Evaluation of the reformed scheme
Natural rate of conversion to the new rules
The political drivers of the reformed scheme
The origin of the percentages in the reformed scheme
The 21st Century is making the reformed scheme obsolete
Will the reformed scheme achieve its objectives?
Related topic - The "Shared Care" flaws in the reformed scheme
Related topic - Can child support schemes ever work?
Home & weblog
Blog archive & site history
Site map & search

The political drivers of the reformed scheme

Green Paper: "The well-being of the growing number of children whose parents have separated depends on them receiving the financial and emotional support of both parents, wherever they live."

Green Paper: "Child support is about families as well as finances."

Green Paper: "Children have a right to care and support from both their parents, wherever they live."

White Paper: "We will put the need for regular and reliable payments of maintenance at the heart of the new system and we will introduce effective and prompt sanctions for those who try to avoid providing for their children."

What happened to those non-financial statements - "emotional support", "families", "care" - in the year between the Green Paper and the White Paper?

This section is still under development


The nature of the reformed scheme, and the features included in it and omitted from it, are more determined by political and administrative advantage than by the nature of children and parenting.

To see what this means, imagine an alternative. In such an alternative scheme, it would be possible to trace each feature of the scheme to some aspect of bringing up children. For example, the amounts paid would be identifiably related to research into how much children cost. If the costs changed at particular ages, this would be reflected in the formula. If the expenditure on children changed in a particular way according to circumstances (such as income or others costs), then this would show in the formula, accompanied by references to the evidence.

In fact, the current scheme shows many of these features. It is based on benefit rates, which are supposed to be based on research into the costs of staying alive for adults & children. (It then screws up by attempting some wealth-sharing features).

The reformed scheme shows none of this. It is as though every feature were chosen for what is politically or administratively expedient. None of it properly relates back to research or analysis of the topics concerned.


The percentages

This summarises a more detailed article about the percentages.

The basic percentages in the reformed scheme are 15% / 20% / 25% of NRP's net income for 1 / 2 / 3-or-more qualifying children. These were not properly based on research.

Everyone knows that the 1st child doesn't cost 3 times as much (15% instead of 5%) as the 2nd child. And the 3rd child child doesn't cost the same (5%) as the 2nd child. And the 4th child doesn't cost nothing. Research, benefits rates, and personal experience, all say this. So why these percentages? Expediency! Besides, what research anywhere in the world says that one child costs 15% of one (or both) parent's net income to bring up?

The Green Paper (consultation paper) [1] was issued about 6th July 1998. This was the first "official" announcement of the 15% / 20% / 25% formula. (There has been leaks before then). It cited 2 pieces of research:

1. "One recent study found that on average parents spend about £3,000 per year on one child. This is equivalent to almost £60 per week. This study also found that many parents go without necessities for themselves in order to protect their children’s standard of living.
‘Small Fortunes’ Middelton
(sic) et al 1997"

The Small Fortunes research [3] is probably the most credible research available in the UK into the cost of bringing up children. Sue Middleton (not Middelton), from the Centre for Research in Social Policy, Loughborough University, has been pursuing a particular "survey methodology" for years, and is a regular contributor of evidence to government policy. It identified an average expenditure (during 1995) at about £57 per week (of which 10% came from people other than the parents and some came from Child Benefit). At 2001 prices this is probably about £65 per week. But - there is no hint whatsoever in the research that this rises as a percentage of net income! Indeed, in evidence to the Social Security Select Committee in November 2000, Sue Middleton said: "We also found, at first sight strangely, that average spending did not vary greatly with income of a family; about 20 per cent from the bottom quartile to the top income quartile."

2. "Overall, there are indications that the cost of a child represents between 20 per cent and 30 per cent of the budget of a family with one child.
‘The Costs of Children and the Welfare State’ Dickens, Fry and Pashardes, 1996"

The Dickens et al research [4] is based upon published surveys (the Family Expenditure Surveys of 1970 - 1986). It tries to identify how much children cost compared with adults, hence the proportion of household spending on children. I conclude from their Table 3.1 that the expenditure on a young only-child in a household with 2 adults is 10.4% of the total expenditure on the 3 people, and for an older only-child is 14.7%. This doesn't say "X% of income is spent on a child", unless it is assumed that "income = expenditure on people", hence no expenditure on anything else, no savings, or no drawing upon savings for expenditure, etc. But even is this were true, the range appears to be about 10% - 15%, not what the Green Paper says.

The White Paper (policy paper) [2] was issued about 1st July 1999. It didn't refer to Dickens et al, only to Middleton et al:

It says "The proposed base rate of 15 per cent of their income is roughly half the average that an intact two-parent family spends on a child". (That statement appears to be innumerate).

But Middleton never identified a formula for expenditure based on percentages on income (see her evidence to the Select Committee quoted above)! The suspicion has to be that the policy drew incorrect conclusions from Dickens et al (who actually appear to say that the average is 10% - 15%) then wrongly attributed them to Middleton et al. Frankly it doesn't matter who they cited - none of the quoted research leads to a formula about the costs of children based on income.

Baroness Hollis showed her thinking on Wednesday 22nd July 1998 when giving evidence about the Green Paper to the Social Security Select Committee soon after the Green Paper was issued:

(Mr Pond) Could we move on to the question of the formula now and I think all of us are delighted that there is going to be a less complex method of dealing with this. I think the Minister is quite right to explain that we need that sort of formula. Can I ask about the percentages that have been put in, the 15, 20 and 25 per cent, why those particular figures were chosen as the appropriate percentages for the first, second and three plus children?

(Baroness Hollis) I think there is both a rationalised answer and an instinctual answer. The instinctual answer was that it felt about right from what we knew. It was supported by a fair amount of evidence about what net income people get left with in, say, Australia, New Zealand, Canada, the United States, as well as by some of the research that has gone into, for example, the Middleton Study, Small Fortunes which illustrated what parents in different financial income bands spend on their child support. On average about a third of a couple's income tends to go on the support of their children, so 15 per cent from one parent in a separated family seemed about right and it fitted our perceptions. But it was also that at the moment you have a system with a protected formula, in other words a protected income in which we build in all the complexities, and then there is effectively a 50 per cent deduction rate thereafter, with all of the problems of disincentives and so on that arose. We felt that by going for a 15 per cent figure, which was lower than the current figure, we left him with more income, as I say, to pay for all the things that we will not be building into the formula, that still could allow us to finance the maintenance disregard, still effectively have an assessed maintenance figure than was higher than the average maintenance currently paid so that parents with care should not on the ground be actual losers for the most part, and kept it cost neutral from the Treasury point of view. It was a mixture of what felt right, informed by the research that was available, informed by the balancing act between the three parties if you like. That was why we came to that figure. If we dropped the figure down to say 10 per cent for him, either the parents with care lost out or the Treasury, that is, other parents as taxpayers were having to pick up that responsibility. If we increased it we felt it was unreasonable for him given the extra responsibilities we were asking him to carry, so it was in that sense a negotiated figure between those pressures.

Second families in the formula

To be supplied

Shared care in the formula

To be supplied

Press releases

Rough justice

Tax on fathers


Instinct, feeling, incorrect conclusions drawn from Middleton, innumeracy, perceptions, cost neutral for the Treasury, "what he can afford", balancing act ....

Don't look for a rational relationship between the attributes of children and parenting and the features of the reformed formula!

This doesn't mean the reformed scheme is totally wrong. If catering for political expediency stops it upsetting too many people, that may be good. If administrative expediency ensures that is responds accurately and in reasonable time, that may outweigh a logical formula. (And there is a possibility that making 2 serious compensating mistakes in the analysis has led the government to nearly the right percentage anyway!)

But it does mean that the reformed scheme must be judged on grounds other than simply whether the formula matches research into expenditures on children, etc. Perhaps one of the grounds needs to be whether it will operate without too much protest or even terrorist action.

It also means that press releases and other political statements must be judged accordingly! (They don't always bear any useful resemblance to what they are announcing).

References, etc

[1] Children First: a new approach to child support
(CSA reform Green Paper)
Command paper Cm 3992 July 1998

[2] A new contract for welfare: children's rights and parents responsibilities
(CSA reform White Paper)
Command paper Cm 4349 July 1999

[3] Small Fortunes: Spending on children, childhood poverty and parental sacrifice.
Sue Middleton, Karl Ashworth and Ian Braithwaite
Published by the Joseph Rowntree Foundation.
ISBN 1 85935 032 1

[4] The Costs of Children and the Welfare State: An Empirical Analysis based on Consumer Behaviour
Dickens, Fry, Pashardes
Discussion paper series no 466, December 1996
Department of Economics, University of Essex


There appear to be consistent arithmetic flaws in this consultation process:

"The proposed base rate of 15 per cent of their income is roughly half the average that an intact two-parent family spends on a child" (White Paper)

"On average about a third of a couple's income tends to go on the support of their children, so 15 per cent from one parent in a separated family seemed about right" (evidence to Social Security Select Committee)

But remember: (third of X) + (third of Y) = third of (X + Y). So if the intact family spent a third of their joint income, why shouldn't each of the separated parents also spend a third of their net income? You don't just go and halve the number!

In fact, since the correct amount from [4] is more like "10% - 15% of the intact household's total expenditure on people is spent on their child", then if "income = expenditure", a more justifiable percentage would be about 10% for a young child and 15% for an older child. Has the combination of incorrect reading of the research and innumerate derivation of the formula from that research led to nearly the right answer? (But remember that [3] doesn't suggest that percentages apply).

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