How Hong Kong's Courts and Registries Ended Up Buried Under Duplicate Image Files — and What It's Costing the City

Hong Kong's government digital infrastructure is carrying tens of thousands of duplicate image files across at least a dozen public-facing document systems, a problem that administrators at the Land Registry in Queensway and the Companies Registry in Queensway Tower have been quietly grappling with since at least 2022. The issue is not dramatic — no data has been lost, no breach has occurred — but the cumulative drag on retrieval speeds, storage procurement and staff time has become a tangible cost as the city pushes to position itself as a smart-city rival to Singapore.

The duplication problem matters now because the Hong Kong government's Digital Policy Office, established in July 2023, has set a rolling deadline to migrate legacy document repositories onto its central cloud platform by the end of 2026. That migration is forcing bureaux to audit what they actually hold. What they are finding, according to publicly filed tender documents on the Government Logistics Department's procurement portal, is that image-based PDF scans of paper forms were systematically saved multiple times at different stages of processing workflows — once on intake, once after verification, and again after stamping — with no automated deduplication step built into the original systems.

How the Problem Accumulated

The roots go back to the early 2000s, when individual bureaux procured their own document management systems under the then-Information Technology and Broadcasting Bureau, with little coordination. The Inland Revenue Department in Wan Chai, the Immigration Department in Wan Chai Tower, and the Land Registry each ran separate platforms with separate storage conventions. When those systems were updated in the mid-2010s, migrated records were frequently duplicated rather than consolidated, because the migration scripts matched files by name rather than by content hash.

A 2019 audit by the Office of the Government Chief Information Officer estimated that across surveyed systems, between 18 and 24 percent of stored image files were exact or near-exact duplicates. That figure, cited in a Legislative Council panel paper on IT infrastructure tabled in January 2020, translated at the time to roughly HK$34 million in excess storage capacity being maintained annually across the surveyed departments. The Covid-19 pandemic then accelerated digitisation without resolving the underlying architecture, adding new layers of scanned documents — court filings, licensing applications, vaccination records — on top of the existing redundancy.

The Judiciary's own e-filing system, which expanded significantly in 2021 to cover the High Court on Queensway and the District Court on Argyle Street in Kowloon, ingested large volumes of scanned exhibits and affidavits that legal firms submitted in both searchable PDF and image-only TIFF formats, sometimes simultaneously. Court administrators flagged the resulting storage anomalies in internal memos, but no cross-departmental working group was convened to address the issue until the Digital Policy Office came into existence.

The Path to Resolution

The current remediation plan, outlined in a tender notice published on GovHK in March 2026, calls for a phased deduplication exercise using content-addressable hashing tools. The contract, valued at up to HK$12.8 million, covers seven departments in the first phase and is expected to conclude by March 2027. The Land Registry and the Companies Registry are in scope for phase one; the Judiciary's systems are slated for phase two.

Organisations and individuals who regularly interact with those registries — law firms in Central, property agents on Des Voeux Road, and company secretarial firms clustered around Exchange Square — should expect intermittent slowdowns in the eSearch and e-Registry portals during scheduled maintenance windows that will run on selected Saturday mornings from September 2026 onward, according to the published project timeline.

The broader lesson the Digital Policy Office appears to be drawing is structural: future document intake systems procured under the new Government Cloud framework will be required to run deduplication checks at the point of ingest, not retrospectively. Whether that requirement survives contact with individual bureau procurement cycles, which have historically favoured vendor lock-in over interoperability, will determine whether the city is still having this same conversation a decade from now.