The Productivity Effects of Cross-border Data Flows: Evidence from Japanese firm-level data

This paper examines the effect of initiating cross-border data flows on firm productivity, using original survey data from Japanese manufacturing and service firms collected in 2019 and 2021, merged with annual productivity measures over 2019–2022. The survey identifies new entrants into cross-border data transfers, enabling a difference-in-differences design that compares “switchers” to firms that either do not collect data or collect data only domestically. We estimate the average treatment effect on the treated using regression-adjustment, inverse probability weighting, and doubly robust AIPW DID estimators, controlling for exporter status, multinational affiliation, R&D intensity, and ICT cost intensity. The results show that firms with higher initial productivity are more likely to start transferring data internationally, which is consistent with self-selection patterns documented in the export- and FDI-related literature. Entry into cross-border data flows is associated with significant productivity gains, which become particularly pronounced in the year after entry. These findings provide rare firm-level evidence from Japan, while also offering broader insights for data-governance debates by highlighting the potential productivity costs of overly restrictive cross-border data regulations.