Paikalliset kaupankäyntiaikojen vaikutus Bitcoin tuottoihin

Abstract

In this thesis, I investigate the existence, magnitude, and evolution of regional overnight return anomalies in Bitcoin. Using high-frequency BTC/USDT data spanning August 2017 to March 2026, I construct session and overnight return components for three regional pseudo-sessions, based on local stock exchange trading hours – Asia (TSE), Europe (Xetra/LSE), and the United States (NYSE/Nasdaq). I document that Bitcoin exhibits a significant overnight premium in the Asian pseudo-session over the full sample, which persists after controlling for realised volatility, trading volume, and illiquidity. No significant premium is found for Europe or the US over the full sample. Notably, holding Bitcoin exclusively during Asian trading hours would have produced a negative cumulative return over a period in which the price of Bitcoin appreciated significantly. Second, the premium is not constant over time; it is strongest during speculative, retail-dominated phases and compresses during bear markets, consistent with the Adaptive Markets Hypothesis. Third, I examine how the introduction of US spot Bitcoin ETFs affected the premium using a difference-in-differences analysis, finding that the overnight premium increases in the US region, while the Asian overnight premium collapses to statistical insignificance. Finally, I find that a simulated trading strategy based on the Asian overnight premium underperforms a passive buy-and-hold benchmark under standard retail fee structures, consistent with the limits-to-arbitrage hypothesis. The findings extend the overnight drift literature to a continuously traded asset and provide direct empirical evidence for geographically segmented price discovery in cryptocurrency markets.