M₀: The Deviation-from-Mean Matrix
How the deviation-from-mean matrix is constructed and what it does.
NOTES / QUANTITATIVE FINANCE
Factor models, portfolio construction, covariance estimation, and the linear algebra behind them.
17 pieces
How the deviation-from-mean matrix is constructed and what it does.
Deviations from means.
A compact derivation of the variance of the sample mean.
The multivariate case of the Newey-West covariance matrix.
The most prolific result in statistics.
A covariance matrix that incorporates heteroskedasticity and autocorrelation.
Deriving the partitioned least-squares estimator by partialling out one block of regressors.
The full derivation of the normal equations for partitioned least squares.
Why the beta of the benchmark component equals one, and what that implies for active weights.
Optimizing a portfolio relative to a respected benchmark.
Supplying alpha to a benchmark portfolio.
Full setup and solving for CWLS in the context of factor models.
Understanding what it means to "partial out" variable effects.
A short reference on projection matrices and why idempotence implies projection.
A technique for approximating the covariance matrix.
A working definition of alpha, its role in portfolio returns, and its formula relative to a benchmark.
Visualizing the workhorse model for regression analysis.