Including a Risk-Free Asset in a Covariance Matrix
Summary
The document explains why a risk-free asset cannot always be added to a risky-asset covariance matrix by appending a row and column of zeros. That construction makes the matrix singular, which can prevent calculations that require a matrix inverse. A covariance matrix including the asset therefore needs a variance assumption, which means modeling some source of rate uncertainty rather than treating the return as perfectly fixed.
The appropriate treatment depends on the investment horizon and instrument. A short-term floating-rate security may have interest-rate risk as its rate resets, while a fixed-term bill that matures at the horizon can have zero return variance over that horizon. In the latter case, the response recommends handling the risk-free allocation separately from the risky-asset covariance matrix, with additional care if transactions must be included in the optimization. For longer horizons, assumptions about rate dynamics, such as mean reversion or links to inflation, may be needed. No single model is prescribed.
Key ideas
- Appending zero covariances and variance creates a singular covariance matrix.
- Modeling a nonzero variance for a nominally risk-free investment implies assumptions about rate risk.
- A fixed-term security matched to the investment horizon may have zero return variance at that horizon.
- For longer horizons, the risk-free rate may require an explicit model of its dynamics.
- Separate allocation decisions can be appropriate, though transaction constraints may complicate optimization.
Tags
Full text
# "Adding" risk-free asset to covariance matrix after the fact # "Adding" risk-free asset to covariance matrix after the fact Given a covariance matrix that was calculated from the returns of a number of risky assets. Is there a way to "add" a risk-free asset to the covariance matrix without calculating its covariance with all the other assets? I thought about simply adding a column and row filled with zeros that seems wrong as there should be some correlation between risky and riskless assets. Just wondering whether there are any shortcuts... ## Answer by John (score 2, accepted) https://quant.stackexchange.com/a/14968 If you add a bunch of zeros, then your covariance matrix will be singular, which could lead to problems depending on what you're doing (anything that involves inverting a matrix will have problems). If you have to have a covariance matrix that includes the risk-free rate, then you need to provide a variance for the risk-free rate (i.e. you're no longer strictly assuming it's risk-free) at a minimum. This is like assuming that you're investing in a short-term floating rate security (and you're implicitly making assumptions about the process of the rates). For instance, if you have a 6 month horizon, you could assume that there is no risk of default in any given month, but that the rate might re-set each month. You still aren't risking default, but you do have some interest rate risk. If you are using a fixed term risk-free security, like a 1 month T-bill, and your fixed term matches your investment horizon, 1 month in this case, then it will have zero variance at the horizon. In this case, you probably shouldn't be using the risk-free rate in a covariance matrix. Separate out the decisions. This becomes a little bit trickier if you're trying to account for transactions (b/c you have to include both decisions in the optimization then). If you have a longer horizon, then you can make assumptions about the dynamics of the risk-free rate, such as mean-reversion or its relationship to inflation.
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.