Choosing Benchmark Weights for Black–Litterman
Summary
Black–Litterman needs equilibrium weights for a market portfolio to infer implied returns. When the investment universe contains a mix of assets such as stocks and bonds, market capitalizations may be unavailable or hard to compare, so the document recommends finding a suitable benchmark portfolio with known allocations.
A large fund with low tracking error to the intended universe can serve as a proxy. If only broad asset-class allocations are available, the approach can support implied-return estimates for subclasses rather than individual securities. The document offers endowment portfolios as an example of potential allocation data and notes that an arbitrary benchmark can also be used. The choice depends on the investor’s objective; the resulting implied returns reflect the selected benchmark and its weights, not a uniquely determined market portfolio.
Key ideas
- Black–Litterman requires weights for a market portfolio or a close benchmark.
- A fund with low tracking error to the target universe can provide a practical proxy.
- Broad allocation data supports implied returns at the asset-class or subclass level.
- The benchmark choice shapes the implied returns produced by the model.
Tags
Full text
# Market weights for Black-Litterman # Market weights for Black-Litterman I'm trying to implement Black-Litterman for an arbitrary selection of assets. One of the input for BL is the "Equilibrium market capitalization weights for each asset". In most examples I've seen, the assets are represented by different market sector so the market capitalization is based on the sector share. What should I use as market weight for an arbitrary selection of assets (stocks, bonds etc)? ## Answer by zuiqo (score 5) https://quant.stackexchange.com/a/7593 For Black-Litterman you need to have weights of the market portfolio, or a close benchmark, where you know the asset allocation. The common approach in practice is to find a large fund with a low tracking error for your investment universe. Depending on your goal it may be enough to have the allocation over classes of your fund, which will allow you to extract expected returns for the subclasses, rather than individual securities. One way I have seen in practice is to use the endowment funds of Harvard, Yale, etc., as there is some data available (from their investment offices). That is the reason you rarely see a more specific implementation, just like you describe. So the point remains, you need to find some asset allocation for the market portfolio. ## Answer by Joshua Ulrich (score 3) https://quant.stackexchange.com/a/7587 It can be difficult to obtain market capitalization for some types of assets. Instead, you can use the weights to an arbitrary benchmark portfolio. That would be like backing out the returns that would result in you investing in the benchmark portfolio if you don't have any views.
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.