The No-Further-Claims Clause in a Hong Kong Divorce Agreement: Can Maintenance Be Reopened Later?

The question many people ask

A great many people are holding a piece of paper like this: signed years ago at the divorce, with a line in it saying neither party will claim maintenance from the other again. Circumstances have since moved — a job lost, an illness, or an ex-spouse who has done well — and a friend's flat 〈well, you signed it〉 has been allowed to settle the matter.

The statute does not say that. Section 14(1)(a) of the Matrimonial Proceedings and Property Ordinance (Cap. 192) provides that where a maintenance agreement includes a provision purporting to restrict any right to apply to a court for an order containing financial arrangements, that provision shall be void. Paragraph (b) of the same subsection then says that only that provision falls: the rest of the agreement stays binding. So "you signed it, that's that" is not a proposition the Ordinance supports — but not supporting it is not the same as opening a door, because which door is open depends directly on which piece of paper you are holding.

This article answers four questions:

• what the no-claims clause is actually worth — s. 14 says two things, not one;
• which instrument you are holding — an agreement, an order, a nominal order, or an order dismissing the claims: four kinds of paper, four answers, and it is not a matter of preference;
• how much has to have changed — three Ordinances, three thresholds, and one of them expressly counts a change the parties had already foreseen;
• whether remarriage or delay shuts the door — four clocks, and what each is attached to is not what most readers expect.

This article states the law in general. It is not an answer about your document. What the courts have said about the weight an agreement carries is set out below from the Court of Final Appeal and the Court of Appeal, quoted with paragraph numbers.

Many people believe a signature ends it

A common belief is that in a settled divorce the obligation to pay exists only if it was agreed, and that if it was not agreed it cannot be claimed again after the divorce.

That is the proposition s. 15(6) denies. ⚠ s. 14(1)(a) is a narrower answer than it looks: it opens "If a maintenance agreement includes a provision purporting to restrict any right to apply", so where there is no agreement at all the section has nothing to operate on. It voids a restricting term in a written agreement; it does not create the right to apply where none was agreed.

And the official guide does not deal with it. The Judiciary's own How to Apply for a Divorce guide covers divorce and ancillary relief, but does not explain varying an order, consent orders, setting aside, or maintenance agreements.

What this actually is: s. 14 says two things, not one

Section 14(1) is a pair, and reading either half alone gets both halves wrong. Cap. 192, s. 14(1), English text:

In one line: the void thing is the clause, not the agreement.

Why it has to work that way. If a no-claims clause brought down the whole agreement, the section would hand the stronger party a demolition tool: write the clause in, then rely on its own invalidity to escape the maintenance obligations sitting beside it. Severance in paragraph (b) is what stops paragraph (a) being usable by the party it was written against. That half is rarely explained to anyone.

Second, what counts as a "maintenance agreement" at all. Section 14(2) defines it, and the definition contains one condition and one non-condition:

• Writing is a condition. An oral 〈neither of us will ever claim〉 is outside s. 14 altogether — there is nothing for the section to make void, because there is no agreement in writing. ⚠ What does NOT follow is that the reader holding a signed document is better placed. Being outside s. 14 establishes nothing about whether an oral understanding can bar a statutory application, or whether it is enforceable at all — those turn on other doctrines and on the provision the relief is sought under.
• Timing is not a condition. The definition expressly covers an agreement made "whether before or after the commencement of this Ordinance", and an agreement containing financial arrangements "whether made during the continuance or after the dissolution or annulment of the marriage". An agreement signed years after the decree sits inside s. 14 exactly as one signed during the marriage — which is a great many readers' own facts.

Third, s. 15(6) closes off a misreading before anyone reaches it:

Section 14(1)(a) voids the ouster clause; s. 15(6) declares that neither section shuts any other door.

So what is the agreement worth? This is the answer, and it comes from two judgments

Established from the statute: the clause is void (s. 14(1)(a)); the rest of the agreement binds (s. 14(1)(b)); no other door is shut (s. 15(6)).

And the question the statute does not answer — how much weight a court gives the agreement — is answered from the judgments themselves. Both are anonymised family cases, and they are kept anonymised as the courts did.

• **SPH v SA, FACV 22/2013, Court of Final Appeal (Ma CJ, Ribeiro PJ, Tang PJ, Bokhary NPJ and Lord Collins of Mapesbury NPJ), heard 12 May 2014, judgment 9 June 2014**, on appeal from CACV 99/2012.
• **L v C, CACV 169/2006 & CACV 181/2006 (heard with CACV 182/2006), Court of Appeal (Stock JA, Yuen JA and Hartmann J), heard 5–9 and 12–13 February 2007, judgment handed down 25 May 2007, on appeal from HCMC 5/2006 and HCMC 1/2003. 106 pages. ⚠ There is a second, later judgment in the same appeal, dated 19 March 2008, and it is dealt with separately at the end of this section — it is not** a second helping of the law on agreements.

Paragraph references below are each judgment's own paragraph numbers.

The Court of Final Appeal's own statement of Hong Kong law

SPH v SA took the occasion to state the law. At paragraph 4 it describes where such agreements started —

— and where they have arrived:

Those two clauses carry the whole answer between them, and readers pull in the wrong direction from each. Not definitively binding; substantial legal effects.

At paragraph 30 the Court of Final Appeal set out what the Hong Kong Court of Appeal had already decided in the second case:

and, in the same paragraph, what it takes to get out:

At paragraph 39 the Court of Final Appeal stated the position for Hong Kong:

and at paragraph 40 it endorsed the Court of Appeal on the point that used to defeat these agreements altogether — that the old rule that agreements

Then the limit on all of it, at paragraph 34, which is the sentence a reader most needs and least expects. ⚠ Attribution, because it decides how this and the next four quotations are to be read: paragraphs 32, 33 and 34 of SPH v SA are the Court of Final Appeal's account of what the UK Supreme Court decided in Radmacher v Granatino, and paragraph 39, quoted above, is the Court of Final Appeal adopting those principles as the law of Hong Kong. What follows is therefore Hong Kong law by way of that adoption, in the Court of Final Appeal's own words; it is not this site restating Radmacher.

Read that against the top of this section. The agreement is not a lock and it is not a scrap of paper. It is an input the court weighs, and the court decides.

What makes an agreement carry full weight, at paragraph 33 — read with the attribution note just above:

And the converse, in the same paragraph: the court should give effect to an

Two further things from paragraph 32 — the same account, adopted the same way — and they matter to a reader comparing paper. First, an agreement can be held to even where the court would otherwise have ordered differently:

Second, the same principles apply whichever kind of agreement it is:

So a pre-nuptial agreement and a separation agreement are not tested by different rules. That is stated by the Court of Final Appeal at paragraph 39 in its own words as well, where it says it sees no reason to distinguish between the two.

The Court of Appeal in L v C: the framework, and where the burden sits

This is the judgment closest to most readers' own paper, because the agreements in it were made after the parties separated and divided assets between them. The whole framework is at paragraph 37. It opens with the proposition the Court of Final Appeal later adopted:

and immediately states the limit on it, which is the same limit s. 14(1)(a) states in statutory form:

Then the sentence that places an agreement inside the statutory exercise rather than above it:

The statute there is Cap. 192, s. 7 — the same seven matters this article sets out in full further down. So the agreement is one of the "circumstances" s. 7 sends the court to.

The same paragraph explains why a court looks hard at the circumstances of signature:

⚠ That last line is the Court of Appeal's own warning: what your own agreement is worth needs a solicitor's assessment on the facts.

The starting position, at paragraph 38:

And at paragraph 42, where the burden lies — which is the question a reader in distress most often has backwards:

So the person asking the court to depart from the agreement carries the burden, and independent legal advice at the time makes that burden heavier rather than lighter.

What will and will not get a court to intervene, at paragraph 47:

and at paragraph 52, what the court is actually looking at:

Which disposes of the thought most readers arrive with — that a bad bargain is by itself a reason to reopen. At paragraph 59:

and at paragraph 52, on an exercise that does nothing but compare values:

Paragraph 37 also lists the questions a court asks about the circumstances of signature. (This site's summary.) The questions the Court of Appeal names are: whether legal advice was available at the time of the negotiations and the conclusion of the agreement; whether the negotiations and the conclusions reached were attended by undue pressure, including self-induced pressure; whether some factor in the relationship between the parties was at work, such as the exploitation of a dominant position or the use of children as a weapon against one of them; whether judgement was impaired by emotion, fear, or a misapprehension of the factual or legal position; whether the parties have acted upon the agreement reached; and whether a fresh unforeseen event has arisen since the agreement was made, by reason of which it may be unconscionable to hold the parties to its strict letter.

One more distinction from paragraph 54, and it decides a great deal, because agreements very often cover only part of what a couple owns. Where an agreement merely allocates particular assets, the court will normally respect the allocation without treating those assets as removed from the pool. It is different where the parties meant to take them out of the pool for good:

And why any of this is framed as principle rather than pure discretion, at paragraph 39:

The second L v C judgment, 19 March 2008 — what it is, and what it is not

There are two judgments in CACV 169/2006 and they do different jobs. The later one says so on its own face, at its paragraphs 2 and 7:

The application being varied is an order nisi as to costs. ⚠ So the 19 March 2008 judgment is a costs ruling.

It does decide one thing, at its paragraph 26, and it belongs to the disclosure section below rather than to this one:

What these two judgments do not decide

⚠ The limits of these two judgments:

• **SPH v SA was not an ancillary-relief hearing.** Its own paragraph 1 says what it was:

> "This is an appeal from a judgment of the Court of Appeal refusing a stay of matrimonial proceedings in Hong Kong."

and its paragraph 78 says what happened:

> "We therefore dismiss the appeal."

The agreements in it had not been ruled on. Paragraph 5 records:

> "There have, of course, been no findings of fact, and the following account must be read in that light."

So the statements of principle above are the Court of Final Appeal stating the law; they are not the outcome of a weighing exercise on any facts.

• The Court of Final Appeal expressly left one class of agreement open. Agreements that do nothing but alter a foreign matrimonial property regime — which is not what most Hong Kong readers hold — were, in the words of paragraph 49, a matter that

> "was not argued on this appeal"

and the judgment says nothing in it is intended to pre-judge that question.

• **L v C was decided on facts nothing like most readers'. It concerned two commercially sophisticated parties with assets in the hundreds of millions, agreements made after separation with independent legal advice on both sides, and a judge's finding that neither had been under undue pressure. The Court of Appeal's own words — the subject matter is peculiarly fact-sensitive — are the bound on reading across from it to anything else.**
• Neither judgment interprets s. 14 or s. 15 of Cap. 192. They are about the weight of an agreement in the ancillary-relief exercise. The statutory routes this article describes below are a separate question, and no sentence above should be moved onto them.

Full and frank disclosure: where a great deal of later argument starts

An article about reopening has to say something about what each side disclosed when the agreement or the order was made, because that is where much of the later argument begins.

The Family Court seminar paper of 8 November 2024, at paragraph 8, describes it as good practice for a consent summons to record that the terms were agreed on the understanding that there had been full and frank disclosure of each party's financial means. That is a description of drafting practice, given by a judge in a paper whose own page 1 disclaims it as personal views — it is not a rule, and it is not a practice direction.

One — an order. Non-disclosure or misrepresentation of material facts at the time the order was made is one of the four recognised grounds on which a court may set a consent order aside. That comes from HCMP 276/2012 at paragraph 10, set out with its bounds in the section on setting aside below — one first-instance judgment, decided on an unopposed application.

Two — an agreement that never became an order. ⚠ *This piece comes from SPH v SA. At paragraph 33, quoted in full above, an agreement carries full weight only if each party entered into it of their own free will, without undue influence or pressure, "having all the information material to his or her decision to enter into the agreement". So incomplete information at the time of signature goes to the weight the agreement carries — it is not a separate application and it is not a rule that the agreement falls. And L v C* at paragraph 47, quoted above, places the court's attention on the circumstances in which the agreement was reached rather than on the bargain's arithmetic.

Three — what a court does about non-disclosure once it finds it. L v C is direct about this, at paragraph 68, where the Court of Appeal said that what is not well understood is the actual consequence of such conduct, namely

and it goes further at paragraph 68 by recording that in cases of this kind questions of perjury may arise and that judges will bear in mind the option of referring papers to the appropriate authority. Then the sharpest sentence of all, from Yuen JA at paragraph 191:

Four — and this cuts the other way, against a party who alleges non-disclosure late. At paragraph 189, on what a court needs before it can find non-disclosure and put a figure on it:

In that case the allegation was made too late in the hearing to be explored, and the trial judge's refusal to put a figure on the undisclosed assets was upheld as entirely justified. ⚠ So "they hid something" is not a submission; it is a case that has to be built in advance, on documents.

Five — costs. The later judgment in the same appeal, 19 March 2008, records at paragraph 26 that a party's material non-disclosure is litigation misconduct that the costs jurisdiction reaches, and that Order 62 rule 7 is not confined to the receiving party:

⚠ What these judgments do not settle. What makes a fact "material" for the third ground in HCMP 276/2012 is not defined in them. Who bears the burden on a contested application to set an order aside for non-disclosure, and how much has to be shown, are likewise not stated in them — the burden stated above, at L v C paragraph 42, is the burden on the party asking a court to depart from an agreement, and it must not be moved onto the set-aside route. By when a set-aside application has to be brought is not stated either, beyond the one point CACV 92/2011 makes about the appeal route, which is quoted in the variation section below. Whether material information was withheld when a particular agreement or order was made is a question of fact, and it is a different question from the change-of-circumstances route this article describes below.

Which instrument are you holding? (check it against yourself)

This is the most important section in the article, because every test below branches from it — and it is not a matter of choice. Readers usually assume there are two kinds of paper. There are four kinds of provision, and the fourth looks most like a full stop — but one document can carry more than one of them.

A note on the table. Row two speaks of the class listed in s. 11(2); the table covers the orders Cap. 192 empowers under ss. 4, 5 and 6. Section 11(2) also names orders made under s. 3, s. 6A and s. 8. Row three's Cap. 16 orders have their own provisions; the table gives the s. 7(1) threshold.

⚠ And the four rows are four kinds of provision, not four kinds of document. One sealed consent order can carry several of them at once — an ongoing periodical-payments provision, a lump sum, undertakings given to the court, and a paragraph dismissing whatever claims remain — so which row you are in is decided paragraph by paragraph, not by the heading on the front page. The seminar paper quoted below makes the practical point in its own way: it tells practitioners that a dismissal or a full and final settlement cannot sit alongside an ongoing periodical-payments order, and that where nothing in the order disposes of the ancillary-relief claims there is no finality at all. So read the sealed order, and read every operative paragraph of it — not only the settlement agreement that came before it.

The fourth kind of paper follows a different route. The Principal Family Court Judge (Acting) put it to practitioners in a Family Court seminar paper of 8 November 2024, at paragraph 7:

The same paragraph names three things — De Lasala v De Lasala [1980] AC 546, Wong Oi Han v Sin Wai Chung [2012] 3 HKLRD 142, and Practice Direction SL10.3. The Hong Kong judgment and the Practice Direction are set out in the next section. ⚠ What follows from that split: setting aside a final order is a different route from varying one. The next section describes the first; every section after it describes the second. The grounds, the time limits and the procedure are not shared between them, and nothing said about variation carries across.

And page 1 of the same paper carries its own disclaimer:

That is: a judge's seminar paper, disclaimed by its author as personal views. Not a practice direction, and not a judgment.

Setting aside a consent order: the grounds, the route, and the Practice Direction that governs it

This is the fourth kind of paper. Of the things the seminar paper named, two are set out here: a Court of First Instance judgment and Practice Direction SL10.3.

1. The judgment

Wong Oi Han v Sin Wai Chung, HCMP 276/2012, Court of First Instance, Poon J in Chambers, judgment 11 May 2012.

The case is also reported at [2012] 3 HKLRD 142; the footnote to Practice Direction SL10.3 spells the respondent Sin Wai Cheung.

Where to find it in the judgment: the facts are at paragraphs 1 to 4; the consent order is not simply a contract at paragraph 9; the four grounds at paragraph 10; the procedural question at paragraph 11; the Hong Kong position at paragraphs 14 to 16; the shut route and its narrow exception at paragraph 20; the mixed order at paragraph 22; the application of the exception at paragraph 23; the disposal at paragraph 24; and the judge's closing remark at paragraph 25. Those are also the paragraphs Practice Direction SL10.3's own footnote cites — see the end of this section.

2. The facts, and why they are unusual in a way that matters

The applicant filed a divorce petition in the Family Court in FCMC 9658/2010 on 29 July 2010. The parties settled ancillary relief and the Family Court made a consent order on 30 November 2010: he was to transfer his shares in one company to her, she was to transfer her shares in a second company to him, and he was to pay her HK$150,000 a month in maintenance after decree absolute. Decree absolute followed on 1 February 2011.

Both sides then discovered they had been mistaken — she about the net worth of the company she was giving up, he about the profitability of a Mainland factory and therefore about his ability to pay that monthly sum. They renegotiated, agreed a fresh package built on a lump sum secured by a share mortgage instead of monthly maintenance, and she issued an originating summons in the High Court on 14 February 2012 asking for the consent order to be set aside and the ancillary-relief matters sent back to the Family Court. The husband did not contest it.

And the shape of the order mattered to the judge. It carried a periodical-payments provision, which section 11 covers, and other provisions, which section 11 does not; but the judgment records that

so the whole of it was approached as one thing.

⚠ Hold on to that. This was an application both sides wanted, on mistakes neither side disputed. How far a case decided on those facts travels is the bound stated at the end.

3. What it decides on substance: a consent order is not simply a contract, and there are four grounds

On the first point:

Where the order falls inside section 11 of Cap. 192 — the variation power this article's next section describes — the court may vary or discharge it. And where it does not, which is the position for the dismissal paragraph most readers are actually worried about:

That is the answer. Four grounds, and note the third: non-disclosure or misrepresentation of material facts at the time the order was made is one of them.

The judge's own view of the state of that law is worth having, because it is a caution from the bench and not from us:

4. What it decides on procedure — and this is what the case was actually about

After setting out four English routes, the judgment turns to Hong Kong:

The re-hearing route is not available here at all:

And the route most people would reach for first — issuing a summons in the divorce file that already exists — is shut, with one narrow exception:

Then the application before the judge was measured against that, and because it was uncontested and the mistakes were not in dispute, it fell inside the exception:

The point is worth stating twice. She had grounds nobody disputed, and her application was dismissed — on procedure. She had gone to the High Court by originating summons when, on these particular facts, the narrow route back to the original court was the right one. Its outcome turned entirely on which piece of paper was issued where.

5. And there is now a Practice Direction that governs the route: SL10.3

*Practice Direction SL10.3, Guidance on Setting Aside a Consent Order on Ancillary Relief.* Its own commencement and date:

It defines what it applies to — and the definition is wider than the phrase "consent order" suggests, because it reaches part of an order:

It also marks off the case it does not deal with, which is the case where the complaint is that the court itself erred:

Paragraph 3 is the operative one, and it is the same rule as the judgment above, now in the Judiciary's own words:

And "the following procedure" is a fresh action, in a choice of two courts, commenced by one of two originating documents:

Two consequences the practice direction states in terms, and neither is obvious. First, the proceedings are not private in the way the divorce was:

Second, the action is an ordinary civil action and is case-managed as one:

And it says what happens after a ground is established, which is not automatically the end:

The practice direction's own footnote closes the loop back to section four of this article, because the authority it cites for paragraph 3 is the judgment set out above, at the very paragraphs quoted there:

6. What this establishes, and its limits

What it establishes. One: a consent order on ancillary relief takes its legal effect from the order, not from the agreement behind it. Two: where such an order does not fall within the section 11 variation power, the recognised grounds for setting it aside are fraud, mistake, non-disclosure or misrepresentation of material facts at the time the order was made, and new events invalidating the basis of the order. Three: the route is a fresh action in the Court of First Instance or the District Court, by writ where there is substantial dispute as to facts and otherwise, where appropriate, by originating summons; a summons in the existing proceedings is normally not permissible and is available only in the narrow uncontested or undisputed-mistake case; those proceedings are usually open to the public and are case-managed as an ordinary civil action.

⚠ The limits.

• This is one first-instance judgment, decided on an uncontested application where both sides wanted the same outcome. It is not a threshold for a fought case. On timing: CACV 92/2011 records at its paragraph 23(2) that the appeal route against a consent order usually requires leave to appeal out of time. The four grounds are stated in the judgment by reference to a textbook, and the judge himself called the area ripe for reform.
• Practice Direction SL10.3 tells you the procedure. It does not tell you the substantive law, and nothing in it says what makes a case of non-disclosure or of mistake good enough.
• On non-disclosure specifically, this section states one point only: non-disclosure or misrepresentation of material facts at the time the order was made is one of the recognised grounds on which a consent order may be set aside. What makes a fact material, who bears the burden, how much has to be shown and what time limits apply are outside this section — see the disclosure section above.
• How much weight a court gives an agreement is a different question, answered by the two judgments in the section above on what the agreement is worth. Those judgments go to the weight of an agreement in the ancillary-relief exercise, not to the threshold for setting an order aside. Nothing in this section changes that section, and nothing in that section changes this one.

Which orders can be varied, and which cannot? The line is not capital against income

Section 11 is not a provision that says every maintenance order can be varied. It has a gate, in s. 11(1):

The gate is the phrase "an order to which this section applies". Section 11(2) then defines that class, and it defines it exhaustively — the subsection opens "This section applies to the following orders, that is to say—", followed by six lettered limbs, with no residual limb and no "and any other order".

Set that list against the orders Cap. 192 actually empowers (ss. 4, 5 and 6 only — see the bound stated above):

(Source: Cap. 192, ss. 4, 5 and 6.)

And s. 11(5) does not merely omit capital — it forbids it:

So the variation power cannot be used to turn an income order into capital.

Now the finding, and it cuts across the usual account. Most people remember the line as 〈capital is final, income is variable〉. Three Ordinances written across three decades do not draw that line.

Guardianship of Minors Ordinance (Cap. 13), s. 10(4) — the exclusion, in the statute's own words:

Separation and Maintenance Orders Ordinance (Cap. 16), s. 7(1) carries the same shape:

Note the qualifying words both Ordinances use — "where all such instalments have been paid". A lump sum payable by instalments is excluded only once the instalments are all paid. An unfinished instalment order remains variable. Cap. 192 draws the same line by a different route: it puts s. 4(2)(b), the instalment machinery, inside s. 11(2)(b).

So the operative distinction is finished versus unfinished, not capital versus income. The mechanism follows from what a variation power is: s. 11 is a power over the court's own subsisting order, and where nothing is outstanding there is nothing to vary. A paid lump sum is a completed transaction; a half-paid one is a live obligation. A reader who has been paid three of eight instalments is not outside s. 11 the way a reader paid in full is, and would never work that out from "capital is final".

⚠ Read that narrowly, because it is narrower than it sounds. What s. 11(2)(b) admits is an order made by virtue of s. 4(2)(b) — the provision for paying the sum by instalments. The lump sum itself is ordered under s. 4(1)(c), which is not in the s. 11(2) list, and s. 4(2) is expressed to operate "Without prejudice to the generality of subsection (1)(c)". So being inside the class means the instalment arrangements are something s. 11 can reach. It does not mean the total figure can be reopened. That last question — whether, and on what test, a Hong Kong court will disturb the overall quantum — is decided by case law. That judgment is the subject of the next section.

Variation: what the court requires, from the Court of Appeal

The statute says what s. 11 can reach. It does not say what it takes to move it. That is decided by case law, and the decision usually cited on it is this:

**CH v MEH, CACV 92/2011, Court of Appeal (Cheung JA, Kwan JA and Yam J), heard 20 December 2011, judgment 10 January 2012, on appeal from FCMC 1969/2007. Anonymised, and kept anonymised here.** Paragraph references are the judgment's own paragraph numbers and sub-numbers.

The facts, briefly, because they are close to a reader's own. A consent order made after a successful Financial Dispute Resolution hearing required the husband to pay a lump sum by three instalments and monthly maintenance. He paid the first instalment; the 2008 global financial crisis then reduced the value of his assets, and he applied to have the second instalment discharged — that is, to cut the total down — and to reduce the maintenance. The judge below refused to reduce either, but rescheduled the remaining instalments over several more years. He appealed and lost.

The two routes, and which is which.

At paragraph 23(2), on changing a lump sum payable by instalments:

Those are the same two routes HCMP 276/2012 identified — a fresh proceeding, and an appeal to a higher court — seen from the other end. ⚠ And note the third sentence: an appeal against a consent order usually needs leave to appeal out of time.

Consent does not put an order beyond variation.

At paragraph 23(4):

This is a direct answer to a question most readers ask, and it answers it in the direction most readers do not expect: an order being made by consent does not put it outside the s. 11 variation power. ⚠ Jurisdiction is not the same as outcome. What the court then requires is the next point, and it is demanding.

How far the power goes, and how it is exercised.

At paragraph 24:

So it can reach the amount, not only the timetable. But in the same paragraph:

And the principle behind the caution, at paragraph 26(2) — the Court of Appeal said that

application as those under the appeal-out-of-time test, because

is what supports that view.

The Court of Appeal's own formulation of the test, at paragraph 26(3), set against the Barder test for an appeal out of time, which it describes as requiring

and then says of the variation route that the anticipated circumstances

and, at paragraph 26(4), that the two come to the same thing:

⚠ This is the threshold s. 11(7) does not contain, supplied by the Court of Appeal, and it is a high one. Read it with the section above: the words of s. 11(7) carry no threshold; the case law does.

⚠ But read the scope of it just as carefully, because the judgment draws a line the statute does not. At paragraph 25(1), on periodical payments, the Court of Appeal says there is no similar stricture, and states the test instead as:

and records, citing AEM v VFM, that

So the two are not on the same footing. The demanding threshold at paragraph 26(3) is stated about reopening the overall quantum of a lump sum payable by instalments. A periodical-payments order is expressly said not to carry the same stricture. (That is how CACV 92/2011 records AEM v VFM.)

Why a lump sum by instalments is treated that way, at paragraph 26(5):

and, at paragraph 23(1), the judgment states the power to order a lump sum

The Court of Appeal relied on De Lasala v De Lasala [1980] AC 546 for that proposition.

What does not count as a change.

This is the part of the judgment that speaks most directly to a reader whose ex-spouse's fortunes have moved. At paragraph 26(6), on the collapse in the husband's asset values:

and then the general point, which does not depend on that particular crisis:

and it runs in both directions:

⚠ So on this judgment, assets going up or down in value is not, by itself, the kind of change that reopens the quantum of a lump-sum order. The appeal was dismissed.

The bounds on this section.

• CACV 92/2011 is one Court of Appeal judgment on one set of facts. The husband had a substantial and recovering earning capacity, which the judge below relied on; the judgment does not lay down what any other payer must show.
• The threshold at paragraph 26(3) is about the overall quantum of a lump sum payable by instalments. ⚠ It is not stated as the test for varying a periodical-payments order, which paragraph 25(1) puts on a different footing. Do not move it.
• It is a variation case, not a setting-aside case. It says what the two routes to changing a lump sum order are; it does not describe the fresh-action route in HCMP 276/2012 or in Practice Direction SL10.3.
• It does not decide what makes a fact material for non-disclosure, and nothing in it is used above for that.
• *The other cases named in CACV 92/2011 — AEM v VFM, AEM v VFM (No. 2), HCTT v TYYC, CSL v WWK, G v G, Barder v Caluori, Westbury v Sampson, Shaw v Shaw, Myerson v Myerson and Cornick v Cornick — are referred to here only as that judgment refers to them.*

How much has to have changed? Three provisions, three thresholds, and one blank

The same act — asking a court to change a maintenance arrangement — runs on three different statutory thresholds in Hong Kong, and which one applies is decided by the document you hold, not by how bad your situation is.

(1) Cap. 192, s. 11(7) — the order route.

Read as a whole subsection: there is no threshold word in it. No "material", no "substantial", no "exceptional", no "unforeseen", and no time limit. The change is named as an inclusion within "all the circumstances", not as a precondition. ⚠ That is a statement about the wording of the subsection, not about what a court requires.

That is an observation about the words of the subsection, and it is worth making — but it is only half an answer, and the other half is not in the statute. The question most people ask is 〈how big does the change have to be〉. Neither s. 11(7) nor s. 7, to which it points, answers it. ⚠ The answer is in the case law. CACV 92/2011, set out in the variation section above, supplies the threshold for reopening the overall quantum of a lump sum payable by instalments: the anticipated circumstances "must have changed very significantly or when it is unjust or impracticable to hold to the original order." For a periodical-payments order, the same judgment says at paragraph 25(1) that there is no similar stricture and states the question instead as whether, in all the circumstances and having regard to any changes, it would be appropriate to vary. And on consent: paragraph 23(4) says "Section 11 does not impose any restriction on variation of consent orders." Setting aside an order is a different and narrower route again, described in its own section above. Do not read this section as telling you that a small change will be enough.

(2) Cap. 192, s. 15(2)(a) — the agreement route, and this is the counter-intuitive one.

(The two ellipses mark limb (b), quoted in the children's section below, and paragraphs (i) and (ii). Nothing else is omitted from the subsection.)

Note the parenthesis: (including a change foreseen by the parties when making the agreement). A change the parties foresaw when they signed is expressly included, not excluded.

And read the words the subsection closes on, because they carry two limits the rest of this section has to state. The power is given "subject to subsections (4) and (5)", and it is to be exercised having regard to "the matters mentioned in section 7(3)". Section 15(4) is dealt with below under remarriage; s. 15(5) is dealt with in the children's section; s. 7(3) is a statutory list, not a blank — see (5) below.

(3) Cap. 16, s. 7(1) — the separation-and-maintenance route, on a visibly higher threshold, which requires cause to be shown "upon cause being shown on fresh evidence".

Set side by side:

(A threshold is what gets a reader through the door. It is not the same thing as what the court may then order, and on the agreement route the two are expressly separated: the power is given "subject to subsections (4) and (5)". Section 15(4) caps a spousal periodical-payments term at the payee's remarriage; s. 15(5) routes a child's term through s. 10(1) and (3). Both are set out below.)

(4) And one further thing decides whether a route is open at all, and it has nothing to do with how bad your circumstances are: where your ex-spouse now lives. The agreement route carries a residence gate, in s. 15(1):

Read the words: "each of the parties" — so both of them; "either domiciled or resident" — so one of the two suffices. Both parties must therefore each satisfy one of the two. If an ex-spouse has emigrated and given up Hong Kong domicile, the s. 15 route may simply not be open — which is a great many readers' actual position.

Cap. 16 is drafted the opposite way. Section 7(2) of the same section:

One subject, two Ordinances, one door open and one gated: Cap. 16 says expressly that its jurisdiction survives a party living outside Hong Kong; Cap. 192 s. 15 requires both parties to have a Hong Kong connection. The contrast produces a counter-intuitive result — the reader holding only an agreement, which looks like the weakest paper, is the one most exposed to an ex-spouse's emigration, while those who obtained an order find that s. 11 states no residence or domicile gate at all.

(5) And the "matters" both routes send the court to are written in the Ordinance. They are not a blank. This is the point at which most accounts stop and say the rest is judicial discretion. Section 11(7) does not admit "any change"; it admits "any change in any of the matters to which the court was required to have regard when making the order". For an order under s. 4, those matters are fixed by s. 7, which is headed:

and which opens:

The seven matters, in the statute's own words:

And the agreement route is sent to the same chapter by name: s. 15(2) closes "including, if relevant, the matters mentioned in section 7(3)" — the further matters that apply where the child is not the child of the party being ordered against.

Two things follow, and they must be held together. What the court has to look at is statutory — a reader can read s. 7(1)(a)–(g) and see whether anything on that list has changed since the order was made. What weight the court gives any of them is not: s. 7 states the matters and states no weighting, no formula and no threshold. There is no threshold word in s. 11(7), and there is none in s. 7 either — which does not mean a court requires nothing before it will vary an order.

Four routes; three of them have a threshold stated in a statute, and this table carries those three. The fourth — applying to set aside an order that dismissed the claims — has its grounds and its procedure described in its own section above, but no statutory threshold.

Nominal maintenance: it looks like a full stop and it is a comma

Section 11(2)(b) admits "any order made by virtue of section 4(1)(a) or (b) or 4(2)(b)" — the subsection says nothing whatever about amount. Section 11(7) attaches no threshold in its wording to varying such an order. So to s. 11, $1 a month and $10,000 a month are the same instrument. ⚠ Being the same instrument is a statement about which orders s. 11 can reach. It is not a statement about what a court will do with one.

A judge and the Government's own commissioned researchers reached that conclusion from two different directions.

(1) From the bench. The Family Court seminar paper of 8 November 2024, paragraph 32:

The weight is stated with it: page 1 of that paper disclaims it as the speaker's personal views (quoted above); it is neither a practice direction nor a judgment. The proposition itself rests on s. 11, not on the paper.

(2) From the data. The Government's commissioned study puts it this way:

(3) The scale, with its ceiling in the same sentence. The Census and Statistics Department's Thematic Household Survey Report No. 61 records that up to 20,800 people — up to 37.7% of the 55,100 who would receive maintenance — were on nominal maintenance of $1 from an ex-spouse, or were awaiting the court's ruling on the mode of maintenance payment.

The words "up to" cannot be dropped. The published cell is a composite: it holds both the people on a nominal $1 and the people still waiting for a ruling on how payment is to be made, and the two cannot be split. So:

• upper bound: 20,800 people, 37.7% of 55,100.

The fieldwork ran from October 2015 to January 2016, and that is the date the figure carries — not 2022 and not today. (A November 2022 government study re-analyses the same survey data; it is not a new enumeration.)

Remarriage, and years of delay: four clocks, and what each is attached to

The four things readers fear most are four different machines, each bolted to a different door. The Ordinance writes them in four separate places.

(1) Remarriage. Cap. 192, s. 9(4):

Read precisely:

• It bars applications under s. 4, s. 6 or s. 6A — the original ancillary-relief powers. Section 11 and s. 15 are not in that list. (That is true of the s. 9(4) bar and it is not the answer to 〈does remarriage end my maintenance〉. The practical effect on a subsisting spousal periodical-payments order is supplied separately, by s. 9(2)(a)/(b) and s. 9(3), which end the maximum term at remarriage — set out below.)
• It bars the party who remarries, in respect of that party's own application, and against the former spouse only.
• It says nothing about the payer's remarriage. The bar follows the person who remarries, not the person receiving.
• Arrears already due are not swept away by it. Section 9(3), on orders made in judicial separation proceedings:

And s. 9(4) is a bar on applying, which is not the same question as what remarriage does to an order the reader already holds. That is answered by s. 9(2), and it is answered on the reader's own facts — because s. 9(2)(a) and (b) are the divorce-and-nullity limbs. Section 9(1) says the term specified in a s. 4(1)(a) or (b) order may not run "longer than the maximum term"; s. 9(2) defines that term:

Read against the reader's own position: if you obtained periodical payments on a divorce, the longest term the court could have specified ends at your own remarriage. That is a cap on the order, not a bar on an application — the two are different machines, and it is s. 9(2), not s. 9(4), that is bolted to the reader's own door.

The judicial-separation limbs are the other two, and they carry no remarriage limb at all:

And the agreement route carries the same cap, at s. 15(4). Where a court alters an agreement by inserting spousal periodical payments or increasing their rate, the term it may specify "shall not exceed"—

So on both routes the payee's remarriage is a ceiling on the term, and on neither is it a bar on asking. (s. 9(2) and s. 15(4) cap what the court may specify. What a court does with a subsisting order on an application under s. 11 is a separate question.)

(2) Twelve months. Cap. 192, s. 12(1):

This is not a limitation period that destroys the debt. It converts enforcement from something done as of right into something done with the leave of the court. Section 12(2) then sets out what the court may do:

Three things follow. It bites on enforcing arrears, not on applying to vary — s. 11 is not among the listed orders. The clock runs to the start of enforcement proceedings, and separately for each instalment of arrears. And the court may refuse, may impose conditions, or may remit the arrears in whole or in part — so arrears neither only ever grow, nor are they certain to be recovered in full.

(3) Six months — the only hard clock here. And before the clock, two gates, both in s. 16(1). Section 16 reaches an agreement only where the agreement itself provides for the payments to continue after a death, and only where the party who died was domiciled in Hong Kong:

So a reader whose agreement says nothing about what happens on a death, or whose ex-spouse died domiciled somewhere else, is not on this clock at all — there is no six-month deadline to miss, because the door the six months times was never open. Only once both gates are passed does s. 16(2) bite:

That is not a calendar date; it runs from an event. The event is the first grant of representation to the deceased's estate — not the date of death. So "six months after your ex-spouse dies" is not merely imprecise: it names the wrong event. And the subsection carries "except with the permission of the court" on its face, so even the hard clock has a door the court controls. Section 11(6) sets the identical clock and the identical escape for secured periodical payments orders.

(4) The Limitation Ordinance names no matrimonial cause of action — and that is a narrower statement than the one you will usually see. Cap. 347 contains no provision addressed to maintenance, divorce or ancillary relief.

But the chapter's general provisions are subject-matter-neutral on their face, and three of them have to be named here:

• s. 2(1) defines the word the whole chapter turns on: action "includes any proceeding in a court of law".
• s. 4(4), which is about judgments and says nothing about what the judgment was for:

• s. 40, the saving, which disapplies the chapter where another Ordinance prescribes the period:

Why those three are not trivia here. Cap. 192 s. 28AA(3)(a) deems arrears of maintenance a judgment debt — but it deems this "for the purposes of section 50 of the District Court Ordinance (Cap. 336)", and for those purposes only. And Cap. 192 s. 12(1) is a leave rule, not a prescribed period, so it is not obvious that s. 40 takes this reader out of Cap. 347 either. Whether s. 4(4) reaches the enforcement of a matrimonial money order is a question of interpretation.

⚠ So do not assume the Limitation Ordinance does not reach any of this. What a reader can rely on is this: the time limits written for this reader are in Cap. 192, not in Cap. 347 — the 12-month leave rule at s. 12(1), the s. 9(4) bar, and the six months from first grant of representation at s. 16(2) and s. 11(6).

One conclusion here has to be stated carefully. On the words of the Ordinance, delay past 12 months does not extinguish the arrears: it makes enforcing them conditional on the leave of the court. ⚠ That is not the same as saying leave is a formality. Section 12(2), quoted above, lets the court refuse leave outright, and lets it remit the arrears in whole or in part — so on the statute's own words a long delay can cost a reader the money, and not merely the right to enforce as of right. How readily leave is given, and what has to be shown to obtain it, is decided by Hong Kong case law — CSL v WWK is the decision usually cited on it. Do not read this section as an assurance that delay can simply be explained away. The mechanism the statute uses is nonetheless clear: the arrears in s. 12 arise under an order the court itself made, so the legislature left the delay to the court to weigh rather than to a limitation period to extinguish — which puts the delay in front of a judge who may or may not be persuaded by it.

Interest on arrears of maintenance is in Cap. 192 itself, and it does not wait for anyone to obtain a judgment. The section is s. 28AA and it is headed "Interest on arrears of maintenance". Subsection (2) states an entitlement:

Subsection (3) then does the work that 〈reduced to a judgment〉 would otherwise have to do — by deeming it:

Read the three limbs together. Cap. 192 supplies the entitlement; Cap. 336 s. 50 supplies only the calculation; and the date interest runs from is the date the maintenance was due, not the date of any later judgment. So a reader who has not been to court to convert arrears into a judgment has not thereby lost the interest.

And the debtor has an answer written into the same section. Section 28AA(7) lets a judgment debtor who considers he has reasonable grounds apply by summons not to pay the interest; s. 28AA(8) then directs the court to "take into account all the circumstances of the case", on a list that includes whether he has a reasonable excuse, whether he evaded service, his past record, whether he explained himself, and his ability to pay. Interest under s. 28AA is therefore neither automatic nor discretionary in the loose sense: it is an entitlement with a statutory answer to it.

The rate, though, is in neither section. District Court Ordinance (Cap. 336), s. 50(1):

No figure appears in either chapter. The rate is whatever the Court orders in the case, or, failing that, a rate the Chief Justice determines by order — a power exercisable at any time without amending the legislation. For the same reason, s. 12(3) leaves the manner of a leave application to rules of court, which can change without amending Cap. 192.

A child's maintenance is a separate question, and it is locked down harder

Many readers assume that 〈neither of us will claim〉 sweeps up the children's money too. The provisions do not read that way.

(1) A child's maintenance is inside s. 14's own definition. Cap. 192, s. 14(2), defining "financial arrangements", expressly includes:

So a provision purporting to restrict the right to apply for a child's maintenance falls inside s. 14(1)(a) and is void on the same terms.

(2) Section 15(2)(b) is a free-standing limb, and it requires no change at all:

Set that against s. 15(2)(a): paragraph (a) needs a change in circumstances; paragraph (b) does not. The court need only be satisfied that the agreement does not contain proper financial arrangements for a child of the family.

But (b) runs into the same closing words as (a), and for a child it is s. 15(5) that matters. The power is given "subject to subsections (4) and (5)", and where the court alters an agreement by inserting or increasing periodical payments for a child of the family, s. 15(5) directs that, in deciding the term:

So the limb requires no change in circumstances, and the term it can produce is still routed through s. 10.

(3) There are two further, wholly independent routes. Cap. 13, s. 10 is a route that depends on no divorce and no agreement; and Cap. 16, s. 7(1) carries the higher "fresh evidence" threshold described above.

(4) The proposition that parents cannot bargain away a child's maintenance rests on the statutory text set out above.

Why the Hong Kong law looks like this: one borrowed sentence, and one decision actually taken

This section will not change the outcome of anyone's case. It explains why the clause is so securely void — and which parts of the Hong Kong position were decided and which were simply never revisited.

(1) The sentence that voids your clause is not Hong Kong's sentence. Matrimonial Causes Act 1973 (England and Wales), s. 34(1):

Word for word the same as Cap. 192, s. 14(1). The Hong Kong text differs only in its cross-references (ss. 15 and 16 for ss. 35 and 36) and in carrying [cf. 1970 c. 45 s. 13 U.K.] at the end. Many people assume the anti-ouster rule is judge-made across the common-law world. It is not: it is statutory on both sides, and it is the same statutory sentence.

(2) Neither legislature has amended it. In Hong Kong, s. 14 carries only an editorial revision note (E.R. 3 of 2018), while ss. 10, 11, 15 and 16 beside it have each been substantively amended several times. In England, s. 34 of the 1973 Act carries no amendment, while s. 31 beside it has been amended many times.

So the clause is not void because a court might take a dim view of it. It is void because two legislatures wrote the same sentence and neither has touched it in more than half a century. A judge-made rule can be distinguished on its facts; this one would have to be repealed. Those are different kinds of security, and the reader has the second.

(3) England then built three things around that sentence; Hong Kong built none of them.

MCA s. 25A(1), the duty itself:

(Punctuation as in the revised text.)

On these three, Hong Kong's position is inertia in a copied text — the 1970 wording was taken in 1972 and this corner was never revisited.

(4) But on s. 14 itself the opposite happened, and it is documented four times.

• 1992 — the profession asked for it. The Law Reform Commission sub-committee records that the Law Society suggested that parties who reach a mediation agreement without independent legal advice risk having the court set it aside unless s. 14 were amended. (As recorded by the sub-committee.)
• December 1998 — a sub-committee consultation paper declined, at Recommendation 15.128: "We do not see the need to amend section 14 of the Matrimonial Proceedings and Property Ordinance (Cap 192) to remove any apparent obstacle to enforcement of a mediation agreement. This section provides that a provision in a maintenance agreement restricting the right to apply to a court for an order concerning financial arrangements is void." But that paper states its own status on its page 2: "does not represent the final views of either the Sub-committee or the Law Reform Commission, and is circulated for comment and criticism only." So it may not be cited as a Commission recommendation.
• January 2002 — the Commission's Report on Guardianship of Children did not deal with s. 14; the question was taken up in another report.
• March 2003 — the Commission's own Report, The Family Dispute Resolution Process, Recommendation 24: "We do not see the need to amend section 14 of the Matrimonial Proceedings and Property Ordinance (which provides that a provision in a maintenance agreement restricting the right to apply to court for an order concerning financial arrangements, is void)." And paragraph 6.61 records the consultation outcome: "On consultation, all the respondents who commented on this recommendation expressed support for it, except for one respondent who submitted that section 14 should be amended for the avoidance of any doubt regarding the enforcement of mediation agreements."

So "Hong Kong never legislated a clean break" and "Hong Kong deliberately kept the anti-ouster rule" are two different kinds of fact, and writing them in one breath makes the second disappear. The first is a borrowed sentence left where it was. The second is a named body, in a published report, refusing a request from the profession — with the consultation outcome on the record.

(5) And three facts must travel together, or the third turns the first two into a falsehood.

• Cap. 192 contains no clean-break duty of the s. 25A kind.
• Cap. 192 contains no s. 31(7A)-style power to capitalise on a variation — and s. 11(5) affirmatively bars a lump-sum order on an application to vary a periodical payments order. Read the preposition.
• But Hong Kong courts do capitalise periodical payments. A Hong Kong family-court tables site records the practice. (This site's summary:) the method is used where the parties, or the court, would rather end a continuing income stream and settle on a single capital sum. The payee receives a fund instead of the payments, invests it, and draws it down, the calculation being set so that the money is spent by the end of the period the payments were meant to cover — normally the payee's lifetime. *It takes its name from the parties in the English case Duxbury v Duxbury [1987] 1 FLR 7 (CA), and the same source records that it has been in use in the Hong Kong courts since C v C* [1990] 2 HKLR 183

The mechanism is s. 4(1)(c), which Cap. 192 does contain:

That is a practitioner source's account of Hong Kong practice. And capitalisation belongs to the original ancillary-relief exercise, not to a variation application — s. 11(5) says so.

A right on paper is not money in hand

The Ordinance gives a right to apply. It does not guarantee payment. This section is not an argument against applying; it is the denominator.

Census and Statistics Department, Thematic Household Survey Report No. 61, fieldwork October 2015 – January 2016:

• Of the 34,300 ever divorced or separated persons who would receive maintenance, 40.5% (some 13,900) did not receive it in full.
• Of those 13,900, only some 1,700 (11.9%) had taken legal action to recover the arrears; the remaining 12,200 (88.1%) had not.

Each base is named in the same sentence as its percentage. The 40.5% is of 34,300 — a figure that already excludes the 20,800 on nominal maintenance or awaiting a ruling. The 88.1% is of 13,900.

(A November 2022 government study, recomputing from the rounded headcounts, gives 12.2% / 87.8%; the figures above are the Census and Statistics Department's own 11.9% / 88.1%.)

These figures are not a reason to give up; they separate two questions. Whether the provisions open a door is one question. What happened after other people went through it is what the survey measures — and it has no necessary bearing on the facts of any individual case. Read it alongside s. 12(2): on a leave application the court may remit arrears in whole or in part, so arrears are neither guaranteed to grow nor guaranteed to be recovered.

The question many people ask: is there anything I can work out myself?

The legislation supplies no formula that predicts an individual's award; amounts depend on the case. But two things are calculable, and both are arithmetic rather than prediction: how the population figures relate to each other, and why certain figures must never be added.

Working the numbers

Every figure below has already been cited above.

(1) The two big populations overlap — so they cannot be added.

Of the 301,100 who had neither applied nor intended to apply for a maintenance order, 4,600 (1.5%) reported that an ex-spouse would pay under a maintenance agreement, leaving 296,500 (98.5%) with no such agreement. Separately, 55,100 people would receive maintenance — and that 55,100 includes the same 4,600.

• So 301,100 + 55,100 = 356,200 is a wrong step: the 4,600 are counted twice.
• This is not pedantry: without it, the two bases get merged and the resulting percentage means nothing.

(2) The 48,000 figure sits on a base of 296,500, not 301,100.

Some 48,000 people — 16.2% of those 296,500 — gave as their reason that both parties had agreed not to require maintenance of each other.

• 48,000 ÷ 296,500 = 16.2%, which reconciles.
• But 16.2% must not be added to the other reasons. The table states that multiple answers were allowed, and the four published reasons already total 55.9 + 16.2 + 11.7 + 8.4 = 92.2% with further rows still to come. A multiple-response percentage is not a share of a partition.
• The same table also carries suppressed cells (statistics not released because of large sampling error), so it is not a complete cross-tabulation.

(3) Writing out the "up to a third" figure shows why it is a ceiling.

20,800 ÷ 55,100 = 37.7%. But the 20,800 holds two groups: those on a nominal $1, and those still awaiting a ruling on the mode of payment. The two cannot be separated. So:

• upper bound: 37.7%;
• lower bound: not derivable.

(4) Gender: two figures, both true, about two different things.

• 28.2% of female ever divorced or separated persons had applied or intended to apply for a maintenance order, against 2.2% of males.
• But in the same report's reasons table, more men (27,800) than women (20,200) gave mutual agreement as the reason for not applying — about 58% of that 48,000 is male.
• So the people who ask are overwhelmingly female, and among those who gave mutual agreement as their reason for not applying, men were the larger group. ⚠ That cell is a survey answer, not a legal waiver: a response that both sides agreed maintenance was not required establishes no written agreement, no court order and no informed relinquishment — and this article's own thesis is that some no-claims provisions are void.