Beyond data collection: APRA's data modernisation agenda
In April this year, APRA announced it had chosen an international consortium to deliver its new Data Collection Solution – a piece of technology that would be used by around 4,500 companies across Australia’s financial sector to meet their data reporting requirements.
The Data Collection Solution will replace APRA’s current Direct to APRA (D2A) tool, which was developed in 2001 and is rapidly approaching its end of life. The choice of vendor was the culmination of an intensive program of industry engagement and requirements gathering aimed at ensuring the new system would best meet the requirements of APRA and financial sector companies into the future.
What’s not widely known is that the Data Collection Solution is only one part of a broader data modernisation program designed to ensure APRA keeps pace with advances in data, analytics and technology. APRA is also overhauling its processes and infrastructure for storing, analysing and delivering data, with future benefits for both industry and the wider public. Collectively, the data modernisation program represents the largest and most complex investment in new technology that APRA has undertaken since it was first formed 20 years ago.
APRA’s broader data strategy
APRA collects data to support its core supervisory work and inform policy development.
Analysing the data enables our frontline supervisors and risk specialists to gain the insights they need to gauge the soundness and stability of the institutions they oversee, and ensure the financial interests of bank depositors, insurance policyholders and superannuation members are protected. APRA provides some of the data it collects to a number of other government agencies, including Treasury, the Reserve Bank of Australia (RBA), the Australian Bureau of Statistics (ABS) and the Australian Securities and Investments Commission (ASIC). The availability of accurate, timely and sufficiently granular data is therefore essential to safeguard the soundness and stability of the Australian financial system.
With the ageing D2A system reaching its end of life, we identified an opportunity to launch a wholesale upgrade of our data collection and analytic capabilities. We developed a data strategy based around the data lifecycle: source, use, manage, share and govern. This is consistent with peer agencies, such as ASIC. The strategy is reflected in our four separate project streams tackling data collection, storage, access and innovation.
Due to be available in 2020, the new Data Collection Solution will replace D2A. We are currently working with industry to support their readiness for the new solution.
Once implemented we will focus our efforts on harnessing increased capability from the new solution. We expect in the coming years that the new solution will allow APRA to:
- take a more flexible approach to the types of data we collect as our interests evolve. For example, we will be able to gather more granular data to support supervision. APRA is currently taking steps in this direction by partnering with ASIC on data collection topics of joint interest, including ASIC’s recurrent data collection pilot of loan level information on residential mortgages;
- define requirements through data models rather than forms to reduce the reporting burden at the same time as increasing flexibility in the use of the data;
- make changes to reporting requirements as needed in the future. When we do, we will look at ways to use the data for multiple purposes, creating opportunities across government to reduce duplication.
Our policy priorities to update the reporting requirements in the short-medium term will be aligned with our policy priorities as outlined in our information paper released in February 2019.
In 2018, we completed the build of a new Enterprise Data Warehouse (EDW) to replace our outdated data warehouse, which had limited capabilities and was built on multiple technologies. The new EDW has created a single source of truth, and centralises entity submitted data with other enterprise data sources. We are progressively transitioning data feeds in and out of EDW, including preparing for the new Data Collection Solution in 2020, and decommissioning legacy systems. The overlay of revised data governance and processes will further ensure that we can be confident in the quality of the data going in and out of this central source of truth.
To illustrate the transformation, imagine what data looked like 20 years ago. APRA’s systems were set up to collect a defined selection of form-based data, which grew over time in line with regulation. Our data warehouse was expanded by bolting on to the original setup to meet changing needs. Fast-track to now, with vastly greater processing power and capabilities of data systems. Having overhauled our data storage, we can now look forward with our new EDW to embrace data opportunities into the future
Once the data is securely and centrally stored, our analytical and supervisory teams must be able to access, model and visualise data accurately and easily. We are developing new standardised and strategic insight reports to produce the information required.
We are also introducing new tools and capabilities to enhance our analysts’ ability to develop customised reports to suit current and future trends, and to allow deeper analysis of data. By reducing the time needed to prepare data for analysis, we can focus our analysis on areas that make a difference to the Australian financial system and the Australian community.
APRA is also investigating different methods of providing data to external users of APRA’s statistical publications, including regulated entities, industry analysts and the general public. APRA’s longer term goal is to present data in more accessible and flexible ways that enable greater usability by a broad audience of stakeholders.
We have set up a new data science capability – APRA’s Innovation Centre Laboratory (ICL) - with a small team of data scientists who partner with different parts of the organisation on experimental data research. The ICL is an environment to experiment with information and test answers to business questions, and we’ll soon have a new platform to share this capability across our organisation.
One example of an innovative approach to data analysis is the use of artificial intelligence techniques such as natural language processing and machine learning to process and analyse documents from entities more effectively.
Towards a data-driven future
Whether you’re a regulator, a business, or an investor, making sound financial decisions requires timely, reliable and detailed information. The rise of advanced data analytics has seen an explosion in the use of algorithms across the business world, including the financial sector, yet this trend is only in its infancy.
Like the companies we regulate, APRA cannot afford to fall behind the technological curve. Our data strategy sets up a road map for APRA to evolve how data and analytics enhances supervision. Like many other organisations who rely on data from legacy systems, we have recognised the significant amount of resources needed to manipulate our data. This problem has largely driven our decision to undertake the data modernisation program. Replacing repetitive, time-consuming, manual data manipulation methods will allow our supervisory teams to concentrate on investigation and action rather than processing data; our risk specialists will be able to better model strengths and weaknesses in our financial system; and our data scientists will be able to train models to draw out thematic insights. The end goal is to improve the quality of data collected, boost the efficiency of report production, provide better, faster, deeper insights, and deliver greater transparency to financial analysts, policy makers, and the general public.
We will provide updates on the data modernisation program on our website and in this publication in the future. Read more about the implementation of the new Data Collection Solution.
This article was first published in APRA Insight - Issue 1 2019.