Predicting the outflow of household deposits based on the intensity of search queries

Abstract

The subject of the study is the intensity of targeted search queries as a leading indicator of bank deposits outflow. The goal is to propose a mechanism for accounting information about the dynamics of search queries, able to signal changes in the volumes of deposits of individuals. The study was conducted using time series analysis models. Statistical data of Rosstat, Bank of Russia, searches in Yandex wordstat, Google Trends for the period from February 2009 to May 2022 were used. The relationship between the intensity of targeted search queries and household decisions to withdraw money from deposits and bank accounts was revealed. An assessment of the short-term predictive ability of search queries on dynamics of deposits was carried out. The use of statistical indicators on the dynamics of targeted search queries as a leading indicator of the outflow of funds of the population from deposits in commercial banks is substantiated. It was revealed that the use of the intensity index of targeted search queries as a signal indicator of the outflow of the placed funds by the population increases the accuracy of forecasting on the horizon in 1 month by 0.15–0.25 p.p. to assess the dynamics of ruble deposits and by 0.20–0.35 p.p. to assess the dynamics of foreign currency deposits. The use of information from search queries for the management of commercial banks is especially useful in the event of a threat of a sharp outflow of deposits of the population. The obtained results indicate the prospects of using textual information, including targeted search queries in order to form leading indicators of deposits outflow of the population. Preventive identification of negative trends associated with the outflow of deposits of the population can ensure the credit institution’s stability against negative macroeconomic influences.