Links to One Paper and One Speech

Affordability, vibes, and the political economics therein! Let's just say I'm setting you up for a very fun weekend.

Paul K, as is his wont, raises good points

The next link I’ll share is a big, long speech I gave yesterday—the Phillip Gamble Lecture at UMass (Paul gave that lecture a few years ago)—wherein I drilled down on an issue that Paul raises today and that G. Elliot Morris (GEM) has also been doing interesting noodling on: using the price-level shock to explain the affordability, or vibes, gap.

But first, this link is to the paper Paul references this AM, forthcoming from SIEPR called The Way We Were: Price-Level Shocks and Consumers’ Memories (this is an unedited draft., co-authored with the excellent Daniel Posthumus, who just got accepted into a PhD program, giving me hope for the future of econ! Congrats!!). In that paper, findings from which I’ve often featured up in here, we first derive what we call the “affordability gap”—the post-pandemic difference between actual and predicted UMich econ sentiment—and we show that by including a variable that captures the level shift in prices, the gap goes away. GEM makes a similar case.

Here’s the key figure, from my talk yesterday (updated from the paper’s version). The little figure is just a pullout of the end of the time-series in the big one. “UMICHSENTF1” are predictions from the basic model; “UMICHSENTF2” adds the gap-closing price-level-shock variable:

Paul’s not so sure that the jump in the price-level explains as much of the gap as we think it does, and I too have reservations, which ftr, I’ve been up front about in all these scribblings. I’ll get to those, but first, I’m sure Daniel and I (and GEM) are, to some degree that’s hard to pin down, right about this.

Maybe it’s me, but I find the extremely simple figure below tells the simple story about what’s going on here:

Inflation climbed the mountain at record speed, then fell just as quickly, though never back to the Fed’s target, which is another very interesting and important point which I write about here, and which Paul also needs to weigh in on. But while many economists, including myself and those at the Fed, where touting the downside of the mountain on the left, consumers were struggling with the unprecedented-in-recent-memory shock in what things cost, the figure on the right. And remember, I was in the front lines in those years, so you should trust me on this one: People remembered their old prices, and Keynes-damnit, they wanted them back!

So, why then am I sympathetic to Paul’s “I’m-not-so-sure?” There are two reasons, with the first, which is slightly technical, bigger than the second. When I write about the level-shock, or trend-break variable, shown below—the one that closes the gap—I remind technical readers that it’s “non-stationary,” which is a statistical term meaning that its average and variance are not roughly constant. It undergoes a level shift.

Here’s what that means intuitively and why it’s a statistical problem in a model like this. Virtually any variable that jumped at this period—COVID cases, ships stuck in ports, hours spent working from home—would have a similar effect (though ours is better as, relative to these others, it both jumps up and stays up, thereby helping to explain the ongoing gap; this too, however, is a limit of its explanatory utility; it will never close!). We’re plugging in a variable that jumps up and stays up over the very period the affordability gap opens, so of course it will explain the gap. But so will a lot of other things that jump up around then, so, as Paul says, we’re in the world of correlation, not causation.

Second, and I’m less wound up about this than are many others, the UMich sentiment index has some issues. It recently switched its data-collection methods, and it’s known for reflecting partisanship over actual econ sentiment. But there’s definitely an important signal in there. It’s not a coincidence that survey respondents, still recovering from the pandemic shock, followed by the tariff shock, sent the index to its lowest level on record after the latest price shock, the one from Trump’s war-of-choice in Iran (which raises the highly concerning issues of cascading shocks dislodging the inflation anchor).

So, much more to come on this interesting area of inquiry as we work to improve the modelling. But we’re happy to put it out now because we’re confident it’s revealing something real and important.

BTW, you know what would be fun? If Paul, GEM, and I all spent 30 mins batting this around on one of their videos.

The Political Economy of the Affordability Agenda

Here’s a link to my big, long speech I gave yesterday at UMass for the Phillip Gamble Memorial Lecture (here’s the accompanying deck). I put a lot of work into it and incorporated great suggestions from many others in this debate, so if you’re interested in a holistic treatment of the economics and politics invoked by this highly salient and consequential issue, please give it a read.