Time Series Factor Models

• Pre-defined and user defined risk models
• True stepwise regression (forward and backward)
• User defined factor selection metrics based on the stepwise regression
• Factor regression by least squares (LS), robust or fat-tailed L1 method
• Non-linear predictor functions: lagged variables (explanatory and dependent factors), square function & up and down market beta
• Option for stress-test scenarios
• Reports include: exposures, standard errors, confidence intervals, p-values, R-squared and Residual Variance; charts for residuals: time series, QQ-plots, histograms, probability densities and residuals vs. fitted values and response vs. fitted values;
• Comparison reporting & charting between changes in any calculation settings

Principle Component Factor Analyses

• Fully automated computation of “blind” factors with high explanatory power
• Reports: explained variance by factor, cumulative explained variance and Eigen values
• Charts for residuals: time series, QQ-plots, histograms and probability densities
• Option to store principle components as generic risk factors and use them in time series factor modeling

Statistical Factor Model Factor Analyses 

• Fully automated computation of “blind” factors with high explanatory power
• Least squares, robust and L1 regression computation of factor returns
• Scenario generation of factor returns, residuals and asset returns with full array of generalized fat-tailed distributions.
• Reports include: factor exposures, specific risk factors variations

Highlights