Published : 6 years ago, on 22/08/2018

Aberdeen Standard Investments has launched a new fund that utilises machine learning to identify sources of potential returns.

The Aberdeen Global (AG) Artificial Intelligence Global Equity SICAV, launched in Luxembourg, is the product of a collaboration between Aberdeen Standard Investments’ Quantitative Investment Strategies (QIS) team and Mitsubishi UFJ Trust Investment Technology Institute (MTEC)/Mitsubishi UFJ Trust and Banking Corporation (the Trust Bank) in Tokyo, Japan – a centre of excellence in robotics, artificial intelligence and financial technology.

The Fund embeds machine learning techniques within the investment process and will use a variety of quantitative techniques to time its investments.

These investments will be based on  ‘factor premia’ – those sources of risk such as value, quality, momentum, small size and low volatility that can provide investors with persistent risk-adjusted excess returns.

Martin Gilbert, Co-CEO at Standard Life Aberdeen, comments:

“For active investment management firms, the ability to use machines to read and understand vast amounts of data in order to forecast market moves more accurately has spawned innovation and a resurgence in active quantitative investment approaches. To benefit from this AI-driven innovation, and to complement our highly successful active fundamental strategies, we are proud to have collaborated with the MTEC – Japan’s leading and most prestigious financial technology think-tank – and the Trust Bank to develop this active quant Fund.

Junichi Narikawa, President of Mitsubishi UFJ Trust Investment Technology Institute (MTEC), added:

“This is the first time in MTEC’s 30-year history where we have collaborated with an entity in Europe and are pleased to work with a world-class investment firm of the calibre of Aberdeen Standard Investments. We have worked with their Quantitative Investment Strategies team in London and Edinburgh over a two-year period, and  developed a number of innovative AI-models to identify and capitalise upon patterns in global equity markets in order to dynamically time factor premia to generate alpha.”

David Wickham, Global Head of Quantitative Solutions at Aberdeen Standard Investments, comments:

“Recent innovations in AI, combined with rapid advances in computational power, have enabled us to harness machine learning techniques to dynamically time factor premia. This is an innovative AI-powered approach to factor timing, that enables us to systematically determine the weightings to each factor within the new global equity Fund and also allows us to time the relevant individual metrics used within those factors. We can now bias our portfolio towards the factors best suited to today’s market environment and continue to evolve the factor exposures as the market changes through time. 

“This new technique builds upon our existing diversified multifactor investing strategies – namely SMARTER Beta¹ and BETTER Beta²that we have successfully employed for over a decade.  We believe this is a unique approach within the mainstream investment community. At present, most AI products are either thematically focused, investing in well-known and relatively expensive AI-related companies – or ‘big data’ focused where investment managers attempt to extract alpha from unstructured data sources using machine learning techniques. We’ve created an elegant approach with great return potential by embedding machine learning within the investment process to enhance factor timing.”

This new capability is an extension of the firm’s existing factor investing strategies, amounting to USD 49* billion in assets, including the recently launched proprietary SMARTER Beta multifactor equity indices and funds and the BETTER Beta range of enhanced indexation funds. Each of these strategies embed an ‘ESG Inside’ methodology to exclude those companies engaged in producing controversial weapons and, where applicable, companies that are deemed to be experiencing severe ESG controversies as rated by its ESG data partner Sustainalytics.