Bitcoin, Ethereum, and the Ambiguity of Price Discovery: A Multi-Measure High-Frequency Analysis

This paper examines whether Bitcoin’s (BTC) price discovery relationship with Ethereum (ETH) is state-dependent, using four years of one-minute Coinbase trade data (2022–2025) and multiple rolling window price discovery, spillover, and predictability measures. Results prove to be metric-dependent: Granger causality from BTC to ETH strengthens significantly in High volatility regimes, the Hasbrouck information share moves in the same direction without reaching significance, and the Diebold–Yilmaz net spillover measure moves significantly in the opposite direction, favoring ETH. No evidence emerges of BTC leading ETH in price discovery during tranquil markets, or of delayed cross-asset adjustment after extreme shocks. By contrast, BTC and ETH show statistically distinguishable intraday volatility regimes, with both troughing during Asian hours and rising sharply at the European and U.S. overlap and ETH’s realized variance exceeding BTC’s every hour. Short-horizon return predictability after extreme moves holds for ETH, but only following positive shocks, with no comparable BTC effect. Findings point to intraday seasonality and asset-specific momentum as the market’s more robust regularities.