A very small team with finance and data science skills can implement a global systematic investment process that follows this general playbook: data gathering, universe definition, research and testing, signal generation, portfolio construction, and trading.
The first step in the process is data. Systematic investment management on a global scale requires large volumes of data, which needs to be available on demand. This can include market data like price and volume traded, classical data like company fundamentals, country and sector classifications, corporate actions, and novel data like tagged news, social media sentiment, satellite imagery of shopping mall parking lots, or commercial shipping lanes. Real-world data is raw, dirty, incomplete, and often suffers from lookahead bias. For example, a company may have a fiscal year-end of 31 December and report its earnings on 31 March the following year. Once the company financials are publicly available on 31 March, data vendors capture the information and backdate the data points to 31 December to align with the fiscal year-end, introducing lookahead bias. Systematic investing requires an extremely high-quality proprietary database and meticulous data engineering to ensure clean, point-in-time data, excellent coverage, and an analytics platform to facilitate easy data access.
The next step is defining the universe and requires some practical considerations. Leading on from the previous step, coverage and depth is critical. Any defined universe should have sufficient data to test and implement systematic investment strategies across geographies, sectors, and market capitalisation. Liquidity and tradability are crucial, and hard-to-trade instruments should be eliminated from the universe as these can degrade any potential upside due to their higher market impact. Another consideration is universe size and stability, as this can influence turnover and introduce unnecessary churn at the expense of potential future return.
The third step in the typical workflow of a systematic investment manager is research and alpha discovery. Typically, initial testing and development are conducted on a historical dataset sample, excluding some markets and time periods. For example, the sample may only include North America and Canada for the period covering 1985 to 2010 and limited to testing valuation characteristics taken from published financial statements. Suppose the hypothesis is not rejected during initial testing. In that case, the test is repeated using the markets and historical periods kept out of sample, all developed global markets, for example, over the period 1980 to 2020. These results are methodically and robustly tested for their sensitivity to changes in assumptions or scenarios.
In today’s markets, rarely is any single source of excess return significant enough to be the sole basis of any investment strategy. This is even more relevant in the context of a diversified global investment universe. The next step is signal generation and involves combining the signals that add value during the continuous research and discovery phase discussed above. These alpha sources must prove to be persistent, pervasive, robust, investable and economically sensible. To be persistent, it must deliver long-term positive drift, and performance should not be limited to a particular timeframe. To be pervasive, it must hold true over various regions, countries and sectors. To be robust, it should not be affected by changes in how it is defined. Being investable is extremely important, and it must be harvested cost-effectively to benefit investors. Furthermore, it must be sensible and easy to explain. Systematic investment strategies aim to combine the various sources of excess return in an objective and repeatable way, removing the uncertainty often associated with non-systematic fund managers.
The next step is portfolio construction and risk management. This is where the “ideal” meets reality. During this step, systematic fund managers aim to manage risks like market, geographic and sector exposure, and liquidity risk, to name a few. This step is also where the strategic objective is defined; for example, maximising each unit of return while simultaneously minimising each unit of risk. Imposing limits such as target risk or the maximum allocation to a single country, sector or company is also applied at this stage. When these objectives and limits are combined with the excess return signal in the previous step, systematic fund managers can optimise the trade off between risk and return and generate one or more target portfolios. Finally, it is possible to continuously monitor and test the target portfolios in a non-live environment to assess real world impact and performance prior to implementing in a live portfolio. This meticulous scientific method is in stark contrast to the ad hoc approach of a typical non-systematic fund manager, where the allocation to any country, sector or company is usually subjective and unscientific.
The final step is the trading and execution phase, which aligns the current portfolio with the new target portfolio. Global systematic fund managers, operating at scale with small teams, are nimble and agile and able to trade and execute across markets and multiple exchanges across time zones with accuracy and precision. These attributes ensure that clients always benefit from the best execution and lowest market impact.
COVID-19 has heightened emotions like never before as investors confront unparalleled levels of uncertainty. There is still a long way to go before financial markets and economies are likely to settle into any kind of equilibrium. Systematic investing provides a fact-based, rational and repeatable way of making it through these exceptionally volatile and challenging times for investors. BM
This article was written by Mario Fisher, Head of Equities | Chief Data Scientist
- Prescient Investment Management (Pty) Ltd is an authorised financial services provider (FSP 612).
- The information contained herein is provided for general information purposes only. This information does not constitute a solicitation, recommendation, guidance or proposal, and the service provided is not intended nor does it constitute financial, tax, legal, investment or other advice. Whilst reasonable steps are taken to ensure the accuracy and integrity of information contained herein, Prescient accepts no liability or responsibility whatsoever if any information is, for whatever reason, incorrect. Prescient further accepts no responsibility for any loss or damage that may arise from reliance on information contained herein. There is no guarantee in respect of capital or returns in a portfolio. Please note that there are risks involved in buying or selling any financial product, and past performance of a financial product is not necessarily indicative of the future performance. The value of financial products can increase as well as decrease over time, depending on the value of the underlying securities and market conditions.