Seed Weekly - Portfolio Construction and Optimisation

It is one thing for a fund manager to have an investment philosophy and process that selects shares that typically outperform the market. But often “gains” made at the stock picking level are mitigated using poor portfolio construction methodologies. Increasingly we are looking more closely at specific methodologies that managers use in constructing their portfolios.

Unless the stock picking ability is taken all the way through to the portfolio construction, total portfolio returns can be dramatically reduced where the manager has too low a weight to a winning position, or conversely too high a weight to a high conviction but high risk position. Or a fund manager may have what appears to be a well-diversified portfolio, but this could include a concentration in a range of shares with the same return driver – i.e. they may all be local importing companies that come under pressure as the rand weakens.

So just what is portfolio construction and optimisation? It is a method of determining the allocation of selected shares (or asset classes, or funds) in a portfolio, so that they work together to maximise the overall returns and / or minimise the risk.

Some of the various methodologies used to construct a portfolio include the following:

• Benchmarking against an index. Many equity fund managers start the process of construction with the weights as determined in the index – for example the JSE Top 40 weights. Preferable shares are overweighed and vice versa, those that do not meet certain investment criteria are underweighted relative to the index.

• Equally weighted. This is a methodology often used as an alternative to an index tracking portfolio, where the weights to the shares in the index are equally weighted, instead of the traditional highest weight to the largest market cap share on the stock market.

• Many active fund managers adopt what is known as a Barbell strategy to portfolio construction. This strategy has its origins in constructing a bond portfolio, but can be loosely defined as allocating a portion of the portfolio, say 60%, to relatively low risk, but low yielding assets, example, money market. The remaining 40% is then allocated to higher risk and higher return expectation assets. An example being a range of small cap shares.

• Increasingly, fund managers are looking at optimising portfolios for various risk measures. These risk metrics may include allocating the biggest weight to the shares with the lowest volatility and vice versa. There is investment merit in this approach, because the performance of shares that display lower than market volatility is often superior to the market.

Investors may come across the term, modern portfolio theory, which was developed in the 1950’s and is a method of constructing a portfolio, which will maximise the expected return, for any given level of risk assumed; where risk is defined as volatility of the portfolio’s return.

Where individual investments have similar return characteristics, but these returns are not perfectly correlated then combining them in a portfolio can reduce the overall portfolio’s volatility (i.e. risk measurement), without detracting from the performance. I.e. a more efficient portfolio can be constructed.

Using a portfolio optimiser

This can also extend to combining different types of unit trusts that have different risk and return characteristics. At Seed Investments we have developed a portfolio optimiser that assists us in determining the percentage mix that we allocate to different funds, when combining these into a portfolio.

The optimiser works by allocating the selected range of fund’s historical performance into the mix. It then runs through all the various combinations and provides an ideal output across different risk criteria. An investor does not need an optimiser to determine which combination of funds gave the highest return over a period of time – that is straightforward.

However when trying to optimise for downside standard deviation, minimised drawdowns, maximising Sharpe and Sortino ratio’s, then an optimiser tool is essential. The output is then used as a guide as to the final portfolio outcome, knowing that the past cannot be fully replicated into the future.

As with so many aspects on investing, there is no one “magic bullet” even for portfolio construction. Very often the best is for a manager to have a defined process that makes sense and for that manager to then remain diligent to that process without unnecessary wavering, which too often results in increased uncertainty and underperformance.

Kind regards,

Ian de Lange

Tel +27 21 914 4966
Fax +27 21 914 4912
Email info@seedinvestments.co.za

Seed is hiring: Visit the Seed Analytics LinkedIn profile to view vacancies.

www.seedinvestments.co.za

Thu, 28 Apr 2016