Blog article
See all stories »

Global Asset Management in 2020

Amid unprecedented economic turmoil and regulatory change, most asset managers have afforded themselves little time to bring the future into focus. But the industry stands on the precipice of a number of fundamental shifts that will shape the future of the asset management industry. One such shift is globalization.

More than 40% of asset managers in developed countries looking to other countries for their long term future believe the most important geographical area of focus will be the South America, Asia, Africa, Middle East region, i.e. the SAAAME markets. This will provide opportunities for existing global asset managers to tap new pools of wealth and significantly expand their franchises. With a prediction of that AuM, globally, will rise to around $101.7 trillion, this opportunity will equally provide the backdrop for a number of fast-growing SAAAME-based competitors and region-specific regulations and challenges to manage and that will not only take on the global managers in SAAAME regions, but in developed markets as well. – Research by PwC

This article exemplifies how the global AuM opportunity presents a criss-crossed network with several moving parts, across asset classes and geographies. It presents the analysis of how principles of triadic closure & transitivity along with location-based features help make a projected AI solution of 2020, a matter of today.

The analytical framework ensures that tomorrow’s asset managers in global tier 1 banks are able to service their clients while accounting for accuracy, credibility, transparency, compliance to regional laws & regulations and time to decision.

A Pan-Regional Multidimensional Problem

Regulators are expected to leverage technology to pressurize asset managers who will continue to broaden their product ranges. By 2020, full transparency over investment activity and products will exist at all levels leaving no place for non-compliant managers to hide as regulatory, tax and alternative information’s reciprocal rights will extend across the globe. The worldwide network will be criss-crossed with a network of Tax Information Exchange Agreements (TIEA) encompassing all offshore financial centers into global tax data-sharing arrangements.

Disconnected Information

Asset manager serving globally will have to build extensive ‘Know your Customer’ (KYC) and anti-money laundering (AML) systems in order to capture key tax data provision, not only to tax authority where the manager and the fund reside, but also to each tax authority where investors reside. Local AML rules will include tax avoidance (and indeed aiding tax avoidance) as a money laundering offence, so asset managers’ customer handling teams will be required to be trained to spot and test for investor’s wealth to determine it has been generated by tax avoidance.

This will result in reconciliation nightmares for back office to align tax systems to geographies and grapple with a huge jigsaw puzzle of varying tax residency definitions for potential investors, as well as different bases of taxation of investment income and capital gains in each jurisdiction.

By 2020, four distinct regional fund distribution blocks will have formed allowing products to be sold pan-regionally. North Asia, South Asia, Latin America, and European blocks will develop regulatory and trade linkages with each other, which will transform the way asset managers view distribution channels. Management fee structures are also expected to evolve significantly as they will vary in shapes and form across misaligned distributors and investors situated globally.

Massively misaligned distributor-investor network will require massively connected systems. Global banks will need solutions that can take pan-regional handbooks and translate them in to machine-legible language easing the burden of meeting regulations while enabling asset managers to service clients in a timely manner.

A massively connected system of such order will need scalable knowledge graph solution providing the peripheral brain needed at the enterprise level. Overall, by 2020, there will be a significant focus on technology that can manage connected attributes to the nth order and handle it at scale to support asset managers even beyond just making the best use of data and provide new product solutions that are both tailored and interactive.

Framework for Contextual Learning

The concepts of triadic closure & transitivity have been around for decades but only recently have they come to prominence with the advent of affordable computing and AI. A graph database enhances AI by providing actionable context. Adding that rich layer of peripheral information around an AI solution helps produce better recommendation and outcomes for clients looking to make smarter long term investment decisions.

Given the end goal of serving a client, three key areas emerge where connected data helps draw better insights for an AI assisted Asset Management function.

Connected feature extraction - Drawing connected features from a prediction machine can be challenging. Current set of Machine Learning methods tend to simplify a problem with vectors, matrices and tensor spaces built for one degree of relationships overlooking more complex 2nd or 3rd degrees of predictive relationships and network data. With graph, you can do it without altering your current models of data pipeline and help with inferring relationships or where relationships don’t exist or sparse relationships using secondary and tertiary connections.

To find information on an unknown entity within a network of information in the quickest way requires taking all relationships out of the given set of nodes and extracting them numerically before applying them in a Tensor. AML, Credit Fraud, Compliance and investigations are among the most common examples asset managers will need to leverage a graph-based AI system.

Graph accelerated Machine Learning - Optimizing models for better context and efficiency presents the largest burnout in AI community. Additionally the time it takes to iteratively train a model is directly correlated to lack of quality and related data. So revisiting traditional practices such as table joins, space and directional matrices with efficient graph structures help with models like matrix factorization that thrive on connectedness. Connected data also helps with subgraph filtering to accelerate otherwise stalled Machine Learning pipelines such as community detection, clustering etc.

Transparency - Context is essential for credibility and ‘explainability’ and a huge value-add for asset managers as simple-to-explain products will benefit them with less time spent on explaining strategies to clients. Graphs introduces transparency (provides lineage of when, where and why certain data was accessed), explainable predictions (associating nodes in neural networks to a labeled knowledge graph allowing for traversing related documents to an explanation), and explainable algorithms (constructing tensors from graphs using weighted relationships) leading to credible neural network algorithms. These benefits also help responding to audits in a timely and efficient manger.

A Connected Geo-inclusive Intelligence Framework

PWC’s research, stated earlier, reveals that 40% of asset managers do not seek social, mobile, behavior or location data to build targeted products to reach more clients. Exploring alternative sources of data is still considered a frontier within Time Series-heavy modeling culture in asset management, however, this trend is quickly shifting.

Yazann Romahi, Managing Director and Chief Investment Officer at JPMorgan Chase states there are interesting avenues of building better valuation requiring cross sectional data more than just Time Series data. These are areas where building long-term investment strategies on hundreds of portfolio companies requires analysis of thousands of news articles, millions of customer sentiments, company asset/operation locations and their satellite images, billions of demographics and hundreds of alternative forms of risks.

These sets of information collected at massive scale would need to be analyzed within a distributed computing environment and placed on a graph to serve instant dashboards for asset managers to render a comprehensive universe of information to recommend the best investment strategy for clients.

While technologies like Big Data and AI help with scale and cost efficiencies providing stronger investor management and CRM capabilities, Graph based frameworks will deliver the flexibility and depth to enable investor reporting, disclosure accuracy and completeness and compliance with the plethora of overlapping tax, due-diligence, fee, regulatory and several other peripheral and subsidiary requirements that are to emerge with the oncoming globalization.

Asset Managers of Tomorrow

The ability to introduce several degrees of freedom for a global asset manager by eliminating the burden of lookups and synthesizing connected attributes has traditionally been a drag to the asset management industry as a whole.

Having data points available at the speed of thought enables asset managers to focus on doing what they prefer, which is looking after their client’s needs, whereas, lack of information or having to spend minutes to look up facts and figures can diminish a client relationship.

Emerging technologies like Smart Beings have already forayed into the territory of decision making for business and it’s only a matter of time when a graph-enabled and geo-inclusively intelligent technology will become a commonplace in the offices of asset managers.


If you are an Asset Management professional or have comments or questions please feel free to reach out to me at


To see other content and research from Chirag:



Comments: (0)

Member since




More from member

This post is from a series of posts in the group:

Banking Strategy, Digital and Transformation

Latest thinking in respect to Banking Strategy, Digital and Transformation. Harnessing our collective wisdom to make banking better. Ambrish Parmar

See all

Now hiring