Setting τ in the Black-Litterman Model
Now we will consider how to select a specific value of τ in more detail. A brief survey of the
literature will be helpful. Given the previous discussion, we will focus on authors who use the
Canonical Reference Model in this section.
From Black and Litterman (1992)
Because the uncertainty in the mean is much smaller than the uncertainty in the return
itself, τ will be close to zero. The equilibrium risk premiums together with τΣ determine
the equilibrium distribution for expected excess returns. We assume this information is
known to all; it is not a function of the circumstances of any individual investor.
He and Litterman (1999) propose considering τ as the ratio of the sampling variance to the
distribution variance, and thus it is 1/t. They use a value of τ of 0.05 which they describe as
“...corresponds to using 20 years of data to estimate the CAPM equilibrium returns.”
As described previously, τ is the constant of proportionality between Σμ and Σ. We will examine
three ways in which we might select the value for τ. It is important to remember that τ is a
measure of the investor's confidence in the prior estimates, and as such there is a subjective
factor involved in its selection.
We will consider three methods to select a value for τ
• Estimate τ from the standard error of the equilibrium covariance matrix
• Use confidence intervals
• Examine the investor's uncertainty as expressed in their prior portfolio
First, we will approach the problem from the point of view of He and Litterman (1999). If we
were using regression techniques to find π using formula (3), then Σμ would be the sampling
variance or square of the standard error of the regression. Formula (14) is the expression for the
standard error where the residual is normally distributed which is assumed in the model.
(17) Std Err=
n
In the Black-Litterman model we do not use a regression approach to find the prior estimate of
the mean return, we solve for it using equilibrium techniques that have no clear standard error