Big data financial transactions and GDP nowcasting: The case of Turkey

Abstract We use aggregated information from individual‐to‐firm and firm‐to‐firm transactions from the Garanti BBVA Bank to simulate domestic private demand and estimate aggregate consumption and investment for Turkey's quarterly national accounts in real time. We show that these big data variables successfully nowcast official consumption and investment flows. To further validate the usefulness of these indicators, we include both indicators among others which are generally used in gross domestic product (GDP) nowcasting and evaluate their contribution to nowcasting power of Turkish GDP by combining both linear and nonlinear models. The results are successful and confirm the usefulness of consumption and investment banking transactions for nowcasting purposes. These big data are valuable, especially at the beginning of the nowcasting process, when the traditional hard data are scarce. Accordingly, this information is especially relevant for countries with longer statistical release lags, such as emerging markets.