Geng Integrity Report — Bray & Mendelson (2012), 'Information Transmission and the Bullwhip Effect: An Empirical Investigation', Management Science

Academic fraud report · Geng Detector

Summary

Verdict: cleared (no fraud indicators detected). DOI 10.1287/mnsc.1110.1467. The Geng review examined Bray and Mendelson's Management Science paper (submitted 11 June 2010, accepted 16 September 2011, online 9 March 2012) for image reuse, fabricated data, statistical/logical inconsistencies, and timeline anomalies, and found none. Figures 1–4 contain only standard economics charts (line plots, box plots, density contours) with no duplicated panels. Table 2 (mean bullwhips across Retail, Wholesale, etc.) shows heterogeneous means and standard errors, including non-significant sectors, consistent with real firm-level data. Table 4 robustness checks yield plausible parameter shifts (e.g., β0 moving from 9.98 to 7.76, 7.09, 5.73) while preserving sign and significance — a pattern hard to fabricate by hand. The reported retail bullwhip of −3.23 (insignificant), the '65% of firms bullwhip' figure, and Table 5's 0.65 fraction are mutually consistent. COMPUSTAT data (1974–2008) is appropriate for a 2010 submission. Limitations: this assessment is based on the manuscript text; raw data and code were not independently re-run.

Verdict

Cleared. No fraud indicators detected across all six checks.

Key findings

Figures 1–4 are standard statistical charts (line, box, density contours); no image reuse detected.
Table 2 shows realistic cross-sector heterogeneity with both significant and non-significant bullwhip estimates.
Table 4 robustness checks produce plausible parameter variation (β0: 9.98 → 7.76, 7.09, 5.73) consistent with actual re-estimation, not hand-fabrication.
Reported retail bullwhip (β = −3.23, insignificant), '65% of firms bullwhip', and Table 5's fraction (0.65) are internally consistent.
Timeline (submitted 2010, accepted 2011, published 2012) and COMPUSTAT data window (1974–2008) are coherent; no temporal anomalies in citations.

Evidence highlights

Paper title: 'Information Transmission and the Bullwhip Effect: An Empirical Investigation'
Authors: Robert L. Bray, Haim Mendelson
Journal/Year: Management Science, 2012
DOI: 10.1287/mnsc.1110.1467
Submission: 2010-06-11; Acceptance: 2011-09-16; Online: 2012-03-09
Data source: COMPUSTAT, 1974–2008
Table 2 sectoral bullwhip means exhibit varied standard errors and significance markers.
Table 4 parameter β0 shifts across alternative specifications: 9.98, 7.76, 7.09, 5.73, retaining sign and significance.
Retail segment total bullwhip estimate: β = −3.23 (non-significant).
'65% of firms bullwhip' claim matches Table 5 fraction of firms = 0.65.

Notes

Confidence is moderate-to-high based on textual analysis; raw datasets and estimation code were not re-executed.
Non-image, statistics-based empirical papers are inherently harder to falsify from the manuscript alone.
Data source (COMPUSTAT) is a publicly licensed standard database, reducing reproducibility concerns about proprietary inputs.