How we build portfolios

All portfolios at pfolio are built using the Portfolio Builder, that provides a specific process to build a diversified portfolio while meeting various criteria.

This article explains this process on a high level. Separate articles zoom in to specific aspects of the process and provide more detail.

Broadly speaking, we break down the portfolio building process into a design phase and a building phase.

The portfolio design is like the blueprint of the portfolio and defines the rules about how the portfolio is build.

The portfolio building then applies those rules, selects assets and allocates to them to construct a portfolio. This happens every rebalancing period (monthly).

The portfolio design can be broken down into three parts.

The portfolio design steps (click for details)

Portfolio setup

The portfolio setup defines basic portfolio parameters such as the portfolio name and currency.

Asset selection

The asset selection defines the universe of assets from which the portfolio can be constructed. Users can include lists of assets, then use filters to apply criteria (e.g. only ETFs from developed markets).

Asset allocation

Defines the rules for asset selection and allocation per rebalancing period (monthly).

Based on the user’s settings, the algorithm selects a set of assets from your universe of assets and calculates the allocation to each asset from the selected set of assets.

The portfolio is constructed from a performance portfolio and a hedge portfolio (unless the hedge portfolio is disabled).

The purpose of the performance portfolio is to provide high returns. The purpose of the hedge portfolio is to reduce risk. Combining both portfolios aims to provide a balanced portfolio with a high return-to-risk ratio.

The portfolio building steps

• Portfolio optimisation: Settings for the portfolio optimisation, i.e. rules about how to allocate to assets once assets are selected for the period
• Constraints: Constraints for the optimisation problem, i.e. constraints about how to allocate to assets, e.g. do not allocate more than 20% to any asset‍
• Goals: Goals for the asset selection, i.e. rules about how to select assets per period

Selecting assets by goals

Per rebalancing period, a number of assets are selected from the universe of assets to become part of the current portfolio. Assets are selected by goals.

The image below shows an exemplary goal that would select the four assets with the highest Sharpe Ratio (weight 1) and mean return (weight 2) over the last 250 trading days.

Example: Goal to select 4 assets based on Sharpe Ratio and mean return

Users can add multiple metrics per goal. Adding multiple metrics per goal selects assets that rank high for all added metrics.

Users can add multiple goals per performance and per hedge portfolio. Adding multiple goals is useful for selecting different sets of assets, e.g. one set of high returning assets and one set of low volatility assets.

Allocating to assets based on portfolio optimisation

Once the assets for the current rebalancing period are selected in the previous step, a portfolio optimisation is performed to determine the optimal allocation to the selected assets.

How the portfolio optimisation is performed is determined by the settings in the Portfolio Optimisation and Constraints tabs.

• Modern Portfolio Theory: Allocates to assets based on Markowitz's Mean-Variance Optimisation, i.e. allocates to assets to maximise expected returns for a selected level of risk
• Hierarchical Risk Parity: Allocates to assets based on López de Prado's Hierarchical Risk Parity algorithm
• Equal Weight: Allocates equally to assets

Portfolio optimisation with Modern Portfolio Theory

The image above shows an exemplary portfolio optimisation with Modern Portfolio Theory. The labelled dots display return and volatility over the lookback period of the assets selected for the portfolio. The coloured dots represent the performance of various possible portfolios that can be constructed from the selected assets.

The curved line on the left edge of the coloured dots is the Efficient Frontier, made up of the portfolios that have the highest return for a given level of volatility. The theoretically optimal portfolio is the point on the Efficient Frontier with the highest return per unit of volatility, i.e. the maximum Sharpe Ratio.

Combining the performance and hedge portfolio

Once the portfolio optimisation of the performance portfolio has determined the asset allocation, a second portfolio optimisation is performed (provided the hedge portfolio is enabled).

Example: Asset allocation split by performance and hedge portfolio

The second portfolio optimisation determines the allocation to the assets in the hedge portfolio and the share of the performance portfolio vs. the hedge portfolio, thus allowing for the creation of an overall portfolio.