Wanted fries and a milkshake and a cheeseburger, and you get nothing like it. I don't know, we'll see. Maybe Steve Miran, who's joining us now, Hudson Bay Capital Senior Strategist Stephen Miran. He had huge jobs, he still does, but he was recently a Fed Governor and a CEA Chair at the same time for a while, I think. Were you, or did you have to give the CEA up immediately? How did that work? You had to give it up, I guess.
Well, I wasn't performing the role of CEA. I was on leave from the White House.
But you were playing it on TV. No, I'm kidding. That's an old playing doctor. So welcome. It's good to see you. So many things that you're saying here in some of your comments would give us a way of not hiking, even though a lot of people think we do need to hike. And I think you need to explain why you think that, that the core is actually 2.5% and that the PCE statistically isn't being calculated correctly because of too much stock market emphasis, so it's actually below.
Look, let me put it this way: Core CPI is running about 2.5%, right? Historically, that's a totally normal number, okay? That's not an abnormal number. Historically, CPI runs about 40 basis points above PCE because of differences in the weighting methodology. So if that historical relationship, the historical norm, were holding, it would be 2.1%.
It's not, it's 3.5%, isn't it?
It's not. The relationship between CPI and PCE is inverted by about a point, and that point of inversion is largely measurement error. When you look at the components of PCE that are elevating it above that sort of normal relationship, about 45 basis points of core PCE on a year-over-year basis is this portfolio management services, which basically just tracks the stock market. Stock market moves higher, it mechanically brings PCE inflation with it—approximately 45 basis points to core PCE from that. That was about half of the month-on-month increase in core PCE that we got yesterday was from this alone. Basically the stock market moves up and that mechanically brings inflation with it. That's not consumer prices. That's not real. There's a lot of other error from software measurement, which is conflating quality improvements with price increases. So for example, Excel now comes with an LLM in it, with AI in it, the price goes up. The inflation data are mistakenly counting that as a price increase when it's actually a quality improvement. These two categories together, almost 70 basis points. Now here's what's interesting: The Bureau of Economic Analysis is going to start revising the methodologies in these components because they recognize that there's measurement error here, and those revisions are going to start coming out a little bit more than a month from now. So the PCE data are about to be revised down in the near future to recognize that there's measurement error in these components. It would be a really, really strange thing to start hiking just as the data are going to be revised down by a lot.
I can hear the outcry now that we're just cooking the books to make the affordability crisis look like it's in everybody's mind. That's what you're going to hear: "Oh, you really can't afford things when people know that they can't."
So that would be a legitimate argument if there were corroborating evidence of high inflation from CPI, but there's not. The CPI data more closely correspond to what people actually spend money on; it captures out-of-pocket expenditures. The PCE data capture imputed prices that people don't actually spend on, like those portfolio management fees. Look, every year there's a news article about how ETF and mutual fund fees come down. Fees go from six basis points to five basis points. These things have been in trend deflation for 20 years, and yet they're one of the single biggest contributors to core inflation. It's just pure measurement.
But all this together means they didn't raise the last two times because of the good inflation numbers, or for whatever reason. Maybe they weren't always good, but they haven't gotten worse. So you think there's no reason for Warsh to get more hawkish.
Look, I think that if you—
If you think that holding was the right move in June, holding was the right move in July, which is what the Fed did—
And then you continued to get better inflation data, which we've gotten now a couple of months of much better inflation data, it would be weird to sort of say, "Okay, before I thought it was appropriate to hold. Now we got good inflation data and now I think it's appropriate to hike." There's no reaction function that gives you both a hold in June and July and a hike in September.
What do you think the chances are that this argument is convincing to the rest of the board? You were there, I was there, and I'm sure you've made this argument before. This is not a new argument, at least from you—maybe new to us.
I did make this argument, and I think that the response was, "Well, we've chosen to target measured PCE, right?" That's what the response is, right? "And so our credibility hinges upon targeting measured PCE, even if there's measurement error in there." Now, for me, that's not satisfying because there's always measurement error in how you measure inflation—always going to be the case—and the target is chosen with that in mind. That's why we have a 2% target instead of a 0% or 1% target. It's chosen with that. But if the measurement error grows over time, you're implicitly moving your inflation target around. So if the measurement error in portfolio management services is constant, fine, built into the target. If it gets bigger because the stock market doesn't go up by the historical 6%, it went up by 20%—
You moved your target down.
You were unsuccessful in persuading the board that this was the way to think about it. Do you think that Kevin, A, agrees with you, and B, will have any better shot at changing this dynamic?
So I don't know if Chairman agrees with me or not. However, what I will say is that BEA clearly recognizes there's measurement error here, and the Bureau of Economic Analysis clearly recognizes that the way they were doing things previously was inadequate, and that's why they've chosen to revise their methodologies, right? And so we'll be getting those a little bit more than a month from now. BEA recognizing there's a problem means that will feed through into the measured inflation data that we get. And so about a month from now, two months from now, we're going to see downward revisions to core PCE. That could be 20 basis points, it could be 40 basis points, there's a range. I don't know exactly how it's going to shake out, but they recognize that this stuff is real.
And we'll go on from this. Hiking rates to compensate for statistical noise is effectively asking Americans to lose their jobs because of a measurement error. I don't think the Fed should do that.
My other point, and I've been making this recently, was if the Fed's tools—raising rates—will not address what's causing this inflation, which you say isn't really as bad as it seems, but if let's say it is the war, let's say it is whatever one-off it is, if it's not going to solve that, then why try to do it and cause someone to lose their jobs when it isn't even going to be effective? And this is sort of playing into that argument.
Well, here's what I'd say: The Fed's got two mandates, right? It's got maximum employment and it's got stable prices. It's got to weigh those off against—
Warsh thinks there's a one mandate, it sounds like.
