The Validity of international gold movement statistics

[Front Matter and Series Introduction]: Title page and series introduction for Oskar Morgenstern's study on the validity of international gold movement statistics. It outlines the purpose of the 'Special Papers in International Economics' series published by Princeton University's International Finance Section. [Table of Contents and List of Tables/Charts]: Comprehensive table of contents and lists of tables and charts included in the paper. The structure indicates a focus on testing the validity of gold statistics through sample tests, hypothesis formulation, and exploration of discrepancies between national reports. [Author Biography and Introduction to Gold Statistics]: Biographical sketch of Oskar Morgenstern and the introduction to the paper. Morgenstern argues that gold movement statistics, despite being considered highly trustworthy components of balance of payments data, are often unreliable. He emphasizes that the accuracy of economic data must be evaluated based on its intended use in theory and policy. [Uses of Gold Movement Statistics]: Discussion on the primary uses of gold statistics in long-run behavior analysis and international economic interaction. Morgenstern highlights the difficulty in separating monetary from industrial gold and references historical critiques by Ferraris and others regarding the massive discrepancies between the export reports of one country and the import reports of another. [The Test of Validity: Hypothesis and Sample Description]: Morgenstern formulates a common-sense hypothesis: the export statistics of country A should match the import statistics of country B, allowing for minor costs and lags. He describes the sample data from 1900, 1907, 1928, and 1935 for the UK, US, Germany, France, and Canada, explaining the selection of these specific years based on business cycles. [Data Tables: Annual and Quarterly Gold Movements]: Detailed statistical tables (Table 1 and Table 2) comparing annual and quarterly gold movements between pairs of countries. The tables include calculated ratios between the two sources of data, illustrating the frequent and significant lack of agreement between national reports. [Analysis of Statistical Failure and Discrepancies]: Analysis of the results shown in the tables and charts. Morgenstern observes that the discrepancies are 'catastrophic' for the classical gold standard period. He discusses the lack of a statistical theory to resolve conflicting observations and notes that correlations between gold movements and exchange rates often fail due to the poor quality of the underlying data. [Further Explorations and Cumulative Discrepancies]: Examination of net gold movements and cumulative discrepancies between specific country pairs (UK-Germany and US-UK). Morgenstern demonstrates that discrepancies often persist over time and do not self-correct. He concludes with Table 3, providing a statistical test of the significance of the algebraic differences in quarterly gold and silver statistics. [Notes and Methodology for Gold Flow Statistical Comparisons]: This segment provides the methodological notes for the preceding tables, defining the two types of gold flow differences (exports minus imports and vice versa) and justifying the use of small sample sizes in calculating standard deviations. It notes that average differences in reporting are significantly larger than can be explained by logistical factors like freight or insurance, suggesting inherent faults or biases in the statistics. [Explanation of the Differences and Further Difficulties]: Morgenstern explores the causes of statistical discrepancies in gold movements, citing factors such as the non-scientific nature of economic data collection, spurious time lags, and deliberate secrecy. He highlights issues like false declarations (mixing gold and silver), incomplete coverage of traveler-carried gold, and the complexities of transit and 'earmarking'—where gold ownership changes without physical movement—which render official trade figures unreliable for detailed economic analysis. [Summary and Conclusions on the Validity of Gold Statistics]: The author concludes that the lack of trustworthy gold statistics undermines the empirical foundations of international trade theory and business cycle analysis. He argues that while the data might suffice for broad, long-range trends, they are inadequate for studying cyclical behavior or precise capital movements. He calls for economists to acknowledge the shakiness of these data rather than continuing to build theories on unreliable statistical manipulations. [Appendix: Sources and Notes for Tables 1 and 2]: This appendix details the primary sources and technical notes for the gold movement data of the United Kingdom, France, Germany, the United States, and Canada for the years 1900, 1907, 1928, and 1935. It specifies the units of measurement (pounds, kilograms, francs, marks, dollars), conversion methods used for cross-country comparisons, and specific discrepancies found within the official reports of each nation. [Canada Data Sources and Princeton Studies in International Finance List]: The final segment provides source notes for Canadian gold statistics and lists recent publications from the Princeton International Finance Section, including works by authors such as Raymond F. Mikesell, Gottfried Haberler, and Thomas C. Schelling covering topics like bilateral payments, trade theory, and the GATT.