Preprint.
ISSN/ISBN: Not available at this time. DOI: Not available at this time.
Abstract: Benford’s law is widely used as a diagnostic tool in applied contexts, relying on the distribution of leading digits to assess conformity with expected logarithmic frequencies. While its theoretical properties are well understood for scale-invariant mechanisms and exponential-type models, less attention has been devoted to the combined role of distributional asymmetry and tail behaviour in data-generating processes. This paper studies Benford compliance under the Skewed Generalized Error Distribution (SGED). We assess conformity using classical goodness-of-fit measures (chi-square and MAD) and a multivariate inferential strategy based on the Non-Parametric Combination (NPC) methodology. Monte Carlo evidence illustrates how skewness and tails interact in shaping digit frequencies and highlights systematic differences between univariate diagnostics and NPC-based global inference. The results support a model-aware interpretation of digit laws and provide practical guidance for rigorous Benford assessment in skewed and heavy-tailed environments.
Bibtex:
@misc{,
author = {Gianfranco Piscopo and Valerio Ficcadenti and Massimiliano Giacalone and
Maria Longobardi},
title= {Permutation-Based Inference for Benford’s Law under Skewed Generalized Error Distributions},
year = {2026},
url = {https://researchportal.lsbu.ac.uk/ws/portalfiles/portal/22004330/STPA_-_PIFIGILO.pdf},
}
Reference Type: Preprint
Subject Area(s): Statistics