Portfolio optimization software is a tool used by many asset management professionals to estimate the potential returns of an investment portfolio and allocate investments in a more efficient manner. But is it really effective, and what features matter the most?
Modern portfolio theory as stated by Markowitz suggests the optimum portfolio is one which lies upon the edge of the Efficient Frontier. That’s fine in theory, but can the typical investor actually build an efficient portfolio with limited capital and without incurring a lot of transaction costs? Any optimization of a portfolio should take into account transaction costs to move from the base or “current state” portfolio and the one that lies upon the Efficient Frontier.
The most important activity that portfolio optimization software does is calculate the covariance matrix between instruments or businesses. The covariance matrix is the heart of the assumption that diversification can produce risk-return benefits, and therefore that an investment portfolio can be optimized at all. However, one deficiency of this reliance on the covariance matrix is the reality that correlations are transitory. In other words, correlations between instruments or businesses vary over time and the lookback period you select to calculate them. This means that any optimization you do is only a snapshot. It also means that correlations may break down or shift at any time, so your optimized portfolio as of yesterday may not be the most optimum one for tomorrow.
There are multiple ways to create portfolio optimization software, including nonlinear, quadratic, and mixed integer programming. We won’t go into these techniques here. For the typical sub-million dollar portfolio of an individual investor, the actual technique used does not matter as much as the ability to shift allocations regularly with low transaction costs. After all, this is an imperfect science.
Portfolio optimization software comes in all shapes and sizes, from hugely complex bank-wide super portfolio simulators that handle every conceivable type of investment and derivative, to easy-to-use Excel spreadsheets made for common stocks only. All of these systems work for their purpose, which is to increase the diversification benefit as much as possible while keeping expected returns as high as possible. Compared to throwing darts at a stock list, investing in your Top 5 Favorites, or pyramiding returns from the best performing business line, the simulations done in an optimizer procedure really reduce risk. Although diversification benefit tends to go away significantly during market crashes, it is still significantly greater that holding only a few investments in related companies or industries that inherently move together and may correlate to 1 during market stresses.
When selecting an optimization tool, these are the important things to look for: ability to specify multiple input parameters such as maximum and minimum return, expected discount rate, benchmark return, maximum and minimum volatility, capital invested, reinvestment of dividends and other cash flows, number of simulation runs, and other factors.
As long as you pick the tool with the right capabilities, you will see great benefits from your portfolio optimization software.