briefings |

Indefinite Informal Asset Freezes – The Beginning of the End?

William Redgrave


For the first time, a court has ruled that the operation of a “no consent” regime is unlawful. Under such regimes the decision of a police officer to withhold consent to move assets, following a suspicious activity report, generates an indefinite “informal freeze” of assets, without any court order or oversight. In Tam Sze Leung and others v Commissioner of Police [2021] HKCFI 3118, the Hong Kong High Court ruled in December 2021 that the Hong Kong no consent regime is unconstitutional.

This is a striking development. Judicial review challenges in recent years to comparable no consent regimes in Hong Kong, Guernsey and Jersey have all been unsuccessful – most recently in Jersey, in Prospective Applicant v Chief Officer of the States of Jersey Police [2019] JRC 161.1 In the Leung case the facts were different in one significant respect from the previous challenges, and the court used this to justify the difference in outcome.

The court took account of the way no consent regimes operate in practice. It concluded that the Hong Kong police acted unlawfully when withholding consent in order to freeze assets. This decision may open the way to courts in other jurisdictions taking a similar approach, and could hasten the end of the indefinite informal freeze.

The Informal Freeze

The “informal freeze” is a product of modern anti-money laundering (AML) legislation. It results from an understandable desire to ensure that funds suspected to be connected to crime can be preserved while the suspicion is investigated, in case criminal proceedings are brought. The ultimate aim is post-conviction confiscation. If the funds are not immediately frozen then confiscation may prove impossible.

However informal freezing happens at a very early stage, when no one has been charged, let alone convicted. Often there is not even a criminal investigation in place at the time the funds are “frozen” – and there may never be one. The suspicion may be wrong and there may be nothing unlawful about the funds at all. And the “freezing” is not ordered by a court. It results from the decision of a police officer alone.

This is an entirely separate process from the statutory system of court orders applied for by prosecution agencies, restraining funds pending a criminal prosecution. Restraint Orders are made by a court after a careful review of the evidence and submissions from prosecution and defence, balancing the public interest in restraining funds for possible confiscation against the interference in the (as yet unconvicted) person’s right to deal with their own property. The court periodically reviews the order to ensure that it is not in force for too long and that the prosecution is proceeding diligently, and may discharge the order if things are taking too long or the evidence no longer justifies restraint. None of this applies with an “informal freeze”: a police officer takes a decision, without having to give any reasons, and the money is frozen indefinitely, with no requirement for any court input or oversight at all.

How can funds be frozen without a court order? AML laws require financial service providers such as banks and trust companies to report suspicions of money laundering (dealing in criminal property) to the local police Financial Intelligence Unit (FIU). These suspicious activity reports (SARs) are considered by the FIU, often in conjunction with agencies in other relevant countries. The FIU then decides whether or not to give consent to the service provider to deal with the funds normally (i.e. to act on their client’s instructions). If the FIU withholds consent, then the financial service provider will invariably decide to block the funds and refuse to move them whatever their client says.

Why? Because the service provider is taking the risk of being convicted of money laundering if it pays the suspect funds to the client and the funds turn out to have been criminal property. Making a SAR and receiving consent from the FIU provides a defence to a prosecution for money laundering: without that defence, they are much safer blocking the funds, so they invariably do so. Avoiding committing a crime entitles the service provider to ignore its client’s instruction, leaving the client – who may well be entirely innocent of any wrongdoing and may not face any charges – unable to access their funds, indefinitely. And they may receive no explanation for it. Thus the funds are informally frozen. No court has ordered it, but the “no consent” from the police has had the same effect on the funds.

In most jurisdictions the obvious potential unfairness of this position led to strict time limits being imposed, so that consent is deemed to have been given if no court Restraint Order has been applied for within a certain period (varying between countries, from one day to six months). The service provider is then protected from prosecution, so it has no reason to continue blocking the funds.

However such time limits were never imposed in Jersey, Guernsey or Hong Kong, three small jurisdictions with relatively large finance industries.2 In those jurisdictions funds can be, and have been, informally frozen for many years, without a conviction or even a prosecution.

Previous attempts to challenge the informal freeze

The lawfulness of such a “no consent” regime was tested in Guernsey in 2011 in Chief Officer, Customs & Excise v Garnet Investments Limited[2011] 19/2011. $36m had been informally frozen in a bank for over six years, and there had been no prosecution, or even a criminal investigation, for the suspected offence (holding public funds stolen by the former president of Indonesia). At first instance the Royal Court of Guernsey ruled that informally freezing the funds without a criminal investigation for so long was disproportionate and thus unlawful. However the Court of Appeal overturned the decision.

Strikingly, the Guernsey Court of Appeal ruled that there was as a matter of fact and law no such thing as an “informal freeze.” It held that the purpose of the relevant statutory provisions was not to give the police the power to freeze funds by withholding consent but, conversely, to permit the police to choose to give consent when they wished to, to avoid alerting suspected criminals. It was the financial service providers, not the police, whose actions froze funds. The consent regime, it held, had two purposes. The first was to encourage the filing of SARs. As to the second:

27       Second, the consent regime gives the police the operational freedom to grant relief from criminal liability in circumstances where it is considered to be in the interests of law enforcement so to do.  Thus consent may be granted to avoid a suspected criminal becoming aware of the suspicions that are harboured in relation to him.  This objective is also reinforced by the existence of offences in connection with tipping off (see section 41).  Consent may also be granted so as to permit a controlled transfer to take place so that funds can be traced for investigative purposes.

 This led the court to conclude that:

38     True it is that in practice the process of reporting suspicious transaction and seeking consent might be used by the reporting institution for delaying possibly difficult decisions as to whether to transact and may also in practice provide a period in which the police may consider whether they wish to commence an investigation or seek any restraint orders, but the practical utility of the hiatus that is created whilst an application for consent is pursued does not mean that this was the sole or dominant purpose of the consent regime and does not support the argument that this was the intention of the legislation.

 39       For the reasons set out above we do not consider it was.  In our opinion the principal purpose of the consent regime was to provide an opportunity to the police to give an exemption from criminal liability by consent but only where it was in the interests of law enforcement to do so; it was not to create an informal mechanism to be used by the police for freezing funds.

 40       It follows that we do not accept the basic premise contended for by Garnet and accepted by the Lieutenant Bailiff as to the statutory purpose of the consent regime to the effect that the FIS is able to deny a person access to their property by refusing to give consent and yet not seek judicial oversight of that refusal by applying for a restraint order’.

 41       In our judgment, it is not the FIS that is denying Garnet access to its property and preventing judicial oversight, it is the impact of the width of the criminal law and its chilling effect upon the person holding the fund, namely BNP.”

Thus the Court ruled that the legislation did not set out to create an informal freeze power was not what the legislation was for. As will be seen below, in Leung the Hong Kong court agreed with this. The difference between the two cases is that in Garnet the court proceeded on the basis that as the power did not exist it was not being exercised, whereas in Leung the court held that although the law did not set out to create such a power, the police were nevertheless using the consent power in order to freeze assets, which was an unlawful use of the power.

The Court of Appeal in Garnet did accept that the operation of the no consent regime affected property rights, and that any interference in such rights must be proportionate in order to be lawful, in line with Article 1 of Protocol 1 of the European Convention on Human Rights. Yet it decided that the absence of a time limit on the effect of the no consent was justifiable in Guernsey, largely because Guernsey had limited resources to investigate SARs. It held that the six-year no consent in that case was not disproportionate to the goal of tackling money laundering.

The reasoning in Garnet was adopted by the Hong Kong Court of Appeal in Interush v Commissioner of Police [2019] HKCA 70.

In the Jersey case of Prospective Applicant, also in 2019, the applicant urged the Royal Court of Jersey not to adopt the Garnet analysis that the police were not freezing funds. It was submitted that that approach simply did not reflect the way the no consent regime operated in reality, at least in Jersey by 2019. The statistics on consents given, and the Jersey police’s own statements, revealed a very different picture. The vast majority of SARs in Jersey (over 99% at that time) resulted in the granting of consent, demonstrating that this was the norm not the exception.3 And it was acknowledged by the Jersey police in their own official statements that the power to refuse consent led to a freeze, which was why it was rarely done. The Jersey police had publicly described the position in this way:

No Consent is only applied where absolutely necessary and a suspicion that the funds may represent the Proceeds of Criminal Conduct.”

“A small minority of SARs submitted to the JFCU receive ‘no consent’.  This is not a formal ‘freeze’ or restraint on the account, although in practice the transactional effects are the same.”

However the Royal Court of Jersey considered itself effectively bound by the appellate decision in Garnet. It held that “evidence of how the no consent regime is used today does not, in our view, undermine the Guernsey Court of Appeal’s conclusion as to the legislative purpose of the consent regime.”

In the Jersey case the applicant had been unable to access his funds for over 20 months. The FIU had withheld consent because of suspicions arising from allegations made in a class action civil lawsuit in the US, alleging market manipulation. There was no criminal investigation in existence anywhere, let alone any charges; a civil regulatory investigation had begun but had not led to criminal involvement; and the evidence was that most US class actions of that kind were struck out or failed as they lacked evidence. The applicant therefore submitted that there was no obvious justification for the funds to be frozen at all. However, applying Garnet, the court did not regard the decision to withhold consent as having caused the freeze, so it did not quash it.

As for proportionality, the Royal Court did not find the maintenance of the no consent to be disproportionate. However it noted that the facts were rather different from those in Garnet and cautioned that, as more time passed without a criminal investigation, the withholding of consent could become disproportionate and therefore unlawful.

Leung – significance of police triggering SARs

In Leung, the first instance Hong Kong court considered itself able to reach a different conclusion from the Hong Kong Court of Appeal in Interush as to the lawfulness of the no consent regime. This was because the facts were importantly different. The difference lay not in the basis for the underlying suspicions as to criminality, nor the length of time the funds had been frozen, but the procedure adopted by the police.

In Leung, four different banks each made separate SARs in relation to suspected money laundering by members of a Hong Kong family. The suspicion they reported was that funds in the bank accounts were the proceeds of market manipulation. However the banks had not each independently become suspicious as a result of their own knowledge or researches. They made SARs because the Hong Kong FIU had approached each of them and confidentially informed them of a criminal investigation into market manipulation. Communications from the FIU to the banks included the following:

  • “Please hold the below accounts and file STR4 asap as LNC5 will be issued.”
  • “In order to facilitate the SFC’s6 investigation, please suspend the operation of the account … Suspicious Transaction Report (STR) to JFIU is urgently requested.”

On these facts the proposition that the freezing of the accounts was an act of the financial service providers, rather than the police, was unsustainable. The banks had only made SARs because the police had urged them to do so.7

This practice, of the police informing financial institutions of investigations in order to generate a SAR, is not improper; as the court affirmed, it is the police’s duty to alert financial institutions to potential offences and to remind them of their legal obligations.However if the police do this then they cannot maintain the line that the consent power is not being used to freeze assets.

The court noted that the Commissioner of Police had changed his stance from that taken in Interush: this time he acknowledged (contrary to the Garnet formulation) that the power to withhold consent created an informal freezing regime, and was used for that purpose. LNCs were issued and maintained in order to “prevent dissipation of the assets”.9 He claimed that the power existed to provide a “temporary stop-gap measure” to ensure the money was still there when a Restraint Order was sought.10 (As the court noted, periods of over six months or more do not “leap out” as temporary stop-gap measures.)11

The court examined the detailed safeguards in the Hong Kong legislation on Restraint Orders, as compared to the unregulated yet draconian effect of the no consent regime as it was being operated. It concluded that the legislature cannot have intended the consent power to be used to generate an indefinite informal freeze. Given the careful protections in the legislation governing Restraint Orders, it was “implausible that the legislature could have simultaneously “consciously enacted” a secret, informal and unregulated asset freezing power of the kind which the Commissioner now asserts he enjoys”.12 Thus the police were operating ultra vires the legislation, and unlawfully. They were using a power for a purpose other than that for which it was created. Further, because there was no time limit on the power, the interference in property rights was a disproportionate interference with individual rights, and thus unlawful for that reason too.13

It remains to be seen whether this decision will survive appeal. It may be overturned. Alternatively a reappraisal at appellate level, in this or another case, could conceivably go further than the present decision, ruling that the no consent regime is not only unlawful in cases where the police have instigated a SAR, but – now that the cat is out of the bag that the consent power is used to freeze assets – that the Garnet analysis adopted in Interush should be abandoned, and that the entire no consent regime is unlawful whoever triggered the SAR.

Such an outcome may be unlikely but it would arguably make the legal landscape rather more logical. The unreasonableness of informal freezes depends on many factors other than who triggered the SAR. For example the arguments in favour of maintaining the no consent were notably weaker in the Jersey case of Prospective Applicant than in Leung. In the Jersey case there was no criminal investigation in existence at all; in Leung there was, indeed it was the basis for the SARs. In the Jersey case the assets had been blocked for over 20 months by the time the case came to court; in Leung, the period of the informal freeze was a mere 10 months, after which a court Restraint Order was obtained so the informal freeze came to an end, and by the time of the hearing the position was academic.

There would be an arguably perverse result if the decision in Leung is limited to cases where the police triggered the SAR, while in other cases the Garnet / Interush reasoning still applies and the use of the consent power to freeze assets is lawful, unless it is shown to be disproportionate. An informal freeze would most easily be ruled unlawful not on the basis of factors such as the weakness of the reasons for suspicion, nor how long the freeze had lasted, nor whether a criminal investigation was in existence, but simply on whether or not the police had triggered the SAR. If they had not, then – unless the applicant could persuade the court that the continuation of the freeze was disproportionate in the circumstances – the freeze will be lawful. And courts have seemed reluctant to find disproportionality, in the absence of time limits. It leaves all concerned in a difficult position, as neither the police nor the financial institution nor the customer can say for certain whether the court will decide that on a particular set of facts the line of disproportionality has been crossed. Litigating it is a potentially expensive risk for everyone.

Yet cases where the police trigger a SAR because of an ongoing criminal investigation are rather more likely to result in a conviction than those where SARs are merely the result of a bank’s internet researches, or a civil claim, or the odd behaviour of a customer. The existence of a criminal investigation would seem to provide a stronger justification, for at least a brief informal freeze, than exists in cases where no such investigation is under way. Why should an informal freeze be harder to challenge when there is no criminal investigation?

The Hong Kong court in Leung has approached the case on the basis of the way in which the consent power is in fact used, rather than the unrealistic and technical analysis adopted in Garnet and followed in Interush and Prospective Applicant. The court was able to do so on the facts of Leung, as the police could not deny that they used the no consent power specifically in order to freeze the assets. Yet in reality the power is used for the same purpose – and with the same effect – in cases where the SAR was not instigated by the police.

Whenever a SAR is made, the police know that for as long as they withhold consent it is virtually certain that the assets will be effectively frozen, perhaps for many years, without any of the court oversight and time limits that would apply if there were a prosecution in prospect or under way and the funds were restrained by court order. If any law enforcement or regulatory agency around the world tells the FIU that they are interested in the case, the FIU can help by withholding consent. They can then relax (as can the other agencies), safe in the knowledge that the funds are not going anywhere. Meanwhile the unconvicted owner of the assets is entitled to no information, and has no protection from the impact of this freeze. He can only challenge it by bringing a judicial review of the police, or a civil claim against the service provider – imposing a costly, difficult and cumbersome burden on him, in the absence of the ongoing court supervision that would exist with a Restraint Order.

When there are no time limits on the freezing impact of the no consent power, the customer may paradoxically even be worse off if there is no criminal prosecution than if there is one. If there is a prosecution, then at least there will be a determination of guilt or innocence one day and the funds will then either be released or confiscated. If the funds remain frozen for years with no prosecution, but the police maintain no consent because they retain a suspicion, the money remains in limbo indefinitely. The quantity and quality of evidence required to make a police officer suspicious may be well below that which a court would require before freezing funds.


In its realistic assessment of the practical use of the no consent power, Leung is a welcome corrective to the strained and technical analysis in Garnet over a decade ago. It may prompt a wider judicial re-evaluation in other jurisdictions of how this power is really used, when it is unlawful to do so, and where the balance should be struck between fighting financial crime and respecting individual rights.

It would surely be preferable, for all those involved in these cases, for there to be far greater certainty as to when the use of the consent power is lawful. The obvious way to achieve this is to impose tight statutory time limits on the freezing effect of no consents, as most countries have long done. Once the time limit had passed, if there were no Restraint Order financial institutions would be free to operate as normal, not be left stuck in the middle, fearing being prosecuted for money laundering if they move the funds or sued by their client if they do not. This would not only tilt the balance back in favour of innocent customers; it would also create a clear structure for the police to appreciate what they need to do and by when. It would incentivise investigators to move quickly to apply for Restraint Orders. It would reduce the need for expensive litigation over whether a freeze is disproportionate.

The Royal Court of Jersey has more than once encouraged the Jersey legislature to bring in time limits. Fifteen years ago it said of the indefinite informal freeze14:

19       This is clearly capable of causing great hardship and unfairness.  There may never be a prosecution, yet the bank may retain its suspicion.  The result may be that a person, against whom no criminal charges have been brought and where there lies only a suspicion, finds his assets informally frozen without there even having been any court order to achieve this.  Furthermore, the freezing of the account may continue for an indefinite period.

20       It is hard to reconcile this situation with the carefully structured protections provided in respect of a saisie15, which are clearly intended to ensure that funds are not frozen indefinitely or for an unreasonably long period in the absence of criminal charges.  The potential injustice of the situation was recognised in the United Kingdom where the relevant legislation was amended in 2002 so as to provide that the police have seven days from the STR in which to respond.  If no response is given they are deemed to have consented to the bank dealing with funds in question.  If they respond within the seven days and refuse consent, they have a further thirty one days in which to apply for a restraint order (the equivalent of a saisie).  If they have taken no such action at the expiry of thirty one days after their refusal of consent, the bank may safely proceed16. …   However, no such amendment to the 1999 Law has been made and we must therefore wrestle with the resulting difficulties.”

The Court commented that the amended UK legislation struck a fair balance between the competing interests and concluded “… we would urge that immediate consideration be given to introducing amendments similar to those which have been introduced in the UK”.

Maybe, now the Hong Kong court has finally recognised that the consent power is being used deliberately to freeze assets, the risk of a similar ruling in Jersey will make such statutory reform more likely.

1 William Redgrave acted for the applicant in that case

2 In 2014 there were 13 countries in this position.

3 Since 2019 the Jersey FIU has moved towards neither granting nor withholding consent until the reporting institution has provided sufficient information for it to make a decision either way, which may take many months.

4 Suspicious Transaction Report – the same as a Suspicious Activity Report (SAR)

5 Letter of No Consent

6 Securities and Futures Commission – Hong Kong market regulator

7 Para 76

8 Para 84

9 Para 62

10 Para 85

11 Para 88

12 Para 78

13 Paras 159-60. The no consent regime was also held in para 117 not to be “prescribed by law” as required under the Hong Kong constitution, because the scope of the consent power was so ill-defined and safeguards against its abuse were inadequate.

14 Chief Officer v Minwalla [2007] JLR 409

15 The Jersey name for a statutory Restraint Order pending a criminal prosecution

16 More recently the UK legislation has been amended to provide for a no consent to have effect for longer periods in certain circumstances, but the principle of deemed consent after a certain period remains.