Abstract

The ECB data quality processes for granular data on financial transactions would be analysed in a three-step approach. First, how to detect and correct wrongly reported data by comparing individual observations with transactions in the same dataset being reported by the same or different reporting agents. This first step would cover daily processes to ensure correctness and completeness of reporting, including anomaly detection and management of revisions. Second, how to check and correct data by pairing a transaction with the opposite transaction reported by the counterparty under the same data set. Here also the challenges and benefits of the Unique Transaction Identifier (UTI) would be discussed. Third, how to compare and analyse micro data from two different data sets in order to identify inconsistencies and ensure quality. Experience with the ECB Money Market Statistical Reporting (MMSR) and European Market Infrastructure Regulation (EMIR) datasets, also in relation to the European Securities Financing Transactions Regulation (SFTR) would be described.