Price Volume Correlation Trend Momentum Factor
factor.formula
1. **Calculate the intraday price-volume correlation coefficient:** For each trading day, calculate the Pearson correlation coefficient between the stock's minute-level closing price and trading volume.
2. **Construct a time series of price-volume correlation:** Take the intraday price-volume correlation coefficients of the last 20 trading days (or a specified time window) to form a time series \( \{p_1, p_2, ..., p_{20} \} \).
3. **Calculate correlation trend momentum:** Perform linear regression on the above time series, with time \( t \) as the independent variable and correlation coefficient \( p_t \) as the dependent variable, and obtain the regression coefficient \( \beta \). The regression coefficient \( \beta \) represents the trend momentum of price-volume correlation over time.
4. **Cross-sectional standardization and neutralization:** On the regression coefficients \( \beta \) of all stocks, market capitalization and traditional price-volume factors (such as 20-day reversal, 20-day turnover rate, and 20-day volatility) are neutralized to obtain the final price-volume correlation trend momentum factor value.
in:
- :
The Pearson correlation coefficient of the stock's minute-level closing price and trading volume calculated on the ( t )th trading day. This correlation coefficient measures the strength and direction of the linear relationship between price and trading volume on that day.
- :
The regression coefficient (regCoef) of the linear regression of the price-volume correlation time series ( {p_1, p_2, ..., p_{20} } ) on time ( t ) represents the slope of the price-volume correlation over time. When ( \beta ) is positive, it means that the price-volume correlation increases over time, otherwise it decreases, reflecting the market's expectations for the change in the price-volume relationship of the stock.
- :
The residual term (resTerm) of the regression model represents the part that the regression model cannot explain, that is, the deviation between the actual correlation coefficient ( p_t ) and the linear regression fitting value.
- :
Time variable, which indicates the serial number of the time series. Here, it is 1 to 20, representing 20 trading days (or a specified time window) to be traced back.
factor.explanation
The lower the value of the price-volume correlation trend momentum factor, that is, the stocks whose price-volume correlation weakens over time, the higher the future short-term returns tend to be. This may reflect the market sentiment shifting from consistent trading to potential trend reversal. This factor uses high-frequency price and volume information to capture the momentum signals contained in the market microstructure. The principle is that when the price-volume synchronization weakens, it may indicate that funds are beginning to leave the market, or the market is beginning to form a new understanding of the fundamentals of the stock, which may lead to a subsequent rise or fall in the stock price.