The Concentration of AI Talent as an Industrial Strategy: A Cross-Country Panel Data Analysis Applied to Financial Services as an Industry

Abstract

Artificial intelligence is redrawing comparative advantages in financial services, while many studies remain descriptive or focused on a single country, without testing whether the observed gaps reflect distinct industrial strategies. In this context, our objective is to establish whether the concentration of AI talent in finance is due to a simple global trend or to differentiated national choices. Empirically, we conduct an observational, comparative, and longitudinal study on 10 OECD countries monitored annually between 2016 and 2025 (N = 100 country-years). The dependent variable is the share of professionals trained in AI in finance (AI_pct, harmonized definition). The dynamics is captured by a linear time trend, supplemented for robustness by annual dummies; heterogeneity is modeled via a random effects GLS with country intercepts, clustered standard errors, and the Hausman test does not reject the RE option. On the data side, we mobilize a single, harmonized public source (OECD.AI) and anchor the analysis in 18 scientific references. The results indicate an average increase of approximately +0.2263 percentage points per year (significant), an overall average of 2.86%, and persistent gaps between countries (e.g., Israel ≈ 4.08% vs. the United States ≈ 2.27%), stable when the trend is replaced by time fixed effects. In sum, the rise in AI skills is general, but is part of national trajectories consistent with industrial strategy; hence implications for upskilling, data and model governance, and state-market coordination, subject to a limited scope and the absence of causal identification.