Estimating Equity Factor Models Across Different Market Hours

Summary

The document raises a problem in estimating short-term statistical factor models from equities trading in markets with different hours. When one market closes before another, their daily returns cover different time intervals. A Japanese stock may respond to U.S. market moves while Japan is closed, so matching calendar-day returns can understate the relationship between the two markets.

The author asks how researchers and practitioners handle this timing mismatch in theory and practice. The text does not provide an answer, method, empirical comparison, or cited literature; it only frames the issue. Any solution would need to consider the timestamps and information intervals represented by each return series, but the source itself does not evaluate such approaches or establish how large the bias is.

Key ideas

• Daily returns from markets with different trading hours cover different information intervals.
• Naively aligning calendar-day returns may understate cross-market relationships.
• The question focuses on short-term statistical factor estimation across U.S. and Japanese equities.
• The document poses the issue but supplies no method, evidence, or literature review.

Tags

# short-term statistical factor models for equities with different trading hours


# short-term statistical factor models for equities with different trading hours












I wonder if there are existing theory/literature about estimating a short-term statistical factor models for equities with different trading hours.

For example if we are estimating a universe with US and Japan equities, they don't close at the same time, so their daily returns are never "aligned." You can imagine Japanese stocks probably move along with US stocks, when Japan is closed and US is open. So the real correlation between them, would have been higher than observed correlation (if you just compute the naive daily returns correlation)

How is this problem handled in theory/practice?

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.