go-v.orgdoi

The Effects of National AI Readiness and Data-Transfer Regulation on the Distance Elasticity of Digital Services Trade

Young-Soo Moon, Bok-Jae Park · container not stated · 2026

  • DOI 10.46396/kjem..112.1
  • Publisher OpenAlex
  • Type article
  • Issued 2026

Not cached yet. This record was fetched live for this request and is stored after the lookup. Reopening the page should serve the cached copy.

· OpenAlex · article · 2026

Abstract

Purpose: This study examines whether distance friction persists in digitally deliverable services (DDS) trade, whether exporter AI readiness is associated with weaker distance effects, and whether data-transfer regulation and sectoral heterogeneity condition this relationship across exporter partner portfolios. Research design, data, and methodology: We combine BaTIS, the Oxford Government AI Readiness Index, OECD DSTRI, and CEPII Gravity into a 2020–2024 panel of 127 exporters, 203 importers, eight DDS sectors, and 911,232 observations. A PPML gravity model includes exporter-year, importer-year, and sector fixed effects.

Because the treatment variables vary only at the exporter-year level, exporter-level clustering is the primary inference criterion, with pair clustering reported as a reference. Robustness checks use an alternative distance measure, a reported-only subsample, and an exporter block bootstrap. Results: Distance remains costly (baseline log-distance coefficient: -0.5123, p<.001). Under pair clustering, the AI-by-distance interaction is positive and marginally significant (+0.0849, p=.055), while the pooled triple interaction with data-transfer regulation is significant at +0.5856 (p=.0088).

Neither is significant under the primary criterion (p=.225 and .161) or the block bootstrap (p=.310 and .697); both are therefore reported as tentative rather than confirmatory. The sector-specific estimate is strongest for charges for the use of intellectual property (+2.6314); it remains significant under exporter clustering, alternative distance, and the reported-only subsample (p≤.0003), but not under the percentile block bootstrap (p=.224). Applying the pooled H1 coefficient to 2024 partner portfolios, a descriptive post-estimation exercise, yields a positive far-minus-near reallocation gap for all 127 exporters.

Korea’s gap is 1.747%p (95% bootstrap CI [-1.31, 4.59], 45th of 127) and Japan’s is 1.406%p ([-1.05, 3.70], 88th), but the ordering holds only in the 84.5% of draws with a positive H1 estimate. Implications: The evidence is consistent with, but does not confirm, an association between AI readiness and weaker bilateral distance effects; the sectorally concentrated pattern in charges for the use of intellectual property is the more robust finding, so data-governance implications are sector-specific rather than DDS-wide. Country rankings are descriptive and conditional on H1.

Open the resolved locator · 10.46396/kjem..112.1

Metadata licence: OpenAlex metadata (CC0)