Inflation’s descent will be hard, too

Good morning. Today is Fed day. Everyone expects Jay Powell et al to say “no rate increase today, but maybe in a month”. We expect this too, and think it would be a sensible decision, given yesterday’s inflation report (about which more below). More excitingly, Unhedged’s first podcast episode, hosted by Ethan and featuring the irrepressible Katie Martin, dropped yesterday. Sign up! Also, email us with ideas before we run out: robert.armstrong@ft.com and ethan.wu@ft.com.

Inflation: falling, but to where?

Because we like tempting fate, last month’s letter on CPI inflation was called “Inflation inflection”. The point was that inflation’s trajectory had shifted for the better, even if the problem wasn’t yet resolved. The great shelter disinflation had begun, as indicated by two months of falling CPI rents; and goods prices, other than for used cars, had stopped rising.

To our relief, both facts remained true in yesterday’s May CPI report, further confirmation that the inflection point is real. Yes, there was an uptick in core inflation. As one dejected economist, Stephen Stanley of Santander, wrote: “It should be pretty clear that core inflation is trending at about a 5 per cent clip, with very little downward momentum . . . Sadly, the data suggest that the Fed has made very little progress in getting underlying inflation back down.”

We don’t agree. Take out used cars and rents and inflation looks a hell of a lot better:

Someone in the comments section is already midway through an 800-word screed about how the Unhedged bubble-denizens think prices are falling so long as you don’t eat, use energy, rent a home or drive a car. But this misses the point. The purpose of data adjustments is to pick turning points in the underlying rate of inflation, not to wave away the problem of rising prices. And there is, at the moment, very good reason to put aside rents and used vehicles.

The Manheim used-car price index, which is based on wholesale car auction data, closely tracks the CPI used cars and trucks component most of the time. But an unusually large demand pop in January was captured first by Manheim in February and March, and is only now showing up in CPI. This is why CPI used cars has shot up 4.4 per cent in each of the past two reports.

But as more car inventory has come to market, the Manheim index has fallen fast. Soon enough, CPI used cars should follow. To illustrate, the chart below shows Manheim shifted forward one month:

The shelter story is a familiar one by now: CPI captures both new and existing leases and takes time, perhaps nine to 12 months, to reflect real-time market conditions. May’s report was the third in a row to show a markedly slower pace of rent inflation (don’t be fooled by the superficial jump in CPI shelter, which reflects volatile hotel prices, not rents). The chart below, showing the year-over-year trend in CPI rent and the more timely Zillow rent index, suggests that there is more good news to come:

There is, however, a warning in the Zillow new rentals index above. If you look at the month-to-month numbers, Zillow comes down sharply at first but has more recently re-accelerated:

It’s not that CPI rental inflation is going to soar again, but rather that there’s a limit to how low inflation can fall on its own, without growth falling first. Say CPI shelter stabilises around a 4-5 per cent annual rate (versus today’s roughly 6 per cent rate), which the Zillow index suggests is plausible. That’s still higher than the long-run 3 per cent average, and too high for the Fed.

The point goes beyond shelter. Unless conventional economic models are simply wrong (possible!), strong growth sets a floor for inflation. This is because falling inflation raises consumer purchasing power, sustaining the spending that lets companies pass along price increases. As Neil Dutta of Renaissance Macro put it to us yesterday:

If your landlord says ‘I need to cut your rent to keep you in this unit’, you just basically got a tax cut. In other words, all the disinflation I think we’re likely to see is akin to a tax cut for households.

I keep coming back to this idea. Ultimately, price inflation is likely to slow more rapidly than wage inflation. Then you have an increase in real wages. What does that do? It’s good for growth. What does that mean? It’s going to be good for demand. If it’s good for demand, it’ll at a minimum keep companies from cutting prices.

This raises the prospect that inflation will fall more on its own — but not all the way to the Fed’s 2 per cent target. To get to 2, growth probably has to weaken. So how is the Fed supposed to feel about inflation that is running at, say, 3.2 per cent? How about 2.9? Should central bankers sacrifice growth, or their hard targets? (Ethan Wu)

Falling earnings

Here is a chart that might worry you, if you were an anxious type:

That’s the annual change in nominal gross domestic product growth and S&P 500 earnings growth. I made the scales different for the two series to make the relationship between them more legible, despite earnings being much more volatile. The zero line is the same for both and earnings growth, the blue line, has just crossed it. The spooky thing about this is that past recessions (the shaded areas in the chart) have been preceded by earnings going negative.

There are, however, two arguments against taking the recent fall in earnings all that seriously. First, an earnings decline is a necessary, but not a sufficient, precondition for a recession. In the mid-cycle slowdowns of the late 1990s and 2015-16, earnings breached negative territory and the economy avoided recession.

More importantly, the reason that earnings are falling now is that they were unnaturally high during the pandemic, as both demand and profit margins leapt. In the two decades before the pandemic, earnings growth compounded at about 6.4 per cent (indeed, over whatever multi-decade period you choose, earnings rise at roughly that pace). Even after falling in recent quarters, earnings are still ahead of that pace over the past three years:

There are some companies that saw demand pulled forward for their products during the pandemic, and their sales and earnings are going to be hit hard (I wouldn’t want to be selling RVs right now). This need not be an economy-wide phenomenon, though. Yes, the chart shows that in the past, periods of above-trend earnings are followed by a nasty period of below-trend earnings. But this is just the economic cycle; earnings peak just before the cycle turns. And it’s not clear that what we have had in the past few years is a cycle at all. It may have been an anomalous profit boom financed by high fiscal spending, and the next phase is a reversion to the long-term trend, rather than a crash below it.

One reason for worry is that a decline in sales, margins and earnings — even from an anomalous and unsustainable high — might be self-perpetuating. When companies see profits fall, they cut jobs. A weaker job market not only reduces aggregate demand directly; it scares everyone, too. This leads to still lower demand and profits, and so on. But I’m not at all sure something like this has to happen this time around. But it’s worth thinking about.

One good read

It’s the summer of pork, people. Armstrong does a nice all-day pork shoulder on the Weber grill; email for the recipe.