The magnificent seven and bitcoin: Time-varying predictability, dynamic conditional correlation, and multi-scale co-movement
Year: 2026
Type: article
Abstract: ABSTRACT This study examines the dynamic relationship between Bitcoin and the Magnificent Seven stocks across three distinct dimensions: conditional correlation, directional predictability, and time-frequency co-movement. Using weekly data from 29 January 2017 to 19 July 2026, the analysis applies pairwise DCC-GARCH models, a control-augmented time-varying Granger predictability framework, and wavelet coherence analysis. In the time-varying predictability models, common market and macro-financial conditions are controlled for using the Nasdaq-100 Index, VIX, the U.S. Dollar Index, and the 10-year U.S. Treasury real yield. To assess whether the results depend on the choice of the technology-oriented Nasdaq-100 benchmark, the NYSE Composite Index is employed as an alternative broad-market control. In addition, sensitivity analyses are conducted for the BTC-META and BTC-NVDA systems using alternative window lengths and lag structures. The findings reveal pronounced firm-level heterogeneity and episodic variation in Bitcoin-MAG7 linkages. The strongest evidence of directional predictability is obtained for the BTC-NVDA and BTC-META systems, whereas the results for AAPL, AMZN, GOOG, MSFT, and TSLA are weaker, more short-lived, or more sensitive to the choice of broad-market control. The sensitivity analyses further indicate that predictability from BTC to NVDA is relatively robust to alternative window lengths and lag structures, while predictability from NVDA to BTC is more sensitive to lag selection. The DCC-GARCH results point to positive average conditional correlations across all pairs, with the highest average correlations observed for TSLA, AMZN, and MSFT, and the greatest correlation persistence particularly for TSLA and GOOG. The wavelet coherence results show that co-movement is concentrated in specific time and frequency regions, strengthening particularly during the 2020 market disruption and in selected periods after 2022. Overall, the findings demonstrate that conditional correlation, directional predictability, and time-frequency co-movement capture distinct dimensions of the Bitcoin-technology stock relationship and should not be interpreted as interchangeable measures.
Authors: Fuzuli Aliyev, Neman Eylasov, Nijat Gasim
Language: English
Subfield: Finance
Domain: Social Sciences
Open Access status: gold
Estimated APC paid: $1,930
OpenAlex ID: W7222445708