Abstract

The Cash Balance of Payments (BOP) statistics compiled by Bank Negara Malaysia (BNM, the Central Bank), is aligned to the International Transactions Reporting System (ITRS) stated in the International Monetary Fund's (IMF) Balance of Payments and International Investment Position Manual, Sixth Edition (BPM6) Compilation Guide. The Cash BOP statistics form an integral part of the External Sector Statistics (ESS) managed by BNM. While the Department of Statistics Malaysia (DOSM, the national statistics office) is the official compiler of Malaysia’s BOP data which is currently compiled and published on quarterly basis, monthly Cash BOP statistics play a crucial role in complementing the BOP as it serves as early indicators for the flow of funds and provide preliminary data for monthly analysis, in addition to supporting surveillance and policy-making functions at BNM. In addition, the data will also supplement the compilation of BOP data, which are currently sourced from International Merchandise Trade Statistics (IMTS), surveys, and administrative records. This paper explores the complexities of compiling Cash BOP statistics, with a focus on Malaysia's statistical experience following the BPM6 methodology. The paper discusses the challenges encountered in data compilation, particularly in managing voluminous cash transactions, and outlines the approaches taken to ensure data accuracy and credibility. These approaches include deploying statistical methods for outlier detection and employing unsupervised machine learning to improve the accuracy of outlier detection. Additionally, the paper delves into exploring alternative data capture/sources to augment the comprehensiveness of Cash BOP statistics. This paper aims to deliver valuable insights to practitioners and policymakers in economies particularly countries that maintains ITRS.