Equity Market Dependence and Hedging Effectiveness Across China, Türkiye, and the United States: Evidence from Market–Clock Alignment, Currency Denomination, and Geopolitical Event Windows

This study examines whether correlation dynamics among the CSI 300, BIST 100, and S&P 500 improve hedging performance under alternative currency denominations and market–clock alignments, including during the geopolitical event window beginning on 28 February 2026. Market-specific GARCH models and CCC, DCC, and asymmetric DCC (aDCC) models with multivariate Student-t innovations are estimated using 1199 daily return observations on common trading days from 12 March 2021 to 17 July 2026, expressed in local currencies and US dollars. Among the dynamic models, symmetric DCC is preferred to aDCC. However, the Engle–Sheppard tests do not reject constant conditional correlation in either the local-currency specification (p = 0.7457) or the USD-denominated specification (p = 0.7519); the DCC news coefficient is also statistically insignificant, and CCC has the lower BIC under both currency denominations. Mean correlations are weakly positive and are highest for the BIST 100–S&P 500 pair. After Holm correction within each currency basis, structural breaks are supported for both CSI-related pairs in local-currency terms, but only for the CSI 300–BIST 100 pair in USD terms; no break is supported for the BIST 100–S&P 500 pair under either currency basis. Over the 85-day out-of-sample period, DCC produces the lowest mean QLIKE loss under both currency denominations. Relative to CCC, the DCC advantage remains significant after Holm correction at the 5% level in USD terms, but not in local-currency terms. Nevertheless, the 90% Model Confidence Set retains DCC, CCC, and the static covariance model under both currency denominations. DCC also provides the highest gross hedging effectiveness in all six pair–currency comparisons, whereas the static benchmark records a higher realized cumulative net hedge P&L in five of the six comparisons under the illustrative base-cost profile; however, this non-risk-adjusted 85-day measure is sensitive to realized sample drift and does not establish general static-strategy superiority. Overall, the evidence indicates limited and denomination-dependent differences in performance and does not establish the unconditional superiority of DCC.