Regulators and banking leaders at a roundtable hosted by Regulation Asia and Nasdaq discussed strategic opportunities and operational hurdles of the GDR 3.0 initiative.

The Hong Kong Monetary Authority (HKMA) has called for deep collaboration with the banking industry to navigate its ambitious Granular Data Reporting (GDR) 3.0 initiative, a multi-year overhaul of its data collection model.

At a strategic roundtable hosted by Regulation Asia and Nasdaq on 30 July, senior bankers and regulators convened to discuss the path forward. The dialogue revealed that while the industry broadly accepts the strategic necessity of the shift, significant challenges remain around data governance, cross-border complexities, and the cultural changes required within financial institutions.

The GDR 3.0 initiative, announced in April 2026, represents a fundamental move away from aggregated, template-based returns towards direct, granular, and potentially on-demand data submissions. For the HKMA, the objective is to create a more data-driven and technology-empowered supervisory regime.

Strategic opportunity

HKMA representatives speaking at the event emphasised that GDR 3.0 should not be viewed as just another compliance exercise, but as a strategic opportunity for the entire banking sector. Alvin Li, Head of Supervisory Technology at the HKMA, positioned the initiative as a core component of the regulator's new 'Risk Data Strategy'.

"We don't want this to be just another compliance reporting exercise," he said, urging banks to build a data foundation that serves not only GDR 3.0 but also their own future business and risk management needs. "We're hoping that the investment is not just for reporting to us."

The regulator is already leveraging granular data from the earlier GDR 1.0 and 2.0 phases, having built an internal AI model using loan data to forecast credit quality. The vision for GDR 3.0 is to scale this capability across more data areas.

Li connected the initiative to the broader evolution of the financial industry, including the wider adoption of AI and the strategic business model reviews currently underway at banks.

A robust, well-governed data layer, he argued, is not only a key component to meeting the objectives of GDR 3.0, but also the essential foundation for future AI-enabled services and data-driven business models that can better serve customers.

Subbaiyan Vaithinathan, Global Head of Product & Solution Engineering at Nasdaq, reinforced this view, placing the HKMA's initiative alongside similar transformation programmes by other banking regulators in Europe, the US, and other parts of Asia.

He said the era of aggregated reporting is "slowly fading away”, presenting an opportunity for banks to break down internal silos, improve decision-making, and conduct stress testing much faster than was previously possible.

Co-design process

A key theme of the discussion was the HKMA's commitment to a "co-design" approach, as outlined in the April circular. This is intended to bridge the gap between regulatory requirements and the operational challenges facing banks.

Steve Li, Senior Manager of Supervisory Technology at the HKMA, who is leading the GDR 3.0 implementation, explained that this approach was primarily a lesson learned from previous phases. "Our objective is to try to close the gap between regulatory requirements and operational reality," he said.

He outlined a multi-stage engagement plan, which has already begun with informal, bank-specific meetings. This will be followed by a formal, industry-wide consultation paper planned for release around September 2026, which will be open for six to eight weeks.

The goal, he said, is to finalise the implementation plan by the end of the year, with banks expected to begin system and process enhancements in 2027. He emphasised that the industry engagement is a continuous and iterative dialogue rather than a one-off consultation.

In a direct response to industry feedback, Steve Li said the HKMA is also planning a new "implementation assurance" stage. Described as an exploration phase rather than a User Acceptance Test (UAT), this phase will allow banks to submit real production data into a secure, production-ready environment before the official parallel run begins.

This will enable the regulator and the industry to test their systems and fine-tune parameters like validation rules and tolerance bands without the pressure of a formal sign-off by a bank’s accountable officers.

"We hope all these potential issues will be identified during the implementation assurance period, so that both sides can adjust," Steve Li explained, adding that this should lead to a much shorter parallel run of two to three months, compared to over a year in past experiences.

Governance and data quality

The discussion brought to light industry concerns about data governance and quality as foundational challenges. One banker from a global institution noted a need for more "directional guidance" on data governance to prevent a scenario where "everyone is thinking about their own internal way of governing the data."

In response, Alvin Li said the HKMA would aim to share "good practices" and examples of effective governance models for different types of institutions, rather than dictating a rigid, one-size-fits-all framework.

A more fundamental challenge raised was the cultural and procedural shift required to ensure data quality at the source. A participant from a major regional bank explained that under the old template-based system, errors could be corrected with a last-minute "overlay" before submission. With granular reporting, the process is far more complex.

"What could be a single reporting line in the current report, could be a thousand-plus transactions behind it, and getting people to fix it takes time, " the participant said. He explained that this necessitates a move from identifying faulty systems to identifying data owners, a change that requires significant internal adaptation.

This sentiment was echoed by another banker, who noted that the initiative will "pass across so many functional units”, making the task of assigning ultimate ownership a matter of extensive internal collaboration and lobbying.

Subbaiyan Vaithinathan of Nasdaq explained that AI can be leveraged to tackle these very challenges. "AI can help trace lineage across the full lifecycle within the regulatory ecosystem – spanning complex calculations, multiple regulatory versions, evolving data versions, and post-submission adjustments – to produce clean, user-readable lineage and surface anomalies the moment they appear.” he said. “This enables firms to quickly identify governance gaps, pinpoint control weaknesses, and continuously improve the entire data control framework."

Acknowledging the challenges facing the industry and requests for sufficient time to adapt, Alvin Li said this was a key reason the HKMA has pushed to initiate the GDR 3.0 project early, decided on a ‘co-design’ approach, and is already providing the draft requirements to selected banks and engaging with the industry to solicit feedback.

The idea is for the industry to start setting up project teams and secure the required funding, resources and approvals now, rather than waiting for the final package of requirements to be issued.

Alvin Li said the HKMA will work to share additional information and documentation that can help drive forward internal discussions at banks to help them “get the project initiated earlier”.

Global and third-party challenges

The discussion also highlighted the complexities faced by global and regional banks operating across multiple jurisdictions. A participant from an international bank described the challenge of gaining support from its head office, which operates in a jurisdiction with less advanced reporting requirements.

Another questioned whether there was scope for greater coordination between regulators through bodies like the Basel Committee to harmonise requirements. Alvin Li acknowledged that while such efforts are made, the different paces and priorities of jurisdictions make it a "not very easy process."

The issue of cross-border data restrictions was also cited as a key concern. A participant pointed to China's Personal Information Protection Law (PIPL) as a potential barrier to submitting granular data from mainland subsidiaries.

Steve Li stated that the HKMA is aware of such legal constraints. He said the regulator will define the requirements cautiously and work with banks to find solutions to address such constraints.

Alvin Li likewise stated that the GDR 3.0 data collection would be "to the extent allowable by relevant jurisdiction”, indicating a built-in allowance for such legal barriers, potentially permitting aggregated submissions for affected portfolios on a case-by-case basis.

Another participant raised the growing dependency on third-party data providers for processes like KYC and credit evaluation. Alvin Li framed this as a strategic decision for banks, weighing the benefits of outsourcing against the need to retain control over core competencies and manage operational resilience.

The role of AI

The transformative potential of AI was another recurring theme during the roundtable session.

Subbaiyan Vaithinathan of Nasdaq said the availability of mature AI is a key differentiator between GDR 3.0 and past transformation programmes. He cautioned, however, that success in regulatory reporting depends on AI that is purpose-built for the domain, explainable, and auditable - so that every output can be traced, validated, and defended to regulators, rather than treated as an opaque, unexplained result.

The HKMA's Alvin Li stated the regulator sees the GDR 3.0 data foundation as a crucial enabler for banks to adopt AI, particularly in establishing the governance needed to manage the unstructured data required to power generative AI models.

As the roundtable concluded, the consensus was that GDR 3.0 is an ambitious project, where success will depend not just on technology, but on a deeper cultural shift towards more robust data governance and ownership practices.

--

This article and the roundtable were produced by Regulation Asia in collaboration with Nasdaq.