Family Law Week Email SubscriptionAlphabiolabsBerkeley Lifford Hall Accountancy Services

Home > Articles

Applying the CMS Formula in High Income Cases

Joshua Viney, barrister, and Henry Pritchard, pupil barrister, both of 1 Hare Court, explore the circumstances in which it might be unfair to apply the CMS formula.






Joshua Viney, barrister, and Henry Pritchard, pupil barrister, both of 1 Hare Court


In December 2019 Mostyn J handed down judgment in CB v KB [2019] EWFC 78. In his judgment he suggested that the correct approach to the quantum of child maintenance in high income cases, where the non-resident parent's income is in excess of £156,000 gross per annum, was to use the figure provided by the CMS formula as a starting point.

In this article we explore how this may be good for certain cases but suggest that it may lead to unfairness in others.

The statutory jurisdiction to award child maintenance

The court retains jurisdiction over child maintenance in a small number of scenarios. This article focuses on the two of these most often engaged where incomes exceed the CMS threshold: top-cases and written agreement cases.

By the operation of s.8(6) CSA 1991, a case is part of the top-up jurisdiction where:

i. a CMS assessment is in force;

ii. the payer's weekly income (as assessed by the CMS) exceeds the maximum applicable amount of £3,000 per week (£156,000 per annum); and

iii. the court is satisfied that the circumstances of the case make it appropriate.

The court may also have jurisdiction over child maintenance where the parties' consent to the court having jurisdiction by virtue of s.8(5) CSA 1991 and the Child Maintenance (Written Agreements) Order 1993.

Where the courts have jurisdiction, they may order periodical payments pursuant to s.23(d) of the Matrimonial Causes Act 1973, informed by the s.25 factors or pursuant to paragraph 1(2)(b) of Sch 1 to the Children Act 1989, informed by paragraph 4 of the same schedule. The court therefore retains a broad discretion in relation to child maintenance payments, in stark contrast to the far more rigid formula applicable in cases where the CMS has jurisdiction.

The authorities on using the CMS formula prior to CB v KB

Nicholas Mostyn QC (as he then was), in GW v RW [2003] 2 FLR 108, set out to qualify and clarify the broad discretion of the courts in relation to the quantum of child periodical payments in a case where the court had jurisdiction because of the habitual residence of the non-resident parent. The judge noted at [74]:

"… I am of the view that the appropriate starting point for a child maintenance award should almost invariably be the figure thrown up by new child support rules."

The main justification for this, he argued, was that parties should negotiate "in the shadow of the CSA", following the principles outlined in the government's White Paper: A New Contract for Welfare: Children's Rights and Parents' Responsibilities [1999] Cm 4349. Specifically, it would undermine parties' agreements over child maintenance if the formula outcome was too divergent from the agreed level of payments, since an aggrieved party would simply have to wait a year before applying for a calculation, negating the agreement.

The approach of using the formula as a starting point had been considered by the higher courts on three major occasions prior to CB v KB:

i. SW v RC [2008] EWHC 73 (Fam), where Singer J described it as a "useful rule of thumb", albeit obiter, since the case at hand "was never a routine case …", where the jurisdiction appeared to have been based on a prior written agreement and where there had been no assessment.

ii. Re M-M [2014] EWCA Civ 276, where McFarlane LJ held that the judge below had been correctly referred to GW v RW and SW v RC since, although it was a case where the court's jurisdiction was founded on the husband's residence abroad, "… it is informative to consider what the position would be were he to be resident in England and Wales."

iii. In Re TW & TM (Minors) [2015] EWHC 3054 (Fam) Mostyn J again deployed the starting point of the formula and seemed to apply it specifically to top-up cases. It is slightly unclear whether this was actually the ratio of this case, taking into account the judge's exposition on the basis for the court's jurisdiction over child maintenance at [3]:

"In such circumstances it simply cannot be gainsaid that the prior orders were validly made. Indeed, in order perhaps to make assurance doubly sure, it should be recorded that in the case of TM the earlier orders recorded that the Child Support Agency had carried out a maintenance calculation in the maximum amount thereby entitling the court to make an order topping up that calculation under s.8(6). There can be no doubt at all that there were in existence valid prior orders capable of variation and the suggestion that the learned deputy district judge did not have jurisdiction to vary her order is legally devoid of any merit."


This case was the final hearing in an application for financial remedies. The husband was a member of a well-known rock band. The main focus of the judgment was on the valuation of the various streams of income available to the respondent husband. A subsidiary issue was the quantum of child maintenance payable.

Mostyn J held that the appropriate approach for the court to take when quantifying child maintenance was to use the figure provided by the CMS formula as a starting point when determining the maintenance payable in top-up cases. The judge referred back to Re TW & TM at [9]:

"My decision in GW v RW makes it clear that where a court is considering issues of child maintenance the formula is not, so to speak, written in marble but supplies only a starting point. There may be in a case a very good reason why there should be departure from the starting point of the formula. In my opinion the formula should apply even where the earnings of the father are in excess of the £3000 per week maximum provided for in the Act and the Regulations. If the earnings of the father were very much in excess of that then there would be a good reason to depart from the formula downwards, but if the income of the father is not un-adjacent to the maximum then to my mind, subject to other factors, that of itself is not a good reason to depart from the formula." (emphasis added)

Mostyn J developed this approach in his conclusions to CB v KB:

"[49] I suggest that in every case where the gross annual income of the non-resident parent does not exceed £650,000, the starting point should be the result of the formula ignoring the cap on annual gross income at £156,000. For gross incomes in excess of £650,000 I suggest that the result given by an income of £650,000 should be the starting point with full discretionary freedom to depart from it having regard to the scale of the excess.

[50] In this case the relevant factors are:

i) Gross income of the husband: £639,000;

ii) Adjustment to gross income referable to three children in the husband's family: 84%; and

iii) Adjustment to computed sum referable to shared care of three subject children: 62.5%.

These factors lead to a computed sum for child support for the subject children of £50,269 per annum or £12,567 for each child.

[51] Having considered the budget of the wife referable to the children I cannot see any good reason materially to depart from the starting point of £12,567 per child. The figure for each child will be rounded to £12,600 and will be payable until each child completes tertiary education." (emphasis added)

The ratio of the judgment appears to be that as a starting point the court should apply the CMS formula in a high income case, with a cross-check against the stated needs of the child in question whereby the court would require a 'good reason' to materially depart from that starting point. We consider below what might constitute a good reason.

Authorities after CB v KB

Since CB v KB was published it appears to have been followed without comment by Nicholas Cusworth QC in G v T [2020] EWHC 1613 (Fam) who stated:

"[70] As to the level of child maintenance, I have fully in mind the views of Mostyn J as recently expressed in CB v KB [2019] EWFC 78. I also bear in mind the fact that under my order the significant majority of the husband's receipt from all sources over the next 3½ years will be employed in meeting the wife's capital entitlement. However, it also seems appropriate to add to what would otherwise be the appropriate level of child maintenance an additional element for extra-curricular expenses, to avoid future dissention between the parties. Payment should be at the rate of £35,000 pa per child, and of conventional duration."

It was then followed by HHJ Hess in W v H (divorce: financial remedies) [2020] EWFC B10 (following a maximum assessment) again without comment:

"[28](iii) Following the logic of Mostyn J's judgment in CB v KB [2019] EWFC 78 (paragraphs 48 and 49), which both parties have agreed, there should now be a child periodical payments top up order at the rate of £525 pcm now, reducing to £420 pcm from Summer 2020 and to £315 pcm from Summer 2021. Again, this will carry on until Summer 2027 when C leaves school."

Finally, Mostyn J applied his own approach again in his case of OG v AG [2020] EWFC 52: 

"[101] As to the general child support, I refer to my decision of CB v KB [2019] EWFC 78 where I suggested at paragraphs 47-51 that a useful starting point for gross incomes up to £650,000 is the statutory child support formula. Mr Sharp QC and Mr Commins argue that the husband's gross income is at least £130,000. However, that calculation is before the husband's receipt of the substantial pay-out for which I have provided above. I anticipate that the husband will invest his pay-out in his new business. I have already stated that I am satisfied that the new business will generate a generous income for him. In my judgment, a reasonable gross income to attribute to the husband is £200,000 per annum. This gives rise to a liability under the formula of £19,248. The overall child support liability is therefore £26,748, which I round up to £27,000 per annum or £2,250 per month. This will be payable monthly in advance starting on the first day of the month after the receipt by the husband of his pay-out from X. It will be indexed annually by reference to movements in the CPI. It will continue until the conclusion of secondary education. In tertiary education it will fall to 50% of the secondary education rate."

Anecdotally, we have experienced a number of judges in different courts simply applying the CB v KB approach as a rule rather than a tool.

The benefits of a formula in a high income case

It appears that there are broadly two benefits to a formulaic approach. The first is that it is simple. The court, litigants in person and representatives can plug numbers into a calculator and quickly ascertain a ballpark figure for a maintenance liability. In the right case this can reduce litigation and resolve this discrete issue at an early stage.

Second, it can also, in the right case, be fair. Where one party has a significant income and the other party has no income or a very modest income, then the percentage thrown out may well accurately reflect a fair outcome and meet the needs of the child in question.

Concerns about a formula in a high income case

As set out above, whilst for the right case a formula may be appropriate, we would highlight the following interlinking concerns for the wrong case and generally:

(i) The approach disregards the parent with care's income. Where that income is significant, this could lead to unfairness and create barriers to settlement. 

(ii) The approach appears to refocus a division of income on the basis of percentage sharing rather than need. Where the formula exceeds the child's needs, this provides an opportunity for a party to pursue spousal maintenance through the backdoor. Further it could lead to a significant overinflation of children's needs in budgets and/or the co-opting of a parent's needs into a child's budget.

(iii) The approach ignores that Parliament specifically imposed a threshold for contributions towards children, preferring a discretionary approach beyond the threshold. Again, this would indicate a focus on need rather than sharing.

(iv) The approach could lead to disputes about the amount of time a child spends with a parent. This is already a criticism of the CMS system, but given the sums in contention, this problem could be exacerbated in the top up arena. 

If the CB v KB approach is correct, then some of the above concerns may also point to 'good reasons' to materially depart from the formulaic approach.

(i) Disregarding of parent with care's income and resources

Where the parent with care has (i) a significant, comparable or higher income than the non-resident parent or (ii) capital savings that provide for a significant, comparable or higher income than the non-resident parent, then there is a large opportunity for unfairness where that income is ignored.

The disregarding of the parent with care's income is a feature of the CMS formula as it now stands. This is the result of a political compromise 20 years ago predicated on the need for the system to be more administratively workable. This element of the CMS formula is potentially unfair; a fact which the government of the day recognised but considered to be outweighed by the need for the system to be less complex.

The original Child Support scheme enacted following the Child Support Act 1991 had many problems. However, what it did not lack was an attempt to provide for a fair outcome. Chief among the proposals to provide a fair level of child maintenance was that the formula would take into account the incomes of both the parent-with-care and the absent parent (later to be re-named "non-resident parent"). The White Paper which introduced the scheme – Children Come First: the Government's proposals on the maintenance of children (1990) Cm 1264 – introduced this element of the formula thus:

"3.14 Both parents of a child have a liability to maintain that child. Where both parents have an income, it is therefore appropriate to take both incomes into account. Most children of separated families live with one of their parents, usually, but by no means invariably, their mother. Where this is the case, there will of course be no financial transaction between the caring parent and the children, but the caring parent's capacity to contribute can, in some circumstances and at higher incomes, affect the calculation of the absent parent's contribution." (emphasis added)

This principle of both parents' incomes being taken into account for the purposes of the formula was duly enacted in the s.1(1) Child Support Act 1991:

"For the purposes of this act, each parent of a qualifying child is responsible for maintaining him"

However, the early years of the 1993 scheme (which implemented the 1991 Act) were beset with difficulties. The New Labour government set out to reform the system and laid out its proposals in a Green Paper Children First: a new approach to child support (1998). The Green Paper was unsparing in its criticism:

"We believe that the current complicated assessment formula should be scrapped. It seeks to be fair to everyone and ends up with a system that is so complex that no one can tell whether it is fair or not." [Ch.5, para 3]

The solution proposed by the government was to radically simplify the formula [at Ch.6, para 8]:

"The current formula asks for information which should be irrelevant to calculating liability to maintain children. And often, parents object to supplying this information. We want to reduce these objections and encourage parents to comply with the scheme. Specifically:

The new scheme will not take account of the parent with care's income. Every child who lives with a parent shares automatically in her living standards and income. And non-resident parents still have a responsibility to their children, however much support and financial provision the resident parent provides." (emphasis added)

Accordingly, the parent with care's income was to be entirely disregarded.

These proposals were the subject of criticism themselves, most notably by Nicholas Mostyn QC (as he then was), in a submission to the Social Security Select Committee on 5 November 1998. He was scathing of the fact that the income of the parent with care was to be disregarded under the new scheme:

"20. … The second aspect of gross unfairness is the failure of the new scheme to bringing into account the income of the parent with care … This logic is incomprehensible. Why should a non-resident parent pay the same for his children whether his former partner earns nothing or £200,000? A fundamental ingredient of the old scheme was to bring into account the income of the residential parent. It reflects elementary fairness."

It was noted that the taking into account of the parent with care's income had been one of the redeeming features of the 1993 scheme:

"24 … the income of the parent with care has from the very start been a central feature of the formula. For all the criticism made of the present system, whether by politicians, judges, lawyers, commentators, journalists or pressure groups there has never been the slightest murmur of criticism of this fundamental axiom. Why? Because it is so obviously fair."

Nicholas Mostyn QC noted that the outcome of such a system amounted to the child in question simply sharing in the non-resident's income, rather than being supported by it according to their needs:

"24. As to the failure to bring into the reckoning the income of parent with care the Minister's response was simply that if the parent with care had an income of significance then the child would by necessity be sharing in it. It is therefore fair for the father to pay the same proportion of his income whatever the income of the mother. The response to this is that it will be very difficult to persuade the fathers to see it like this."

He concluded his evidence to the Select Committee thus:

"Of course, it is a fact of life that the parent with care will spend a part of her own income on the child. But that obvious truism cannot lead to the conclusion that the paying parent should therefore pay the same whatever the income of his former partner might be. Every child requires a finite amount of support. That obligation should be shared rateably between the parents in the ratios of their respective incomes. Only thus will the system reflect the legal obligation on each parent to support the child in question. The extent to which the parents should then endow largesse on the child from their residual incomes should be a matter for them. A percentage system does not fit easily with these precepts, while for all its faults the present system clearly does … The failure to take into account the income of the non-custodial parent and the failure to impose a maximum are hideously unfair. There is no justification for them." (emphasis added)

The new system, enacted in the Child Support, Pensions and Social Security Act 2000, was introduced as a clear compromise between fairness and administrative workability. This was set out by the Parliamentary Under-Secretary of State, Baroness Hollis of Heigham in a House of Lords debate on 17 April 2000 (Hansard, Vol 612):

"I was asked why we are not taking into account the income of the parent with care … The first point is philosophical: she [the parent with care] is already contributing. The second point is practical and administrative: 96 per cent of parents with care have incomes of less than £100 a week. Only 500 on our casebook of 1.2 million have incomes over £500 a week. At that point, they might be expected to be making a realistic financial contribution … we should have to assess the income of every parent with care and … If that were to be done, instead of seeking three pieces of information, we should have to seek 30 or 40 in order to get an accurate assessment. In that case, noble Lords would be reimporting back into our simple formula some of the complexity that has made the existing agency impossible successfully to administer. Therefore, philosophically, it is not right; administratively, it would be a nightmare; and it is not worth it."

The underlying premise of the Labour government when enacting the new scheme was that it would not be practical to assess the incomes of parents with care, particularly when research suggested that most parents with care had very little income. It would simply not be an efficient use of resources to assess such people. This could lead to unfairness, but it would be justified as a political compromise intended to make the scheme work.

By returning discretion to the court to award child maintenance where a non-resident parent earns more than £156,000 per annum, Parliament has mandated that the court must consider the income and resources of both parties – section 25(2)(a) of the Matrimonial Causes Act 1973 and paragraph 4(1)(a) Schedule 1 to the Children Act 1989. To simply take the CMS formula and apply it, fails to carry out this exercise. 

As set out above, where the parent with care's income and resources are modest, the impact on the fairness of the case may be limited. However, consider the below scenario. This is deliberately at the higher end of the £650,000 range to express the point starkly. 


Mr and Mrs Bloggs both earn £500,000 per annum. They have one child, Bloggs Junior. The family lived modestly during the marriage. After they separate Bloggs Junior lives with Mr Bloggs and Mrs Bloggs broadly equally. There is a slight discrepancy in that Bloggs Junior spends an additional one week a year more with Mr Bloggs than Mrs Bloggs. Mr Bloggs is deemed to be the parent with care. Mr Bloggs and Mrs Bloggs agree that neither of them has a spousal maintenance obligation towards the other and their assets are divided equally.

If we apply the CMS formula to this family, Mrs Bloggs is obliged to pay approximately £24,000 per year to Mr Bloggs. What is the purpose of this payment? Can Bloggs Junior spending 1 additional week with Mr Bloggs really justify an additional £24,000 to meet needs?

(ii) Post-separation sharing of income

The practical consequence of raising the threshold to £650,000 and applying a formula to income below that level is that rather than applying a discretionary needs-based approach to the income above £156,000, there is effectively a 'share' of the income above that rate. Once a child's needs have been generously met, it is difficult to reconcile any additional payment with the Court of Appeal decision in Waggott v Waggott [2018] EWCA Civ 727, where the court held that a party's earning capacity is not susceptible to post-separation sharing as though it were a matrimonial asset (per Moylan LJ):

"121. … is an earning capacity capable of being a matrimonial asset to which the sharing principle applies and in the product of which, as a result, an applicant spouse has an entitlement to share? …

122. In my view, there are a number of reasons why not."

Of course, Schedule 1 authorities make it clear that there are no grounds for applying any Matrimonial Causes Act notions of sharing or conduct to a child maintenance award. 

Applying a formula above £156,000 can raise many problems. Where the formula exceeds the needs of the child in question, this could provide an opportunity for a party to pursue spousal maintenance through the backdoor. This applies equally to both cases brought under Schedule 1 to the Children Act 1989 and the Matrimonial Causes Act 1973. Admittedly, Mostyn J does provide a cross-check against the child's budget but suggests that there would need to be good reason to depart from the starting point of the formula.

However, this in itself could lead to parties overinflating their budgets for their children, to enable them to continue to pursue these sums and/or co-opt their own needs into a child's budget.

(iii) Parliament determined that there was to be a threshold

Put simply, Parliament determined that there was to be a threshold and distinguished between the sum of maintenance to be paid above and below that threshold – below would be formulaic and above would be discretionary. The authors of the 1990 White Paper set out the reasons for this:

"3.28 If the formula were applied to all levels of income, some people with very high incomes would pay large amounts in maintenance. Such large amounts might be seen as unreasonable or inappropriate especially if some other settlement had been made for the children, such as a family trust. The Government is therefore proposing to establish an upper income limit. The formula would apply up to that limit. No further child maintenance would be deducted from income beyond that limit." (emphasis added)

The Secretary of State for Social Security, Tony Newton, emphasised the issue of incomes in the top-up bracket during a debate on 4 June 1991 (Hansard, Vol 192):

"The courts will also retain jurisdiction to cover some elements of maintenance not covered by the formula. The first of these is the child's right to share in higher levels of income. The operation of the formula will be limited at the upper end to prevent excessively high payments of maintenance becoming due, but the courts will be able to award maintenance over and above any formula assessment. The courts will also have a top-up power allowing them to make an order in addition to the normal formula award, for education costs, which are not covered by the formula."

This issue was so uncontroversial that it was barely debated at the time. It stood to reason, as it still does, that percentage payments tied to levels of income would become excessive at some stage when meeting the needs of the relevant child. There must be a finite cap on any child's needs.

The impact of the widespread adoption of Mostyn J's approach in CB v KB is such that the maximum assessment has effectively been abolished, albeit with the quasi-substitution of £650,000 as a new upper limit. It is unclear why the judge alighted on that figure. It may have been because the husband's income in CB v KB was slightly less than this at £639,000.

The result of this approach appears to shift the judicial focus away from a discretionary consideration of need.

(iv) Child welfare

One of the problems with the CMS system is that there is a monetary advantage / disadvantage to (i) being the parent with care and (ii) the number of nights spent with each parent. To a degree this is unavoidable and part of the political / administrative compromise made in the 1990s.

However, the CMS formula was designed to apply to a range of incomes up to £156,000. By taking that cap and increasing it to £650,000, the scale by which a party can gain or lose a financial advantage is increased significantly. For example, on a gross income of £600,000, the liability for the non-resident parent of one child under the formula could range between approximately £50,000 and £29,000, depending on whether they had the child for one night a week as opposed to four nights a week.

This must raise concerns that this could encourage litigation in respect of how much time a child spends with a parent.

In the right case, where the parent with care has a modest income and/or modest resources and there is a careful consideration of a budget to consider whether there is a good reason to depart from the CMS formula, then the approach in CB v KB may well be the fair one. However, we suggest that in the wrong case, by failing to properly consider the section 25 factors, notably the income and resources of both parties and the actual needs of the child in question, the application of the CMS formula could lead to unfairness.