of July 2008, data ROSCs (and updates) containing data quality ratings have been published for some 80 IMF member countries. 1 2 This paper examines the usefulness of Benford’s law to ascertain overall data quality. To this end, it revisits and expands the work by Nye and Moul (2007) who tested certain data for conformity with Benford’s law, focusing on the gross domestic product (GDP) series for members of the Organization for Economic Cooperation and Development (OECD) and African countries. Benford’s law, also called the first digit law, refers to an
Front Matter Page Statistics Department Authorized for distribution by J.R. Rosales Contents I. Introduction II. Benford’s Law and Testing Methodology A. Benford’s Law B. Testing Methodology III. Conformity of Macroeconomic Data with Benford’s Law IV. Comparison of Goodness of Fit Tests Results with ROSC Assessments of Data Quality V. Can Tests of Conformity with Benford’s Law be Considered Reliable Indicators of Macroeconomic Data Quality ? VI. Conclusion Appendix: Country Groups and Data Series Used References Text Tables
Policy in Singapore Eskesen, Leif Lybecker No. 09/9 Distress in European Banks: An Analysis Based on a New Dataset Poghosyan, Tigran; Cihák, Martin No. 09/10 Benford’s Law and Macroeconomic Data Quality Gonzalez-Garcia, Jesus; Pastor, Gonzalo C. No. 09/11 How Can Burundi Raise Its Growth Rate? The Impact of Civil Conflicts and State Intervention on Burundi’s Growth Performance Basdevant, Olivier No. 09/12 The International Diversification Puzzle When Goods Prices Are Sticky: It’s Really about Exchange-Rate Hedging
reported deficit but did not improve public finances in a broader sense. In three countries, the dubious transactions they identify (they do not attempt to specify which involve off-balance-sheet assets and liabilities) averaged more than half a percent of GDP. The transactions were more likely to occur when the deficit rule was in danger of being breached. There is also some suggestive evidence of a different kind of problem. Auditors looking for made-up numbers in company accounts assess the extent to which the numbers in the accounts deviate from Benford’s law