Jason Furman (@jasonfurman) header
Jason Furman (@jasonfurman) profile picture

Professor @Harvard, Director @HarvardBizGov, Senior Fellow @PIIE, teaches Ec10, words @nytopinion, proprietor https://nitter.gq/t.co/4eOTjrd0x4, was Chair of Pres Obama's CEA.

Cambridge, MA
Joined April 2011
Photo by Jason Furman (@jasonfurman)Photo by Jason Furman (@jasonfurman)Photo by Jason Furman (@jasonfurman)Photo by Jason Furman (@jasonfurman)Photo by Jason Furman (@jasonfurman)Photo by Jason Furman (@jasonfurman)Photo by Jason Furman (@jasonfurman)Photo by Jason Furman (@jasonfurman)Photo by Jason Furman (@jasonfurman)Photo by Jason Furman (@jasonfurman)
I'm thrilled to be launching Econvitals today. It's been the main way in which I have been tracking and understanding the economy over the last several months. And I decided I might as well share it with all of you--for free. Has a combo of website, email lists & social media.
48
230
58
1,560
258,382
Jason Furman retweeted
🧵on the AI audit of Saez and Zucman and by implication Piketty (or PSZ together): Since some are objecting to what the computer found, here's a list of biases in this trio's work that were found by humans over the last decade.
4
39
11
144
19,241
1. AI should play a large role before we see anything from anyone so it comes out right in the first place. 2. AI should play a big role in retrospective replications. 3. Piketty-Saez-Zucman work has many flaws that all go one direction. Spreading slop hurts all three causes.
3
5
1
31
3,640
Did no one check any of this before launching the Revolution?
An independent, AI-powered replication audit of the Declaration of Independence (1776) Oh man. You guys. This is pretty serious. We may have made a major policy error based on bad data.
2
1
10
2,909
Jason Furman retweeted
Markets after the GDP and PCE release, by 11:05 ET: 2-year yield -8 bp, 10-year yield -6 bp, dollar index -0.2%, S&P 500 futures -0.1%. Fed funds priced for after Dec 9: -2 bp, from 4.11% to 4.09%, or +22 bp vs the 3.875% target midpoint.
1
1
3
2,357
Best roundup on this issue I've seen:
Not sure what is more striking—how bad the research audited here seems to be, or the thought that a lot of other research might come out similarly shaky. Via @tylercowen / @MargRev – saez-zucman-audit-replicatio…
5
3
34
8,883
BEA has quarterly GDP data back to 1947. Of the 318 quarters it tracks this was the third most negative contribution from net exports (and the second largest was 2025-Q1). That is, much of the consumption, investment, etc was imported--so take 3.3pp of GDP off what those added.
5
4
2
30
7,303
And consumer spending booming, the biggest additions are services--but goods strong too.
1
1
2,137
Normally I think of final sales to domestic private purchasers as a good measure of "core GDP". It excludes the large subtraction from net exports, just looking at what people and businesses spend on consumption and investment. A bit less sure it is right now. We'll see.
3
2,086
Very strong GDP numbers, a touch higher than expected. More later from me, for now you can read @econvitals
↗️ Q3 2026 GDP, advance estimate: real GDP +3.4% annualized (consensus +3.2%, prior +2.2%). Real final sales to private domestic purchasers +4.9% (prior +4.6%). Core PCE prices +2.5% (consensus +2.4%).
1
4
32
5,394
Inflation came in pretty much as expected which is, well, quite high. In Econvitals charts the solid dots are compared to what people thought yesterday, those are all close to zero. The hollow circles are absolute, those are all quite high.
↘️ September PCE: prices +0.49% (consensus +0.51%), +3.6% over the past year (+3.6% expected). Core +0.32% (consensus +0.34%), +3.1% y/y (+3.1% expected). Personal income +0.2% (+0.3% expected), spending +0.5% (+0.6% expected). Saving rate 3.8%.
4
2
1
31
5,610
The PCE-based Ecumenical Underlying Inflation measure for September was 3.0%. Here are the different pieces that go into it, the median of a number of different inflation series, all re-meaned to match the mean of the PCE the Fed targets.
1
12
2,869
One silver line some might take from this: Trimmed mean (which Chair Warsh has cited) was more moderate. Dallas will have the official numbers later but my replications are almost exactly the same so enjoy it now.
2
11
2,595
If you believe that mindless AI evaluations are legitimate then why would you spread the evaluation of Saez & @gabriel_zucman without first subjecting it to its own mindless AI evaluation? I just did and it turns out the critic's errors all go in one direction.
30
73
39
657
98,358
Finally, that headline 97% of the errors go in one direction? Well, that was guaranteed by the method it used. Here is the anonymous site: saez-zucman-audit-replicatio…
1
4
1
81
10,095
I put little effort into this--a short prompt (basically evaluate this evaluation) & some judgmental spot checking of the results. Not the right way to do things. But is a case of two wrongs make a, well, nothing. Back to where we started before the new criticism was published.
6
113
7,375