Hmm. An interesting data puzzle. Jason Furman — a very serious guy — has been arguing that US inflation has accelerated much more than Euro area inflation, suggesting a strong role for fiscal policy. I've been accepting that analysis. But ... 1/
Menzie Chinn, also a very serious guy, working with much the same data, finds that the rise in relative US inflation is fairly small and not statistically significant 2/ http://econbrowser.com/archives/2022/04/how-much-faster-has-american-inflation-been-compared-to-euro-area-inflation
Fwiw, Chinn's data seem to suggest that there was a substantial divergence in mid-2021 that has largely gone away. This is core inflation, so it's not just Russian gas 3/
I've been mucking around with the data, and concluded that I don't understand it well enough to referee this difference. But it seems kind of important. Guys? Others? What's going on here? 4/



A broader post on the US-Euro area inflation differential numbers with commentary on points raised by @menzie_chinn that were flagged by @paulkrugman.
TL;DR: US inflation is *much* higher than euro area inflation. This chart shows it lots of different ways.
To understand the table let's look at 12 month core inflation. Is up 7.4 percent in the United States and 2.9 percent in the Euro area. That is 4.4pp higher in the US.
The US was running 0.7pp higher in the 24 months pre-COVID.
So diff-in-diff is 3.7pp--and roughly steady.
The difference in overall inflation is smaller but meaningful. The reason is that energy prices are up much more over the last 12 months in the euro area (+44) than the United States (+32). In fact, over the the last three months overall inflation is running higher in Europe.
Natural gas prices in Europe are about 4 times as high as in the United States, that is bleeding into core inflation too. Some other factors go the other way, however, like cars are a bigger part of the U.S. consumption basket.https://eia.gov/naturalgas/weekly/archivenew_ngwu/2022/04_21/#tabs-prices-2
Technical note: These estimates use the HICP measure of inflation for the US. Is calculated by the BLS to match the European concept. The main change is taking out shelter (or OER) from U.S. CPI. (A much smaller change is including rural consumption in the expenditure weights.)
Conclusion: These numbers are broadly consistent with my view that demand is playing a big role in US inflation. But U.S. demand is so high it is driving up global goods prices. And Europe has had a lot of demand too. So a differential is not necessary to believe it is demand.
P.S. Why did @menzie_chinn get such different results? I think he was looking at whether any given month was individually statistically significantly different in the US & EA. But we care about the last 12 or 25 months collectively. Best to look at the full period.
P.P.S. Here was a broader recent thread on cross-country comparisons of a range of variables, including inflation.

A 🧵with some international comparisons of G7 economic performance in the COVID recession and recovery:
1. GDP. Strong recoveries everywhere, generally stronger in the United States.




追伸:なぜチンはあのような異なる結果を得たのか? 彼はすべての所与の月それぞれにおいて米国とユーロ圏が統計的に有意に異なるかを見ていたのだと思う。しかし我々が気にするのは直近12か月や25か月全体の差である。全期間を見るのが最も良い。


Here's my interpretation http://econbrowser.com/archives/2022/04/four-questions-and-four-answers-us-and-euro-area-core-inflation or why @jasonfurman and I get different answers - I use inflation, he uses growth rates in price levels, and neither is "the" only way to go...
私の解釈は、http://econbrowser.com/archives/2022/04/four-questions-and-four-answers-us-and-euro-area-core-inflation だ。ファーマンと私が違う答えを得たのは、私はインフレを使ったのに対し、彼は価格水準の伸び率を使った。どちらかが「唯一」の正しい方法というわけではない。


Not sure I agree. If you asked which country grew faster over the last decade France or Belgium you would not ask whether the forty quarterly growth rate differences in the growth rates was statistically significant.
Among other issues is measurement error, including seasonal adjustment errors, some of which are serially correlated. You would need a complex specification for the 40 quarterly growth differences. Or you could avoid that and do the 10-year difference.
To make it more extreme, if we had daily growth rates would we do 3,650 of the daily differences? Might not show a star significant difference btwn Italy and US growth. But a decade of cumulative data would.
極端な話、仮に日次成長率があったとすれば、3650の日次の差を見るというのか? イタリアと米国の成長率でも星が付くような有意な違いは出ないだろう。しかし10年の累積データならば出る。


The Q I'm answering is: Does the inflation differential increase (with statistical significance) after the onset of the pandemic? Your calculation answers the Q: Did the differential trend growth in P levels increase w/onset of pandemic. I'm agnostic on which one's a "better" Q


But you still need to deal with the issue of correlated errors and auto correlated dynamics.
Like VAT holidays in Europe that came and went lower your T stat.
We know CPI differential was 0.1 before and 1.3 after. We know sampling errors on both. Pretty confident T stat>2.


From a quick look @menzie_chinn appears to be focusing on 3 month inflation. Natural gas prices have skyrocketed in Europe relative to the US in the last few months, driving inflation higher there over that period.
If you look over the last 12 months or since the beginning of the pandemic European inflation much higher.
E.g., over the last 12 months core inflation (measured on a roughly comparable HICP basis:
US: 6.7 percent
EA: 3.0 percent
Will look more and share tomorrow.
P.S. arguments about whether inflation is primarily demand or supply don’t rise and fall with the US-Europe inflation disparity. Europe also had a lot of monetary and fiscal support and excess saving from reduced consumption in 2020 and 2021.


*2:cf. ここ

*3:cf. BLS説明([2022/6/12]リンク先をpdfからhtmに修正)。