Integrity review: Tariffs and Goods-Market Search Frictions (IFDP No. 1437, 2026)

Academic fraud report · Geng Detector

Summary

Verdict: No evidence of academic misconduct. This is a theoretical macro/international-trade paper (DSGE-style model with search frictions) by Federal Reserve Board staff, containing only formulas and code-generated figures. All six 'Geng-style' forensic checks were inapplicable or returned clean. Reference timeline is internally consistent (2025–2026 working papers). Calibration in Table 2 reproduces selected targets exactly (e.g., exporter failure rate 40% vs 40%) because those are calibration targets, while other moments (US consumption 12,727 vs 9,673; US GDP 18,707 vs 15,708; CH capacity utilization 74% vs 84%) show substantial gaps, indicating the authors did not over-fit. Parameter choices (β=0.50, σ=4.00, η=0.33) follow standard literature conventions. The only artifact is garbled glyphs in figure captions, attributable to PDF text-extraction encoding rather than tampering. Confidence in the clean verdict is high; limits include inability to audit replication code or raw calibration data files.

Verdict

No misconduct detected. The paper is a standard macro/trade theoretical contribution with no image-based, statistical, or referencing irregularities. The 'Geng six-style' forensic checks did not trigger any fraud indicators.

Key findings

Image duplication (Check 1): Not applicable — paper contains no biological or microscopy images, only formulas and vector figures.
Photoshop artefacts (Check 3): Not applicable — all figures are code-generated.
p-hacking / statistical anomalies (Check 4): Not applicable — no sample-based inferential statistics; analysis is model-based calibration.
Reference timeline (Check 6): Consistent. Paper dated April 7, 2026; cites Eaton, Jinkins, Tybout & Xu (2026) and Alviarez et al. (2025a, 2025b), Flaaen et al. (2025) within plausible date range.
Calibration fit (Check 2 variant): Mixed, and therefore credible. Selected moments match exactly by construction (e.g., Log-log imports distance −0.9 vs −0.9; US firms exporting to CH 6% vs 6%; US–CH exporter failure rate 40% vs 40%), while other moments show large deviations (US consumption 12,727 vs 9,673; US GDP 18,707 vs 15,708; CH mfg. capacity utilization 74% vs 84%).
Parameter plausibility (Check 5 variant): Standard literature values (β=0.50, σ=4.00, η=0.33); no suspicious over-fitted constants.

Evidence highlights

DOI: 10.17016/IFDP.2026.1437 — Federal Reserve International Finance Discussion Paper No. 1437, "Tariffs and Goods-Market Search Frictions," Pawel M. Krolikowski & Andrew H. McCallum, April 2026.
Table 2 (Page 41) shows both close matches (calibration targets) and sizeable misspecifications, consistent with honest reporting.
Table 1 parameters align with canonical values used in search-and-matching trade literature.
Figure caption text contains garbled control characters, typical of LaTeX/PDF text-extraction artefacts; no evidence of deliberate obfuscation.

Notes

Discipline mismatch means most image-manipulation heuristics are inert here; the assessment relies on logical, parametric, and reporting consistency rather than pixel forensics.
Replication code and underlying calibration files were not independently executed; residual risk of unreported data tuning cannot be fully excluded without the authors' replication package.
The garbled figure captions should be corrected by the publisher or in a revised posting, but do not constitute an integrity issue.