Latest / The Federal Reserve Podcast with Fexingo: Interest Rates, FOMC Meetings, and Monetary Policy / Why the Feds Inflation Forecasts Keep Missing
Transcript
- Lucas: So the core PCE price index for May came out this morning — three point four percent year-over-year. That's the highest reading since October 2023. Luna: And the Fed's own forecast from March was calling for core PCE to end 2026 at two point eight percent. We're barely halfway through the year and already blowing past that. Lucas: Right, and this isn't a one-off miss. Look back at the past two years — the Fed has consistently had to revise its inflation forecasts upward. The December 2024 dot plot had the median FOMC member projecting core PCE at two point four percent for 2025. It came in at three point one. Luna: So what keeps going wrong? Is it the models, the data, or are they just too slow to update their assumptions? Lucas: I think it's partly a structural issue with how they think about inflation persistence. For a long time, the dominant view inside the Fed was that the pandemic-era inflation was driven by transitory supply shocks — goods bottlenecks, energy spikes, that kind of thing. Once those faded, inflation would come back down naturally. Lucas: And goods disinflation did happen. But the problem is that services inflation — things like rent, medical care, auto insurance — those categories stayed sticky. And they're much larger components of the core PCE basket. Goods are maybe a quarter of the index; services are over sixty percent. Luna: So you're saying the Fed underestimated how much of the inflation was coming from the services side, which is more tied to domestic labor costs and less to global supply chains. Lucas: Exactly. And there's a second layer: the neutral rate of interest, or r-star. The Fed's estimates of r-star have been creeping up. Longer-run estimates from the June dot plot show the median FOMC member now sees the neutral rate at around three percent, up from two point five percent a couple years ago. Lucas: If r-star is higher, then the current fed funds rate of three point six three percent isn't as restrictive as the Fed thought. Monetary policy may not be doing enough to cool demand. Luna: That's a pretty fundamental miss. If you're wrong about r-star, you're wrong about the whole stance of policy. Lucas: Right. And the data we're seeing today — the ten-year breakeven inflation rate has fallen to two point eighteen percent, down from two point twenty-one earlier this week. That suggests the bond market isn't panicking about inflation expectations becoming unanchored. But the actual realized inflation is running hot. Luna: So the market trusts the Fed to eventually bring inflation down, but the Fed's own forecasts keep being wrong. That's kind of a weird disconnect. Lucas: It is. And it raises a question about the FOMC's reaction function. Chair Powell has said they need 'greater confidence' that inflation is moving sustainably toward two percent before cutting rates. But if they keep missing on their forecasts, how do they ever gain that confidence? Luna: Maybe the bar for cutting rates is just higher than they're letting on. Or maybe they're waiting for a clear signal in the labor market. Lucas: That could be. The factory job cuts we saw in early June — near financial crisis levels according to S&P — that might be the kind of signal that would shift their focus from inflation to employment. But so far, the labor market overall still looks fairly tight. The unemployment rate is still historically low. Luna: So if they're not cutting rates soon — and the inflation data isn't cooperating — what's the path forward? Do they have to raise rates again? Lucas: I think a hike is unlikely at this point. The fed funds rate is already above five percent in real terms if you use the breakeven rate as a proxy for expected inflation. That's restrictive by most measures. The issue is more about time: how long do they need to hold rates at this level before inflation actually comes down? Lucas: And that's where the forecast errors really matter. If they think inflation will fall to two point eight percent by year-end, based on the same models that have been wrong, they might hold rates steady. But if inflation stays at three point four percent or higher, then by December they'll be in a tough spot — behind the curve again. Luna: I want to step back for a second. If today's episode gave you a clearer picture of why the Fed's inflation forecasts keep missing — and why that matters for your portfolio or your business — then it's working exactly as we intend. The way we keep this show ad-free and focused on the substance is through listeners who find it valuable enough to support. If that's you, buy me a coffee dot com slash fexingo. That's it. No strings. Lucas: And we mean that. Every contribution goes straight to keeping the research and production independent. So thank you if you've ever chipped in. Lucas: Now, back to that forecast methodology. One thing I've been digging into is how the Fed's staff model — the FRB/US model — handles the pass-through from shelter costs to overall inflation. Historically, shelter has a long lag, and the model assumes that the recent slowdown in market rents will eventually feed into official inflation measures. Luna: But that hasn't happened yet, right? The owners' equivalent rent component of CPI is still running hot. Lucas: Correct. OER is still at around five percent year-over-year. And the model predicted it would have fallen to four percent by now. So there's another miss. The question is whether the lag is just longer than expected, or whether structural changes in the housing market — like more single-family rentals and institutional ownership — have changed the relationship. Luna: If it's structural, then the models need to be recalibrated. But the Fed is conservative about changing its models. They don't want to overreact to a few data points. Lucas: Right. And that conservatism is part of why they've been behind the curve on this inflation cycle. During the 2021-2022 inflation surge, they were slow to tighten because the models said inflation was transitory. Now they're slow to acknowledge that the models may need updating again. Luna: So is the solution to change the models, or to change the way the FOMC uses them? Maybe rely more on real-time data and less on the structural forecasts. Lucas: Governor Waller has actually suggested something like that. He's advocated for a more data-dependent approach that looks at a broader set of indicators — including things like the Cleveland Fed's trimmed mean inflation, which strips out outliers. That measure has been running around three percent, which is still above target but less alarming than the headline. Lucas: But the problem is, if you ignore your model entirely, you lose any forward-looking guidance. You're just reacting to the past. And the Fed's credibility rests partly on being able to say 'we expect inflation to be X in two years.' Luna: So they're stuck between a model that's been wrong and a purely backward-looking approach that offers no guidance. Lucas: Exactly. And that's why the next few months are so critical. If core PCE stays at three point four percent or higher for another quarter, the FOMC will have to seriously debate whether their framework is still valid. Some members — like Kashkari — have already started questioning whether two percent is the right target at all. Luna: Raising the inflation target would be a huge step. That would effectively admit they've lost the battle. Lucas: I don't think they'll do that. But they might quietly shift their horizon. Instead of targeting two percent 'over time,' they could say it's a 'longer-run goal' and accept that the next couple years will see inflation around two point five to three percent. That would be a de facto target change without formally admitting it. Luna: And the bond market seems to be pricing that in already. The ten-year yield at four point three nine percent is well above where it would be if the market expected inflation to average two percent over the next decade. Lucas: Right. The ten-year yield implies an average inflation expectation of around two point two percent, but that's a market-based expectation, not a forecast. And the market has been wrong too. The real question is whether the Fed will adjust its framework before the next recession hits, when they'll need room to cut rates. Luna: If inflation stays sticky, they won't have that room. They'll enter the next downturn with rates at three point six percent and inflation at three plus percent. That's not a great place to be. Lucas: No, it's not. And that's the risk that keeps me up at night. The Fed's forecast errors aren't just academic — they have real consequences for how much policy space we have when the economy turns.