지정학적 이벤트에 따른 K-방산 주가 상관 네트워크의 구조적 변화 분석

초록

This study constructs dynamic correlation networks from daily returns of 20 Korean defense stocks (2015–2026) to analyze structural changes in network topology across geopolitical regime shifts. Network density varies dramatically across regimes, from extreme sparsity during the inter-Korean détente to a sharp surge during the COVID-19 pandemic. Market-adjusted residual correlation analysis reveals that the COVID-19 surge is predominantly market-wide, whereas the increase during the Russia–Ukraine conflict largely reflects sector-specific co-movement. Robustness checks across multiple thresholds, correlation measures, and window sizes confirm invariant temporal ordering. Stocks with higher defense revenue shares exhibit systematically larger density fluctuations. The study provides evidence that market-wide shocks homogenize sub-group structures while sector-specific shocks preserve block differentiation, and validates a GMM-based regime classification.

키워드