Latest / Inflation Explained with Fexingo: CPI, Prices, and the Cost of Living for Everyday People / The Inflation Gap Between New and Used Cars That CPI Misses
Transcript
- Lucas: So the May CPI data came out a few weeks ago, and there's a really interesting split happening inside the transportation numbers—new cars are still getting more expensive, but used cars finally started to drop. Luna: Yeah, I saw that. Used cars fell about 0.4 percent month over month, while new cars ticked up 0.2 percent. But that gap feels like it should be bigger, right? Lucas: It should, and that's exactly the puzzle. The CPI methodology actually masks how much prices have diverged. The index uses a substitution bias assumption—when car prices rise, the BLS assumes consumers shift to cheaper models or delay purchases. But in reality, if you need a car, you're not necessarily switching to a cheaper brand. Luna: Right. So the CPI is effectively smoothing out the spikes, making inflation look more moderate than what people actually experience at the dealership. Lucas: Exactly. And that's why the core CPI—excluding food and energy—still sits at 336.1, up from 335.42 in April. The headline number is 334.0. Those are still elevated, but the used-car drop is helping keep them from rising faster. Luna: Meanwhile, the Fed's preferred gauge—the PCE price index—hit 131.5 in May, up from 130.94. Core PCE is 130.1. That's a slightly different basket, but the same trend. Lucas: And the 10-year breakeven inflation rate—that's the market's expectation for average inflation over the next decade—dipped to 2.23 percent. So bond traders are saying, yeah, inflation is cooling, but it's still above the Fed's 2 percent target. Luna: Let's talk about that substitution bias specifically. How does it work for cars? Lucas: Here's the mechanics: The CPI tracks a fixed basket of goods, but the BLS updates the basket weights every two years, and they also allow for 'substitution' within categories. If the price of a Toyota Camry goes up, the model assumes some buyers will switch to a Honda Civic or a used car instead. That dampens the measured price increase. Luna: But that assumption doesn't hold when supply is constrained. After the pandemic, chip shortages meant new car inventories were lean for years. People couldn't just switch to a cheaper new car because there weren't enough of any model. Lucas: Precisely. And even now, in mid-2026, inventory levels haven't fully normalized. The average new car transaction price is still around $48,000, according to industry data. So the CPI's substitution adjustment is understating the true cost increase for anyone who actually bought a new car. Luna: And for used cars, the opposite is happening. Prices peaked in 2022, then fell through 2023 and early 2024, but they've been bouncing around. Now May's data shows a decline again. But the way CPI captures that—it's an average across all used cars, not accounting for the fact that older, cheaper cars are seeing bigger drops than late-model used cars. Lucas: Yeah, the BLS groups all used cars together. So if a 2018 Honda Civic drops 2 percent but a 2023 Toyota RAV4 drops only 0.5 percent, the average is 0.4 percent. But the typical buyer looking for a reliable used car is often pushed toward newer used cars because financing is tighter on older ones. Luna: So the CPI is kind of averaging away the pain that specific buyers feel. Lucas: Right. And this isn't just an academic quibble. The Fed watches core PCE more than CPI, but both have similar methodological quirks. The Fed Funds rate has been at 3.63 percent since June, and the effective rate is flat. The Fed is waiting to see if inflation truly settles. But if the official numbers are smoothing out real price pain, they might cut rates too soon or too late. Luna: Let's bring in a real example. Say you're a listener who needs a car right now. The CPI says new car inflation is running at about 2.4 percent year over year. But actual transaction prices are up more like 5 percent from two years ago. That gap matters. Lucas: It does. And it's one reason why consumer sentiment remains low even as headline inflation cools. People feel the pinch at the dealership, but the official numbers tell them, 'Hey, it's not that bad.' That disconnect erodes trust in the data. Luna: If today was actually useful to you, the way these stay ad-free is listener support. If you've gotten something out of walking through the economy with us, you can help keep that going. Lucas: Yeah, it's buy me a coffee dot com slash fexingo. Just a way to say this kind of detailed breakdown matters. And we really appreciate it. Luna: Alright, back to cars. So Lucas, what's the fix? Should the BLS change how it handles substitution? Lucas: Some economists argue for what's called 'superlative indices'—like a Fisher index—which don't assume substitution but use current consumption patterns. The BLS actually publishes a supplementary CPI using a geometric mean that partially addresses substitution, but it's not the headline number. Luna: So the headline CPI is always going to be a bit of a lagging, smoothed indicator. Lucas: Exactly. And that's fine for some purposes—like indexing Social Security payments—but for understanding your personal cost of living, it's better to look at the detailed components. For cars, the 'new vehicles' index and 'used cars and trucks' index are published separately. You can see the divergence. Luna: And those show that new car prices are still climbing, even if used cars are dropping. So if you're in the market, you're probably better off buying used right now, if you can find a good one. Lucas: Assuming financing works for you. The Fed's high rate means auto loan rates are still around 7 percent for new and 11 percent for used. So the monthly payment might not feel cheaper even if the price is lower. Luna: That's the real-world inflation that no single number captures. Lucas: Right. And that's why we keep coming back to these methodology questions. The numbers are useful, but they're maps, not the territory. Luna: And the map for cars right now shows a split that the headline CPI flattens. Good to keep in mind. Lucas: Alright, next time we'll look at how the PCE handles medical costs differently from CPI—another big gap. Thanks for listening.