Custom equity portfolios are disrupting the way institutional investors such as pension funds, sovereign wealth funds and endowments invest. Now, these investors can improve investment results by fully aligning their portfolios with ESG and investment policies.
In this article we demonstrate how to use SigTech’s end-to-end platform to research, create and ultimately run custom equity portfolios in-house, with minimal disconnect between theoretical model portfolio and live trading. Watch our demo recording at the end of the article and see how to build custom equity portfolios using the SigTech platform.
Investing is about creating a product that 100% corresponds with the investor’s requirements, not about finding an existing product that offers the best possible fit. With custom equity portfolios, investors can tailor their investments according to their specific investment and ESG policies.
In order to build a custom equity portfolio, investors need to conduct quantitative research to create the underlying investment strategy, as well as having the processes in place to efficiently implement it. In this guide, we will demonstrate how to create three different equity portfolios step-by-step, as well as illustrating the importance of having the right technologies to integrate research and production in one system.
The main objective of backtesting an investment strategy is to capture a realistic insight into the strategy’s performance as if it had been traded live in the financial markets. The backtest also serves as a guide for the strategy’s expected future characteristics. To achieve trustworthy backtests, research must utilise validated and curated historical point-in-time data and properly mimic a realistic trading environment. Failing to address these issues (for example by using a simple timeseries approach) will result in performance discrepancies between backtested model portfolios and portfolios implemented live.
The solution to this problem is to use a unified platform, though historically this has been hard to achieve. Research and production environments are often written in different languages, making the transition from one to another tricky, and often requiring double implementation of a strategy. The SigTech platform represents a new and novel approach: offering an end-to-end platform that uses the same data and technology infrastructure for both research and production.
In this article, we will show how, in four straightforward steps, an integrated platform like SigTech can deliver bespoke, custom equity portfolios.
Receive our latest blogs, trading strategy ideas and insights
Step 1: Define the investment universe
The first step in creating custom equity portfolios is defining an investment universe. In this example, we apply a universe filter based on geography and market capitalisation. On a given rebalancing date, we select the 500 largest US stocks by market capitalisation. It is also possible to apply filters based on fundamentals (e.g. dividend yield > 2.5%), sectors (e.g. technology stocks only), country or region and technical indicators (e.g. applying a momentum score).
We also apply an ESG sector exclusion by ruling out stocks in the oil and gas industry. Exclusions are fully customisable on a sector, country and single stock level. In addition, the user can upload a list of stock identifiers to the SigTech platform, if desired. As shown in the graph below, the number of excluded stocks during the backtesting period varied between 20 to 40. The main reason for fewer exclusions over the last couple of years is the underperformance - and thus on a relative basis lower market capitalisation - of energy stocks that did not belong to the largest 500 stocks anymore.
Additionally, historical total return time series for all stocks in the investment universe need to be constructed. Most backtesting tools apply an approximate “weights times return” approach that does not take vital real-world aspects into consideration. SigTech takes a different approach: modelling the full trade lifecycle at instrument level. It treats historical corporate actions as they arose historically and adjusts the simulated portfolio’s stock and cash positions accordingly. In this example, we choose to reinvest dividends in the dividend-paying stock, apply a 30% withholding tax on dividend payments, and participate in all rights issues. In terms of modelling transaction costs, we apply a model where the transaction cost is represented as a combination of spread and market impact.
Step 2: Create the investment strategy
The next step is to construct an investment strategy that is customised to the investor’s unique requirements. Using the SigTech platform, strategies can either be constructed from scratch, or by using building blocks created by SigTech’s research team as a starting point. The python code used within these building blocks is accessible to end users and can thus be altered in any way the user sees fit, resulting in large efficiency gains in terms of time spent to develop and test an investment strategy.
In this example, we create three different strategies:
1. Market capitalisation-weighted strategy
The portfolio weights of each stock is a function of its relative free-float adjusted market capitalisation (MCAP), i.e. the methodology applied is the same as that used by traditional indices such as the S&P 500 and STOXX Europe 600. MCAP-weighted strategies tend to exhibit a tilt to the large cap and growth risk factors, as well as a skew against value.
2. Equal-weighted strategy
An equal-weighted strategy fully eliminates the relationship between price and weight by giving all stocks in the portfolio the same weight. In this example, with a starting investment universe of 500 stocks, each stock will therefore get an approximate weight of 0.2% at every rebalancing date. Compared to the MCAP portfolio, this strategy will exhibit a much larger tilt to the small cap risk factor.
3. Minimum volatility strategy
A minimum volatility strategy applies an optimisation to minimise the portfolio’s total risk level. When running the optimisation, constraints such as minimum and maximum weight for a given stock, as well as the lookback window for estimating covariance matrices need to be specified. In this example, we use a six month lookback window and apply a maximum weight of 2%. The strategy overweights stocks exhibiting low volatility and low relative correlation.
Step 3: Analytics
To review the various strategies historically, the SigTech platform gives access to a vast number of customisable analytics tools ranging from traditional risk/return statistics to graphical representations and factor contribution analysis. As an open platform, the user can fully customise which types of analysis to focus on.
Over the backtesting period, the market capitalisation- and equal-weighted portfolios exhibit the highest average returns. However, from a risk-adjusted perspective (i.e. Sharpe ratio), the performance of the three indices are actually aligned, which is explained by the lower risk level of the minimum volatility strategy.
Next, we show the annual returns of the three portfolios. In 2021, on the backdrop of a resurgence of the small cap and value risk factors, the equal-weighted portfolio has performed most strongly, whereas the minimum volatility strategy is the weakest performer in line with the current risk-on environment.
Other examples of analytical tools available on the SigTech platform include how correlation and tracking error structures change over time and the impact from the ESG exclusion list.
We can access detailed portfolio allocation data for any given day of the backtesting period showing the exact instrument the strategy would have been invested in. In table 3, the current sector allocation is displayed. It shows, amongst others, that the minimum volatility strategy - in line with expectations - exhibits a high allocation to conservative sectors such as consumer non-cyclicals and healthcare, whereas the MCAP strategy currently has a high exposure to the technology sector.
Step 4: Production
The last step is - from the investor’s perspective - the most important step: namely to implement the investment strategy in a real portfolio. This step is often more challenging than expected, commonly resulting in a disconnect between the performance of the theoretical model portfolio and the live portfolio. And inevitably, the impact is more often than not detrimental to the performance of the strategy.
SigTech’s end-to-end platform minimises this risk of negative slippage by using a unified environment for research and production. This ensures that the same data and technology infrastructure is used throughout the strategy development and production lifecycle. In addition, backtesting is conducted by replicating the full lifecycle at the instrument level, and trading costs are modelled in great detail within the research environment, serving to further bridge the gap between simulation and reality. With a few lines of code, a strategy can easily be deployed from research to a live environment, where seamless execution is offered by integrating trade order files with a portfolio or order management system.
Investors no longer have to settle for one-size-fits-all index products. New technologies, such as SigTech’s, are leading the way in empowering asset owners to build their own custom equity portfolios in-house, and improve investment results by fully aligning their portfolios with ESG and investment policies.