The Feb core PCE forecast for 2024-Q1 was 2.1%, or "mission accomplished." The actual was 3.7%, a red warning sign.

Full history from Survey of Professional Forecasters forecasts/actual for 2024-Q1:

Feb-23: 2.5%
May-23: 2.5%
Aug-23: 2.6%
Nov-23: 2.7%
Feb-24: 2.1%
Actual: 3.7% The 1.6pp forecasting error two months ahead is larger than all but one forecasting error from when these forecast data start (in 2007) and COVID. And the previous similar-sized (but opposite sign) error was 2008-Q4.

GDP growth came in a bit below expectations at a 1.6% annual rate in the first quarter.

But much of the slowdown was in non-inertial items like inventories (-0.35pp) and net exports (-0.86pp). The better signal of final sales to private domestic purchasers was 3.1%. Consumer spending was strong, up at a 2.5% annual rate. Business fixed investment was a touch on the weak side, up at a 2.9% annual rate with business structures down. But residential investment was very strong, 13.9% annual rate.

If the Fed cuts rates by Sep or possibly even this yr, it's much more likely because they get bad news about the employment side of their mandate than because they get reassuring news about the inflation side.

The reason: even if inflation falls will take time to be convincing. A bit of math helps. Suppose core PCE is 0.3% in March & 0.157% each month after (consistent with 1.9% annual rate). The 6-month / 12-month annualized rates would be:

April: 2.9% / 2.6%
May: 3.0% / 2.5%
June: 3.0% / 2.4%
July: 2.4% / 2.5%
Aug: 2.2% / 2.5%
Sep: 1.9% / 2.4%

My story for what has happened to inflation and what it means going forward.

TL;DR: Underlying inflation fell from 4.0-4.5% to 2.5-3.0% as labor markets loosened (through openings down not unemployment up). Last mile will be much harder than to date.

A shortish 🧵 Let me state at the outset, this is a story. Other stories that fit the data too. I'm not sure about this. And even if I had the model right, there are still unexpected shocks.

Plus no story fits the very high frequency data which has lots of noise & measurement error.

Core CPI coming in very hot for the third month in the row. The numbers are not kind to the thesis that January was a seasonal anomaly.

12 months: 3.5%
6 months: 3.2%
3 months: 4.6%
1 month: 4.6% The bottom-up forecasts of inflation coming down were premised on much larger reductions in shelter growth than we've seen. It has, as widely expected, come down from its 9% annual rate high but plateaued around 5.5% ar.

We know what the Fed meant by "transitory" in 2021 because the SEP produced explicit forecasts. These forecasts had three properties:

1. Rapid return to 2% inflation

2. Period of inflation < 2% (not just asymptote down to 2%)

3. Happens without any increase in fed funds rate 1 & 3 have been discussed a lot. But 2 was also a feature--basically saying we would have mean reversion in the price level not the inflation rate. (E.g., 3% inflation followed by a period of 1% like you might see with an oil shock, not 3% falling to 2%).

https://x.com/jasonfurman/status/1463531410535464965?s=20

For the 3rd time in the last 9 months core PCE inflation came in well above the Fed's target (albeit a touch below expectations). Annual rates:

1 month: 3.2%
3 months: 3.5%
6 months: 2.9%
12 months: 2.8% You see a similar pattern with various ways of removing house--core ex housing, core ex housing and used cars and core serves ex housing and used cars.

All of which is because of the resurgence of core services ex housing.

GDP revised up to a 3.4% annual rate. GDI came in even higher, at 4.8% (first time >GDP since 2022-Q2). Led by strong consumption, non-residential structure investment and government purchases. My preferred signal of PDFP (which is C + non-inventory I) up 3.3%. Strong.

Here is what the different ways to measure the level of output look like compared to CBO's pre-COVID projections.

Note we're basically at those projections because labor is much higher than they expected which more than offsets productivity being slightly weaker than expected.

Core CPI came in high for the eighth month in a row.

Annual rates:

1 month: 4.4%
3 months: 4.2%
6 months: 3.9%
12 months: 3.8%

For perspective, the 3/6/12 month rates higher than any time from 1992-2019. Inflation remains unusually high. Shelter continues to drive inflation up. It moderated from its extreme pace in January (which was a widely understood quirk), but is not moderating to nearly the extent the optimists had hoped.

A soft landing jobs report. Tilts the balance of worry ever so slightly away from inflation and towards recession. But overall things still looking good.

275K jobs added. Hours up. But unemployment rate from 3.7% to 3.9% while participation flat. Wage growth only 0.1%. Note that even while job growth has been very strong the general trend in hours has been down.

A meticulous examination by @WendyEdelberg & @taraelizwatson of a huge issue in interpreting the recent past and near future of the macroeconomy: immigration.

It's tricky because of huge differences in estimates of immigration (from 1.1-3.3m for 2023).

hamiltonproject.org/wp-content/upl…
They use the higher (better?) CBO estimate. It explains why there are 2.7m more jobs than CBO had expected pre-pandemic (using the establishment measure).

With higher immigration steady-state job growth was 160-230K instead of 60-130K. Economy was less hot than some thought.

So much of what people think about China's economy, trends in manufacturing jobs, and global surpluses/deficits is both factually confused and economically confused. This excellent @BaldwinRE piece confines itself to some of the factual issues. linkedin.com/pulse/fact-che… 1. Exports were component of China's economy. But as it has grown relative to the global economy it could not keep expanding exports at the same rate--no one could have absorbed them. So now its exports are about the same share of global production as US exports.

Core PCE inflation picked up a lot in January. Annual rates:

1 month: 5.1%
3 months: 2.6%
6 months: 2.5%
12 months: 2.8%

Not a surprise relative to what we thought yesterday. But high relative to what might have been hoped a month ago. There was no one story behind this. Numbers picked up a lot with different treatment of houses, goods, used cars and the like.

Inflation was wacky in Jan (possibly due to seasonal adjustment). Big increases in services like medical, transportation and shelter. In fact core services increase was at the 99th percentile.

Big goods falls like used cars and apparel. 97th percentile of decline for core goods. This is a continuation of the recent pattern--but more extreme.

Bottom-up optimists think the services increase was a quirk & will mean revert (e.g., shelter coming down).

Bottom-up pessimists think the goods declines have been unusual & will mean revert (less goods deflation).

Core CPI comes in hotter than expected, 0.4% in the month of January--which is a 4.8% annual rate.

I'm not a big fan of second derivative forecasting but those of you who are should be worried. Annual rates:

12 months: 3.9%
6 months: 3.6%
3 months: 4.0%
1 month: 4.8% Overall inflation also picked up but was below core as already (relatively) high food prices rose further but already low energy prices fell for the fourth straight month.

A blockbuster jobs number--with the caveat I'll come back to about hours being down.

353K jobs added in January. Upward revision to 333K in December. Yes one worries about seasonals/noise but still a 289K monthly pace for 3 months.

Urate unchanged at 3.7%.

Wages up A LOT. Average hourly earnings were up 0.6% in January, which is a 6.8% annual rate. That is basically tied for the highest month in the pre-COVID data (data starts in 2006). 3 month average is 5.4%.

that is what you would expect in an economy with ~4.5% inflation.

But...

What should the Fed do now? I think two mistaken approaches:

Dovish fallacy: Inflation is fine so why can't rates be like they were in 2019?

Well inflation is fine in large part because rates are high. Hawkish fallacy: Inflation could reemerge. Plus growth is really strong. So should not cut rates.

Well, none of that is an argument for raising real rates, which is what has happened with the decline in inflation. Plus inflation is genuinely lower.

To my surprise & delight we have landed softly. Risks we veer off the runway into resurgent inflation or recession are real but balanced & not unusually high. Fed deserves a lot of credit & should start cutting soon.

A 🧵w/ my @aei comments yesterday. aei.org/events/soft-la… The slides from my discussion are here. Will summarize in this thread. The panel also had great comments by @jc_econ, Nathan Sheets, Desmond Lachlan and was really well framed and organized by @steven_kamin. So would watch it all. dropbox.com/scl/fi/nj512ds…

The last four monthly core CPI prints were (annual rates) 3.4%, 3.9%, 2.8% and 3.5%.

So why isn't the Fed more worried about inflation?

It's because they target a different measure of inflation, the PCE, and for core PCE the last prints were 1.3%, 3.8%, 2.0% & forecast 1.8%. 3 big reasons for difference:

1. Shelter has a smaller weight in the PCE than the CPI. Even though shelter growth has fallen is still > than non-shelter.

2. CPI medical services growth high due to lags/quirks, PCE uses different data/methodology.

3. CPI generally runs higher.

Strong jobs report across the board:

--199k job (above expectations)
--Unemployment rate down & participation up
--Hours up
--Strong wage growth We're now 2 million jobs and 2 million employed above CBO's pre-pandemic projections.

What are the chances that inflation reignites? I think about 25%. This short 🧵explains.

First, I'll define "reignite" as core PCE inflation > 3.0% in 2024. I respect different definitions.

This chart is one model of possibilities in 2024, explained in the next tweet. The previous based on regressing inflation for t+2 to t+14 on inflation over t-6 to t using data since 1959. For the current moment that amounts to using the 2.5% inflation over the last 6 months to predict inflation in 2024. The plot is based on actual errors which are skewed.