Abstract:

Purpose: This study investigates temporal causal relationships in EU macroeconomic systems using Bayesian network analysis to identify transmission mechanisms and policy timing implications.

Methodology: Bayesian networks with Hill Climbing algorithms analyze EU MIP data covering 27 countries over 23 years. Cross-sectional analysis (264 observations) and temporal analysis with 2-period lags (184 observations) examine 8 key macroeconomic indicators including GDP per capita, building permits, corporate debt, and international investment positions.

Findings: Cross-sectional analysis reveals 16 causal relationships with real GDP as primary driver and corporate debt as transmission hub. Temporal analysis identifies 242 connections dominated by 2-period lags. Building permits predict economic growth after 2 periods, while corporate investment affects GDP and external balance with substantial delays.

Implications: Two-period lag dominance suggests EU macroeconomic policy requires sustained implementation (≥2 periods) rather than short-term interventions. Results provide actionable insights for coordinated policy design within EU's MIP framework.

Originality: Novel network-based approach to temporal causality in EU panel data, complementing traditional econometric methods.