Is there any way to set to set a Value at Risk (VaR) constraint or a Probability of Loss constraint on the optimization? For example, I would like to construct an optimal portfolio that would evaluate a 15% value at risk given the specified expected returns and covariance in my case file.

**Answer**

No, this constraint extends the non-linear portfolio constraints. There are a few methods to approach this problem.

The Windham Software allows you to specify a portfolio risk constraint. Since VaR and Probability of Loss results are a function of portfolio risk and return, one can iteratively arrive at an acceptable level of VaR by constraining the portfolio risk during the optimization process.

Alternatively, you can also use the risk aversion and increments to find the portfolio with the highest return that matches your downside risk tolerance. It is an iterative process that may require some searching. We suggest a starting risk aversion of 3 and a step size of .5. Find the portfolio that is closest to you VaR objective on the Exposure to Loss screen. On the Portfolio Attributes screen you will be able to find the risk aversion level of this optimal portfolio. Specify a new risk aversion level back in the optimization screen, and iteratively search while reducing your increments to .1. Repeat as necessary reducing increments further until your specific VaR objective is met.

Category:Understanding the Software -> Optimization

### Categories: