stupid.

Which is exactly why it works. A position you can restate completely is the only kind you’re actually capable of refuting. The kind you can only restate as a straw man, you’ve never really beaten.

An exercise for today:

Pick something you’re currently arguing about, or already have a settled view on. It doesn’t have to be investing — “should the kid go to cram school,” “should I change jobs,” “why does my mother always talk to me like that.” Take a piece of paper. Write two things.

First: in your own words, state the other person’s position in three sentences — written well enough that if they read it they’d say “yes, that’s exactly what I mean.” No downgrades allowed: no “they just don’t understand,” no “they’re only protecting their ego.”

Second: write one sentence — what would have to happen for me to admit they were right?

The moment you can’t write that second sentence is the whole point. It means what you’re holding isn’t a judgment. It’s an identity. Judgments can be overturned by evidence; identities can’t. And nearly every argument that never ends is the second kind.

Raw Markdown

This article is an educational discussion of investment method. It is not advice to buy or sell any individual security, offers no target prices, and does not analyze any current holding. Investing carries risk; make your own decisions or consult a qualified professional.