Latest / The Macro Memo with Fexingo: Daily Conversations on Inflation, GDP, and Federal Reserve Policy / The Growing Gap Between Core CPI and Core PCE
Transcript
- Lucas: So we got the core PCE reading for May on Thursday — 3.4 percent year-over-year, the highest since October 2023. That's the number that got all the headlines. But here's what I find more interesting: core CPI for May was also 3.4 percent, but the two indexes have been drifting apart over the last few months, and that divergence tells a pretty important story about what the Fed is actually looking at. Luna: Right, because the Fed's target is based on PCE, not CPI. So when CPI runs hotter, the market tends to overreact. Lucas: Exactly. And that gap has been widening. Core CPI is now running about 0.6 percentage points above core PCE — that's bigger than the historical average. The CPI index level for May was 336.1, while core PCE was 130.1. The annual rates are 3.4 versus 2.8 percent. That half a point matters a lot when you're trying to read the Fed's next move. Luna: And we should mention — this show is ad-free and listener-supported. If walking through the economy with us has made something click, you can help keep it that way at buy me a coffee dot com slash fexingo. Lucas: Yeah, appreciate that. So why the gap? Two main reasons. First, PCE accounts for substitution — when the price of beef goes up, CPI assumes you still buy the same amount, but PCE assumes you might switch to chicken. That makes PCE more realistic in periods of price volatility. Second, the weights are different — CPI gives more weight to shelter, which has been sticky. Luna: Shelter is the big one. Owners' equivalent rent in CPI is still running near 5 percent annually, right? Lucas: Yeah, around 5.1 percent in the May CPI. That alone adds about 1.8 percentage points to core CPI. In PCE, shelter is about half the weight, so it drags the index down less. That's the mechanical explanation. But the real question is whether the Fed sees the narrowing gap as a sign that inflation is converging or diverging. Luna: And the breakeven rate dipped to 2.20 percent last week, even after the core PCE print. So bond markets are pricing in less long-term inflation risk than the CPI suggests. Lucas: Right. The 10-year breakeven — the difference between nominal and inflation-protected Treasury yields — fell about three basis points on the week. That tells me the market trusts the Fed's preferred measure more than CPI. But the effective fed funds rate is still 3.63 percent, and the interest on reserve balances is 3.65. No cuts priced in for months, even with GDP growth slowing. Luna: Real GDP growth annualized was 2.1 percent in Q1, up from 0.5 percent in Q4 last year. So the economy is still expanding, but at a modest pace. That's the kind of environment where the Fed can afford to stay patient. Lucas: And patient they are. The last FOMC statement basically said they need to see sustained evidence that inflation is moving toward 2 percent. Core PCE at 2.8 is still well above target. So even if CPI is a bit noisy, the PCE trajectory is the one that matters. Luna: One thing I dug up — the Bureau of Economic Analysis actually revised the PCE methodology last year to better capture online shopping and digital services. That might explain some of the divergence too, because CPI hasn't fully incorporated those changes. Lucas: That's a great point. The BEA updates PCE weights quarterly using actual transaction data. CPI updates its weights every two years from the Consumer Expenditure Survey, which is slower. So in a period when spending patterns shift rapidly — like post-pandemic — PCE is going to be more current. That's one reason the Fed chose it as its target back in 2012. Luna: So if you're an investor, looking at CPI headlines and panicking about 3.4 percent might be overblown. The real inflation story is closer to 2.8. Lucas: That's the takeaway. Now, that doesn't mean the Fed is about to declare victory. Core PCE has been stuck between 2.7 and 2.8 since February. Services inflation ex-housing is still running at around 3.5 percent. And the job market — nonfarm payrolls hit 159 million in May, job openings surged to 7.6 million, up from 6.9 million. That's a tight labor market that could keep wage pressure up. Average hourly earnings rose to $37.50, up 0.2 percent month-over-month. Luna: And initial jobless claims dropped to 215,000 last week. So layoffs are low, hiring is steady. The Fed can't cut rates into a tight labor market with inflation still above 2.5 percent. Lucas: Right. The unemployment rate held at 4.3 percent in May, which is low by historical standards. So the Phillips curve story — low unemployment leading to wage inflation — still has some legs. But the interesting thing is that wage growth has actually moderated. Average hourly earnings are up 4.1 percent year-over-year, down from the 5 percent plus we saw in 2022-23. That's consistent with productivity growth absorbing some of the wage increase. Luna: So the gap between CPI and PCE — is it likely to close, or could it widen further? Lucas: I think it narrows over the next few months. Shelter costs in CPI are finally showing signs of deceleration — the May data had owners' equivalent rent rising only 0.3 percent month-over-month, the smallest increase in over a year. If that trend continues, the CPI rate will drift closer to PCE. But it's not a given. The other big component is medical care services, which has been volatile. PCE uses a different source for medical prices — it includes Medicare reimbursement rates, which have been cut, while CPI uses out-of-pocket costs. That adds about 0.2 percent to the gap. Luna: There's a paper from the San Francisco Fed that found the cpi pce gap tends to widen when inflation is high and narrow when it's low, because of the substitution effect I mentioned. So if inflation comes down, the gap should close. Lucas: That's consistent with the data. In 2021, when inflation was spiking, core CPI was running about 1.5 points above core PCE. Now it's about 0.6. As we get closer to 2 percent, the gap should shrink to maybe 0.2 or 0.3. So really, the CPI headline is a lagging indicator of the progress the Fed has already made. Luna: What does this mean for the rate path? The fed funds futures are pricing in the first cut in Q1 2027. Do you think that's realistic? Lucas: I'd lean toward Q1 2027 as well, unless the labor market cracks. The economy is still adding jobs — nonfarm payrolls grew by 172,000 in May, roughly in line with the pre-pandemic trend. But the manufacturing sector is showing some weakness. Factory layoffs spiked earlier this year, and the ISM manufacturing PMI has been below 50 for several months. If that softness spreads to services, the Fed might have a reason to cut sooner. Luna: But core PCE at 2.8 percent doesn't give them much room. They'd need to see a sustained drop toward 2.5, maybe even 2.4, before they signal anything. Lucas: Agreed. And the new Fed chairman — who took over in March — has emphasized data dependence. They're not going to be swayed by one month of CPI noise. The May core PCE print was a reminder that the last mile of disinflation is the hardest. We've been hovering around 2.8 percent since February. That's not a trend. Luna: One more thing — the nominal GDP reading for Q1 came in at $31.87 trillion, up from $31.42 trillion in Q4. That's a 2.9 percent annualized growth rate in nominal terms, but real growth was only 2.1 percent. So the difference is inflation. And that inflation drag is exactly what the Fed is trying to squeeze out. Lucas: Right. The GDP deflator — which is actually the broadest measure of inflation in the economy — came in at 2.0 percent annualized in Q1. That's pretty close to the Fed's target. So if you look at the GDP numbers, inflation looks better than either CPI or PCE suggest. That's the paradox. The Fed uses PCE, but the GDP deflator is telling a slightly different story. Luna: Because the deflator includes investment and government spending, not just consumption. It's a different basket. But it's not the Fed's target. Lucas: Exactly. And that's why you have to be careful which inflation measure you're looking at. If you're a bond trader, you watch PCE. If you're a consumer, CPI matters more for your cost of living. If you're a policymaker, you look at the whole picture. The gap between CPI and PCE is basically the difference between how economists think inflation works and how consumers experience it. Luna: That's a good way to put it. So for now, the Fed sees 2.8 percent, the market sees 2.2 percent in the breakeven, and consumers see 3.4 percent. All three are right, in their own way. Lucas: And that's the macro story of mid-2026. The economy is resilient, inflation is sticky but not accelerating, and the Fed is on hold. The cpi pce gap is a useful reminder that inflation isn't one number — it's a set of lenses. And the lens you choose changes the picture completely.