Weekly Wrap-Up: Inflation, Econ Vibes, and How the White House Decided to Go to War

Consumer sentiment hit its lowest level on record last week.

I know I’m always saying this past week was especially consequential, but don’t blame me if events keep coming at us with such speed and fury!

47.6

The datapoint of the week was 47.6. That’s the preliminary April reading of the nation’s economic sentiment from the UMich survey. I tweeted about this on Friday, and offered many germane caveats of which you should be aware (e.g., it was taken pre-ceasefire, though the ceasefire appears fragile this AM). But none of them alter the conclusion that people are pretty shook up with what’s going on in this country and, in this context, its economic impact.

What’s that? Not everyone’s unhappy? Trump supporters still revere the Orange Menace? (Also apparently happy: Lauren Sanchez Bezos.) You’re right, but MAGA’s gonna MAGA, and there’s nothing we can do about that. And, in fact, Trump is losing ground with supporters. Twenty-one percent of women who voted for him now disapprove vs. 12% for men. In the UMich survey, Rs support for Trump, while still elevated, is down 10 index points over the past two months. To me, those disapproval numbers are ridiculously small, given Trump’s record, but they’re as weak as they’ve been.

But what matters electorally is neither the never- or always- Trumpers. It’s the decisive middle group that tends to lean anti-incumbent when they find their affordability concerns unmet by the latest politician they believed would help them. And those folks are as fed up with Trump as they were with Biden.

But didn’t inflation peak at 9% under Biden and at least so far, a smaller 3.3%—that’s March’s yr/yr CPI, juiced by an 21.2% jump in the gas price, the largest monthly spike on record—for Trump?

For one, when Biden left and Trump took over, in Jan25, the yearly inflation rate was 3%, and its jiggled around from the mid-2s to ~3% ever since. Also, prewar, goods inflation was already on the rise, as per tariffs, and groceries too, as per tariffs and deportations (leading to immigrant outflows) of farm workers. My point is that, as much as the Trumpies would wish otherwise, people’s reference point isn’t Biden’s 9% in June of ‘22. It’s Trump’s 3% when he got here. It his economy now and, at least in March, inflation is up from when he took office.

Now remember, that decisive middle group—the anti-incumbent swingers—pulled the lever for Trump based largely on his claim that he’d get prices and interest rates back down to where they were during Trump 1. Instead, he’s pushing hard the other way, and while the link between presidents’ actions and economic outcomes can often be hard to see, if the case of this war, making that linkage couldn’t be simpler.

Trump ignored warnings that his and Israel’s invasion would shut down the Chokepoint of Hormuz, and now oil’s up ~$30 a barrel and gas is up >$1/gallon. Anyone can connect those dots, which helps to explain 47.6.

But, you know my methods, Watson. I don’t over-torque on one month. These recent datapoints are very important, but they’re noisy, monthly data. Should a ceasefire stick and the SoH reopen, both inflation and sentiment will improve.

But I don’t think they’ll improve that much, especially sentiment/consumer confidence/econ vibes. And that’s because, as I’ve long argued, it’s not only inflation people are unhappy about; it’s the level of prices.

Here’s an update of work I’ve been doing with Daniel Posthumus, from a forthcoming paper from SIEPR. It shows the full UMich series, along with a two predictions from two statistical models, based on inflation, unemployment, real consumer spending, and the S&P 500. We trained the model on data from 1990-2019 and then forecast and backcast. As you see, both models fit the data well, even back to the 1950s.

At least, they do until COVID and the shock to prices (see pullout figure to the right). Then the first model says we should be a lot happier, damnit!

To explain the gap, Daniel and I added this variable, which we think of as a trend-break to the price level. We run a smooth trend through the CPI price level through 2019, and extend the trend’s 2014-19 growth rate. Think of that shaded area at the end as a quantifying of the extent to which people were shocked/surprised/angered by the jump in the price level. We plug the log difference between trend and actual into the model, which, as you see in the pullout above, tracks the actual decline in sentiment.¹

Of course, the price-level shock isn’t the only thing delivering the lowest sentiment read on record. There’s chaos, threats to the rule-of-law, the horrible and shocking fact that the killers (and the killers’ bosses) of Renée Good and Alex Pretti continue to evade justice, and now a war that, at least at this moment, has left America weakened and isolated (listen to this on that; I found it powerful).

I’ll get back to all that in a moment when I comment on a striking insider report on how the Trump admin got to the decision to go to war. But as they say in golf, “don’t lose the hole” (which I take to mean “don’t get so hung up on reading the green that you lose sight of the cup”). Or, if you prefer, take a close shave with Occam’s Razor. Affordability—the lack thereof—is a highly consequential issue for American families right now, and not just low-income families. This goes up the scale. And for a lot of folks, that’s a matter of what things cost, i.e., the price level.

I also highlighted a new problem for folks in this space, again, if it sticks. The weaker job market has led to slower nominal wage growth (that’s a trend, not a data point), and thus with the March inflationary spike, real hourly pay fell in March and hardly grew over the past year.

In other words, both sides of the affordability challenge are in trouble right now: prices and wages.

How The Decision To Go To War Was Made

The other big splash last week came from this article, by the very-well sourced Maggie Haberman and Jonathan Swan (I speculate on that sourcing below).

You’ve gotta read it—I listened to it while working out; highly recommended. I won’t recount the zigs and zags. Instead, I’ll offer what the kids might call “a hot take.”

I’ve heard many experienced voices, including the guys in the excellent Opinions podcast I linked to above, calling the process described in the piece “remarkable” and “shocking.”

As someone who’s been involved in White House processes—obviously, not to go to war, but I’ve been in the room for big ones—what surprised me was how normal this process was. We understandably view the Trump admin as completely bespoke, but this process, granularly described, was quite familiar. The president is considering an action, and the staff weighs in. Opinions differ, as they did here, and there’s definitely too much telling the boss what he wants to hear (that’s always the case, not just in this example).

But key voices warned Trump that the Strait would do what it did, that regime change wasn’t going to happen, and that this wasn’t the Maduro raid, meaning that the likelihood of getting bogged down was high. No question, other key voices were all: “if you’re for it, Mr. President, then it must be the right call.” And General Dan Caine, who Trump appears to trust on such matters, comes across as particularly ineffective in that he never took a side.

But the process described is a pretty normal one. What went wrong wasn’t due to the process. It was the inevitable result of the guy at the top hearing only what he wanted to hear, having far too much faith and confidence in his deeply uninformed instincts, not to mention his gullibility re what Netanyahu was selling him.

One last political point. There’s a good question as to where Hab/Swan got all this detailed info. This White House is actually not particularly leaky, and they were careful to keep some known leakers out of the room. But my guess is this mostly came from VP Vance’s team. If that’s right, it was due to his desire to make sure his opposition to the operation is known to the base. Not my business—talk about not having a horse in this race!—but making the boss look bad is a dangerous game, JD.

1. Statistically-minded readers will correctly point out that this price-spike variable is non-stationary (look it up if interested).