Equity Valuation and Portfolio Management

KEY POINTS

• High-frequency data are different from daily- or lower-frequency data. Whereas low-frequency data typically comprise regularly spaced open, high, low, close, and volume information for a given financial security recorded during a specific period of time, high-frequency data include bid and ask quotes, sizes, and latest trade characteristics that are recorded sequentially at irregular time intervals.
• The differences affect trading strategy modeling, introducing new opportunities and pitfalls for researchers. Numerous data points allow researchers to deduce statistically significant inferences on even short samples of high-frequency data.
• Different sampling approaches have been developed to convert high-frequency data into a more regular format better familiar to researchers; nevertheless, diverse sampling methodologies result in data sets with drastically dissimilar statistical properties.
• The short-term dynamics among equities are also significantly different from those at lower frequencies, creating new frontiers in portfolio management research.