Maryline Stowe's website kindly invited me to do a guest post. I did this about adoption statistics. A response with questions has been written by Lucy R on the transparency project blog
I have written this blog post to answer the questions and reconcile the statistics. All figures are figures for the English jurisdiction. Most (if not all) are based upon the SSDA903 return. The SSDA903 return is an annual electronic return from Children's Services Authorities to the Department for Education. It includes details of each change of legal status and each change of placement status for children in care. I have had additional analyses done by the DfE statisticians - which is what the source of the spreadsheets that I have uploaded is.
The most widely known analysis from the SSDA903 return is the Statistical First Release. This is normally issued in the September after the end of the Financial Year that it is relevant to. Hence the 2014 SFR can be found in a PDF form here, but the 2015 release has not yet been published. I use the excel format files as they are easier to produce analyses from.
I shall put in italics what appear to be questions or other elements from Lucy's blog post and then answer them (or respond as appropriate).
But what other reasons were there for children leaving care?
More than a third (34% or 10,300) of those 30 thousand or so children ceased to be looked after because they had gone home to parents or relatives.
The first point to make is to distinguish between "care" - what for precision I call "compulsory care" and "looked after children".
Looked after children include all the children looked after by the local authority whether they are in care compulsorily as a result of a placement order, police protection, interim care order, EPO, care order or other compulsion or (additionally) those who are in for a form of respite care.
Table A2 in the SFR gives the following summary figures for 31st March 2014.
Care orders, 39,930, freed for adoption 60, placement orders, 9,260, Section 20, 19,230, detained for child protection 50, youth justice 300.
The S20 category is complex and there are abuses of S20, but this cohort of children is substantially different to those who are compulsorily in care. Hence when I ask the DfE to analses compulsory care I ask them to exclude these children. From a court perspective S20 does not involve the court.
To summarise, 8% of children were adopted non-consensually, whilst 1/3 went home or to family, 9% were adopted without opposition, and 17% were made the subject of SGOs or residence orders. This, of course, is conflating children who are voluntarily put into care by their parents and then taken from care and the compulsory care system involving public family law. We know that the number of care orders and the like each year is around 11,500. Hence from the table (C1) of children who start being "looked after" where for 2014 there are 30,340 we can estimate that the total number coming into the care system as "looked after children" on S20 was around 19,000.
The number starting to be looked after in any one year will not match the number ceasing being looked after. However, whatever way one wishes to consider this the S20 cohort should be looked at separately to those in compulsory care. Particularly if one wishes to consider the operations of the courts one could not mix up the statistics by including children who are not subject to court proceedings of any form.
Only 16% of under fives reunited with parents?
John Hemming goes on to say that the proportion of under fives reunited with parents is only 1,300 of 8,200 (16%). For comparison, the percentage for return to parents across all ages is 34%, so if right this would be disproportionately low. Again, we cannot see how Mr Hemming’s analyst has extracted those figures, and will ask for clarification. We cannot presently verify this assertion.
The info comes from DfE. If you ask them for a copy of the stats they send to me then you can confirm this. Decisions are not made on the basis of quotas
Let me put it this way. There are quotas. Increases in the quotas cause changes in the numbers of children adopted. How does this happen if decisions are not influenced by quotas.
Further Comment 31/7/15 following response from Transparency Project (Lucy Reed)
http://www.transparencyproject.org.uk/adoption-statistics-again-part-2/ was written in response to the above. My comments follow
3/9/15 - I have also uploaded some stats from 2011 (I have not found more recent ones) which look at disposals in public family law by outcome and numbers of children. here these demonstrate that there are only a small number of S20 children actually subject to care proceedings which don't end up in a care order.
3/9/15 - Further comments.
The reason I have put up the court stats is because of the argument by Lucy R that I am missing out a large number of children on S20 who are represented in care proceedings. My view is that there are likely to be a small number of children, possibly around 100, but that this does not have a significant effect on the argument about the figures. Notwithstanding this issue furthermore the legality of outcome targets for those providing expert opinion remains permanently in question. (and the issue of repeat player prejudice and Experts on retainers).
My point is that a very high proportion of applications for care orders result in a care order. Hence even if there is no ICO at an early stage, the children concerned will be picked up in my figures when the care order is Granted.
The SSDA903 return operates on a financial year so that the 2011 figures are 1st April 2010 to 31st March 2011. The court stats are on a calendar year basis so actually the 2011 court figures are more like the 2012 SSDA903 figures.
Looking at applications for a care order. There were 10,844 orders made, 4 orders refused, 260 orders of no order and 263 applications withdrawn. That is 95.38%. Hence under 5% of cases might not be trapped in my figures.
Secure accommodation is irrelevant to this issue and obviously if a care order is discharged then there was a care order. Supervision orders are not relevant as obviously the children are being supervised with their parents. They do not go in care. If those become care orders then they appear under the care order figures. Contact, refuse contact authority, education supervision and Child Assessment orders are not relevant to the issue.
An EPO is of course relevant. Because they are short term, however, and normally result in an ICO I don't think it is worth separately considering them. If proceedings continue on after this they will be trapped in the care order stats.
Recovery orders are not relevant nor the parental responsibility, S8, residence, contact, prohibited steps or specific issue orders.
An SGO will follow care proceedings and is, hence in those statistics.
My argument is that it is care proceedings that we should be concerned with and although there may be a small number of cases of care proceedings where the threat of an ICO is used to keep a child in care under S20, the number is easily under 5%. My experience of cases, however, is that I don't see cases happening where S20 goes directly to SGO. Those would be an inappropriate use of S20 anyway as it is not voluntary where there are care proceedings.
I would accept the argument that it would be better to have a more exhaustive analysis of the nature of the looked after children population. However, there remain limits as to what I can get DfE to do. They should really do more work on this themselves.
I have written this blog post to answer the questions and reconcile the statistics. All figures are figures for the English jurisdiction. Most (if not all) are based upon the SSDA903 return. The SSDA903 return is an annual electronic return from Children's Services Authorities to the Department for Education. It includes details of each change of legal status and each change of placement status for children in care. I have had additional analyses done by the DfE statisticians - which is what the source of the spreadsheets that I have uploaded is.
The most widely known analysis from the SSDA903 return is the Statistical First Release. This is normally issued in the September after the end of the Financial Year that it is relevant to. Hence the 2014 SFR can be found in a PDF form here, but the 2015 release has not yet been published. I use the excel format files as they are easier to produce analyses from.
I shall put in italics what appear to be questions or other elements from Lucy's blog post and then answer them (or respond as appropriate).
But what other reasons were there for children leaving care?
More than a third (34% or 10,300) of those 30 thousand or so children ceased to be looked after because they had gone home to parents or relatives.
The first point to make is to distinguish between "care" - what for precision I call "compulsory care" and "looked after children".
Looked after children include all the children looked after by the local authority whether they are in care compulsorily as a result of a placement order, police protection, interim care order, EPO, care order or other compulsion or (additionally) those who are in for a form of respite care.
Table A2 in the SFR gives the following summary figures for 31st March 2014.
Care orders, 39,930, freed for adoption 60, placement orders, 9,260, Section 20, 19,230, detained for child protection 50, youth justice 300.
The S20 category is complex and there are abuses of S20, but this cohort of children is substantially different to those who are compulsorily in care. Hence when I ask the DfE to analses compulsory care I ask them to exclude these children. From a court perspective S20 does not involve the court.
To summarise, 8% of children were adopted non-consensually, whilst 1/3 went home or to family, 9% were adopted without opposition, and 17% were made the subject of SGOs or residence orders. This, of course, is conflating children who are voluntarily put into care by their parents and then taken from care and the compulsory care system involving public family law. We know that the number of care orders and the like each year is around 11,500. Hence from the table (C1) of children who start being "looked after" where for 2014 there are 30,340 we can estimate that the total number coming into the care system as "looked after children" on S20 was around 19,000.
The number starting to be looked after in any one year will not match the number ceasing being looked after. However, whatever way one wishes to consider this the S20 cohort should be looked at separately to those in compulsory care. Particularly if one wishes to consider the operations of the courts one could not mix up the statistics by including children who are not subject to court proceedings of any form.
Only 16% of under fives reunited with parents?
John Hemming goes on to say that the proportion of under fives reunited with parents is only 1,300 of 8,200 (16%). For comparison, the percentage for return to parents across all ages is 34%, so if right this would be disproportionately low. Again, we cannot see how Mr Hemming’s analyst has extracted those figures, and will ask for clarification. We cannot presently verify this assertion.
The info comes from DfE. If you ask them for a copy of the stats they send to me then you can confirm this. Decisions are not made on the basis of quotas
Let me put it this way. There are quotas. Increases in the quotas cause changes in the numbers of children adopted. How does this happen if decisions are not influenced by quotas.
Further Comment 31/7/15 following response from Transparency Project (Lucy Reed)
http://www.transparencyproject.org.uk/adoption-statistics-again-part-2/ was written in response to the above. My comments follow
- SSDA903 is a return to DfE from Childrens Services Authorities and is stored in a database. Hence it is no in itself published although various extracts of data are published. I can forward the email I got from DfE to Lucy Reed or Sarah Phillimore if they wish to try to argue that the stats are not provided by the departmental statisticians.
- Lucy makes a valid point that some sets of proceedings commence with S20 placement with the local authority. My "compulsory care" statistics pick up those that then change to an ICO or Care Order. It is true that in a small number of cases the application does not result in an ICO or Care Order. From memory it is around 3% of applications (according to MoJ statistics) that do not result in an ICO or Care Order. It remains that an ICO is granted on "reasonable grounds" rather than proof and hence the analysis I have had requested is valid without those additional 3% of cases.
- There are questions about the abuse of S20. However, any useful analysis of the care system has to distinguish between child protection and other cases. I don't think you can sensibly lump all the cases where the LA looks after the child and treat them as being the same.
- Quotas - I gave a link to the figures for Merton with all the details about their targets. Quotas do have an affect on managerial priorities. What is the sense of setting targets if you don't intend hitting them. Having a target is materially different to counting something. Senior management in the LA tell the junior management to hit the targets.
- "John Hemming’s figures suggest that a very high 78% of children over 5 who left care through adoption were adopted unopposed compared with an overall figure of 9%." - I personally am not claiming this.
- Cases and Children. One application for a care order can affect more than one child. Hence you will get different figures. SSDA903 is child based and I have had agreement from Cafcass to work on stats to see what agreement we can get between their analyses and those of the department.
- It remains that the outcome of adoption from care without parental consent is commonplace rather than exceptional. There is, in fact, no real threshold distinction between dispensing with parental consent and simmply adopting from care. That is not in accordance with international law. We can argue about the 3% of cases which start with S20, but do not end with a care order of any kind, ubt that does not make a substantial different to the point that the system as a whole is unlawful.
3/9/15 - I have also uploaded some stats from 2011 (I have not found more recent ones) which look at disposals in public family law by outcome and numbers of children. here these demonstrate that there are only a small number of S20 children actually subject to care proceedings which don't end up in a care order.
3/9/15 - Further comments.
The reason I have put up the court stats is because of the argument by Lucy R that I am missing out a large number of children on S20 who are represented in care proceedings. My view is that there are likely to be a small number of children, possibly around 100, but that this does not have a significant effect on the argument about the figures. Notwithstanding this issue furthermore the legality of outcome targets for those providing expert opinion remains permanently in question. (and the issue of repeat player prejudice and Experts on retainers).
My point is that a very high proportion of applications for care orders result in a care order. Hence even if there is no ICO at an early stage, the children concerned will be picked up in my figures when the care order is Granted.
The SSDA903 return operates on a financial year so that the 2011 figures are 1st April 2010 to 31st March 2011. The court stats are on a calendar year basis so actually the 2011 court figures are more like the 2012 SSDA903 figures.
Looking at applications for a care order. There were 10,844 orders made, 4 orders refused, 260 orders of no order and 263 applications withdrawn. That is 95.38%. Hence under 5% of cases might not be trapped in my figures.
Secure accommodation is irrelevant to this issue and obviously if a care order is discharged then there was a care order. Supervision orders are not relevant as obviously the children are being supervised with their parents. They do not go in care. If those become care orders then they appear under the care order figures. Contact, refuse contact authority, education supervision and Child Assessment orders are not relevant to the issue.
An EPO is of course relevant. Because they are short term, however, and normally result in an ICO I don't think it is worth separately considering them. If proceedings continue on after this they will be trapped in the care order stats.
Recovery orders are not relevant nor the parental responsibility, S8, residence, contact, prohibited steps or specific issue orders.
An SGO will follow care proceedings and is, hence in those statistics.
My argument is that it is care proceedings that we should be concerned with and although there may be a small number of cases of care proceedings where the threat of an ICO is used to keep a child in care under S20, the number is easily under 5%. My experience of cases, however, is that I don't see cases happening where S20 goes directly to SGO. Those would be an inappropriate use of S20 anyway as it is not voluntary where there are care proceedings.
I would accept the argument that it would be better to have a more exhaustive analysis of the nature of the looked after children population. However, there remain limits as to what I can get DfE to do. They should really do more work on this themselves.
Comments