Portfolio construction (optimisation)

Building efficient portfolios

In theory an investment portfolio should be put together in such a way that the expected performance from any holding is proportional to the incremental risk from that position. On the face of it this is obvious – it would make no sense to take a risk that is out of proportion to the expected return, or not to take more risk if the expected return is very large by comparison with the marginal risk.

However, in practice, the simple logic breaks down to the extent that both the expected return and the risk are uncertain. EMA aims to help managers with both of these issues. In relation to returns, EMA has developed the Alpha Toolbox which is a robust means of testing alpha indicators. On the risk front, EMA’s FASTVaR utilises a proprietary multi-factor GARCH process to generate a daily updated forward looking covariance matrix and residuals scalar, to reduce the reliance on purely historic risk relationships.

Once the expected returns and risk model are in hand, EMA offers an optimisation tool – “Optema” – to combine these into the most efficient portfolios.

Optema can be operated via and API or a user interface. It produces outputs in dynamic Excel workbooks that have links to the optimisation routines so that, once an optimisation task has been set up, the Excel workbook can be easily tweaked to experiment with different settings.

Typical objective functions are to find the minimum risk portfolio or the “efficient frontier” showing the trade-off between portfolio risk and a measure of expected return. The solution may be constrained at the portfolio level, for example with limits on sector or country weights, or at the asset level to ensure broad diversification.

Example features:

  1. Equity, Fixed-income or mixed-asset portfolios
  2. Can use EMA or user defined risk-models
  3. Limit number of holdings in final portfolio
  4. Generate shortlist of best trade ideas
  5. Target tracking error or build full efficient frontier
  6. Customisable – system or user-defined constraints
  7. Correct treatment of residual risk of composites
  8. Apply turnover constraint or transaction costs
  9. Black-Litterman adjustment for alpha uncertainty
  10. Tweak option on output allows quick and easy experimentation

The system is easy to use and with the forward-looking risk model clients have found that it produces superior portfolios.