Addressing the Relatively Autonomous Relationship Between Child Maltreatment and Child Protection Policies and Practices

Child protection policies cannot be understood as simple reactions or responses to the phenomena of child maltreatment. They have their own dynamics and determinations to the point where they can be seen to operate quite independently of the phenomena that they are supposed to be trying to prevent. A whole range of political, cultural, and sociological influences come to bare which have the effect of deflecting such policies and systems from recognising the nature of the social problem of child maltreatment. The article develops the argument by summarising a number of both historical and comparative studies of child protection. It also draws upon prevalence studies to demonstrate that the problem is much bigger and more complex than is often assumed. In conclusion, it outlines ways in which this challenge can be addressed.


Introduction
For many years, I have attempted to analyse, critically, the development of child protection policy and practice in England (see in particular Parton 1985Parton , 1991Parton , 2005Parton , 2014. Initially the approach was historical, but in more recent years, it has become much more explicitly comparative by trying to locate developments in England in a much wider international context (Lonne et al. 2009;Gilbert et al. 2011a;Stafford et al., 2012;Parton 2017). Increasingly I have come to recognise that child protection policies cannot be understood simply as reactions to or responses to the phenomena of child https://doi.org/10.1007/s42448-019-00022-9 * Nigel Parton n.parton@hud.ac.uk 1 School of Human and Health Sciences, University of Huddersfield, Huddersfield HD1 3DH, England maltreatment. They have their own dynamics and determinations to the point that they seem to operate quite independently of the social problem, which, it is assumed, they are trying to prevent and respond to. A whole range of political, cultural and sociological influences come to bare which have helped establish child protection as an institution in its own right and which varies according to both time and place. This article develops this argument by drawing on a number of both historical and comparative studies of child protection. In the process, it will pose the question as to how far child protection policies, in many Western jurisdictions, seriously attempt to tackle the social problem of child maltreatment. In doing so, it will draw upon a number of prevalence studies to demonstrate that the nature of the problem is much bigger and more complex than is often assumed. In conclusion, it will outline a 'social harm' perspective and demonstrate how such a perspective is important for reforming such policies in the future.

The Politics of Child Protection in England
My initial interest in these issues was prompted by my experiences as a social worker in a local authority child welfare agency in the North of England in the early/mid 1970s. The period was punctuated by an event which generated enormous media, public, and political concern-the manslaughter of a 7-year-old child by her step-father in early 1973 while under the supervision of social workers. This lead to a high-profile public inquiry (Secretary of State for Social Services 1974) and the introduction of a range of new policies and procedures. Local authority social workers were henceforward required to take lead responsibility for responding to a newly discovered problem-child abuse-that appeared to require new skills, new forms of intervention, and new organisational arrangements to deal with it. However, I was uneasy about what was happening to both clients and social workers. I was especially concerned that the nature of this new problem did not seem as clear-cut or as self-evident as was suggested in the new training and practice guidelines that were being introduced. Nor was I convinced that many of the new policies and practices were as humane as was assumed. They seemed to have a number of unintended consequences that were deleterious for all concerned-not least for many of the children and parents who were on the receiving end. I felt the need to develop a more critical analysis of what was happening with the hope of formulating new insights and perspectives as the basis for developing new ideas for taking policy and practice forward.
I wrote The Politics of Child Abuse (Parton 1985) to engage with these issues. Part one of the book tried to explain why the problem of child abuse had emerged as such a crucial one for welfare practitioners in England from the early 1970s onwards, what the nature of the problem was, how this had changed, and how this had impacted on policy and practice. It became clear that the modern re-discovery of child abuse in England was heavily influenced by developments in the USA and that the publication of the paper by Henry Kempe and his colleagues in Denver, Colorado, in the Journal of the American Medical Association in 1962 on the 'battered child syndrome' was key (Kempe et al. 1962).
The term was specifically chosen to appeal to as wide an audience as possible, including relevant professionals, particularly paediatricians. There was no reference to any possible legal, social, or deviancy implications. The problem was medicalised so that doctors and medical technologies, particularly the X-ray, could be used to identify old and hidden injuries, which were seen as key in diagnosing the problem. Not only did the issue receive wide media and political coverage in the USA (Nelson 1984), it also attracted the attention of a number of paediatricians in England, as well as the NSPCC (the National Society for the Prevention of Cruelty to Children) which, in 1968, set up the Battered Child Research Unit to carry out research and promote awareness of the problem. However, I argued that it was the public inquiry into the death of Maria Colwell, which was established by the Minister responsible for social services, Sir Keith Joseph, that provided the catalyst for mobilizing public, professional, and political and, crucially, media concern about the issue. While the report was critical of the work of individual professionals, its main conclusion was that 'it was the 'system', using the word in its widest sense, which failed her' (Secretary of State 1974 p86). The reforms which were introduced were premised on an individualised disease model of child abuse (Parton 1985). As Melton (2005) has since argued, the original designers of modern Anglo/American child protection systems made two fundamental errors. Both emanated from assumptions which underpinned the notion of 'the battered child syndrome'; both the scope and the complexity of the problem of child abuse were underestimated. It was assumed that 'the problem of child maltreatment was reducible to 'syndromes'in effect, that abusive and neglectful parents were either very sick or very evil and that they could thus be characterised as 'those people' who were fundamentally different from ourselves' (Melton 2005, p11).
Between the publication of the Colwell Inquiry Report in 1974 and 1987, there were over 30 further inquiries into the deaths of children in the UK while under care or supervision of local authorities. The overriding conclusions were that there were failures by professionals in sharing information and using the full authority of the law vested in them resulting in a failure to protect the children from significant harm and death.
However, in the summer of 1987, a rather different scandal hit the headlines. This time 121 children were kept in a General Hospital in the county of Cleveland in the northeast of England against the wishes of their parents on emergency 'place of safety orders' on suspicion of being sexually abused and based on, what came to be seen, as very dubious diagnostic methods. The government established a high-profile public inquiry, which reported the following year (Secretary of State for Social Services 1988). Not only was this the first scandal into possible professional overreaction-intervening too early and too authoritatively-but also the first time where the actions of paediatricians and other doctors, as well as social workers, were put under the microscope and subject to very public criticism. Thus, by 1988 it seemed that the state in the guise of a number of health, welfare, and criminal justice agencies was both under-intervening and over-intervening. This was a major challenge and helped prompt the passage of the Children Act 1989, which put in place a fully reformed legislative framework, which, among other things, tried to ensure that professional practice was much more accountable and transparent.
Governing the Family (Parton 1991) provided a detailed political analysis of the processes involved in establishing the new sets of balances and concluded that there was something of a revision in the discourse of child protection as a result. Whereas previously the concern was with diagnosing and preventing the 'disease' or 'syndrome', increasingly the emphasis became investigating, assessing, and weighing 'forensic evidence'. The focus was on the need to identify cases of 'high risk'-where 'high risk' was concerned with both identifying and protecting children from significant harm while also protecting the privacy of the family from unwarrantable state interventions (Parton 1991). Such a characterisation was further substantiated in a subsequent collaborative study, based on a secondary analysis of social work files (Parton et al. 1997), where it was established that the overriding concern of child welfare professionals was to try and identify high-risk cases, differentiate these from the rest, and allocate scarce resources accordingly.
A major focus of both The Politics of Child Abuse and Governing the Family was to make explicit how a variety of social, political, and cultural factors had influenced the development of child protection policies in England. The aim was to demonstrate how child protection, together with the idea of child abuse itself, had been shaped by political and social factors over time. This theme was central in two subsequent studies Safeguarding Childhood (Parton 2005) and The Politics of Child Protection (Parton 2014).
What was also evident was that child protection scandals also became something of a proxy for much wider political debates about a range of issues concerned with the work and organisation of a variety of health, welfare, and criminal justice agenciesparticularly social work and more recently the police-and the direction of social and public policy services more generally. Such scandals have been used, politically, as vehicles for bringing about major policy changes well beyond the more specific concerns of child protection.
For example, the Every Child Matters programme of reform (Department of Education and Skills 2003) introduced by the New Labour government of Tony Blair in England was said to be the government's response to the public inquiry into the death of Victoria Climbie in 2000 (Lord Laming 2003). However, it was clear that the scandal was used by the government as a vehicle for bringing about much wider changes to the organisation of children's services at both the central and local government levels and which had been previously clearly indicated in the Public Expenditure Review the previous year (HM Treasury 2002;Parton 2005).
Similarly, in late 2008, the cases of Peter Connolly and Shannon Matthews were explicitly used by the Conservative Party as examples of the 'broken society' and as, apparently, providing clear evidence of the failures of the changes introduced by Every Child Matters and more fundamental failures with the New Labour social policy reforms. In many respects, the Conservative Party's policies for the reform of welfare when it came into coalition government under the Premiership of David Cameron in 2010 can be seen to have been forged in the immediate fallout from the child abuse scandals in late 2008 (Parton 2014).

The Growing Crisis in Child Protection in Anglo/American Jurisdictions
Increasingly I became aware that some of my concerns about the development of child protection policy in England were also reflected in other jurisdictions, particularly the USA, Canada, and Australia (Lonne et al. 2009). Henry Kempe et al.'s identification of the 'battered child syndrome' (Kempe et al. 1962) had an immediate impact in the USA. In 1963 the Children's Bureau issued a model reporting law whereby certain health and welfare professionals were required to report cases of actual and suspected child abuse to designated public authorities, and all 50 US states adopted such a law by 1967 (Hutchison 1993). This was followed by the first national child protection legislation, the Child Abuse and Prevention and Treatment Act (CAPTA) in 1974, which, among other things, required states to have such mandatory reporting laws in place. Most Australian states also introduced mandatory reporting from the late 1970s onwards (Ainsworth 2002).
Having established the new systems to protect children, the next 30 years witnessed a huge increase in the number of children being reported in the USA from 9563 (0.1/ 100,000 children) in 1984, to 3,126,000 (47.0/100,000 children) in 1996 (Lonne et al. 2009, p.26). However, while the proportion of reported cases which were 'substantiated' as child abuse was over 60% in the 1970s, the figure had dropped to well below 40% by the early 1990s. Similar trends were evident in Canada (Trocme et al. 1995). While the way statistics were collated varied between Australian states, the rate of growth of child abuse reports was even greater than in the USA. For example, in the state of Victoria, reports of child abuse and neglect increased more than 5000% between 19775000% between /19785000% between and 19935000% between /19945000% between from 517 to 26,622 (Parton et al. 1997.
In England, there were no comparable statistics. The only statistics available which cover the last quarter of the twentieth century relate to the numbers of children on a child protection 'register'. A child's name was placed on a register where, following an investigation and a multidisciplinary case conference, it was felt that the child continued to be at risk of suffering abuse or neglect and should be subject to a 'child protection plan'. The numbers of children on registers in England quadrupled between 1978 and 1991 from 11,844 to 45,300.
While mandatory reporting legislation was particularly associated with the growing pressures on the child protection systems in the USA and Australia, the UK never had such legislation even though it also began to experience similar problems. It seems the problems emanated from the key assumptions which had underpinned the systems from the outset. This original failure to recognise the full scope and complexity of the problem was exacerbated by the subsequent broadening of the concept of child abuse. While the notion of 'the battered child syndrome' proved the dominant underlying metaphor for many years, the category of child abuse was subject to various 'mouldings' (Hacking 1992) and 'diagnostic inflation' (Dingwall 1989, p.29), so that by the late 1980s, it included emotional abuse, neglect, and sexual abuse as well as physical abuse and was no longer focused only on babies and young children but included young people up to the age of 18.
By the mid-1990s, there were a number of authoritative reports which were arguing that child protection systems in the USA, UK, Canada, and Australia were, at best, out of balance, or, at worst, in crisis and in need of reform. It seemed child protection systems had developed wide 'nets' in which were caught a whole variety of concerns about children and that certain sections of the population, particularly the poor, single parent households, and certain minority ethnic groups, were at much greater risk of being caught in the nets than others and with very little prospect of appropriate help being forthcoming (DoH 1995;Thorpe 1994;Waldfogel 1998).
By the early/mid 1990s, therefore, evidence was emerging in all the Anglophone countries that there were significant problems with their child protection systems.
Public inquiries, authoritative official reports, and research were all pointing to significant challenges that needed to be addressed.

Comparing Child Protection Systems
It is in this context that researchers and policy makers in the USA, Canada, the UK, and Australia began to look elsewhere to see if other countries approached these issues differently and whether they had any more success. Some of the earliest research was carried out at Brunel University in London and compared child protection systems in England and France (Cooper et al. 1995). The researchers talked with child protection workers in both countries and studied the respective histories and operation of their child protection systems. They became very aware that there were major cultural differences between the two countries and that this could be seen to permeate all areas of law, policy, and practice. In particular, while the French system seemed to be infused with both an optimism and trust of both families and social workers' abilities to look after children, this was not the case in England where pessimism and distrust seemed to dominate.
This was followed by a much bigger study (Hetherington et al. 1997) which analysed child protection systems in the Belgium Flemish community, the Belgium Francophone community, France, Germany, Italy, Netherlands, England, and Scotland. Compared with the situation in the other European countries, the English system was characterised as one of increased managerialism with a culture of defensiveness where accountability had been reduced to checking whether procedures had been followed and where there was a pervasive anxiety about things going wrong.
At the same time, another project was being lead by Gilbert (1997) in the USA. While the research complemented that lead by Cooper, it was prompted primarily by the rapid increase in reports of child maltreatment in the USA. Researchers were recruited to analyse and compare the child protection systems in nine countries-Belgium, Canada, Denmark, England, Finland, Germany, the Netherlands, Sweden, and the USA. It was thought that a major reason for the upsurge in reports in the USA and the subsequent strains in the system arose from the mandatory reporting system and the vague definitions of child abuse evident. Comparing the USA system with systems elsewhere would provide one way of testing out these assumptions.
However, the key finding proved to be that there were important and more wide ranging variations between the countries concerning the extent to which systems emphasised a child protection or family service orientation, and these did not depend on whether there was a mandatory reporting system in place. The two orientations were distinguished along four dimensions.
The first was the way that the problem of child abuse was framed. In some systems, abuse was conceived as an act which demanded the protection of children from harm by 'degenerative relatives', whereas in other systems, abuse was conceived as a problem of family conflict or dysfunction which arose from social and psychological difficulties but which responded to help and support.
Secondly, and depending on how child abuse was framed, the response operated either as a mechanism for investigating deviance in a highly legalistic way, or as a service response to a family's needs. As a result, thirdly, the child welfare professionals functioned either, in the child protection orientation, in a highly adversarial way, or, in the family service orientation, in a spirit of partnership-particularly with parents. Finally, while there seemed to be a high rate of voluntary arrangements with parents in making out-of-home placements with the family service orientation, in the child protection orientation, the majority of out-of-home placements were compelled through the coercive powers of the state, usually in the form of court orders. However, the use of mandatory reporting laws did not appear to be linked to either the child protection or family service orientations.
The countries were grouped into three broad categories: 1. Child protection United States, Canada, England 2. Family service-mandatory reporting Denmark, Sweden, Finland 3. Family service-non-mandatory reporting Belgium, Netherlands, Germany The research clearly suggested that there were important differences in the way Anglo-American child welfare services were organised and the way they responded to concerns about child abuse, compared with northern European and Nordic countries. The researchers argued that while the details of different programmes and policies were important, the ways different systems operated were influenced by wider cultural and societal conditions.
The two orientations seemed to provide something of a benchmark for discussions and analyses in both research (Cameron et al. 2007a, b;Freymond and Cameron 2006;Harder and Pringle 1997) and various government reports. For example, the two orientations were drawn upon for analysing the child protection systems in Scotland (Scottish Executive 2002) and Australia (Bromfield and Holzer 2007).
Bringing these various studies together, a clear picture began to emerge regarding the nature of these two orientations and how the countries studied differed and is summarised in Table 1. More recently, the Gilbert research (1997) has been updated by comparing ten countries-the same countries as before plus Norway. The overall conclusion (Gilbert et al. 2011a) was that, while the two original orientations-child protection and family service-were still relevant, they needed to be revised in the light of the developments in the various countries during the intervening 15 years up to 2008/2009. The findings suggested that the approaches had become more complex. Countries previously identified with the child protection orientation, for example England and the USA, had taken on some of the elements of the family service orientation. At the same time, there was also evidence that those countries that had previously operated according to a family service orientation had made efforts to respond to increasing concerns about harm to children. This seemed to be the case in all the Nordic countries, with the possible exception of Sweden, and all the north European countries studied.
It was also possible to discern the emergence of a new approach-a child-focused orientation (Gilbert et al. 2011b). This orientation focused on the child as an individual with an independent relation to the state. It was not restricted to narrow concerns about harm and abuse; rather, the object of concern was the child's overall development and well-being. It aimed to go beyond protecting children from risk to promoting their wellbeing. If for any reason there was concern about a child's development, the state sought to intervene to offer support or intervention which could be more authoritative if this was required. With a child-focused orientation, the state takes on a growing role in terms of providing a wide range of early intervention and preventive services and also attempts to take into account the views and wishes of the child. Gilbert et al. (2011b) suggest that two major and somewhat contrasting lines of influence shape the emergence of the child-focused orientation. On the one hand, it has been influenced by ideas, related to 'the social investment state', and on the other hand, it has been influenced by a growing priority allotted to the processes of 'individualisation' as these apply to children and hence emphasise children's rights. However, these two lines of influence do not sit easily together and can lead to tensions, which signifies that the childfocused orientation can take different forms in different jurisdictions.
The idea of 'social investment' emerged in the 1990s as an ideal promoted by the OECD and the EU, among others. According to this view, investment in children takes on a strategic significance for a state keen to equip its citizens to respond and adapt to global economic change in order to enhance individual and national competitiveness. In this respect, trying to ensure that all children maximise their developmental opportunities, educational attainment, and overall health and well-being becomes a key priority for social and economic policy. This is a future-oriented approach, which considers childhood as a preparation for adulthood, so that investment in children in the present is designed to ensure that they will develop into productive and law-abiding adults. In contrast, the rationale for policies and practices which emphasise children's rights perceives children as individuals in the here and now and, while different, equally as valuable as adults. These policies are concerned with the quality of children's childhood, stating that it is a social justice issue to make sure that children are treated with respect and given a loving upbringing. Children are not seen so much as future workers, but as current citizens.
Overall, the child-focused orientation puts an emphasis on parent's obligations and responsibilities as carers so that services are provided to promote children's needs and well-being, often with and via the parent's, but in return demands positive 'outcomes' for the child.
The three orientations identified by Gilbert et al. (2011b) can be summarised as shown in Table 2.
The orientations range along a continuum from a more laissez-faire neo-liberal approach which emphasises the night-watchman functions of government to the more social democratic approach. The three orientations can be seen to parallel Esping-Andersen's (1990) often-cited classification of liberal (Anglo-American), conservative (Continental), and social democratic (Nordic) welfare state regimes. These comparative studies clearly demonstrate the importance of locating child protection policies and systems in their relevant political, cultural, and social contexts.
More recently, it has been suggested that, rather than focusing on the specific structures or components of the systems, the analysis might shift to the essential social values, laws, and culture which inform the different systems (Connolly et al. 2014). A theoretical typology is developed based on two value dimensions-Individualism and Collectivism, and Authoritarianism and Permissiveness. From these two dimensions, four types of child protection systems emerge: Authoritarian Individualism, Authoritarian Collectivism, Permissive Individualism, and Permissive Collectivism. The essential characteristics of the four types are:

& Authoritative Individualism
Here the child protection system focuses upon identifying and assessing individual children who are 'at risk' of abuse, punishing individual perpetrators, and removing children from harmful situations. Early intervention is targeted at 'high-risk' vulnerable families who are offered 'interventions' to bring about change with sanctions applied if they do not comply or there is no improvement. State-sanctioned agencies act to protect individual children and draw on the full weight of the courts, including the removal of children from families and transferring parental rights. This is very similar to the Gilbert, Parton and Skivenes' child protection orientation.

& Permissive Individualism
While the focus is again upon identifying vulnerable children and their families, the emphasis is upon supporting them in order to enhance their overall well-being. Here early intervention aims to provide unconditional support to children and families who volunteer for services and where little compulsion is used. The state acts to mediate and bring about positive change and is very similar to Gilbert, Parton and Skivenes' family service orientation.

& Authoritarian Collectivism
This primarily focuses on intervening in and regulating collective societal behaviour towards children and young people. The main emphasis is upon legal and cultural change through tight regulation of communities and organisations and clear standards of behaviour for organisations and communities to follow. While the rights of the child are likely to be emphasised, the system focuses on community priorities. The role of the state is primarily concerned with regulating communities and organisations.

& Permissive Collectivism
This emphasises the support of communities to improve the well-being of children and is likely to involve community development and public health approaches. It involves media campaigns and other health promotion-type tools to bring about cultural change. Unlike a more regulatory approach, the system is focused at the community level and encourages a diversity of approaches, with the role of the state somewhat downplayed. One of the very positive outcomes of these various comparative projects is the development of conceptual frameworks, which are able to help explain variations between child protection systems. In doing so, the importance of different social, cultural, and political contexts and values is underlined. As Neil Gilbert has suggested 'the comparative study of how different countries respond to child abuse has advanced over the last two decades, particularly in regard to the general characteristics of the systems they develop for these interventions' (Gilbert 2012, p532). However, I would go further for Gilbert suggests that these different systems have been developed in response to the problem of child maltreatment. But, as I have argued here, such systems have developed their own dynamics so that they seem to operate quite independently of the social problem which it is assumed they are trying to respond to. A range of diverse political, cultural, and sociological factors come to bare. While it might be an exaggeration to say that child protection policies and practices operate in a wholly autonomous way from what is known about child maltreatment, there is clearly a very large element of relative autonomy.

The Realities of Child Maltreatment
In order to identify the most appropriate approach for advancing child protection policies and practices, it is therefore important to have a clear understanding of what we mean by child maltreatment. While definitions will vary across time and place, this does not mean that we cannot develop definitions and approaches which are relevant to a particular time and a particular place: what do we understand by child maltreatment? What is its nature? How big is it? And what laws, policies, and practices are most appropriate for dealing with it? More immediately-how appropriate is the current child protection system for tackling child maltreatment? One way of considering these issues is to attempt to assess the prevalence of the child maltreatment.
An authoritative review of research on the prevalence of child maltreatment in 'high-income countries' (Gibert et al. 2009) concluded that every year between 4 and 16% of children were physically abused and one in ten were neglected or psychologically abused. During childhood between 5 and 10% of girls and up to 5% of boys were exposed to penetrative sexual abuse and up to three times that number were exposed to some form of sexual abuse. The review concluded that the number of cases of substantiated child maltreatment known to official agencies only accounted for a tenth of this total. Child maltreatment was found to substantially contribute to child mortality and had long-lasting effects on mental health, drug and alcohol misuse (especially in girls), risky sexual behaviour, obesity, and criminal behaviour, which lasted into adulthood.
A review of global research published on prevalence between January 1980 and January 2008 (Stoltenborgh et al. 2015) identified 244 publications covering 551 prevalence rates. The overall prevalence rates for self-report studies (mainly assessing maltreatment during childhood) were as follows: 127/1000 for sexual abuse; 226/1000 for physical abuse; 363/1000 for emotional abuse; 163/1000 for physical neglect; and 184/1000 for emotional neglect. In contrast, the studies which analysed the rates where victims had made their abuse known to official agencies were just four per thousand for sexual abuse and three per thousand for both physical and emotional abuse. The authors concluded that 'child maltreatment in all its forms is a global phenomenon of considerable extent, touching the lives of millions of children' (Stoltenborgh et al. 2015, p48) but that much of this is hidden from view.
A study of prevalence in the UK carried out in 2009 by the NSPCC (National Society for Prevention of Cruelty to Children) (Radford et al. 2011) estimated that 18.6% of young people aged 11-17, 25.3% of 18-24-year-olds, and 5.9% of children aged under 11 had experienced maltreatment at some point during their childhood. These figures compare with 42,850 children who were the subject of a child protection plan on 31 March 2012 (just 0.42% of the child population under the age of 18). The research by Radford et al. (2011) clearly confirms what Gibert et al. (2009) andStoltenborgh et al. (2015) established in their reviews of research that only a small proportion of abuse ever becomes known to official agencies.
However, a comparison of the 2009 NSPCC study with the previous NSPCC prevalence study in 1998 (Cawson et al. 2000) suggested there was a general decline in physical abuse from 13.1% (in 1998) to 9.8% (in 2009), sexual abuse from 6.8 to 5%, and verbal abuse from 14.5 to 6%. However, no significant changes were found in the prevalence of neglect. This seems to reflect similar trends in the USA Jones et al. 2006). Radford et al. (2011) suggest that the decline in the UK resulted from a number of factors including the impact of positive economic changes; public health measures, such as early intervention policies; and general changes in attitudes and behaviour including a decline in the use of the physical punishment of children and young people.
Prevalence studies also demonstrate that, while child maltreatment is found in all sections of society, there are clear links with certain social divisions, particularly in relation to class and gender. There is now a substantial body of evidence demonstrating a strong association between family poverty and the likelihood of a child experiencing child maltreatment. As Pelton concluded in a review of over 30 years of studies, 'there is overwhelming evidence that poverty and low income are strongly related to child abuse and neglect as well as to the severity of maltreatment' (Pelton 2015, p30). Just as in child health and education (Donkin et al. 2014), there is a social gradient in child maltreatment so that, at each incremental increase in family socio-economic disadvantage, there is a disproportionate increase in the chances of a child experiencing maltreatment (Berger and Waldfogel 2011;Bywaters et al. 2016).
However, it has become clear that it is material inequality rather than poverty per se which is key in understanding the differences in health and violence in society. Societies that have become more unequal also have the poorest relative of levels of child well-being (Mapp 2011;UNICEF, 2007). Societies which are poorer but more egalitarian have relatively higher levels of good health and less violence because of a higher degree of social cohesion. The relative distribution of income, wealth, and lifestyle were identified as central factors in influencing an individual's sense of worth and whether they felt valued together with the key mediating role played by social capital (Szreter and Woolcock 2003). Societies which are becoming wealthier but more unequal, for example the USA and the UK, perform poorly as a result. The countries that have consistently performed the best in terms of the regular UNICEF assessments of inequality in child well-being in rich countries are from the more social democratically organised countries of Scandinavia-particularly Denmark, Finland, and Norway (UNICEF 2016).
It has been argued that the effect of economic inequality on the level of lethal violence is limited to nations characterised by relatively weak collective institutions of social protection such as welfare states (Savolainen 2000;Messner et al. 2008). Societies in which the economic market is dominant increase social marginalisation and violence, loss of social bonds, egoistic individualism, loss of empathy, and high risk taking-all of which are conducive to higher levels of violence. Such arguments support Pemberton's contention (2015) that solidaristic societies-based on a social democratic state model where there are high levels of public services, high levels of trust, and low levels of inequality-offer the best protections against social harms. The idea of social harm 'acts as shorthand to reflect the relations, processes, flows, practices, discourses, actions, and inactions that constitute the fabric of our societies which serve to compromise the fulfilment of human needs and in doing so result in identifiable harms' (Pemberton 2015, p24).
Social harms encompass avoidable events and are preventable insofar as they are potentially subject to organised human control and grounded in alterable social relations. Such a perspective takes into account corporate and collective responsibility, together with the activities of the state and focuses on structures of inequality, exploitation, and vulnerability. The perspective is directly relevant for our understanding of the social problem of child maltreatment (Parton 2019) for, as I have argued throughout this article, the way child maltreatment is defined is central to the way it is recognised, managed, and prevented. Most approaches are defined narrowly in terms of the physical, emotional, and sexual violence or neglect perpetrated by individual adults, usually parents or those in positions of trust. Professional and societal responses are framed in terms of the protection of the child from adult perpetrators. In the process, such approaches do not permit any recognition of the collective harm and exploitation that is caused by institutions, harmful policies and laws, conflicts, failures of governance, and social disruption. There is now a recognition that, to seriously try and prevent child maltreatment, it is important to try and address the social factors underpinning child maltreatment. This in turn has led to the development of 'whole of community' and early intervention strategies designed to influence a broad network of relationships across the wider community (Bywaters 2019).
Drawing on the typology developed by Connolly et al. (2014) and recognising the complexity and pervasive nature of child maltreatment, it would therefore seem that some combination of the authoritarian collectivist and permissive collectivist approaches should be drawn upon for developing our child protection policies and practices in the future.

Conclusion
The central argument of this article has been that it is important to recognise that the developments of child protection policies and practices are only very loosely connected to responding to the social problem of child maltreatment. Using examples, primarily from England, I have tried to demonstrate they have their own dynamics and determinations which are related to a range of political, cultural, and sociological influences. In the process there has been a failure to recognise the extent and complexity of child maltreatment and what is required to address this. I suggest that prevalence studies demonstrate that child maltreatment is a social problem which causes social harm and that this needs to be placed centrally if we are serious about addressing this in the future. One of the positive outcomes of comparative child protection projects, which have been carried out over the last 20 years, is that they have helped develop conceptual frameworks, which help explain variations between child protection systems and also reforming such systems in the future. Drawing on the typology developed by Connolly et al. (2014), I have argued that some combination of the authoritarian collectivist and permissive collectivist approaches should provide the basis for developing policies and practices in the future.