Ask the Expert Interview with Axioma: Talking Technical: How the portfolio optimizer works

In this discussion with Axioma’s Mellissa Brown and Rob Stubbs, we dig deeper into the workings of the Axioma Portfolio Optimizer. We talk about how the optimizer differs from other strategies, the role of the risk model, and how Axioma’s method can be used as a starting point in the portfolio construction process.

What are the main advantages of using an optimizer over simply implementing a common ESG exclusion strategy?

Melissa Brown: An optimizer can quickly and accurately find the portfolio that most efficientlymeets input criteria. For example, the lowest risk portfolio that meets a given level of sustainability, or the highest exposure to a metric, or set of metrics for a given level of risk. It is able to exploit the relationships between assets or factors (such as industry association) to ensure the portfolio delivers the desired level of diversification at the lowest possible risk.

In contrast, a heuristic approach (eliminating the disallowed stocks and reweighting the rest) is far less likely to deliver an efficient, well-diversified portfolio, because it does not recognize portfolio exposures that can offset each other.

What are the main ways in which the Axioma optimizer differs from its peers?

Rob Stubbs: There are two types of peer optimizers: special-purpose portfolio optimizers and general-purpose optimizers that can be used for any problem, such as optimizing portfolios, supply-chains, fleet assignment, or any type of schedule.

The leading general-purpose optimizers have good underlying engines but have a couple of downsides. First, they are difficult to use because the user must formulate the optimization model themselves from low-level building blocks that require greater experience and skill to manage. Second, their methodologies are not tuned specifically for portfolio construction, so they generally do not perform well on hard problems involving taxes, limits on number of holdings or trades, or threshold holdings or trades, for example. Other special purpose optimizers generally contain weak optimization engines that do not facilitate advanced modelling and do not provide good solutions to difficult problems.

The Axioma Portfolio Optimizer provides the best of both worlds. It is built around a sophisticated optimization engine specifically tuned for portfolio construction problems. This allows for fast, quality solutions to even the most difficult problems and it can easily model portfolio constraints and objectives not found in other optimizers such as threshold trades, taxes, robust expected return, risk in variance or standard deviation, and nonlinear market impact.

Why should investors be so concerned with tracking error–is it an inherently bad thing?

Melissa: Tracking error (perhaps better phrased as active risk, since “error” implies something bad) can give an investor an idea of how much the return of the portfolio is likely to vary from the return of the underlying benchmark. For example, a tracking error of 1% implies that the return of the portfolio will be the benchmark return +/- 1% about two-thirds of the time, and +/- 2% about 96% of the time.

“Tracking error is not inherently bad or good, it is just one measure of how active the portfolio is as compared with the market.”

It can also help define the return one can expect for the level of risk taken, which makes a comparison across different portfolios more meaningful. Tracking error is not inherently bad or good, it is just one measure of how active the portfolio is as compared with the market. In the case of a sustainability portfolio (or any active portfolio), managers typically prefer to use their risk budget where they have an expectation of return. By using the optimizer to reweight the included stocks, more of that budget is freed up for the desired bets.

Is minimizing active risk considered more important when constructing sustainable portfolios?

Rob: There is nothing about sustainable portfolios that makes minimizing active risk inherently more important compared to any other type of active strategy. In most cases, we find that users of the optimizer want to either minimize active risk for a given level of exposure to sources of alpha, or to maximize the exposure to alpha (in this case sustainability = alpha), while maintaining a specific level of active risk.

In either case, investors should prefer to use their risk budget as wisely and efficiently as possible. If they can minimize the impact on active risk of exclusions, there is more risk available to use for tilting toward stocks with better sustainability profiles and away from those companies with poor sustainability characteristics.

The article continues on page 10 of our Special Report: 'Exclusions Evolved: Minimizing tracking error through intelligent optimization'