AlphabiolabsHousing Law WeekIQ Legal TrainingBerkeley Lifford Hall Accountancy Services

Home > Articles > 2015 archive

Questioning the Use of Section 20

Judith Masson, Professor of Socio-legal Studies at the University of Bristol, considers section 20 of the Children Act 1989 within its broader historic, legal and practice context.

Professor Judith Masson, Department of Socio-legal Studies, University of Bristol

Judith Masson, Professor of Socio-legal Studies, University of Bristol

Over the last year there has been increasingly strident criticism of local authorities' use of their powers to accommodate children under the Children Act 1989, s.20. That this has largely been in the form of obiter dicta, which are of course not statements of law, is irrelevant if they begin to create an impression that judicial views can determine how and when s.20 should and should not be used, and what amounts to 'good practice' in child care social work. This is even more important where the presentation of these statements about s.20 lacks any reference to the broader legal framework within which s.20 operates.

Rather than judging the use of s.20 with the keyhole vision provided by care proceedings, those who seek to promote policy or guidance about how and when the provision should be used need to consider it within its broader historic, legal and practice context.

A little bit of history.

The Children Act 1989 was the result of a substantial period of review and open consultation in the form of a Select Committee Inquiry; the interdepartmental Review of Child Law, chaired by Brenda Hale, now Deputy President of the Supreme Court; a further consultative process on draft provisions and full Parliamentary scrutiny, over a period of 12 months. Through these processes Parliament came to its conclusions about what the law should be, re-orienting child law away from the courts with the 'no order' principle, Children Act 1989, s.1(5), and restricting the use of compulsory powers of intervention in family life. Only local authorities (and the NSPCC) were permitted to bring care proceeding. Only care proceedings could be used to restrict parental responsibility for the long term, and then only in accordance with statute. Care and supervision orders could only be granted where the threshold conditions set out in s.31(2) were satisfied. Gone were the court's wide and indeterminate powers under wardship or the inherent jurisdiction to commit children to care on welfare grounds, or to direct how local authorities should care for such children. Indeed, with hindsight, it is very hard to see that such powers could have withstood a challenge under the European Convention on Human Rights, art 8(2).

Following enactment of the Children Act 1989, there were further consultations about regulations and guidance and then a very substantial implementation programme for all those who had to work with the Children Act 1989, within and beyond local authorities. In the years following implementation the Department of Health funded a major research programme involving 24 studies by leading and independent academic researchers to examine social work (and court) practice in the operation of Parts III and IV of the Act. These studies, including a study by Packman and Hall, specifically examining the use of s.20, were brought together and widely disseminated in the publication: Aldgate and Statham, Children Act Now (TSO, 2001). It does not seem that many of those who have expressed opinions on s.20 are familiar with these publications, the subsequent research, practice guidance or the published statistics on looked-after children. Courts are not places for academic analysis of issues beyond the focus of the litigation but policy cannot be developed without this wider understanding. 

Questions and answers about the law relating to s.20

1) Can the court make a full care order where a parent is in full agreement with the local authority plan for their child being looked after under s.20?

It is not possible to know whether, or how often the courts are faced with such a scenario. The only statistics on the proportion of uncontested care proceedings come from research studies, notably Masson et al., Care Profiling Study (2008) and Hunt et al., The last resort (1999). There are, so far as I can establish, no published judgments on this point. Nevertheless, it is clear that a care order should not be made. Children Act 1989, s1(5) states that 'the court shall not make the order or any of the orders unless it considers that doing so would be better for the child than making no order at all.'

An order is only legitimate if it is better for the child's welfare than no order. So we need to consider what difference a care order would make where the parent agrees fully with the local authority's plan.

There is no difference in the duties owed by the local authority to the child under the Children Act 1989. The local authority is under the same obligations to consult the parents, child and others. It owes at least the same duties about placement, contact, education and aftercare. It must review the child's case just as regularly as it must review the care of children subject to care orders. What the local authority does not have is parental responsibility; the person with care of the child can do whatever is reasonable to safeguard and promote the child's welfare (s.3(5)) and so the local authority needs the parent's agreement to the care plan and general arrangements. Written agreements, sometimes termed partnership agreements, are used for this. Using the powers in the Local Authority Social Services Act 1970, s.7(7), the Secretary of State for Education has issued guidance on these matters: Children Act 1989 Guidance  and Regulations Volume 2: Care planning, placement and case review (2015). This document, developed through consultation can be seen as providing an indication of good practice across the range of circumstances where s.20 is used.

2) Is the use of care orders in circumstances where the child's welfare can be met by a s.20 agreement Convention compliant?

The European Convention on Human Rights, art 8(2) requires intervention by the state to be 'necessary in a democratic society … for the protection of the rights and freedoms of others.' This test is generally explained as one of proportionality. A care order gives the local authority parental responsibility for the child (s.33(3)(a)) and curtails parents' parental responsibility (s.33(3)(b)).  This is an infringement of parents' article 8 rights, which must be proportional (Re B [20143] UKSC 33) to comply with the Convention. Where a working partnership can be established for the care of a child without recourse to a care order it is clearly disproportionate for the local authority to seek a care order, or for the court to grant one. It is for this reason that courts refuse to grant ICOs or EPOs when the child is adequately protected during proceedings by a parent agreeing (or not objecting, s.20(7)(b)(ii)) to their child being accommodated under s.20. 

The mere fact that the threshold conditions could be satisfied for a care order makes no difference; satisfying the threshold conditions is of itself insufficient because Children Act 1989 applies further tests in s.1(1) and (5), and the Human Rights Act 1998 requires compliance with the Convention. There are circumstances where the local authority needs the court to make findings of fact or orders so that it can formulate or implement its plan for the child. However, it was never the intention of Parliament that starting proceedings in public or private child law would necessarily result in an order.

3) When is it good practice to limit the use of s.20 to a short term measure pending care proceedings?

Where care proceedings are inevitable it is best for all concerned that they are not delayed unnecessarily. This is, of course, far easier said than done. The appropriate time for starting care  proceedings can more readily be identified with the 20/20 vision of hindsight, especially from the vantage point of the bench than it can be detected in the daily work of children's services, where there are always more demands than resources and 'cases' develop and resolve in a myriad of different and contradictory ways.

Even where it is clear that care proceedings should be brought there can be good reasons for delaying them, which should be balanced against the importance of avoiding delay. For example, a plan to start care proceedings may be postponed where children are in safe care because a parent is too ill to participate or has been recently bereaved. Similarly, a mother's short term imprisonment and expressed desire at the birth for adoption does not make immediate care proceedings essential or even desirable, and it is easy to see that they would be oppressive. The mother's consent will not be valid until she has recovered from the birth, and her incarceration will make it much more difficult for her to participate in the proceedings. Delaying issuing proceedings for a few weeks until her release is unlikely to have a negative impact on the child, particularly if assessments and plans can be prepared.  Of course, the child's temporary safety is not a sufficient reason if delay is likely to impact on his or her long-term care, for example by making a permanent placement more difficult to find or less likely to succeed. Rules based on time periods, rather than the individual assessment of the case, are unhelpful because they may force the prioritisation of cases which are less urgent when all the circumstances are considered.

Local authorities should be sure that care proceedings are necessary before they take such a serious step. This was well recognised in the development of a pre-proceedings process for care proceedings. Cases could be diverted from care proceedings, and where care proceedings could not be avoided, better preparation enabled the courts to complete their work more expeditiously. There are cases where diverting to s.20 accommodation is the right thing to do; rather than starting care proceedings to end s.20 care, any need for care proceedings is ended through use of the power.

In addition, the use of s.20 cannot be seen simply as an action pending care proceedings when it has become so common for children to remain in s.20 for the duration of care proceedings. There are many reasons why children have to be looked after during care proceedings –  they may not be safe a home and the relative carer needs the  support of being a local authority carer during this difficult time for them and the child. The use of s.20 during proceedings reflects a recent move away from the use of ICOs to place children away from home, which was not clearly prefigured in the legislation.

Very many children in care under s.20 are simply not potential candidates for care proceedings. Section 20 is a family support measure – families need support for many reasons, not just where children need protection. In these cases there is no question of the use of care proceedings, indeed it would be an abuse of process and of the parents' and child's art 8 rights for the local authority to start proceedings.

The use of s.20 is not time limited because it has to cater for diverse needs, which include the need for long-term care where a court order is not necessary or not possible. In the context, limiting s.20 to a short-term measure, pending care proceedings is not a marker for good practice, and in some cases it is completely wrong.

4) Is it good practice to discourage parents from removing their child from s.20 accommodation without prior notice to the local authority?

No. The local authority should try as far as possible to create an effective working partnership with parents whose children are in need of services. The content and nature of the working partnership will vary, dependent on the child's needs and change over time as relationships of trust develop (or break down) between the social worker, the parents and the child. The aim always should be to assist the parents to recognise the child's needs at the particular time and for the future, demonstrate their understanding and maximize their contribution to meeting these needs. A notice period can play a part in this because it serves to remind a parent of the potentially disruptive effect on their child of a sudden removal, even if it is a return home.

A notice period also has another important function. It allows the local authority and parent time to plan for the child's return. Research (Farmer and Wijedasa, 'The reunification of looked-after children with their parents: what contributes to stability?' British Journal of Social Work 1612-29 (2013)) identifies the need for support to enable children to return home successfully. Such support must be planned so that it can meet the family's needs. Without some notice it is not possible for the local authority to identify what is required or to make the right support available. The Care Planning, Placement and Case Review (England) Regulations 2010 (SI 2010 No 959), reg 39, added by The Care Planning and Fostering (Miscellaneous Amendments) (England) Regulations 2015 (SI 2015 No 495) recognises the importance of planning by requiring decisions for children to return home after 20 days in s.20 accommodation to be signed off by a nominated officer. Clearly this is not possible if children are removed without any prior warning to the authority.

Discouraging the use of notice periods undermines the idea of partnership – that parent and local authority need to work together to meet the child's needs and promote their welfare. It also makes it less rather than more likely that the child's return home will endure, risking for the child further disruption of return to care.

The Children Act 1989 removed any statutory requirement for parents to give notice before removing their child from s.20 accommodation not because Parliament wanted to encourage parents to act precipitously but because it wanted arrangements to be informal, based on agreement, as they would be within families, rather than by reference to statutory powers. There is no reason why parents should be asked to notify the local authority of their plans for their child's return but it is clear that breach of the agreement is not of itself the basis for curtailing parental responsibility.

5) Must the local authority return the child just because the parent objects?

This most controversial aspect of s.20 was prefigured in the Parliamentary Debates on the Children Bill. The Lord Chancellor specifically said that s.3(5) allowed a foster carer to refuse to hand over a child to a parent who was drunk or incapable, or late at night when a child was asleep. Of course, a statement in a debate, albeit by a Lord Chancellor with responsibility for piloting the Bill through the Lords does not amount to law. However, it may be more authoritative than an obiter statement by any judge. It is well established that such statements can be used to aid interpretation if the meaning of an ambiguous provision comes before the courts. 'Any time' in s.20(8) should be interpreted in the light of the Lord Chancellor's statement.  Additionally, it is clear from s.20(7) that Parliament did not intend a parent who was unwilling or unable to care for their child to object to the local authority doing so. Section 20 is not simply about parents' rights; it is also about the child's right to private and family life. It would be shocking if the law did not allow a child in foster care a good night's sleep at the whim of a parent, or placed him or her at risk of serious harm at the hands of an inebriated parent or one who had no means of providing the basics of care. Such retention of the child could only be very temporary, until the next morning if the parent did not change their mind. At this point the local authority would have to resort to the courts, and the court would have to apply s.31 and the child's welfare for the future.


Approximately three-quarters of children subject to care proceedings are in a protective placement when these proceedings start. Since the Children Act 1989 was implemented the proportion of these placements that have been made on the basis of agreement with parents has increased and the use of emergency coercive powers – EPOs and police protection – has declined. This trend is not the result of the general misuse of s.20 but an intended consequence of reforms to child care law, which aimed to develop working partnerships between children's services and parents, and planned, not emergency, intervention. Although children subject to care orders are the majority of children in the care system, far more children initially enter care by s.20 agreements than through the courts. Policies and practice need to take account of the majority as well as the minority, and be designed to meet the needs of all parties to the process, not just reflect narrow legal perspectives. Hard cases make bad law; and reported cases are a poor foundation for good policy.