Developing AI-driven tools to maintain data standards and empirically validate investment outcomes for the managed accounts sector

The Australian funds management industry is seeing a major shift from collective investment products such as unit trusts to more individually oriented separately managed accounts (SMAs). New financial advice provided by financial planners now directs one in every two dollars of client money into the burgeoning $200 billion managed accounts sector. Responding to the lack of standards across SMAs for naming identical portfolios, calculating and presenting fees, classifying and reporting asset allocations, and tracking the roles of service providers, this collaboration will leverage new AI technologies that have powerful reasoning capabilities over large amounts of unstructured data. The outcomes are AI-enabled tools that provide rich user interfaces with the ability to continually integrate newly developed data standards and facilitate consistent SMA information provision to retail clients and institutional investors such as superannuation funds.

Despite its growing popularity, the SMA market is still struggling with identical portfolios having different names across platforms, fee structures that are impossible to compare, asset allocations that are reported using completely different methodologies, and inconsistent attribution of fiduciary roles between consultants/investment managers and platform operators. These problems pose major risks to financial advisers who are not able to meaningfully compare SMA products when selecting investments and tracking performance on behalf of investors. Led by the major players in the financial advice sector, SMA providers have adopted an industry-initiated set of data standards, the Separately Managed Account Standards (SMARS). The proposed PhD project interrogates gaps in the nascent standards, develops protocols and AI-driven tools for their maintenance, and performs a series of empirical studies utilising the newly standardised SMA data to provide evidence-based validation of investment outcomes, including risk-adjusted and post-fee performance rankings and performance attribution.

Based on rigorous data analytics and econometric studies, Project (a) will identify systemic gaps in the inaugural version of SMARS and develop AI tools for ongoing maintenance and updating of the standards and data collection protocols. Project (b) will improve the accuracy and timeliness of fund classification using machine learning algorithms to track SMA portfolio holdings and quantify the incremental value of implemented techniques compared to the base case of maintaining the status quo. Project (c) will utilize the corpus of clean SMA data produced by the proposal with machine-learning methods to exploit SMA characteristics to investigate the interactions between SMA characteristics and future performance. In Project (d) consistent SMA performance rankings will be produced and publicised. The project will be supported by the Fintech AI Innovation Consortium (FAIC), a new initiative building on the strength of digital innovation technology and AI capabilities of both the Engineering and Business faculties.

Who we are looking for:

• International or Domestic PhD candidate

Top Up Scholarship:

• 29K per annum top up scholarship for 3.5 years. Candidate must secure UNSW stipend.

How to apply: 

• Email Jerry Parwada with a copy of your CV and academic transcripts 29K per annum top up scholarship for 3.5 years. Candidate must secure UNSW stipend.