Different assets require different data

With the Reserve Bank of Australia still struggling to bring rising prices to heel, it might seem jarring to be considering the impact of low inflation on the investment world. Nonetheless, historically low interest rates following the global financial crisis (GFC) explain some of the data issues now faced by investment managers. As central banks tightened monetary policy, investors went in search of yield from an increasingly diverse range of asset classes.

Private markets in particular grew at an unprecedented rate. Pre-GFC, global private assets under management peaked at around US$2.2trn. As of 30 June 2023, that figure stood at US$13.1trn, having grown several times faster than public market capitalisation over the same period. This has resulted in a unique set of data challenges.

Private market investments include assets like real estate, venture capital, private debt, private equity, infrastructure, farmland, and forestry. They are generally valued far less frequently than public assets and less likely to come with a standard data format. For instance, the set of data required to describe shares in Apple say, is far more standardised than the data required to describe a stake in an oil pipeline or a pine plantation. Investment managers must therefore deal with data from a greater variety of sources than ever, many expressed in ad hoc and inconsistent formats. Spreadsheets abound.

The need to drill down into many layers of data

As part of its Superannuation Data Transformation, APRA is contemplating requiring superannuation trustees and their investment managers to provide “look-through” reporting on unlisted (that is, private) assets. Trustees will have to be able to provide data in aggregate, as well as at the individual investment level.

The Australian Government’s climate-related financial disclosure regime looks set to require fund managers with assets exceeding A$5bn to report on greenhouse gas emissions from all sources. Reporting on scope 3 emissions (emissions resulting from activities that an organisation indirectly affects in its value chain) will require investment managers to be able to drill down through many layers of data to understand their underlying asset holding.

Following the invasion of Ukraine by Russia in 2022 and the imposition of western sanctions, Australian investment managers scrambled to divest from Russian-linked assets. Due to complex ownership hierarchies, many struggled to identify these assets amongst their holdings.

All of these problems point to the same need. Investment managers need to be able to look through or drill down into their investment data to understand who owns and manages what assets at all levels of the hierarchy.

Creating this kind of view often presents a significant challenge. As mentioned earlier, different asset classes with very different data formats add complexity. And, like other types of organisations, investment managers have their fair share of legacy systems and processes that make aggregating data difficult.

Why data management platforms offer a solution

Data management platforms (DMPs) are essential to solving investment managers’ data problems. The following aspects of a DMP are crucial:

  • Canonical data model
    The data management solutions article in Novigi’s current Quarterly, talks about the importance of system and source agnostic data models. This is especially so in investment management. Investment management DMPs need a data model that is flexible and extensible enough to accommodate all the variation in descriptive data across different asset classes. Enabling look through is in part a data model problem — the complex hierarchies of ownership need to be logically represented in the model in a consistent and retrievable way.
  • Automating ingestion
    This is a great use case for AI. Where investment managers have access to feeds that deliver data in a consistent format and to a regular frequency, automating ingestion is relatively straightforward. Unfortunately, this is often not the case. Data on private assets is especially likely to be manually generated and unstructured. Using AI to interpret images, recognise text, and comprehend natural language can allow us to automate the ingestion of even this data, where in the past that process would have been manual.
  • Data quality tooling
    Given the range of formats and varying quality of the source data that investment managers contend with, automating data quality screening yields substantial benefits. If data ingestion is to be reliably automated, picking up data quality issues on the way in will allow for manual intervention only when necessary.

While these strike us as the most important elements, obviously investment managers need to think about this as part of a broader DMP implementation and strategy. And in doing so they will invariably find that they need to have appropriate data governance in place, and ensure they are doing the right things to ensure data hygiene and security.

Investment managers face increasingly complex data management challenges, driven by the growth in private market investments and evolving regulatory requirements. As they navigate diverse asset classes and inconsistent data formats, the need for robust data management platforms is critical. By adopting some increasingly best practice approaches to data management, investment managers can effectively address these challenges and gain a better understanding of their portfolios.

Investment data is one of the topics explored in Novigi’s Quarterly Report for Q4 FY2024.