Getting the best trade-off between risk and return

Broadly speaking an “optimiser” is a software tool takes account of both expected future returns (outperformance or alpha) and expected future risk (using a covariance matrix) to obtain the most efficient portfolio subject to some constraints, such as maximum position size or a turnover limit.

Typically the portfolio manager will form a view of expected returns, either as an explicit level of return, or by means of an indicator such as a P/E ratio, and this is then combined with a historic covariance matrix.

EMA’s unique contribution to the challenges faced when building an optimal portfolio is to have developed a method for generating a forecast covariance matrix with a time horizon of up to one month in the future.

By matching up forward looking returns forecasts with a forward looking covariance matrix the EMA optimiser is uniquely well placed to avoid the reliance of many optimisers on historic risk patterns that can lead to inefficient portfolios.

The EMA optimiser can be instructed through an API or set up manually via a user interface. It offers a range of objectives such as minimum risk, target alpha or efficient frontier and allows for the setting of default and user-defined constraints  on the final portfolio and on the costs of trading. A back-testing feature allows for the automation of multiple-optimisations through time. It uses some of the latest mathematical techniques – both quadratic-programming and integer-programming – to enable solutions to the problems typically faced in fund management.

A particular feature of the EMA Optimiser is that it generates an output report in Excel which contains the key optimisation settings. These can be edited and the optimisation re-run, making it very easy to experiment with parameters to achieve the desired portfolio.