Abstract
First and higher order digits in data sets of natural and socio-economic processes often follow a distribution called Benford’s law. This phenomenon has been used in business and scientific applications, especially in fraud detection for financial data. In this paper, we analyse whether Benford’s law holds in economic research and forecasting. First, we examine the distribution of regression coefficients and standard errors in research papers, published in Empirica and Applied Economics Letters. Second, we analyse forecasts of GDP growth and CPI inflation in Germany, published in Consensus Forecasts. There are two main findings: The relative frequencies of the first and second digits in economic research are broadly consistent with Benford’s law. In sharp contrast, the second digits of Consensus Forecasts exhibit a massive excess of zeros and fives, raising doubts on their information content.
Similar content being viewed by others
Notes
Even if that is the case, attempts to replicate the studies mostly fail. McCullough et al. (2006) analysed more than 150 articles from the Journal of Money, Credit, and Banking, but were able to reproduce the results in less than 10 percent of the cases.
For specific digits, e.g. whether there is an excess of fives, the standard normal statistic \( {\text{T}}_{\text{d}} = {{\sqrt {{\text{N}}_{\text{d}} } \left( {{\text{h}}_{\text{d}} - {\text{p}}_{\text{d}} } \right)} \mathord{\left/ {\vphantom {{\sqrt {{\text{N}}_{\text{d}} } \left( {{\text{h}}_{\text{d}} - {\text{p}}_{\text{d}} } \right)} {\sqrt {{\text{p}}_{\text{d}} (1 - {\text{p}}_{\text{d}} )} }}} \right. \kern-\nulldelimiterspace} {\sqrt {{\text{p}}_{\text{d}} (1 - {\text{p}}_{\text{d}} )} }} \) can be used to check whether the observed frequency significantly deviates from its theoretical value.
Recently, Tam Cho and Gaines (2007) proposed the Euclidean distance as a measure to characterize the deviation from the Benford distribution. This measure is independent of the sample size, however, it is lacking a statistical foundation.
The results for third digits have been evaluated as well (overall showing a very good agreement with Benford’s law) but are not reported due to space limitations. An analysis of higher-order digits (which are more likely to be uniformly distributed) is impeded by insufficient digits in most published papers.
For both journals also the possible sequences of the first and second digits (e.g. 14, 73, 86) have been analysed. The results, which are not reported here, show no clear pattern, neither regarding the tendencies of tests (which rejects more often) nor the effects of sample size.
Based on an analysis of the realised CPI growth rates for Germany (10/1989–07/1994).
The Chi2 statistic for the second digits is 11.61 (p-value: 0.24) and for the “first digit after the decimal point” 5.61 (p-value: 0.82).
One promising approach has been proposed by Berlemann and Nelson (2005). They introduce a small-scale experimental stock market which yields the (mean) forecast of inflation rate as well as a likelihood measure for different inflation scenarios. The main idea is to use the market as the best instrument to uncover and aggregate private information.
References
Batchelor R (2001) How useful are the forecasts of intergovernmental agencies? The IMF and OECD versus the consensus. Appl Econom 33:225–235
Benford F (1938) The law of anomalous numbers. Proc Am Philos Soc 78:551–572
Berlemann M, Nelson F (2005) Forecasting inflation via experimental stock markets: some results from pilot markets. Ifo Working Paper No. 10
Camerer CF (2003) Behavioural game theory: experiments in strategic interaction. Russell Sage Foundation and Princeton University Press, New York, NY
Carslaw C (1988) Anomalies in income numbers: Evidence of goal oriented behaviour. Account Rev 63:321–327
De Ceuster MJK, Dhaene G, Schatteman T (1998) On the hypothesis of psychological barriers in stock markets and Benford’s Law. J Empir Finance 5:263–267
Diekmann A (2007) Not the first digit! Using Benford’s law to detect fraudulent scientific data. J Appl Stat 34:321–329
Dovern J, Weisser J (2007) Survey expectations in G7 countries: professional forecasts of macroeconomic variables from the consensus data set. The Kiel Institute for the World Economy, Mimeo
Durtschi C, Hillison W, Pacini C (2004) The effective use of Benford’s law to assist in detecting fraud in accounting data. J Forensic Account 5:17–34
Gallo GM, Granger CWJ, Jeon Y (2002) Copycats and common swings: the impact of the use of forecasts in information sets. IMF Staff Pap 49:4–21
Giles DE (2007) Benford’s law and naturally occurring prices in certain ebaY auctions. Appl Econ Lett 14:157–161
Hamermesh DS (2007) Viewpoint: replication in economics. Can J Econ 40(3):715–733
Harvey DI, Leybourne SJ, Newbold P (2001) Analysis of a panel of UK macroeconomic forecasts. Econom J 4:37–55
Hendry DF, Clements MP (2004) Pooling of forecasts. Econom J 7:1–31
Hill TP (1995) A statistical derivation of the significant-digit law. Stat Sci 10:354–363
Hill TP (1998) The first digit phenomenon. Am Sci 86:358–363
Isiklar G, Lahiri K (2007) How far ahead can we forecast? Evidence from cross-country surveys. Int J Forecast 23:167–187
Judge G, Schechter L (2007) Detecting problems in survey data using Benford’s Law, November 1, Working Paper. University of California and University of Wisconsin
Kuiper NH (1959) Alternative proof of a theorem of Birnbaum and Pyke. Ann Math Statis 30:251–252
Leamer E (1978) Specification searches ad hoc inference with nonexperimental data. John Wiley & Sons, Inc., New York
Ley E (1996) On the peculiar distribution of the U.S. stock indexes’ digits. Am Stat 50:311–313
McCullough BD, Vinod HD (2003) Verifying the solution from a nonlinear solver: a case study. Am Econ Rev 93:873–892
McCullough BD, McGeary KA, Harrison TD (2006) Lessons from the JMCB archive. J Money Credit Bank 38(4):1093–1107
Mochty L (2002) Die Aufdeckung von Manipulationen im Rechnungswesen–Was leistet das Benford’s Law? Die Wirtschaftsprüfung 14:725–736
Newcomb S (1881) Note on the frequency of use of the different digits in natural numbers. Am J Math 4:39–40
Nigrini MJ (1996a) A taxpayer compliance application of Benford’s law. J Am Taxpayer Assoc 18:72–91
Nigrini MJ (1996b) Using digital frequencies to detect fraud. The White Paper (April/May) 3–6
Nigrini MJ (1999) Adding value with digital analysis. Intern Auditor 56:21–23
Niskanen J, Keloharju M (2000) Earnings cosmetics in a tax-driven accounting environment: evidence from Finnish public firms. Eur Account Rev 9:443–452
Osterloh S (2008) Accuracy and properties of German business cycle forecasts. Appl Econ Q 54(1):27–57
Pinkham RS (1961) On the distribution of first significant digits. Ann Math Stat 32:1223–1230
Quick R, Wolz M (2003) Benford’s law in deutschen Rechnungslegungsdaten. Betriebswirtschaftliche Forschung und Praxis 208–224
Reulecke A-K (2006) Fälschungen – Zu Autorschaft und Beweis in Wissenschaften und Künsten. Eine Einleitung. In Reulecke A-K, Fälschungen. Suhrkamp Verlag, Frankfurt am Main, pp 7–43
Roberts CJ, Stanley TD (2005) Meta-regression analysis: issues of publication bias in economics. Blackwell Publishing, Oxford, UK
Schatte P (1988) On mantissa distributions in computing and Benford’s law. J Inf Process Cybern 24:443–455
Schäfer C, Schräpler JP, Müller KR, Wagner GG (2005) Automatic identification of faked and fraudulent interviews in the German SOEP. Schmollers Jahrbuch—J Appl Soc Sci Stud 125:119–129
Schräpler JP, Wagner GG (2005) Characteristics and impact of faked interviews in surveys. All Stat Arch 89:7–20
Tam Cho WK, Gaines BJ (2007) Braking the (Benford) law: statistical fraud detection in campaign finance. Am Stat 61(3):218–223
Thomas JK (1989) Unusual patterns in reported earnings. Account Rev 64:773–787
Tödter K-H (2007) Das Benford-Gesetz und die Anfangsziffern von Aktienkursen. Wirtschaftswissenschaftliches Studium 36(2):93–97
Van Caneghem T (2002) Earnings management induced by cognitive reference points. British Account Rev 34:167–178
Varian H (1972) Benford’s law. Am Stat 23:65–66
Author information
Authors and Affiliations
Corresponding author
Additional information
The paper was written while the first author was a visiting researcher at the Research Centre of the Deutsche Bundesbank. The views expressed in this paper are those of the authors and do not necessarily reflect those of the Deutsche Bundesbank. We thank two anonymous referees and the editor for helpful comments and suggestions.
Rights and permissions
About this article
Cite this article
Günnel, S., Tödter, KH. Does Benford’s Law hold in economic research and forecasting?. Empirica 36, 273–292 (2009). https://doi.org/10.1007/s10663-008-9084-1
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10663-008-9084-1
Keywords
- Benford’s Law
- Fraud detection
- Regression coefficients
- Standard errors
- Growth and inflation forecasts
- Rounding