Where treaty access between Hong Kong and the CIS stands now

Treaty access between Hong Kong and the Commonwealth of Independent States – the loose grouping of post-Soviet economies that spans Russia, Kazakhstan, Ukraine, Azerbaijan, Uzbekistan, Belarus and their neighbours – turns on a question that is frequently misread: Hong Kong does not automatically inherit the People's Republic of China's treaty network, and the CIS jurisdictions that have signed agreements with Hong Kong have done so under terms that now face active scrutiny from tax administrations on both sides. The governing instrument is the bilateral agreement between Hong Kong and the relevant CIS state, each concluded by Hong Kong under Article 151 of the Basic Law as a separate tax jurisdiction. The correct starting point is therefore not Beijing's treaty position but Hong Kong's own, narrower network – and the substance-and-source conditions that Hong Kong's territorial system places on any outbound or inbound payment stream before a treaty rate can be claimed at all.

This analysis examines the commercial stakes, the governing architecture, and the risk points that our desk identifies as live in 2028 – a period when the practical utility of the Hong Kong–CIS treaty interface is being tested by restructuring pressure, heightened documentation requirements, and a shifting geopolitical backdrop that has already changed the payment corridors principals are actually using.

What is commercially at stake?

The answer is straightforward: the difference between a treaty-rate withholding on a dividend, royalty or interest payment and the domestic withholding rate in a CIS state can span fifteen to twenty percentage points or more. For a group running an operating entity in Kazakhstan, Uzbekistan or Azerbaijan – with a Hong Kong holding or intellectual-property vehicle sitting above it – that difference compounds over years of distributions. The treaty question is not academic. It is a cash-flow decision made at the time the structure is drawn, and a compliance risk that runs every year the structure is maintained.

Our cross-border practice regularly sees three commercial configurations where the Hong Kong–CIS treaty question is live. First, a manufacturing or trading group with CIS operations that moved its holding entity to Hong Kong in the period before the current geopolitical disruption, and now needs to assess whether treaty access still holds. Second, a family-office principal with income streams from Kazakhstan or Azerbaijan who is using a Hong Kong holding vehicle as the collection point. Third, a group that previously routed through a European holding jurisdiction and is now re-examining Hong Kong as the surviving accessible common-law vehicle following the disruption of that route.

The commercial pressure on that third configuration is acute. Windows close. The question for any principal in that position is whether a Hong Kong holding entity can inherit, or newly establish, a treaty-compliant position before the restructuring cycle closes.

How does Hong Kong's treaty network reach into the CIS?

Hong Kong maintains its own tax treaty network, negotiated and concluded separately from China's treaties, under the authority granted by Article 151 of the Basic Law. Hong Kong is not a signatory to the Vienna Convention on the Law of Treaties as a state, but its bilateral comprehensive avoidance of double taxation agreements (CDTAs – Hong Kong's term for full double-taxation treaties) and limited tax treaties (covering airline and shipping income only) are internationally recognised as binding bilateral instruments between Hong Kong and the counterparty state.

Within the CIS space, Hong Kong has concluded CDTAs with a subset of jurisdictions. The network is materially smaller than China's. A principal who assumes that a Hong Kong holding entity gives access to the same treaty benefits that a China-based entity would enjoy is wrong on the structure, and the error is consequential. The relevant question is always: does Hong Kong have a CDTA with this specific CIS state, and does that agreement cover this specific payment type at a rate lower than domestic withholding?

Several CIS states have no CDTA with Hong Kong at all. For those jurisdictions, domestic withholding applies in full, and the treaty question falls away. The analytical work concentrates instead on substance, source and the anti-avoidance provisions of the relevant domestic law – a point to which we return below.

Where does Hong Kong's territorial system cut across the treaty claim?

Hong Kong taxes profits on a territorial basis: 8.25% on the first HK$2,000,000 of assessable profits, and 16.5% above that threshold, but only on profits that are Hong Kong-sourced. Profits that arise offshore are outside the charge entirely – unless the foreign-sourced income exemption (FSIE regime, which imposes economic-substance conditions on certain passive-income types received by Hong Kong-resident entities) brings them within scope.

The FSIE regime, in force from 1 January 2023 and since amended, changes the default. Passive income – dividends, interest, disposal gains and intellectual-property income – received by a Hong Kong-resident entity from an associated entity is now subject to a substance test. If the Hong Kong entity does not meet the relevant substance condition, the income is deemed taxable in Hong Kong regardless of its offshore source.

Why does this matter for treaty access? Because the CIS counterpart state's tax administration is looking at the Hong Kong entity and asking whether it has sufficient presence and function to be a resident of Hong Kong for treaty purposes. If the FSIE analysis concludes that the entity lacks substance, the Hong Kong entity's treaty-residence claim is simultaneously weakened on both sides: Hong Kong can assert a tax, and the CIS state can deny the treaty rate. The FSIE regime and the treaty-access question are therefore not parallel tracks; they interact, and the weak point in one exposes a risk in the other.

What does economic substance actually require in the Hong Kong–CIS context?

Substance requirements for Hong Kong entities holding CIS income streams depend on both the FSIE conditions and the treaty's own limitation on benefits or principal purpose test provisions. The FSIE conditions for dividends received by a Hong Kong entity from a related CIS operating entity require that the Hong Kong entity meets at least one of three conditions: economic-substance condition, participation condition, or subject-to-tax condition. In practice, for most holding structures over CIS operating entities, the economic-substance condition is the relevant one unless the entity holds the participation stake directly and the subject-to-tax condition applies in the CIS state.

Economic substance in the FSIE context requires, at minimum, that the entity is managed and operated in Hong Kong. This means board meetings held in Hong Kong with directors physically present, strategic decisions made in Hong Kong, and adequate employees and premises proportionate to the entity's activities. A shelf company with a nominee director signing in Hong Kong in name only does not meet this test. The Inland Revenue Department has issued guidance on the conditions, and its enforcement posture has hardened since the FSIE regime came into force.

From the CIS side, the analysis mirrors this. Kazakhstan, for example, has its own domestic anti-avoidance rules that require a beneficial-ownership inquiry before treaty rates are applied. Azerbaijan and Uzbekistan have similarly introduced or strengthened their domestic provisions in recent years. The convergence of domestic anti-avoidance rules on both sides of the structure means that a thin Hong Kong entity facing scrutiny from a CIS tax administration is exposed from two directions simultaneously.

What foreign advisers sometimes get wrong at this juncture is assuming that a clean Hong Kong incorporation certificate plus a certificate of residence from the Inland Revenue Department is sufficient to satisfy a CIS administration. It is not. A certificate of residence confirms that the Inland Revenue Department regards the entity as tax-resident in Hong Kong; it does not itself address the substance conditions that the CIS state will apply under its own domestic anti-avoidance provisions. Those provisions operate independently of the treaty text, and the CIS administration is entitled to apply them. The distinction between treaty residence and domestic beneficial-ownership or substance analysis is where the practical risk sits.

The comparative read: what separates a defensible position from a vulnerable one?

In our cross-border practice, the positions that survive CIS tax-administration scrutiny share a recognisable profile. The Hong Kong entity has a genuine decision-making function: it makes the decisions that matter to the business, not just the decisions that its advisers have asked it to make. Its directors are people with relevant qualifications and time available to engage. It has banking in Hong Kong, not merely a Hong Kong address for correspondence. And its books reflect the economic reality of what it does, not a compliance fiction constructed after the fact.

The positions that do not survive tend to share a different profile. The entity was set up to hold a CIS operating company because Hong Kong had a treaty with that state, or because Hong Kong was "clean" after the European route closed. No further thought was given to substance. The directors are nominees. Meetings are held on paper. The functional analysis, if one was ever done, shows that decisions are actually made by the ultimate beneficial owner sitting in Moscow, Almaty or Baku.

Consider a practical scenario. A mid-sized trading group with operations in Kazakhstan restructured its holding layer in the second half of 2024, moving from a Cyprus holding entity to a Hong Kong one following changes in the geopolitical environment that affected the utility of the Cypriot treaty position. The group came to us to assess whether the new Hong Kong entity would support a treaty-rate dividend stream from the Kazakh operating company. Our analysis identified three gaps: the directors lacked any time commitment to the Hong Kong entity's affairs, the management accounts had been prepared in Cyprus for the Cyprus entity and had not been migrated, and the banking was still running through a Cyprus account. Those gaps were not fatal, but they needed to be corrected before a dividend was taken. The correction involved appointing two Hong Kong-resident directors with genuine involvement in the commercial decisions, migrating the banking, and convening the first substantive board meeting in Hong Kong with contemporaneous minutes reflecting the decisions made. The process took one quarter. A dividend was taken in the following quarter under the treaty rate, with a clean substance file on record.

That sequence illustrates a wider point: substance is not a state of being, it is a documented condition that can be created, but it requires time and deliberate action. Principals who attempt to take a treaty-rate dividend before the condition is established create the very exposure they were trying to avoid.

How does the principal-purpose test change the analysis?

The principal purpose test (PPT – the BEPS-derived anti-treaty-shopping standard that has been incorporated into Hong Kong's treaties concluded or updated since the OECD's Multilateral Convention came into force) is the mechanism that CIS administrations can and do invoke when they believe a Hong Kong entity was inserted into a structure principally to obtain the treaty rate. The PPT is not a mechanical test: it asks whether, having regard to all relevant facts and circumstances, it is reasonable to conclude that obtaining the treaty benefit was one of the principal purposes of the arrangement.

The standard is deliberately broad. A structure that was designed to hold CIS assets and that has no business purpose other than the holding – no employees, no decisions, no operational function – is vulnerable to a PPT challenge even if it technically meets the residence conditions of the treaty. The PPT analysis requires that the principal demonstrate an objective business purpose for the structure, separate from the treaty benefit. For most holding structures, that purpose exists: centralised treasury, risk-management separation, investor access to a common-law jurisdiction, professional governance. But the purpose must be documentable. It cannot be asserted after the event.

For Hong Kong specifically, the PPT interaction with the FSIE regime creates a nuanced position. If the Hong Kong entity has genuine substance – meeting the FSIE economic-substance condition and therefore not brought into charge by the FSIE regime – that substance is also the strongest answer to a PPT challenge from the CIS side. The substance file that satisfies the Inland Revenue Department is largely the same file that answers the CIS administration. Building one file serves both purposes.

Where does the risk sit now, and what is the directional read?

The risk in the Hong Kong–CIS treaty-access position in 2028 concentrates at three points, and the directional read for each is clear.

The first risk point is network gaps. Several CIS states have no CDTA with Hong Kong. For those states – and for payment types not covered even in the CDTAs that exist – the treaty-access question simply does not arise. The risk here is not a failed treaty claim; it is a planning error made at the structuring stage by a principal who assumed treaty access existed when it did not. Our desk sees this regularly, and the correction is expensive: restructuring a holding layer after payments have been made without the treaty benefit is a multi-year exercise in remediation.

The second risk point is substance deterioration. A Hong Kong entity established with adequate substance can lose that substance over time if no one maintains the condition. Directors become inactive. Meetings are no longer held. Banking drifts. The substance file that was complete in 2024 may be thin by 2028. This is a maintenance risk, not a structural one, and it is manageable with discipline. But it is also the risk that produces the most adverse outcomes, because a lapsed substance condition is not a planning error – it is an avoidance finding waiting for a trigger.

The third risk point is geopolitical and administrative volatility. The CIS space is not a stable regulatory environment. Tax administrations in several CIS states have become more aggressive in their beneficial-ownership and substance inquiries, partly in response to BEPS implementation and partly as a result of their own domestic fiscal pressures. A treaty that existed and was functional when the structure was built may now face a CIS administration that is applying conditions the administration did not previously enforce. The treaty text does not change; the administrative posture does.

The directional read on all three risk points is the same: the safe position is one that is documented, maintained, and capable of being demonstrated to two tax administrations simultaneously. The unsafe position is one that was documented once and then forgotten, or that was never documented at all.

A second scenario illustrates the third risk point. A family-office principal with income from an Azerbaijani energy-services business used a Hong Kong holding entity to receive dividends under the Hong Kong–Azerbaijan treaty rate. The structure had been in place since 2021 with adequate substance at the time. In 2026, the Azerbaijani tax administration issued a questionnaire challenging the beneficial-ownership position of the Hong Kong entity, citing the entity's thin payroll and the fact that the ultimate beneficial owner was resident in a third jurisdiction. We were engaged to respond to the questionnaire. The response required assembling three years of board minutes, director engagement records, banking statements and management accounts to demonstrate that the Hong Kong entity was the economic owner of the Azerbaijani stake, not a conduit for the beneficial owner's personal income. The matter was resolved without a reassessment, but the process took six months and required significant remedial documentation work for the periods where contemporaneous records were thin.

The lesson from that matter is not that the structure was wrong. It was, at its core, a defensible position. The lesson is that defensibility is not the same as defended: the file must be capable of withstanding an inquiry, and that requires active maintenance, not passive existence.

The interaction with withholding-tax planning and the broader structure

Treaty access does not operate in isolation. For groups with CIS exposure through a Hong Kong holding entity, the treaty question intersects with the broader withholding-tax planning position across the structure. The internal question – how the group moves value from the CIS operating entity to the Hong Kong holding entity and then to the ultimate owner – involves multiple withholding points, multiple substance tests, and potentially multiple treaty claims in sequence.

The Hong Kong entity's treaty claim on the inbound CIS dividend is only the first layer. The distribution of that dividend from the Hong Kong entity to its own shareholder – whether a BVI holdco, a Cayman fund, or an individual beneficial owner – engages Hong Kong's own withholding position (which, as a general matter under Hong Kong's territorial system, does not impose withholding on dividends distributed by Hong Kong companies, given that withholding tax on dividends is nil in Hong Kong). But the BVI or Cayman entity's own treaty position, and the ultimate beneficial owner's residence-jurisdiction tax position, are separate questions that need to be mapped as part of the same analytical exercise.

For the broader structuring picture across Greater China and the offshore centres, the withholding-tax and source analysis that applies to cross-border holding structures is developed in our analysis of withholding-tax planning across Greater China. Groups with both Mainland China and CIS exposure running through a Hong Kong holding layer face a compounded analysis: the Mainland–HK interface has its own source and remittance rules, and those rules interact with the FSIE conditions in ways that require a combined, not sequential, read.

The sequence for most groups should be: map the payment corridors first; identify the withholding points in each corridor; assess the treaty availability and conditions at each point; assess the FSIE conditions for the Hong Kong entity in respect of each passive-income type; and then build the substance and documentation file that answers both the Inland Revenue Department and the CIS administration simultaneously. That is the discipline that produces a defensible position.

The objection handler: "our Hong Kong entity is clean – the risk is minimal"

The most common planning myth in this space is that a Hong Kong entity with a clean incorporation record and a certificate of residence from the Inland Revenue Department is "safe" for treaty purposes. It is not, for reasons already set out above, but the myth persists because it conflates two different assessments.

A clean entity – one with no compliance failures in Hong Kong, no outstanding tax, no adverse findings – is not the same as a substantive entity. Cleanliness is a compliance condition; substance is a functional condition. The Inland Revenue Department's issuance of a certificate of residence confirms that the entity is regarded as tax-resident in Hong Kong for the purposes of the relevant agreement. It does not validate the entity's substance. It does not pre-empt the CIS administration's domestic anti-avoidance analysis. And it does not resolve the PPT question, which is a purpose inquiry, not a residence inquiry.

The corollary of this point is that the risk mitigation work is not duplicative of what the entity already has. A principal who has a clean incorporation record and a certificate of residence still needs the substance file, the contemporaneous board records, and the functional analysis. These are not overlapping requirements; they are layered requirements, and each layer serves a different part of the two-jurisdiction inquiry.

For principals who have a tax-review obligation arising from a planned exit or distribution from a holding structure, the interaction between the treaty-access position and the tax review is significant. The review that precedes an exit is an opportunity to close the gap in the substance file before the CIS administration has a trigger. A distribution or sale provides exactly the trigger that prompts an inquiry; the period before that event is the window in which the position can be documented and, if necessary, corrected. Our practice on tax review before a Cayman Islands exit or distribution – set out at our dedicated service page – addresses a parallel set of issues where the exit trigger produces a substance and source inquiry in the holding jurisdiction.

For the full scope of our approach to tax positions for groups with CIS and Greater China exposure, the starting point is our Tax Positions practice.

Related practices

• Holding Structures – structuring the holding layer above CIS and Greater China operating entities
• Private Wealth – succession and asset protection for principals with CIS and Asia-Pacific exposure

Speak with Lockhart & Yip

For a scoped view of your matter, contact info@lockhartyip.com. Discuss your matter →

Related

• Tax Positions
• Tax Review Before Cayman Islands Exit Or Distribution 3
• Withholding Tax Planning Across Greater China Structure Analysis

This publication is general information and does not constitute legal advice. For advice on your situation, contact info@lockhartyip.com.