Advanced Portfolio Construction Techniques › Key Parameters

The key parameters for the portfolio construction models are summarized below. Understanding how to calibrate these parameters is essential for applying the techniques effectively in practice.

• $\delta$: Risk aversion coefficient. Determines the tradeoff between expected return and variance in the utility function. Typical values range from 2 to 4. Higher values produce more conservative portfolios that prioritize risk reduction over return maximization.
• $\tau$: Black-Litterman scalar. Controls the weight given to the equilibrium prior relative to investor views. Typical values range from 0.01 to 0.10. Smaller values make the prior more influential, requiring stronger views to move the posterior significantly.
• $\kappa$: Robust optimization parameter. Defines the size of the uncertainty set (confidence level) for expected returns. Larger values produce more conservative portfolios that hedge against greater estimation error.
• P, Q: View matrices. P maps views to assets, specifying which assets are involved in each view and with what signs. Q specifies the expected returns for those views.
• $\Omega$: View uncertainty matrix. Represents the confidence (variance) in each expressed view. Smaller diagonal elements indicate higher confidence and greater influence on the posterior.
• $\alpha$: Shrinkage intensity. Determines how much the sample covariance is pulled toward the structured target. The optimal value depends on the ratio of assets to observations and is typically estimated from the data.