USMacroBETA

The Revisions Ledger

What the numbers said first — and what they became. First prints vs the value three years later, from the archived record. As of 2026-08 · live regime: Disinflation (benign family).

THE READ

First prints are drafts; this page shows the drafts against the settled record.

Payrolls — Aug first print +162k · 2 rounds + benchmark ahead · typical ±80k · in benign-family months: revised up 193 of 311
Active streak: personal income's last 8 settled observations all revised up.

Every number here is first-print vs the value three years later — the +3y yardstick — from the archived record; nothing is restated from memory.

The dossiers

Nonfarm payrolls FRED: PAYEMS · vintages: ALFRED jobs · m/m Δk latest first print: 2023-08 since 1955

Typical movement (first→settled, k jobs): med +19k · IQR [-37k, +82k] · worst 2020-03: -701k → -1,427k

The lean: revised up 482 · down 335 · unchanged 2 (of 819 since 1955)

Months like this one (regime family at the DATA month; live selector: benign — the Disinflation read):

family	n	med	share up
benign	311	+25k	62% up
stress	44	-60k	25% up
stress-adjacent	150	-2k	49% up

Greatest revisions (own p95 bar; newest 3 of 41):

period	first → settled	Δ	own pct
2021-12	+199k → +566k	+367k	p99.5
2021-11	+210k → +631k	+421k	p99.9
2021-10	+531k → +860k	+329k	p98.9

Personal income FRED: PI · vintages: ALFRED nominal · m/m % latest first print: 2023-07 since 1966

Typical movement (first→settled, %): med +0.07 · IQR [-0.10, +0.21] · worst 1993-01: +0.47 → -3.83

The lean: revised up 366 · down 218 · unchanged 106 (of 690 since 1966)

Months like this one (regime family at the DATA month; live selector: benign — the Disinflation read):

family	n	med	share up
benign	310	+0.07	54% up
stress	44	-0.14	34% up
stress-adjacent	150	+0.03	47% up

Active streak: the last 8 settled observations all revised UP.

Greatest revisions (own p95 bar; newest 3 of 35):

period	first → settled	Δ	own pct
2022-07	+0.22 → +0.94	+0.73	p97.0
2020-10	-0.65 → +0.07	+0.72	p96.8
2020-07	+0.35 → +1.03	+0.68	p96.5

Retail sales FRED: RSAFS · vintages: ALFRED nominal · m/m % latest first print: 2023-08 since 2001

Typical movement (first→settled, %): med +0.07 · IQR [-0.27, +0.37] · worst 2022-01: +3.75 → +1.82

The lean: revised up 139 · down 104 · unchanged 23 (of 266 since 2001)

Months like this one (regime family at the DATA month; live selector: benign — the Disinflation read):

family	n	med	share up
benign	158	+0.05	51% up
stress	27	+0.07	52% up
stress-adjacent	81	+0.07	54% up

Greatest revisions (own p95 bar; newest 3 of 14):

period	first → settled	Δ	own pct
2023-01	+2.96 → +4.37	+1.41	p97.7
2022-03	+0.50 → +2.10	+1.60	p98.1
2022-01	+3.75 → +1.82	-1.93	p100.0

Industrial production FRED: INDPRO · vintages: ALFRED index · m/m % latest first print: 2023-08 since 1928

Typical movement (first→settled, %): med +0.00 · IQR [-0.27, +0.34] · worst 1952-06: -5.61 → -0.84

The lean: revised up 514 · down 474 · unchanged 160 (of 1148 since 1928)

Months like this one (regime family at the DATA month; live selector: benign — the Disinflation read):

family	n	med	share up
benign	311	-0.02	43% up
stress	44	+0.05	50% up
stress-adjacent	150	+0.00	46% up

Greatest revisions (own p95 bar; newest 3 of 58):

period	first → settled	Δ	own pct
2020-04	-11.25 → -13.38	-2.13	p97.3
1975-06	+0.36 → +2.37	+2.01	p97.1
1974-12	-2.79 → -4.48	-1.69	p95.9

Core PCE (the contrast row) FRED: PCEPILFE · vintages: ALFRED prices · m/m % latest first print: 2023-07 since 2000

Typical movement (first→settled, %): med +0.02 · IQR [-0.02, +0.06] · worst 2001-09: -0.90 → -0.51

The lean: revised up 90 · down 25 · unchanged 160 (of 275 since 2000)

Months like this one (regime family at the DATA month; live selector: benign — the Disinflation read):

family	n	med	share up
benign	159	+0.02	32% up
stress	31	+0.01	26% up
stress-adjacent	85	+0.03	36% up

Greatest revisions (own p95 bar; newest 3 of 14):

period	first → settled	Δ	own pct
2022-11	+0.17 → +0.29	+0.12	p96.0
2022-10	+0.22 → +0.36	+0.14	p98.2
2022-07	+0.08 → +0.23	+0.14	p98.9

Real GDP growth FRED: GDPC1 · vintages: ALFRED real · SAAR latest first print: 2023-04 since 1992

Typical movement (first→settled, %): med -0.01 · IQR [-0.90, +0.79] · worst 2008-10: -3.80 → -8.89

The lean: revised up 58 · down 61 · unchanged 6 (of 125 since 1992)

Months like this one (regime family at the DATA month; live selector: benign — the Disinflation read):

family	n	med	share up
benign	79	+0.04	48% up
stress	insufficient history (n=9)
stress-adjacent	37	-0.17	46% up

Greatest revisions (own p95 bar; newest 3 of 7):

period	first → settled	Δ	own pct
2020-04	-32.90 → -29.86	+3.05	p97.6
2015-01	+0.25 → +3.24	+2.99	p96.0
2011-01	+1.75 → -1.29	-3.04	p96.8

Corporate profits (CPATAX) FRED: CPATAX · vintages: ALFRED nominal · % level q/q latest first print: 2023-04 since 1981

Typical movement (first→settled, %): med +0.01 · IQR [-2.67, +2.60] · worst 2009-01: +1.11 → +29.39

The lean: revised up 84 · down 83 · unchanged 3 (of 170 since 1981)

Months like this one (regime family at the DATA month; live selector: benign — the Disinflation read):

family	n	med	share up
benign	103	-0.04	49% up
stress	13	+1.55	62% up
stress-adjacent	52	-0.24	46% up

Greatest revisions (own p95 bar; newest 3 of 9):

period	first → settled	Δ	own pct
2023-01	-6.80 → +3.86	+10.66	p97.1
2021-01	-0.79 → +12.32	+13.12	p98.2
2009-01	+1.11 → +29.39	+28.28	p100.0

GDI (the gap register) FRED: GDI · vintages: ALFRED nominal · SAAR latest first print: 2023-04 since 2013

Typical movement (first→settled, %): med +0.66 · IQR [-1.23, +1.85] · worst 2023-01: +1.70 → +7.05

The lean: revised up 25 · down 16 · unchanged 0 (of 41 since 2013)

Months like this one (regime family at the DATA month; live selector: benign — the Disinflation read):

family	n	med	share up
benign	32	+0.65	62% up
stress	insufficient history (n=1)
stress-adjacent	insufficient history (n=8)

Greatest revisions (own p95 bar; newest 3 of 3):

period	first → settled	Δ	own pct
2023-01	+1.70 → +7.05	+5.35	p100.0
2020-01	-2.70 → +1.26	+3.96	p95.1
2014-01	-1.02 → +3.90	+4.92	p97.6

The great-revisions register 181 admitted at each series' own p95

period	series	first → settled	Δ (native)	own pct
2023-01	Retail sales	+2.96 → +4.37	+1.41	p97.7
2023-01	Corporate profits (CPATAX)	-6.80 → +3.86	+10.66	p97.1
2023-01	GDI (the gap register)	+1.70 → +7.05	+5.35	p100.0
2022-11	Core PCE (the contrast row)	+0.17 → +0.29	+0.12	p96.0
2022-10	Core PCE (the contrast row)	+0.22 → +0.36	+0.14	p98.2
2022-07	Personal income	+0.22 → +0.94	+0.73	p97.0
2022-07	Core PCE (the contrast row)	+0.08 → +0.23	+0.14	p98.9
2022-03	Retail sales	+0.50 → +2.10	+1.60	p98.1
2022-01	Retail sales	+3.75 → +1.82	-1.93	p100.0
2021-12	Nonfarm payrolls	+199k → +566k	+367k	p99.5

Ranked by each series' OWN percentile, never raw units across series; benchmark-class level wedges measured as SPAN, dated at incorporation.

Recognition lags — when did the data admit it

recession	real GDP	payrolls
1990	by 1992-07-30 (floor-bounded: GDPC1 archive begins 1991-12)	1990-09-07 (Aug −75k)
2001	2002-07-31 (−0.60) — 16 months late; NOMINAL GDP never admitted it at any vintage (OBS-C1-10)	2001-04-06 (Mar −86k)
2008	2009-07-31 (−0.72) — 18 months late	2008-05-02 (Apr −20k)
2020	2020-04-29 (Q1 advance −4.78) — the only real-time admission	2020-05-08 (Apr −20,500k)

First prints vs +3y settled, from the ALFRED archives + the site's own vintage store. Counts, never probabilities; descriptive record, never a forecast. Methodology →