Black–Litterman Risk Aversion and Sharpe-Optimal Market Weights
Summary
The document examines why a risk-aversion estimate in a simple Black–Litterman example does not reproduce the returns expected for each asset. It explains that the estimate derived from a chosen portfolio’s excess return and variance reflects that portfolio’s allocation. If those weights do not maximize the Sharpe ratio, applying the estimate to individual asset variances and weights will not necessarily recover the assumed asset returns.
The response illustrates the point by finding Sharpe-maximizing weights for the example and then calculating a corresponding risk-aversion parameter that reproduces the original returns. It also describes the framework’s assumption that investors can borrow at the risk-free rate: investors with lower risk aversion can scale up their allocation to the market portfolio. The result is a conceptual explanation tied to the example’s assumptions; it does not assess real-world borrowing constraints or establish that market weights are always Sharpe-optimal.
Key ideas
- An implied risk-aversion estimate depends on the portfolio weights used to derive it.
- Weights that do not maximize the Sharpe ratio may not reproduce assumed asset-level excess returns.
- Using Sharpe-maximizing weights can make the implied returns match the example’s inputs.
- The framework assumes investors can borrow at the risk-free rate and scale exposure to the market portfolio.
Tags
Full text
# Implied Equilibrium Returns Example # Implied Equilibrium Returns Example I've been trying to work through a simple example using A Step-by-Step Guide to the Black Litterman Model, but I'm having trouble understanding implied risk aversion. Say I have two uncorrelated assets, cash as my stand in for the risk free rate and my market capitalization weight happens to conveniently be equal weighted. According to the paper my implied risk aversion $\lambda$ should be the portfolio excess return (5%) over the portfolio variance (0.0066) which gives $\lambda = 0.05 / 0.0066 \approx 7.6$. However, when I try check my work by calculating the equilibrium excess returns I get an equity excess return (again assuming uncorrelated assets) of $$ r_e = \lambda * \sigma_e^2 * w_e = 9.7 \% $$ rather than the expected $7\%$. Is there a reason the implied risk aversion does give implied equity excess returns? Something to do with requiring fully funded portfolios maybe? Though that doesn't seem to tie out either. ## Answer by Kalev Maricq (score 4, accepted) https://quant.stackexchange.com/a/44395 It's because the model assumes that the market will maximize its Sharpe ratio and your weights don't do that. Essentially, your example assumes investors are irrational in their allocation. If you solve for the weights that maximize the Sharpe ratio, the implied returns will equal the given returns. In your example, the Sharpe Ratio reaches a maximum value of 1.091516 when weights of 7.58% and 92.42% are given to equity and bonds, respectively. This implies a λ of 36.06771. Taking 36.06771*[16%^2,3%^2]*[7.58%,92.42%] gives [7%,3%] implied returns, which match the original implied returns. The risk aversion you solved for is indeed the aversion implied by a 50/50 weighting, but the market participants would be better off using the weights that maximize the Sharpe Ratio and then allocating more or less of their assets to the portfolio vs cash. In the ideal world of the BL framework, market participants can borrow at the risk free rate, so someone with a lower risk aversion than that of the (Sharpe optimized) market portfolio would simply take a leveraged position in the market portfolio.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.