Improved portfolio performance directly benefits both advisors and their clients. A systematic approach to portfolio construction brings time efficiencies, allowing advisers to increase their capacity to win, service and retain clients. Assessing and designing an investment portfolio in an uncertain world requires a resilient and well-thought-out approach. Diversification benefits are enhanced if the portfolio builder can harness high-performing portfolio constituents. Not all expertise is available in-house, so it often makes sense to engage domain experts to solve bespoke problems outright or to give guidance on the use of advanced portfolio construction tools.

We know that rigorous portfolio construction has been key to risk management and long-term sustainability of any investment portfolio ever since Markowitz’s 1952 seminal paper.

Mean-variance is a compelling framework for portfolio selection. It is no wonder that many investors, ranging from university endowments to online robo-advisors, have turned to mean-variance analysis as the primary asset allocation model.

As with any model, simplifying assumptions both increase the model’s utility and detract from it. And these assumptions especially limit the ability of mean variance to model real-world portfolio requirements.

Viking Harbour has access to technology designed to streamline the portfolio construction process. A client wishing to create an investment portfolio can work with Viking Harbour to access the construction tools with which to build robust portfolios of desirable constituents. We can offer solutions with varying levels of sophistication for most encountered portfolio construction problems: AI-clustered risk parity, mean-variance, tactical asset allocation, principal component portfolios, performance metric driven & growth optimal including tracking portfolios for the passive investor class.

Robustness is key to real-world applicability of any solution. To this end, practical constraints are easily incorporated into portfolio problems, covering box type, drawdown, sparsity, cardinality, turn-over/cost, sectorial & tracking error. Using the latest numerical methods in robust statistics and convex optimization gives our clients the ability to benefit from realistic, robust, bespoke solutions.

Risk Parity is a robust and straight forward approach, namely, a portfolio where the constituent strategies deliver the same level of volatility to the terminal portfolio. An elegant characteristic of risk parity portfolios is they are mean-variance optimal if the underlying strategies are un-correlated.

Viking Harbour can apply distinct enhancements on the Risk Parity portfolio approach.

Firstly, *T**ilted Risk Parity*, which allows us to express a view on the relative allocation of the underlying assets. A good example would be for a combination of equity and bond assets. We may wish to limit the total risk contribution of the bonds to no more than 25% of the portfolio. However, because of the differences in volatility between equities and bonds this could correspond to a solution with around 60% notional in bonds. Your awareness of the portfolio characteristics can be enhanced by this explicit demonstration that the risk contribution of the bonds is different to the notional allocation.

Secondly, *C**lustering Risk Parity*, is a generalisation of the volatility equalization procedure. The first step consists of clustering the assets into sub-groups of similarly behaved clusters of assets. The streams within each cluster are Risk Parity weighted to deliver a target volatility on each cluster-portfolio. In the second step, said risk-adjusted cluster-portfolios geared to first equalize volatility between them and then to meet the terminal portfolio target volatility.

Another allocation method is called *Clustering Risk Parity**. *This is an algorithm that begins by grouping individual assets that behave similarly to each other. Assets here can refer to single name equities, bonds, ETFs, etc. The final steps of Clustering Risk Parity involve treating those groups of assets, the “Clusters”, as mini portfolios in their own right. At the last stage, each Cluster (mini-portfolio) is sized to achieve: a) an equal contribution to risk (volatility) from each cluster, and, b) achieve a desired target volatility of the final overall portfolio.

Given Viking Harbour’s expertise and its relationships with experienced technology providers, we can help advisors to navigate the complexity of the portfolio construction process. Even small improvements in allocations and diversification can compound to provide a more robust value proposition for advisor’s clients. And the systematic allocation approach, using well established quantitative technology, can reduce the time required to test new allocations, alternative assets and illustrative scenarios.