Time scales involved in emergent market coherence

摘要: Abstract In addressing the question of time scales characteristic for market formation, we analyze high-frequency tick-by-tick data from NYSE and German market. By using returns on various ranging seconds or minutes up to 2 days, compare magnitude largest eigenvalue correlation matrix same set securities but different scales. For sets stocks capitalization (and average trading frequency), observe a significant elevation with increasing scale. Our results study can be considered as manifestation so-called Epps effect. There is no unique explanation this effect it seems that many factors play role here. One such randomness in transaction moments stocks. Another interesting conclusion drawn our contemporary markets emergence correlations occurs much smaller than more distant history.