Measuring Unmeasurable Soft Information

Abstract

In this study, an index of accumulating soft information is developed using text records of meetings with customer firms from 2003 to 2022 from a Japanese financial institution. The index is unique in reflecting the rich patterns of relationship dynamics. We chose conversation themes and keywords indicative of strong relationships and developed an index based on the frequency of keywords related to these themes. We employ a two-pronged approach in the selection of our themes and keywords, the practice-based way to involve interviews with financial institution officers and the data-driven way to use key financial indicators. The results are robust to the two methodologies. We find that there is a positive correlation between the amount of soft information and the effort devoted to building relationships by financial institution officers. Conversely, no clear relationship is observed between the amount of soft information created under the definition of this study and variables, such as geographical distance from the customer firm and the length of the transaction, which have been considered in the literature to contribute to relationship building. Furthermore, the information collected is independent of the location and channel related to where and how the meeting took place.

Keywords: finance, financial intermediation, banking, relationship banking