The ‘troubled families’ numbers game

“we will turn around the lives of the 120,000 most troubled families in the country”

David Cameron 15 August 2011


**This post was updated on 18 January 2014 after a member of the Troubled Families Team in DCLG raised concerns about its accuracy. These concerns are addressed at the foot of the post. SC**

Concerns about the identification of 120,000 ‘troubled families’ are not new. A number of people have highlighted concerns about the conflation of families with troubles and those who may cause trouble. Ruth Levitas and Jonathan Portes are perhaps the two best examples and Levitas also highlights the role that sampling error and sampling bias should play in the development of an estimate. She states that the figure of 120,000

ignores both sampling error (the probable discrepancy between even a randomly generated sample and the population from which it is drawn) and sample bias (the departure from randomness of the effective sample, principally caused by differential response rates across the population).

She suggests that the number of families experiencing multiple disadvantage at that time is likely to be higher than 120,000 as a result of sampling bias and sampling error (plus or minus 3 per cent) could mean that the figure was ‘as high as 300,000’ (p5)

More recently the government have refused to disclose how they have come up with the number of 400,000 ‘high-risk families’ which will see the TFP extended beyond the life of this parliament.

But, as the TFP is being operationalized and local authorities are beginning to look for and work with these ‘troubled families’, some interesting new numbers are beginning to emerge. Each local authority was allocated a number of ‘troubled families’ to work with and ‘turn round’ by the government. This again is fairly well known but what is less well known is the expectation that local authorities will need to work with more than their indicative numbers of ‘troubled families’. The Troubled Families Programme (TFP) Financial Framework states that

You may not succeed in turning around every family that you work with, and therefore it is likely you will need to work with more families than your indicative numbers (p3)

This seems to be good sense as local authorities are unlikely to achieve a 100% success rate in their work, but it does cast further doubt on the figure of 120,000. At the same time that the government are suggesting that LAs work with more than their indicative number, these same councils are finding that the number of families living in their areas that meet the three central criteria (Crime/ASB, educational non-attendance & worklessness) is surprisingly lower than government estimates. To give some examples:

  • Coventry have identified 203 out of 750 ‘troubled families’ up to year 2 of the programme who meet all 3 central criteria
  • Cornwall, in year 1, identified 65 out of 451 families who met all 3 central criteria
  • In Staffordshire, according to Louise Casey, around 33% of ‘troubled families’ meet all 3 criteria

I need to point out that some of these figures are nearly a year old and so they may have changed as families circumstance have changed but the figures do reflect low numbers of families meeting all 3 central criteria when compared to the indicative numbers of ‘troubled families’ given to each local authority.

Late last year, the Communities and Local Government Select Committee published a report into Community Budgets which included discussion of ‘troubled families’. The report noted that ‘We found that the fourth local criterion was important ad without it the definition was too narrow to allow local authorities to work with families that were in need of help’ (p31).

In practice, local authorities are allowed to use whatever local criteria they like as long as it reflects families that are a high cost to the public purse. They are also allowed to use as many criteria as they like (some LAs are using six or seven different local criteria) and an ability to ‘refresh’ the criteria exists.  In evidence given to the Committee, one LA Troubled Families Co-ordinator stated that ‘It has been really liberating to have locally developed criteria … It has given us the freedom to make sure that local families we want to be part of the programme are part of the programme.’ So, a ‘troubled family’ is basically whatever a local authority wants it to be* and as long as the net is cast sufficiently wide to enable them to turn around their indicative number of families by May 2015.

The similarities with the ‘problem families’ of the 1950s is striking. John Macnicol, writing about those ‘problem families’ highlights that during attempts to count the number of families ‘the scope for arbitrary judgement was enormous’ (1999 p85) and that there was a ‘great danger that, in practice, judgements would be made according to middle class social norms and expectations’ (p87). He goes on to highlight that C. P. Blacker, a key supporter of the ‘problem families’ thesis was forced to acknowledge that ‘the number of problem families deemed to exist in any community will depend on where the line is drawn’ (p85).

It would appear then, that the government has so much political capital invested in the figure of 120,000 that it won’t budge from it and, even as local analysis suggests that the figure may be much lower according to the government’s own criteria, the government are actively encouraging local authorities to work with more than 120,000 families and to define what a ‘troubled family’ is any way they like, as long as the total number ‘turned round’ adds up to or exceeds an arbitrary figure extracted from an out of date survey used for entirely different purposes. This beggars belief and makes a mockery of the Prime Minister’s commitment to turn round the lives of the ‘120,000 most troubled families’.

In writing this blog, I looked for the origin of ‘the number game’ and found, via Wikipedia, that the ‘numbers game’ is also known as the ‘policy racket’ and the game involves ‘policy shops’ and policy banks’ – which all seems quite fitting….


Macnicol, J. (1999) Problem Family to Underclass in Fawcett, H. & Lowe, R. (eds) Welfare Policy in Britain: The road from 1945, Basingstoke: Macmillan

* On Saturday 18 January, someone from the Troubled Families Team tweeted me and told me that the statement above – ‘a ‘troubled family’ is basically whatever a local authority wants it to be’ was ‘not actually true’ and ‘plain wrong’ because ‘troubled families’ still have to meet 2 out of 3 national criteria. I agree, I should have been clearer with my writing and so I am happy to clarify this here.

I would also like to clarify  that two of the 3 ‘national’ criteria allow for ‘local’ interpretation. A footnote to the ‘national’ anti-social behaviour criteria in the TFP Financial Framework states that it is ‘an indicative and not an exhaustive list’ (p4). Recently issued guidance to local authorities states that:

In order to ensure that you are able to include families who are involved in anti-social behaviour, you should use comparable measures or interventions according to your local approach to anti-social behaviour. This may involve working with, and drawing on the information of, a range of local partners including police, landlords and council colleagues to make sure you find families of concern. (my emphases)

The ‘national’ education criteria also appears to have been amended to allow local ‘discretion’ based on ‘common sense’. The guidance from October 2013 states that

We expect that colleagues in local authorities and schools will adopt a common sense approach. Based on evidence of worrying levels of non-attendance, head teachers and Troubled Families Coordinators should apply their professional discretion to identify children whose patterns of attendance are  of equivalent concern to those set out in the existing criterion (my emphases).

I couldn’t find any clarification on what counted as ‘evidence’ or what a ‘worrying’ level of attendance is or what level of ‘concern’ would be ‘equivalent’ so there is a fair amount of room for professional ‘discretion’ and ‘common sense’. The main concern of the government appears to be that whatever evidence or measures are used, they satisfy ‘internal auditors’ – again this amounts to a check at a local level.

So, apologies for the confusion in the original posting. What I could have said, more accurately, is that a ‘troubled family’ is, within reason, basically whatever a local authority and their partners (including police, schools, landlords etc) want it to be’. Glad to have the opportunity to clear this up.

In relation to the third ‘national’ criteria which covers households that also have ‘an adult on Department for Work and Pensions out of work benefits, the last official statistics release from November 2013 show that there were 5.5 million working age benefit recipients in May 2013.

6 thoughts on “The ‘troubled families’ numbers game

  1. Pingback: “In a hurry”: Fast policy and ‘troubled families’ | A Kind of 'trouble'...

  2. Pingback: Social workers as ‘bounty hunters’? | A Kind of 'trouble'...

  3. Pingback: (Mis)Understanding Troubled Families (Part 1) | A Kind of 'trouble'...

  4. Pingback: (Mis)Understanding Troubled Families (Part 3) | A Kind of 'trouble'...

  5. Pingback: Has the Troubled Families Programme itself been ‘turned around’? | A Kind of 'trouble'...

  6. Pingback: The Queen and I | A Kind of 'trouble'...

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s