Unpublished discussion paper.
ISSN/ISBN: Not available at this time. DOI: Not available at this time.
Abstract: Benford’s Law determines the way in which the digits that make up a ‘legitimate’ multi-digit number are distributed. A vast multi-disciplinary literature has illustrated the relevance of this law, and its usefulness in detecting fraudulent reporting. There is a small literature that investigates whether published regression results satisfy Benford’s Law. A handful of these studies relate to such results in economics journals, and they provide evidence that this law is satisfied for legitimate results in this case. This paper extends that literature through formal statistical testing for Benford’s 1st and 2nd digit distributions in all of the regression results reported in New Zealand Economic Papers, from its inception in 1966 to 2025. Both Bayesian and frequentist tests are used, with special attention being given to the issues associated with the extremely large sample sizes. Overall, the results provide substantial additional support for adherence to Benford’s Law in this particular type of data.
Bibtex:
@misc{,
author = {David E. Giles},
title = {Benford’s Law and Regression Results Published in Articles in "New Zealand Economic Papers"},
year = {2026},
url = {https://github.com/DaveGiles1949/My-Documents/blob/master/Benford%E2%80%99s%20Law%20and%20Regression%20Results%20Published%20in%20Articles%20in%20New%20Zealand%20Economic%20Papers.pdf},
}
Reference Type: Other
Subject Area(s): Economics