Shared Care & the child support formula
Child support & Shared Care - overview
Should child support be linked to access?
The "Fair Shares" formula (copy of Hansard)
Issues with the "Shared Care" formula (1998)
Related topic - Case Studies - anomalies in the reformed shared-care formula
Related topic - Letter from Baroness Hollis, minister for child support
Home & weblog
Blog archive & site history
Site map & search

Issues with the "Shared Care" formula (1998)

This is an HTML transcript of a paper I wrote in 1998 while using the name "John Ward".

It is still relevant, and I have been pursuing this analysis since then. It formed the basis for evidence to the Social Security Select Committee. Some details are out-of-date, for example the formula now has a strange feature at the "equal sharing" point, but the conclusions still hold.

Original paper, Word for Windows (RTF) format

"Children First" & Sharing of Care:
Problems with the proposed formula, and a revised proposal.

Barry Pearson

version 2.1
3rd August 1998


This document is a response to the "Children First" Green Paper.

The scope of this document is the method for modifying the maintenance liability where separated parents share the care of their children. (It has nothing to say about the formula where care is not shared, nor how the amount of sharing is decided).

The Green Paper's proposal for handling sharing of care is flawed:

  1. It is surely desirable for both parents to share care, yet the formula behaves very unjustly where care is shared more or less evenly.
  2. An NRP on low-income who wants to share care with the children is financially discouraged from doing so.
  3. It reduces the maintenance liability too slowly & unfairly for the case where PWC & NRP have similar incomes.
  4. In certain cases it encourages the separated parents to lie, play the system, and attempt to reduce the liability and maximise the money obtained from the taxpayers.

The revised proposal is that the formula should be symmetrical. While one parent cares, the other pays. These calculations for both parents are then "netted" to identify the resultant maintenance liability. This overcomes the above flaws.


  1. References
  2. Description
  3. Related comments
  4. Examples

1 References

[1] "Children First" Green Paper for CSA Reform

[2] CSA Annual Report & Accounts 1997-98

[3] Wisconsin USA law

[4] Canada Child Support law

[5] Maryland USA law

2 Description

Reminder of the Green Paper formula

The Green Paper [1] first identifies the liability for the case where only the PWC cares for the children. Then it proposes reducing this liability by one seventh for each night per week that the NRP cares for the children.

This has the following unjustified inconsistency:

  • For the nights when the PWC is caring, the maintenance scheme operates. The NRP transfers money to the PWC for those nights so that both parents co-operate in the upbringing of the children.
  • For the nights when the NRP is caring, the PWC is not part of any maintenance scheme. He/she neither pays nor receives, yet doesn't care either. During these nights it is as if the NRP is a true lone parent with no-one to assist.

The consequences are serious:

  1. There is a major financial step at 3/4 nights, or PWC/NRP status, which will inevitably lead to disputes and injustices. [Example 1].
  2. There is no financial support for an NRP on low-income who wants to share care of the children. [Example 2].
  3. The maintenance liability reduces too slowly & unfairly for the case where PWC & NRP have similar incomes. [Example 1].
  4. In certain cases the separated parents are encouraged to lie, play the system, and attempt to reduce the liability and maximise the money obtained from the taxpayers. [Example 3].

Revised proposal for sharing of care

This document neither agrees nor disagrees with the basic formula in the Green Paper. The following revision would work if a different basic formula applied. It is based on the following principle: "children have to be looked after and paid for, so at any time, a parent either cares for the children or pays whoever does so".

Proposal: "Whatever formula applies for the period when the PWC is caring for the children should also apply mutatis mutandis for the periods when the NRP is caring for the children".

In other words: "Whatever basic formula is thought to be right should be applied consistently, and not just one way".

For example:

  • For each night the PWC cares for one child, the NRP pays one seventh of 15% of the NRP's weekly income.
  • For each night the NRP cares for that child, the PWC pays one seventh of 15% of the PWC's weekly income.

The consequences depend on the amount of sharing and the relative incomes of the PWC and NRP:

  • If there is no sharing of care, this proposal has no effect. The Green Paper formula (or any other) operates unaltered.
  • If the PWC and the NRP earn about the same amount, then the effect is to reduce the maintenance liability by about two sevenths per night. The distinction between being PWC and NRP where there are more-or-less equal amounts of caring becomes relatively insignificant and not worth behaving badly over.
  • If the PWC earns a lot less than the NRP, the liability falls by about one seventh per night, rather like the current proposal in the Green Paper, which is probably financially reasonable from the children's point of view.
  • If the NRP earns a lot less than the PWC, then in fact the net effect is likely to be that the PWC will pay the NRP, which is probably necessary if the NRP gets no benefit assistance for caring for the children for (say) a couple of nights per week.

This scheme is totally symmetrical, hence "fair" (or at least defensible). It tends to be neutral, or possibly favour children while they are with the the poorer-off person, and certainly not favour the better-off person. It appears to favour sharing of care for the children while safeguarding the taxpayer's interests.

Rejected alternatives

The simplest revision to the green paper formula would be to reduce the maintenance liability by two sevenths each night of care by the NRP rather than one seventh.

Unfortunately, this has poor behaviour where the incomes differ significantly. It is approximately right if the separated parents have about the same net income, but otherwise has the unacceptable feature that it favours the better-off person at the expense of the poorer-off person:

  • If the NRP is better-off it reduces too fast, so that a rich NRP caring for 3 nights per week would pay only about 2% of net income to help a poorer PWC care for 4 nights per week.
  • If the PWC is better-off and the NRP is poor it fails to provide sufficient money for the NRP to care for the child.


Other nations also modify the basic formula to take into account the incomes of both parents where care is shared. For example, Wisconsin [3] and Canada [4] both do so where sharing is significant. Maryland [5] uses a "netting" scheme for shared care which appears to be similar to this proposal.

Another case where the basic formula would apply to both parents is where they have split responsibility for 2 or more children. They would presumably both be assessed, and the result netted. The revised proposal is similar to treating shared care in a similar way to split responsibility. This has the benefit of simplicity.

The "simplicity" argument

The Green Paper claims that one advantage of the formula is that it is simple, and predictable for the NRP. (However, it is not predictable for the PWC, because it depends on the NRP's income). This revised proposal is also reasonably simple and predictable.

Where there is no sharing, this revised proposal makes no difference.

Where there is sharing, an additional piece of information is needed - the PWC's net income. Often this is already known - for example, where the PWC is on benefits. It is certainly no more an invasion of privacy for the NRP to be aware of the PWC's income than vice versa. (It is already taken into account with the existing CSA formula).

Now the NRP can predict the maximum liability, not the actual liability. Both parents will use the same tables, picked up at Post Offices, etc, to identify their own portions of the resultant liability.

3 Related comments


Although the terms "PWC" & "NRP" may be suitable for the case where the parents don't share care of the children, they are inappropriate for the sharing case. Alternative terminology does exist, eg. Wisconsin [3]:

  • "Parent" means the natural or adoptive parent of the child.
  • "Parent with less time" means the parent having physical placement of the child less than 182 days a year.
  • "Parent with more time" means the parent having physical placement of the child more than 183 days a year.

This terminology is not perfect, but is more neutral and less emotive than either current or Green Paper terminology.

One reason that this matters is that using terminology which polarises parents is likely to influence the formula. For example, the Green Paper shows signs of being designed for the case where one parent cares all the time and the other cares none of the time, with the shared-care modification then bolted on as an after-thought. Once it is accepted that shared care is common, and in fact to be encouraged, a fairer formula will be desired and sought.

Avoiding significant "steps"

Financial formulae with "steps" (points where small practical differences make big financial differences) lead people to "play the system" and behave badly. Outside the CSA formulae, examples include the working hours and asset valuations which occur in the rules for means-tested benefits, or the income level at which National Insurance is triggered. (This latter is known to be so important that it has been addressed in a recent budget).

Because there are likely to be both financial and personal motives operating in any child support formula, it is important to reduce incentives for people to behave badly. It is likely to be the children who suffer most.

The existing (1991/1995) CSA formula has some key "time steps" - points where small practical differences in time make big financial differences.

One is at 182/183 nights, which is the difference between being a PWC and an AP. The Green Paper retains this, as the 3/4 night step described and criticised in this current document.

Another is at 103/104 nights, where an AP who cares for children for at least 104 nights per year has a reduced liability. The Green Paper replaces this with a series of steps at 1, 2, and 3 nights.

While the reduction of the lower limit from 2 to 1 night per week is welcome, removing the steps entirely, or perhaps making them much smaller (eg. percents of a year) would be less likely to encourage bad behaviour of the form: "I know it makes sense for you to look after the children tonight but that will put you over the 2 night step and reduce my maintenance so I'm taking the children off you".

4 Examples

Example 1: The anomaly where sharing is nearly equal

This anomaly is that, even where the parents split care of the children more-or-less evenly, there is a massive financial gain in being granted "PWC" status, and a corresponding massive financial loss in being granted "NRP" status. This is out of all proportion to any estimate of the cost of bringing up children.

Consider: both parents ("P" and "Q" respectively) earn £465 net income, and have one child, with no other children or other conditions to complicate this example. (These incomes are chosen to make the maintenance liabilities round numbers, but equivalent results are obtained for any other common value).

Green paper:

This shows how much Parent P pays to Parent Q if P cares for that number of nights per week. A negative number means that P receives from Q.

Nights with P: 0 1 2 3 4 5 6 7
P pays Q: £70 £60 £50 £40 -£40 -£50 -£60 -£70

It is seen that at first the difference in payment for each additional night spent caring for the child is £10, about £500 per year. Perhaps this is felt to be the cost of caring for a child for an extra night.

But the extra night between 3 and 4 nights makes a net difference of £80, or about £4000 per year. This could be for just a few nights difference - it is really the effect of having PWC or NRP status, rather than much to do with the cost of caring for a child for an extra night.

Clearly, such a large amount of money for such a small difference is unjust and is expected to cause the parents to behave badly, and especially to do everything possible (including lying if necessary) to be considered the PWC.

Revised proposal:

The revised formula avoids this polarisation. This shows how much Parent P pays to Parent Q if P cares for that number of nights per week (net).

Nights with P: 0 1 2 3 4 5 6 7
P pays Q: £70 £50 £30 £10 -£10 -£30 -£50 -£70

(Eg. at 1 nights, P pays Q £60, Q pays P £10; at 2 nights, P pays Q £50, Q pays P £20).

At the extremes (no sharing) the liability is the same - all of the Green Paper's formula (or any other formula) applies. In between, there is a smooth switch of liability from one to the other.

(An alternative of decreasing liability by two sevenths instead of one seventh per night has a similar effect in the case of parents with similar incomes, but has poor behaviour otherwise, see earlier).

Example 2: Problems for a low-income NRP trying to pay for sharing of care

Everyone agrees that where possible separated parents should share the care of their children. But the formula does not encourage this where the NRP has very low income (a significant minority of NRPs are out of work [2]).

Consider a PWC earning (say) £232, and an NRP on benefits or very low income. The Green Paper states that the NRP should pay £5 per week if care is not shared. Suppose, now, the NRP cares for the child for (say) 2 nights per week.

Green paper:

Presumably the maintenance liability will be reduced to about £3.60 per week. This makes virtually no difference to the PWC, who is relieved of the cost of the child for 2 nights a week, but it makes no provision for the NRP to pay for the child for those 2 nights. (Presumably benefits will not pay extra for those 2 nights, and the NRP is already getting less than "poverty relief" benefit levels because of that £3.60).

Revised proposal:

The revised proposal identifies that the NRP pays the PWC £3.60 for 5 of the nights, and the PWC pays the NRP £10 for the 2 nights. (The latter amounts to two sevenths of what someone on £232 income would pay per week). The PWC pays to be relieved of 2 nights care, and the NRP has about £6.40 above his/her own income to spend on the child.

Example 3: Encouragement to commit benefit fraud

Consider how an out-of-work PWC "P" and a medium-to-high earning NRP "N" can conspire to defraud the taxpayer using the anomaly in the formula.

N earns (say) £465, and so pays £93 maintenance for two children if there is no sharing of care. P is out-of-work, and probably gets little or no Income Support (depending on P's age and the children's ages).

Green paper:

If, instead, they switch roles, with P perhaps retaining PWC status for 4 nights care for one child in order to remain on Income Support, and becoming the NRP for the other child, then N's maintenance liability drops to £40, while P's Income Support rises perhaps between £30 and £40, plus passported benefits (depending on details).

The taxpayer is about £40 worse off, while the parents can share this between them privately and both be better off. (Such undeclared income is benefit fraud).

Revised proposal:

The revised proposal doesn't cure the problem, but makes it less likely.

N, who now claims PWC status for one of the children, would still pay for the nights that P claims to care for that child. This might be £30. So the liability of N has dropped from £93 to £70, not to £40, and the taxpayer is relieved by a corresponding amount.

The fraud could still have its full effect if P now claimed not to care for the child at all. But this is less likely - it might count against that parent in any dispute later. So trying to "play the system" requires much harder choices, and is easier to detect.

Page last updated: 5 July, 2004 © Copyright Barry Pearson 2003