our methodology accurately predicts growth during downturns.
recession backtests support predictive performance against annual DINA files, with acknowledged group-specific errors
Our monthly estimates track the annual tax-data-based statistics well for all groups, including for the top 1%
wage-share comparisons support annual tracking, while the QCEW calibration window remains unclear
the post-Covid period was characterized by significant wage growth at the bottom of the distribution—a break from the trend prevailing since the early 1980s—but also at the very top, with the middle of the distribution experiencing no gains
fine monthly differences lack uncertainty bands and direct monthly validation
Overall, the average pretax income of Black adults relative to white adults is lower in 2025 (45%) than in 1989 (52%).
the temporal comparison depends on survey-matched racial information rather than directly observed race in tax records
The share of wealth owned by the top 0.1% adults increased 2 points from the end of 2019 to the end of 2021, to reach 21.9%—then the highest level recorded in the post-World War II era.
constructed wealth estimates support the increase, with broader concentration trends corroborated by the Distributional Financial Accounts
Derived from the full evaluation — not a separate score.
Strengths
The resource combines public sources into monthly microfiles covering national income and wealth, extending beyond wage-only trackers and aggregate growth statistics.
↳ Introduction; Section 2.1
One- and two-year-ahead validation compares the pipeline with uniform growth and simplified rescaling. Separate recession results and discussion of tax-reform errors clarify where predictions perform differently.
↳ Section 5.3; Tables A1–A3
The paper describes program structure, licensing, execution instructions and a microfile codebook, alongside a regularly updated public website.
↳ Introduction; Appendix D
Limitations
Comprehensive backtests assess annual growth against model-derived annual DINA files. They do not directly establish the precision of fine monthly differences, and released estimates lack uncertainty bands.
↳ Sections 5.1 and 5.3; Section 6.2
The rescaling-only comparison tests the combined benefit of distributional adjustments, not the separate effects of employment adjustment, wage-source blending, CES extension and transfer modeling.
↳ Sections 4.2–4.4; Section 5.3; Table A3
The introduction describes predictions as significantly less biased, although the reported Diebold–Mariano test evaluates squared forecast errors. Broad improvement language also requires the next-9% exception disclosed in footnote 18.
↳ Introduction; Section 5.3; footnote 18
The contribution rests on integrating scattered public sources into national-account-consistent monthly distributions with demographic detail. Section 5.3 supplies historical backtests, recession analyses and relevant benchmarks, while Appendix D documents the construction and replication workflow. Those strengths do not fully resolve the mismatch between annual validation against modeled reference files and the precision needed for monthly applications. The rescaling comparison also leaves individual adjustment components only partly isolated, and broad significance wording requires the qualifications supplied elsewhere in the text.
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Strong3.9
Confidence highThe monthly microfiles integrate national-income totals, distributional information and demographic characteristics into a resource differentiated from existing trackers. The advance meets the upper competent-contribution anchor, but its validation against model-derived annual files does not establish the airtight support required for the highest band.
The documented pipeline, historical backtests, recession analyses and two benchmarks support assessment of execution. Partial component isolation, annual rather than comprehensive monthly validation, and absent uncertainty bands account for the downward movement, while the calibration-window question limits confidence.
The paper separates construction, assumptions, validation and applications in a followable sequence and explicitly qualifies important failures. Broad significance and all-concept wording is less precise than the accompanying tests and exceptions, preventing placement in the airtight band.
The discussion distinguishes the resource from the Atlanta tracker, DFA and BEA approach and explains consequential matching and tax-reform limitations. Dependence on the annual DINA reference and uncertainty in Covid-transfer assumptions receive less interpretive treatment.
Lower confidence on Methodological Rigour, Positioning — domain match limited.
Caveats4 of 4 checks
One reported value does not reconcile, but the key claims do not rest on it.
The paper supplies replication instructions and public data and code locations, but prints two different GitHub account paths without identifying a canonical repository.
Data availability: declared · Code availability: declared
32 of 32 checkable references verified
44 references in manuscript 12 have no canonical index record — counted, but not index-checkable 1 reference confirmed by manual review
References not verified automatically are resolved manually within 24 hours.
No retraction or expression of concern found in Crossref (includes Retraction Watch).
Sources: Crossref (incl. Retraction Watch) ✓
Where this paper’s evidence sits on the path from initial observation to real-world use.
A functioning website, regularly updated estimates, public microfiles and replication instructions support the Deployable readiness stop. Incorporation into official statistical production remains prospective rather than demonstrated.
AI-generated, human-governed. Something look off? Contact us to request a review.
Keep reading
Other papers Nabu has evaluated blind, starting with the same field.
Rodionov, Assylbekov · 2026 · Computer and Information Sciences
Across 30 models and 10 repeated runs, aggregate IFR-a is 0.93 ± 0.004 and aggregate IFR-i is 0.809 ± 0.009.
SupportedBüchi · 2021 · Media and communications
Individual’s digital practices arise within and shape socio-technical structural conditions, and lead to often concomitant harms and benefits.
SupportedDana R. Carney, Amy J. C. Cuddy, Andy J. Yap · 2010 · Psychology
High-power posers experienced elevations in testosterone, decreases in cortisol, and increased feelings of power and tolerance for risk
Tentative