The ‘cumulative impact’ problem in social welfare: some legal, policy and theoretical solutions

ABSTRACT

Notwithstanding its central place in social security policy-making and in the experiences of those on the receiving end of welfare reforms, cumulative impact is an under interrogated and theorised problem in social welfare scholarship. This paper seeks to address this in two ways. First, it draws on the lessons of the comparatively well-developed literature on cumulative impact in environmental studies to: (i) identify different dimensions of cumulative impact and (ii) best practices in undertaking cumulative impact assessments. Second, it argues that having regard to cumulative impact is not just good policy-making, but is also a legal obligation. To discharge the ‘Public Sector Equality Duty’ under s.149 Equality Act 2010, a public authority needs to have due regard to clear cumulative impacts between measures. In the context of social security policy-making, a failure to undertake enquiries to establish cumulative impacts between policies is likely to breach the PSED. Drawing on examples from the UK Government’s Welfare Reform Act 2012 and the Welfare Reform and Work Act 2016 throughout, the paper makes the case for both the importance of addressing cumulative impact in policymaking, and the need for greater interrogation by social welfare scholarship.

KEYWORDS:

• Cumulative impact
• public sector equality duty
• welfare reform
• social welfare
• Equality Act 2010

Introduction

If a public authority introduces two or more policies at the same time which affect the same group of people, should they consider their cumulative impact? This is not a hypothetical question. Overlapping policies are an ‘intrinsic feature’ of the modern welfare state (Harris 2013, pp. 60–67). Any working-age household in the UK in receipt of social security payments has by definition been subject to a litany of interwoven reforms following the Welfare Reform Act 2012, the Welfare Reform and Work Act 2016 and a raft of secondary legislation. As the seemingly evergreen ‘welfare reform’ agenda continues, the impact of one policy initiative inevitably collides with prior, ongoing and future reforms.

This paper is a detailed interrogation of this cumulative impact problem in social welfare law and policy. Notwithstanding its central place in social security policy-making and the experiences of those on the receiving end of welfare reform policies, cumulative impact is under interrogated and theorised in social welfare scholarship. Instead, cumulative impact often (but not always) goes uninterrogated in favour of more siloed critiques of individual policies. In response, this paper has two agendas. The first is to draw on the burgeoning literature on cumulative impact assessments in environmental studies. Here, cumulative impact has been at the heart of disputes since the 1980s. As a comparatively well-developed field of both practical and theoretical insights into cumulative impact, it offers a rich source of material to help interrogate the same problems in the social welfare context. I draw on this material to set out three dimensions of cumulative impact and to outline good practice in cumulative impact assessments.

The second is a narrower argument, focusing on the legal requirements of the Public Sector Equality Duty under s.149 Equality Act 2010 (PSED). I argue that to discharge the PSED, a public authority needs to have due regard to clear cumulative impacts between measures. In the context of social security policy-making, a failure to undertake enquiries to establish cumulative impacts between policies is likely, in many circumstances, to be a breach of the PSED.

The argument is in four sections. First, I draw on the well-developed literature on environmental impact assessments to outline three dimensions of cumulative impact in the social security context: the aggregative (the summing of impacts), the interactive (how one policy can affect others), and the longitudinal (tracing compounding impacts over time). Second, I turn to a more detailed assessment of the UK Government’s welfare reform programme and particularly the failures to undertake cumulative impact assessments for reforms in the Welfare Reform Act 2012 and the Welfare Reform and Work Act 2016. Here, I provide a critique of the Government’s justifications for avoiding cumulative impact analyses in the social welfare context. Third, I argue that in circumstances where a package of measures is introduced by one Government department and these are liable to impact cumulatively on one or more protected groups, regard to the measures’ cumulative impact is required to be compliant with the PSED. Finally, I identify what good practice looks like for a cumulative impact assessment, again drawing on the well-established literature tackling this question in the environmental context. In straddling theoretical, policy and legal concerns, it is hoped that this paper provides both a case for the importance of interrogating cumulative impact more fully within social welfare scholarship, and a starting point for doing so.

Types of cumulative impact in social welfare

In his influential work on complexity in the structure and administration of social security systems, Harris underscores that overlapping interactions between entitlements are an ‘intrinsic feature’ of modern welfare states (Harris 2013, pp. 60–67). A decision for one entitlement may affect another; a claimant may need to navigate multiple agencies (in the UK, particularly the Department for Work and Pensions and HM Revenue and Customs) or even ‘different offices within the same agency’; and reporting changes of circumstances can be labyrinthine (ibid, p. 60–65). Complexity is, therefore, not just a problem of efficiency – it has a real impact on claimants themselves. The ways in which these impacts bite across different populations has been the focus of other work. The Welfare Conditionality project raised complexity of the social security system as a particular barrier to realising entitlement, particularly for veterans and migrants who described the system as ‘baffling’ and a ‘minefield’ (Dwyer et al. 2016, Scullion et al. 2019). A recurrent theme throughout O’Brien’s seminal advice-led ethnography is the ‘default of administrative complexity’ (O’Brien 2017, p. 4), where interfacing legal and bureaucratic obstacles are ‘significantly compounded for EU nationals’ (ibid, p. 25).

The focus here is different to this concern with the ‘complexity’ of social security schema, catalysed by Harris’ work. Instead, this section seeks to explore one consequence of this complexity in detail: cumulative impact. Although critique of the ‘complex network of overlapping systems’ is well-established in social security scholarship (Johnson 1992, p. 14), there has been far less scholarly focus on how overlapping impacts can be theorised and subsequently traced.

The environmental studies literature is particularly instructive when considering how best to conceptualise and measure the ‘cumulative impact’ of policy. There was a turn within the 1980s away from the individual piecemeal assessment of the impact of particular policies, towards a recognition that the ‘collective power’ of policies leads to ‘insidious’, often negative, cumulative impacts on environmental outcomes (Odum 1982, p. 728). Characterised as ‘progressive nibbling’, ‘death by a thousand cuts’, and the ‘tyranny of small decisions’ (Blakley and Franks 2021, p. 6), within the broad-ranging academic literature on evaluating environmental impacts, it is now ‘widely understood that cumulative effects are the only effects that really matter’ and ‘without cumulative effects assessment, environmental impact assessment misses the point’ (ibid, p. 8).

The sustained interrogation of cumulative impact within this literature offers a resource for considering the same problem in the social security context. It is important to start with what ‘cumulative impact’ actually is, and how it can manifest in the social security context. Writing about environmental impacts, Vlachos argues that there are three dimensions in which cumulative impacts can be considered: the aggregative (the summing of impacts across policies), the interactive (the cross-effects of actions), and the longitudinal (the combined effect of measures over time, what he terms the ‘diachronic’ dimension) (Vlachos 1982, 1985). All three can be seen within the context of social welfare policy.

The aggregative dimension takes policies affecting the same class of persons and simply ‘adds them up’ (Vlachos 1982, p. 64). This is similar to concepts of ‘accumulation’ in other literatures, such as access to higher education (see Lambrechts 2020). This approach defines the vast majority of attempts to consider the cumulative impact of welfare reforms. For instance, an individual affected by the UK’s so-called ‘bedroom tax’ policy would face an average deduction to their housing benefit award of between £10-£15 per week (Department for Work and Pensions 2021a). They would also – almost by definition – be affected by the abolition of Council Tax Benefit and its replacement with the Council Tax Reduction Scheme. Although schemes vary by locality, this leads to an average additional contribution of £3 per week for low-income households who may otherwise receive full council tax benefit (House of Commons Library 2020, p. 16). The aggregative approach adds these together. Other examples abound, such as the combined losses caused between reforms to Employment and Support Allowance and the ‘bedroom tax’ (Roberts et al. 2014, p. 14), or reforms to Tax Credits and housing benefit (Social Security Advisory Committee 2014, p. 7). As put by Edwards et al, wherever a package of social welfare measures is introduced, households subject to one policy are likely to ‘be absorbing other welfare reform changes simultaneously’ (Edwards et al. 2013, p. 55).

The interactive dimension goes further by assessing the cross-effects between measures. For instance, in response to the COVID-19 pandemic, the UK Government increased the levels of housing benefit for those in the private rented sector to the 30^th percentile of rents (see Reg. 4 Social Security (Coronavirus) (Further Measures) Regulations 2020 (SI 2020/371)). As these rates had been pegged to 2012 rent levels, this marked a considerable increase in the support available. However, for many households, particularly in high rent areas, this increase in support was entirely illusory due to interactive effects with the pre-existing ‘benefit cap’ policy. For a lone parent with two children renting a three-bedroom property, substantial areas of the country become unaffordable. If they were renting in central London for instance, the maximum housing benefit available following the change in policy would be £441.86 per week. However, given the benefit cap would apply (£23,000 per year for a lone parent in Greater London (Department for Work and Pensions 2021b)), this would leave just £23.28 per week for living costs; (obviously) well below the destitution line for a family of three (Harris et al. 2020, p. 72). The same story is reflected across the country – any household outside of London with a rent higher than £146.18 per week would not receive the benefit of the uplift (ibid).

Finally, the longitudinal dimension looks forward to the ‘‘totality of interactive impacts over time’ (Vlachos 1982, p. 64). This dimension has a lot in common with the broad literature on ‘cumulative inequality’ and ‘cumulative disadvantage’ that has been particularly influential in American scholarship on ageing, race and disability (Meyer and Abdul-Malak 2020, pp. 27–46). Here, if one policy disadvantages a particular group (such as lone parents or people with disabilities), it may in turn increase the risk of further disadvantages (such as the exposure to health risks, harming educational or work opportunities, etc), which in turn increases the risk of even further disadvantages, and so on. The cumulative impact of measures can therefore work like a downward spiral, felt by affected communities across the life-course.

Research on the Welfare Reform Act 2012 reforms illustrate this longitudinal dimension well. Moffatt et al (2016) qualitative study into the effects of the ‘bedroom tax’ underscores that the policy’s impact was far greater than simply the reduction in household budgets. They highlight the long-lasting and overlapping nature of the impacts at play that scar households for years afterwards:

One of the consistent predictions of the impact of current welfare benefit changes is worse mental health and wellbeing and our findings bear this out. Mechanisms which have been proposed to lead to increased health inequalities include decreased incomes, increased food poverty, increased stigmatization and decreased housing security, all of which we observed (Moffat et al, p. 203).

Similar findings are made in a longitudinal qualitative study by Community Links with those directly affected by welfare reforms in Newham, London (Roberts et al. 2014, p. 41). Their work underscores that the ‘the cumulative impact of welfare reform’ makes it ‘hard for people to cope with and respond positively to changes’, with ‘knock-on impacts on their health and wellbeing’ (ibid). Poverty is a well-established driver of cumulative disadvantage, including poorer health which in turns fuels other forms of disadvantage (Lynch et al. 1997). The associated effects of stigma and shame are long-felt and particularly aggravated in the context of a wide-ranging welfare reform programme (Garthwaite 2013).

Indeed, the very nature of these cumulative effects form part of Dwyer and Patrick’s calls for longitudinal approaches to welfare reform research, noting that the ‘deeply negative’ impact of policies such as benefit sanctioning are only truly felt ‘over time’ (Dwyer and Patrick 2021, p. 63). As they argue, the ‘future is an especially rich terrain’ in studies of welfare reform impact – only longitudinal studies can begin to identify how affected individuals’ future planning and activities are affected (ibid, p. 67).

What is striking about the literature on environmental impact assessment is the sophistication with which cumulative impact is assessed across these the aggregative, interactive, and longitudinal dimensions. Regulatory agencies tend to define cumulative impact in ways that cross-cut these three dimensions. The United States Council on Environmental Quality defines ‘cumulative impact’ as:

• The total effect, including direct and indirect, on a given resource, eco-system or human community of all actions taken.

• Effects that may result from the accumulation of similar effects or the synergistic interaction of different effects.

• Effects that may last for many years beyond the life of the action that introduced them.

• Effects that must be analysed in terms of the specific resource, eco-system, or human community affected, and not from the perspective of the specific action that may cause them (Blakley and Franks 2021, p. 5) (emphasis added).

This focus on the aggregative, interactive and longitudinal dimensions of impact from the perspective of the ‘human community affected’ and not from the ‘perspective of the specific action’ responsible, underscores both the importance of social impacts to environmental impact assessments and the methodological sophistication in current practice. Importantly for the arguments that follow, the literature on environmental impact assessments also highlights the disproportionate burden faced by certain groups across these three dimensions of cumulative impact. As Blakley and Franks put it:

… the pace and scale of emergent cumulative effects issues in the Anthropocene era disproportionately affect communities, particularly the marginalized, the disempowered, and the world’s indigenous peoples (ibid, p. 3).

We will return to insights from the literature on environmental impact assessments when asking what effective cumulative impact assessments could look like in the social security context. However, it is first important to interrogate the current (lack of) assessment of these impacts in the UK social welfare context, focusing on the ‘welfare reform’ agenda.

Cumulative impact and the UK Government’s welfare reform agenda

The UK Government’s two flagship welfare reform bills since 2010 – the Welfare Reform Act 2012 and the Welfare Reform and Work Act 2016 – are examples of social welfare cumulative impact par excellence. They contained a smörgåsbord of interwoven reforms, many of which clustered around low-income working age households. The former introduced eight large-scale reforms, including the introduction of Universal Credit (a reform designed to ‘subsume’ a number of pre-existing working age benefits together, affecting millions of families) (Royston 2012), Personal Independence Payments (a wide-reaching reform to payments to support the costs of disability by replacing Disability Living Allowance) (Roulstone 2015), and the Benefit Cap (a ‘limit’ for the total amount of working-age benefits a household can receive).¹ The latter reduced the level of the Benefit Cap, introduced the ‘two-child limit’ in Child Tax Credit/Universal Credit (O’Brien 2018), and introduced various benefit freezes, abolitions, and conditionality.

Across these two pieces of legislation, each individual reform was subject to a separate equality impact assessment (EIA). There were a total of nineteen such assessments during the passage of the Welfare Reform Act 2012, each covering a different policy in isolation. Table 1 details these in full.

Table 1. Equality Impact assessments for the Welfare Reform Act 2012

Benefit cap – equality impact assessment

Social sector housing under-occupation – equality impact assessment

Disability Living Allowance reform – equality impact assessment

Universal Credit – equality impact assessment

Single Fraud Investigation Service – equality impact assessment

Time-limiting contributory element of Employment and Support Allowance – equality impact assessment

‘Tell Us Once’ registration service – equality impact assessment

Fraud penalties and sanctions – equality impact assessment

Youth provisions in Employment and Support Allowance – equality impact assessment

Child maintenance (new scheme) – equality impact assessment

Entitlement to work as a condition for contributory benefits and statutory payments – equality impact assessment

Universal Credit budgeting advances – equality impact assessment

Social Fund localisation – equality impact assessment

Lone parent conditionality – equality impact assessment

Housing Benefit: CPI uprating of Local Housing Allowance – equality impact assessment

Consideration of revision before appeal – equality impact assessment

Conditionality, sanctions and hardship – equality impact assessment

Welfare Reform Bill 2011: equality impact assessments general introduction

Industrial injuries benefits simplification – equality impact assessment

Strikingly, all these EIAs take place in a silo – they analyse individual measures as if the other policies introduced by the Act simply did not exist. For instance, the EIA on Universal Credit carves out conditionality measures central to its operation (such as the claimant commitment) into a separate EIA, does not consider reforms to the Social Fund at all, and has no mention of the ‘Benefit Cap’. The EIA on changes to Disability Living Allowance make no reference to the disproportionate impact of the ‘removal of the spare room subsidy’ (i.e. the ‘bedroom tax’) on those with disabilities, nor the numbers of households likely to lose entitlement to Carers Allowance. There is a failure to consider even the most rudimentary aggregative dimensions cross-cutting these measures, let alone the interactive and/or the longitudinal.

The failure to consider the cumulative impact of this wide-ranging reform package was widely condemned. The Social Security Advisory Committee in particular warned that:

As it is inevitable that a significant number of claimant households are likely to be affected by multiple reforms, this approach to assessing impact does not sufficiently demonstrate the full effect of these changes. It is important that the impact of these reforms is fully understood so that any unintended gaps in support, particularly to the most vulnerable claimants, may be addressed (Social Security Advisory Committee 2014) (emphasis added).

The Parliamentary Joint Committee on Human Rights, following their inquiry on the UK Government’s compliance with the UN Convention on the Rights of the Child, called for a unified assessment of the likely cumulative impact of the Welfare Reform Act 2012 reforms (Equality and Human Rights Commission 2015a). The Equality and Human Rights Commission described a single department being responsible for monitoring and assessing the cumulative impact of spending review and budget decisions in the wake of the reforms as ‘vital’ (Equality and Human Rights Commission 2015b). Indeed, a petition organised by campaigners for carers and people with disabilities to draw ‘attention to the cumulative impact of the reforms on disabled people and carers’ and calling for a cumulative impact assessment was debated in parliament (Machin et al. 2014, p. 16). Introducing the Opposition Day debate based on the petition, Liam Byrne MP called on the Government to ‘publish, by October 2012, a cumulative impact assessment of the changes made by the Government that affect disabled people’, arguing that the Government had ‘tried to disguise and bury the impact of their reforms on disabled people’ by approaching the Welfare Reform Act’s impact ‘with all the finesse of a bull in a China shop’ (HC Deb, vol.566, col.398, 10 July 2013).

The Government’s failure to produce such a cumulative impact assessment led to many other organisations filling the gap themselves. The SSAC refers to evidence by the Joseph Rowntree Foundation, the Centre for Regional Economic and Social Research, the Institute for Fiscal Studies, and Demos (among others) in their 2014 report, ‘The cumulative impact of welfare reform: a commentary’ (Social Security Advisory Committee 2014). The Scottish Welfare Reform Committee, noting the UK Government’s approach only offers ‘a guide to the impacts of each element in the reform package’ undertook its own work to assess the ‘cumulative impact on different sorts of households – on pensioners, couples, lone parents, households with and without dependent children, and so on^’ (Scottish Welfare Reform Committee 2015).

These studies of cumulative impact expose the disproportionate burden faced by groups with protected characteristics. Work by the Centre for Welfare Reform, cited by Dodd, highlights the impact on disabled people of the Welfare Reform Act 2012 reforms:

The cumulative impact of these changes … total a cut equivalent to £4410 per disabled person, or nine times the burden of cuts endured by most citizens … disabled people, 8% of the population [will] bear 29% of all cuts (Dodd 2016, p. 152-153).

The Scottish Welfare Reform Committee instead underscore the disproportionate impact on lone parents (who are themselves disproportionately women), finding that:

Families with dependent children are one of the biggest losers – in Scotland, couples with children lose an average of more than £1,400 a year, and lone parents around £1,800 a year. Because this is the cumulative impact of several individual benefit changes the overall impact has previously been hidden (Scottish Welfare Reform Committee 2015, p. 3).

Notwithstanding these concerns with the failure to address cumulative impacts of Welfare Reform Act 2012 reforms, the same approach was adopted for the Welfare Reform and Work Act 2016. Again, a set of separate impact assessments, siloed for each new policy, was undertaken by the Department for Work and Pensions, without regard for the cumulative impact on certain populations. Table 2 sets these out in full.

Table 2. Equality impact assessments for the Welfare Reform Act 2012

Impact Assessment of Social Rent Reductions

Impact Assessment of the change in conditionality for responsible carers on Universal Credit

Impact Assessment of the Benefit rate freeze

Impact Assessment of Tax Credits and Universal Credit, changes to Child Element and Family Element

Impact Assessment for the benefit cap

Impact Assessment to remove the ESA Work-Related Activity Component and the UC Limited Capability for Work Element for new claims

Impact Assessment for Converting Support for Mortgage Interest (SMI) from a benefit into a Loan

Again, the Government’s failure to consider the cumulative impact of measures led to other organisations stepping in. Perhaps the most instructive is Portes and Reed’s cumulative impact analysis for the Equality and Human Rights Commission (Portes and Reed 2018). Considering the impact of the Welfare Reform and Work Act 2016 reforms alongside those in the Welfare Reform Act 2012, they concluded that:

… in particular, the freeze in working-age benefit rates, changes to disability benefits and reductions in UC rates [are] likely to lead to significant increases in the number of households below a minimum acceptable standard of living. A large number of households in vulnerable groups (such as lone parents and couples with children, and households with disabled adults and/or children) lose substantial proportions of their incomes (over 20% in many cases) from the package of reforms to direct taxes and transfer payments, even taking into account increases in gross incomes arising from [increases to the minimum wage] (ibid, p. 184).

Given concerns about the deficiency of the impact assessments underpinning the Welfare Reform and Work Act 2016, the Equality and Human Rights Commission wrote to the Secretary of State for Work and Pensions (Equality and Human Rights Commission 2016). It was concerned, among other issues, about the failure to adequately assess the impact of the Bill on people who share protected characteristics under the Equality Act 2010, and that the process was not compliant with the PSED (ibid). In reply, the Secretary of State underscored that they used ‘the most robust analysis available to give a good assessment of both the rationale for and the impacts of the reforms’ (ibid).

The Department for Work and Pensions’ approach to impact assessments for the Welfare Reform Act 2012 and the Welfare Reform and Work Act 2016 therefore exists in a parallel reality where individual policies take effect in perfect isolation from one another. The Department for Work and Pensions goes to the effort of undertaking a large number of individual impact assessments, but does so without regard to even the most basic of interactions between the policies. The next section explores the UK Government’s justification for this in more detail.

The Government’s justification for not assessing cumulative impact

The Government’s reasons for not undertaking an assessment of cumulative impact has remained the same across both waves of welfare reform. The Government position is that a ‘cumulative impact assessment would be so complex and subject to so many variables that it would be meaningless’ (HC Deb, col.413, vol.566, 10 July 2013). As put by Mark Hoban MP, a Minister of State within the Department for Work and Pensions

All major welfare reform changes have been accompanied by a published equality impact assessment and these are updated if impacts change. I reiterate that a cumulative impact assessment would be so complex and subject to so many variables that it would be meaningless, helping neither individuals nor policy makers, and it would soon be incorrect and out of date. (HC Deb, col.414, vol.566, 10 July 2013).

The position is put at greater length in a note accompanying the Welfare Reform Act 2012 impact assessments (Department for Work and Pensions 2011). There are effectively three strands to this ‘complexity’ justification. First, the ‘scale’ of the interactive impacts at hand:

The scale of policy change provided for by the Welfare Reform Bill is significant, and is planned to take place over an extended period … Therefore, the impacts build up over a significant period of time, and at a different rate for different measures. To provide a single summary accurately taking account of the different timings would be analytically complex and extremely challenging. To simplify would risk providing a set of misleading impacts (ibid, p. 3).

It is not clear that this complexity prevents consideration of cumulative impacts. Not least as dozens of organisations have already undertaken their own research to address the cumulative impacts of reforms. For example, the Equality and Human Rights Commission (Portes and Reed 2018); the Children’s Commissioner (2015); Contact a Family (2012); and the Scottish Parliament’s Welfare Reform Committee (2015), to name but some. As suggested by the Social Security Advisory Committee, ‘such methodological problems [are not] insurmountable to the extent that headline findings cannot be produced, given appropriate modelling assumptions’ (Social Security Advisory Committee 2014, p. 30). The aim, as its members put it, is not to ‘produce the perfect study’ (ibid, p. 5) – instead, it is to assess the aggregative, interactive, and longitudinal effects of welfare reforms as accurately as possible. I will return to some methodological issues below when considering what a good cumulative impact assessment within the social security context would look like.

The second strand to the justification is the interface between broader fiscal measures and welfare reform:

Moreover the changes to social security benefits and tax credits contained in the Bill take place over a wider context of fiscal change. The impact assessments therefore do not account for wider changes that would impact on households over the period, for example, the aim to increase income tax personal allowances to £10,000 (Department for Work and Pensions 2011, p. 3).

There are two problems with this justification. In common with the arguments above, it is perfectly possible to account for changes in other policy areas where interactive effects are obvious and should be accounted for (such as changes to income tax thresholds). Indeed, this analysis has already been done by the Equality and Human Rights Commission (Portes and Reed 2018). In their report, Portes and Reed undertake a cumulative impact assessment of all tax, benefit, tax credit and UC policies based on age, disability, race and sex, including two-way combinations between these characteristics. Although most reforms are included in their modelling with ‘high accuracy’ (including ‘most parts of the benefit system’ and ‘most parts of the tax system’ (2018, pp. 53–53), other funds – especially those administered locally, such as the Council Tax Reduction Scheme – had less reliable data. Indeed, the report is framed by prior work undertaken on behalf of the Equality and Human Rights Commission, with Portes and Reed underscoring that:

The commission noted that cumulative impact assessment work undertaken [to date] showed that such analysis, while technically challenging, was entirely possible with existing data (Portes and Reed, 2018, pp. 38).

Their findings demonstrate both the possibility of assessing cumulative impact across tax and welfare measures and the power of doing so for groups with protected characteristics. For instance, they find that the cumulative effect of policy changes on tax and welfare between May 2010 and January 2018 was particularly acute for women in comparison with men. Women aged 35 to 44 lost over £2,200 per year on average, compared with less than £550 for men. Likewise, lone parents (who are disproportionately women) in the bottom quartile of the income distribution lost around 25% of their net income (Portes and Reed 2018, pp. 183–184).

Importantly, the Government’s same concerns about the difficulty of accounting for interactions between social security measures and fiscal policy bites even more acutely on the Government’s current approach of addressing each policy in silo. Indeed, it is not only the ‘wider changes’ that fail to be accounted for within individual impact assessments, or even those within the same Government department, but also changes sitting directly alongside them in the same Bill.

Third, the Government argue that cumulative impact assessments would ‘obscure’ the impact of individual policies themselves:

Collectively these factors substantially limit the extent to which a cumulative impact assessment would provide an accurate analysis of the impacts of the Bill as a whole. Moreover, an amalgamated assessment is likely to obscure the impacts of individual policies rather than aid the understanding of those considering the Welfare Reform Bill in Parliament and the wider public (Department for Work and Pensions 2011, p. 3).

However, clearly if the Government is concerned that the impact of ‘individual policies’ would be obscured, they could still undertake the siloed impact assessments they do currently alongside any cumulative impact assessment. Indeed, an appraisal of individual policies is likely to be an important stepping-stone for analysing cumulative impact in any event. I argue therefore, that the Government’s justifications for failing to undertake a cumulative impact assessment of welfare reform measures do not stand up to scrutiny.

Is a cumulative impact assessment a legal requirement?

Having addressed the different types of cumulative impact and argued that the UK Government’s current approach (and justifications for this approach) is deficient, this section turns to the legal obligation to have regard to the cumulative impact of measures. In particular, I argue that to discharge the PSED, a public authority – in some circumstances – must have regard to the cumulative impact of measures.

Before turning to some example cases which turned on PSED considerations, it is worth providing an overview of the duty. At its core, the PSED is a statutory procedural duty which requires all public bodies (such as government departments, local authorities, schools, and so on) (Schedule 19 Equality Act 2010) or private organisations fulfiling ‘public functions’ (s.149(2) Equality Act 2010)² to “have due regard to the need to, eliminate discrimination and advance equality of opportunity for those with protected characteristics, and foster good relations between those with and without protected characteristics.³ It effectively replaced the ‘hotchpotch of different provisions’ (Hickman 2013, p. 326) dealing with specific groups – such as the procedural duties under s.71 Race Relations Act 1976 and s.76A Sex Discrimination Act 1975 – with a single ‘equality duty’. This duty is laid out in s.149 Equality Act, which specifies:

(1) A public authority must, in the exercise of its functions, have due regard to the need to—

(a) eliminate discrimination, harassment, victimisation and any other conduct that is prohibited by or under this Act;

(b) advance equality of opportunity between persons who share a relevant protected characteristic and persons who do not share it;

(c) foster good relations between persons who share a relevant protected characteristic and persons who do not share it.

…

(3) Having due regard to the need to advance equality of opportunity between persons who share a relevant protected characteristic and persons who do not share it involves having due regard, in particular, to the need to—

(a) remove or minimise disadvantages suffered by persons who share a relevant protected characteristic that are connected to that characteristic;

(b) take steps to meet the needs of persons who share a relevant protected characteristic that are different from the needs of persons who do not share it;

(c) encourage persons who share a relevant protected characteristic to participate in public life or in any other activity in which participation by such persons is disproportionately low.

(4) The steps involved in meeting the needs of disabled persons that are different from the needs of persons who are not disabled include, in particular, steps to take account of disabled persons’ disabilities.

(Emphasis added).

A failure to discharge the duty can be challenged through the courts by way of an application for judicial review. There is a sizeable body of case law, described by the Court of Appeal as ‘two lever arch files’ worth (R. (on the application of Bracking) v Secretary of State for Work and Pensions [2013] EWCA Civ 1345 (para. 25) and by the High Court as a ‘feast of decided cases’ (R. (on the application of Rowley) v Minister for the Cabinet Office [2021] EWHC 2108 (Admin) (para. 38)), which has built up around the interpretation of the PSED. A detailed examination of this case law here would leave room for little else. Instead, it is sufficient for present purposes to note four key elements of the duty before turning to the specific issue of cumulative impact.

First, the weight accorded to the duty is highly fact sensitive and varies according to the decision at play. As put by Lord Neuberger in Hotak v London Borough of Southwark Council [2015] UKSC 30, ‘what is required by the section 149 duty will inevitably vary according to the circumstances of the case’ (para. 74). For instance, the duty ‘operates differently’ if dealing with a ‘broad statement of principle’ for a forthcoming policy, than it does for a ‘detailed policy document’ (see R. (on the application of United Trade Action Group Ltd) v Transport for London [2021] EWCA Civ 1197 (para. 71).

As a result, compliance is a question of fact to be assessed by the Court. Challenges have repeatedly emphasised the existence of an ‘important evidential element’ which must demonstrate the ‘recording of steps taken by the decision maker to meet their statutory requirements’, and the need to ‘assess the risk and extent of any adverse impact’ (Bracking, para. 26). This does not have to be in the form of a formal equality impact assessment, but the Courts will assess whether there has been compliance with close scrutiny of documentary evidence put before them (Hickman 2013, p. 339). In other words, the duty does not always need referencing by name in order to be discharged effectively, providing sufficient evidence is available to the effect that regard has been had pursuant to the PSED. In practice, however, public decision makers often produce ‘equality impact assessments’ to indicate that they have had ‘due regard’ to the potential for adverse impact and can provide associated evidence to that effect.

Second, the duty is a personal rather than general one; namely, it is on the relevant minister or decision-maker themselves. What matters is what they knew or took into account, not what their staff or sub-ordinates did (R (National Association of Health Stores) v Department of Health [2005] EWCA Civ 154 (para. 26)). It is not enough that an equality impact assessment exists if no decision-maker has seen it. This consideration can become more complicated in the context of decision-making shared between multiple individuals. In Mark Logan v London Borough of Havering [2015] EWHC 3193 (Admin), fifty-four elected representatives were able to vote on the design of their Council Tax Reduction Scheme – a discretionary, localised replacement for the abolished Council Tax Benefit. Of these, at least twenty were notified of the availability of an EIA on the scheme’s potential impact, but it could not be determined if the others had been aware of its availability or had considered the impact it detailed. The Court determined that this resulted in breach of the PSED (para. 52). The important link is between the decision maker(s) and the regard required by the duty.

Third, consideration of the duty has to occur as an identifiable part of the decision-making process, not following the introduction of the policy as a ‘rear-guard action’ (Bracking, para. 25). The duty is designed to draw attention to the potential risk of adverse impact on protected groups with the intention that moves may be taken to mitigate it – clearly this aim cannot be advanced if regard is only had post-hoc. However, the duty is a continuing one; it is not sufficient to simply discharge it at an early stage (for instance, while setting an initial budget) and then not have regard at a later stage (for instance, when allocating this budget to individual policies).

Fourth, the Court will not assess the weight given to the equality implications within the final decision. The focus is instead firmly on the procedural duty to ‘have due regard’. In other words, ‘the court cannot interfere with the decision simply because it would have given greater weight to the equality implications’ (R. (on the application of Hurley and Moore) v Secretary of State for Business Innovation & Skills [2012] EWHC 201 (Admin) (para. 78). The issue of weighting the considerations mandated by the duty within the final decision is for the decision-maker; the duty is therefore not concerned with the substantive decision itself, but rather with the procedure leading to it. Although s.149(3)(b) (in bold above) states that a public authority must have regard to ‘tak[ing] steps to meet the needs’ of those with protected characteristics, where they are different to those without, the duty does not imply ‘active steps’ must be taken by the decision-maker (as may be the case, for instance, if the PSED required ‘due consideration’) (Butler 2016, p. 31). However, this focus on ‘process’ is still broad-ranging. As the Court put it in Rowley, it includes ‘enquiry, thinking-process and reasoning-process’ (Rowley, para. 39). Despite being a process-focused decision, it still very much demands a detailed assessment of the decision-making process by the Court.

Cumulative impact and the PSED

I argue that in circumstances where (i) a package of measures is introduced by a single Government department and (ii) these are liable to impact cumulatively on one or more protected groups, regard to their cumulative impact is required to be compliant with the PSED. Given the wealth of case law on the PSED, it is helpful to adopt the ‘linked themes’ and ‘virtues’ that bring the legal principles together outlined by Justice Fordham in R (on the application of Katherine Rowley) v Minister for the Cabinet Office [2021] EWHC 2108 (Admin):

“In the present case the following linked themes, regarding the principled application and enforcement of the PSED duty, are of particular significance: (i) importance, (ii) proactivity; and (iii) rigour, together with the recognised virtues of (iv) evidence-based thinking; and (v) legal sufficiency of enquiry” (Rowley, para. 40).

When considering the need for regard to the cumulative impact of measures, two of these are of particular importance: ‘legal sufficiency of enquiry’ and ‘rigour’. The former will in some circumstances mean a public authority must undertake work to assess the cumulative impact of related measures to be compliant with the PSED. In the context of secondary legislation, the duty of enquiry associated with the PSED is broad-ranging. In Bracking, the court endorsed a ‘combination’ of the principles in Secretary of State for Education and Science v Tameside Metropolitan Borough Council [1977] AC 1014 with the interpretation of the ‘due regard’ formulation. Elias LJ summarised the position in R (Hurley & Moore) v Secretary of State for Business, Innovation and Skills [2012] EWHC 201 (Admin) in terms endorsed in Bracking (para. 25) and again by the Court of Appeal in R. (on the application of Bridges) v Chief Constable of South Wales [2020] EWCA Civ 1058 (para. 175), as:

… [the] duty of due regard under the statute requires public authorities to be properly informed before taking a decision. If the relevant material is not available, there will be a duty to acquire it and this will frequently mean than some further consultation with appropriate groups is required … the public authority concerned will, in our view, have to have due regard to the need to take steps to gather relevant information in order that it can properly take steps to take into account disabled persons’ disabilities in the context of the particular function under consideration (Bridges, para. 89-90) (emphasis added).

This same sentiment is expressed in a myriad of different ways within the case law on the PSED. In Bridges, this duty to make enquiries is characterised as the need to take ‘reasonable steps to make enquires about what may not yet be known to a public authority about the potential impact … on people with the protected characteristics’ (Bridges, para. 181). In summary, it is well-established that if relevant material is not available to the public authority to assess their duties under the PSED, there is a duty to acquire it, and this will frequently require further consultation with appropriate groups.

So, can the cumulative impact of multiple measures form part of this duty to make enquires, and if so, in what circumstances? The failure to have due regard to the cumulative impact of measures has only arisen once in the context of a ‘due regard’ duty, in a permission to appeal hearing in Fawcett Society v Chancellor of the Exchequer [2010] EWHC 3522 (Admin). This was a bold challenge to the 2010 Budget in its entirely: an ambition, as Hickman notes, it is ‘difficult to imagine any other ground of judicial review being capable of’ (Hickman 2013, p. 333). The Fawcett Society, a charity which campaigns for gender equality and women’s rights, had calculated that the cuts in the 2010 budget fell disproportionately on women – £5.7 billion of the £8.1 billion in cuts came from women, and 72% of those receiving a real-terms cut in their public sector wage were women (Fawcett Society 2012).

The Fawcett Society challenged the Government’s failure to discharge their duty under s.76A Sex Discrimination Act 1975 to have ‘due regard’ to the need to eliminate unlawful discrimination and harassment and to promote equality of opportunity between men and women (Fawcett Society v Chancellor of the Exchequer, para. 2). In particular, the Government had failed to produce a Gender Equality Impact Assessment for the Budget. As argued by Karon Monaghan QC for the claimants:

… that approach is arguably unlawful because it would cause the cumulative impact of the budget on the equality of opportunity objective to be ignored and would mean that swings and roundabouts would also fall out of the picture (ibid, para. 6) (Emphasis added).

The Government argued that an assessment of cumulative impact was not required. James Eadie QC’s arguments mirrored the justifications criticised above, stating that gender equality objectives are better dealt with ‘by reference to the impact of the various specific individual items’ (ibid) as the Budget:

… contain[s] a number of varied measures, differing in nature, which may be introduced or flagged, the overall aim of the emergency budget made an impact exercise of the Budget as a whole inappropriate, as did “the polycentric nature of the decision-making process”, the number of assumptions that would need to be made and the questionable validity and thus utility of any results of any such attempt (ibid, para. 7).

The Court rejected the Fawcett Society’s arguments, noting that the ‘varied nature of the Budget content’ and the ‘timetable necessary to give effect to it’ made an overview of its cumulative impact a ‘particularly difficult task’ (ibid, para. 8). Instead, it was ‘perfectly possible’ to have regard to the cumulative impact of Budget measures ‘from the accumulation of separate impacts’ (ibid). A consideration of the Budget’s impact line-by-line, ‘cannot arguably be said to be unlawful’ (ibid).

Although this is seemingly a strong endorsement of the Government’s position, the Court draws two important caveats. First, Justice Ouseley endorses the Government’s concerns about the ‘polycentric nature of the decision-making’ leading to the formulation of the national Budget (para. 7). Indeed, it is difficult to conceive of a decision-making process that stretches across more Government departments. However, this same critique does not bite when considering social security decision-making sitting within the Department for Work and Pensions. Where decisions are made out of one department, and in the case of welfare reforms, ordinarily within the same package of policies introduced at the same time, the process cannot be said to suffer from polycentricism. Indeed, the development of policy detail for welfare reforms is largely monocentric inso-far as Government decision-making processes are concerned.

Second, the Court caveats their dismissal of the claimant’s arguments. Justice Ouseley concludes that:

… it is clear that if the analysis of gender equality impacts can be undertaken by consideration of the line items in the budget, the duty is fulfilled or rather not breached by its being dealt with in that way at that stage (ibid, para. 7) (Emphasis added).

It follows, therefore, that if such analysis cannot be undertaken by consideration of items line-by-line, then more may be required. What the cumulative impact assessments referred to above demonstrate, is that such a line-by-line assessment does not allow for the adequate analysis of equality impacts in the context of welfare reform. This because clear aggregative, interactive and longitudinal dimensions of cumulative impact are excluded from the analysis.

The necessity for this approach is underscored by the second theme identified above: ‘rigour’. Exercising the PSED ‘in substance, with rigour, and with an open mind’ is fundamental to discharging the duty lawfully (Hotak, para.75, Bridges, para. 175). This applies not only to ‘a proper and conscientious focus on the statutory criteria’ (Hotak, para. 75) but also a proper appreciation of the impact any such measure has on those with protected characteristics. As the Court put it in R (Hurley & Moore) v Secretary of State for Business, Innovation and Skills [2012] EWHC 201 (Admin) (Divisional Court), as endorsed in Bracking (at para. 25), the Court must be:

… “satisfied that there has been a rigorous consideration of the duty, so that there is a proper appreciation of the potential impact of the decision on equality objectives and the desirability of promoting them” (Hurley and Moore, para. 77) (emphasis added).

In circumstances where one Government department is introducing a package of measures that have aggregative and interactive effects on one another, it is difficult to see how an assessment of their individual impact alone is sufficiently rigorous to discharge the PSED. A failure to have regard to the potential impact of cumulative effects is a failure to have a ‘proper appreciation of the potential impact’ (Hurley and Moore, para. 77) of the decision.

What is clear from the case law on the PSED therefore, is that there is a context-sensitive, but broad-ranging, duty of enquiry and expectation of rigour for the discharge of the PSED. Taken together, where the analysis of individual impacts would be deficient (as in the case of welfare reform measures), and where the decision-making does not suffer from clear polycentricism (as is the case for decisions stemming from one Government department), I argue that a failure to assess the cumulative impact of measures would be non-compliant with the PSED.

What does a good cumulative impact assessment look like?

If an assessment of cumulative impact is required for some forms of social welfare decision-making, what would a good cumulative impact assessment look like in this context? What is required for PSED compliance is minimal – indeed, it is not necessary for there to be a structured ‘impact assessment’ as such, providing there is some documentary record of due regard having been had (Hickman 2013). However, the bare minimum for legal compliance should not be the aim of any policy-maker. Instead, the well-developed literature on cumulative impact assessments in environment studies identifies four characteristics for a robust assessment of the cumulative impact of a policy. Such impact assessments can, in turn, have tangible impacts on the quality of the policy-making process (Blakley and Franks 2021).

First, a robust cumulative impact assessment should involve consultation with affected groups. Referred to in the environmental studies literature as ‘social scoping’, building in ‘participatory, meaningful and transparent’ (International Finance Corporation 2013, pp. 21–22) consultation with affected groups has ‘long been considered essential to good practice’ in the assessment of cumulative impact (Blakley and Franks 2021, p. 10). This has practical benefits. Early engagement of affected groups (in our context, those social security recipients affected directly by the reform) serves a ‘boundary-setting’ function for the rest of the cumulative impact assessment exercise (Baxter et al. 2001, p. 255–256). It can help to identify cumulative impacts that may otherwise be neglected or deemed insufficient, particularly longitudinal impacts that may relate to historic reforms or a household’s ongoing experiences of poverty (ibid).

Clear consultation and engagement may also help to build greater levels of trust in Government policy-making by affected communities. As found recently by the Social Security Advisory Committee, there had been a significant erosion in trust by disabled people in the policy-making process in the Department for Work and Pensions, with:

this erosion of trust stemm[ing] from the cumulative impact of benefit changes which were designed to reduce spending on disability benefits and the numbers of people who were receiving them (Social Security Advisory Committee 2021, p. 25).

Second – and following on from this first point – cumulative impact assessments should be mixed methods, drawing on both quantitative and qualitative approaches. This has been called for by the Social Security Advisory Committee, who have recommended that the Department for Work and Pensions adopt a broader range of methodological techniques to address cumulative impact:

Social research might also be employed to further understand the impact of successive social security reforms. Hereby a range of methods would be employed to better understand the experience of affected claimants, including: surveys, interviews or focus groups (SSAC, 2014, p. 31).

Indeed, environmental cumulative impact assessments have long adopted a mix of qualitative and quantitative approaches, with qualitative processes dominating early practices. As found by Cooper and Canter in their survey of Environmental Impact Assessment practitioners in the US, four of the five favoured methodologies used were qualitative (such as case studies) (Cooper and Canter 1997, pp. 25–26). Qualitative approaches can help to fill gaps in access to quantitative data – particularly where data may not already be held or collected by the Department for Work and Pensions – and provides access to information that quantitative approaches cannot, such as the direct experiences of those affected by the reforms. Put another way, drawing on Patrick’s work, a mixed method approach can help to expose the ‘everyday realities’ of welfare reforms that may otherwise go undetected (Patrick 2017).

Third, a robust cumulative impact assessment process should not simply be a point-in-time snapshot, but should instead involve processes for tracking impact over longer periods – for instance, by seeking the feedback of a sample of affected parties deeper into the life of the policies concerned. As Blakley and Franks put it, a good cumulative impact assessment:

… clearly defines roles and responsibilities for implementation, monitoring, and feedback over the long term. The gravitas of assessing and managing cumulative effects in the [environmental impact assessment] cannot be understated as “each additional disturbance or impact, regardless of its magnitude, can represent a high marginal cost … ” (2021, p. 7).

These arguments have been mirrored in the Social Security Advisory Committee’s recent calls for greater engagement and ‘co-production’ with disabled people, including a recommendation to recruit a large-scale panel of disabled people with experience of social security to consult with regularly on both new reforms and the impact of existing reforms (Social Security Advisory Committee 2021). Tracing impact over time in this way helps to ensure an ongoing dialogue with affected communities, while also uncovering those longitudinal cumulative impacts that are more difficult to assess with more snap-shot processes.

Finally, cumulative impact assessments should be transparent. This simply requires that where impact is assessed, the analysis should be made available publicly. As argued by Campbell, Fredman and Reeves, it is not always the case that the Government is as forthcoming as it could be with the publication of equality impact assessments. They criticise the Government’s failure to publish the equality impact assessment for the Coronavirus Act 2020 as ‘antithetical to the participation dimension of substantive equality’ (Campbell et al. 2020, p. 194).

Indeed, improving the transparency of decision-making processes is a key aim of the PSED – securing improvements to the policy-making process is more effective than simply providing remedies after waiting for the discrimination to occur (Olatokun 2021, pp. 85–86). The lesson of the literature on environmental impact assessment is that good quality cumulative impact assessments do more than provide a veneer of legitimacy to a decision-making process or serve a simple box-ticking exercise. Instead, they can help to improve the policy-making process and identify where existing policies become problematic when interacting with new policies (for a detailed example, see Baines and Taylor 2021).

Conclusion

This paper has had two agendas. The first is to draw on the environmental studies literature on cumulative impact assessments to provide a framework for addressing overlapping policies in the social welfare context. The three key dimensions of aggregative, interactive and longitudinal cumulative impact identified in this literature all bite in social security reform and warrant greater attention in both social welfare scholarship and in policymaking processes themselves. The second has been to argue that having regard to cumulative impact can, in some circumstances, be a legal obligation. This is particularly true in the social welfare context, where in the UK, often one Government department – the Department for Work and Pensions – is responsible for developing and implementing reforms. This is even more acute where reforms are introduced at the same time or as part of package of measures. Here, a failure to undertake work to establish the cumulative impact of measures and have due regard to this analysis, may in turn be a breach of the PSED.

However, it is important to underscore in conclusion that assessing the cumulative impact of policies is not just a legal obligation. Meeting the minimum requirements of the PSED or a tokenistic effort to assess the most rudimentary overlaps is not the aim. Instead, recognising the importance of cumulative impact is about good quality, inclusive, evidence-based policy-making that reflects the experiences of those affected by Government policy. To fail to recognise cumulative impact, is to deny the reality of a policy’s impact.

The cumulative impact of welfare reform measures has been an implicit focus of social welfare scholarship. Patrick’s work in particular, which has highlighted the importance of the realities of welfare reforms (Patrick 2017) and, with Dwyer, the importance of longitudinal approaches (Dwyer and Patrick 2021), touches on those longstanding longitudinal effects of being subject to a seemingly never-ending programme of colliding reform. Indeed, it could be said that it is the cumulative impact of policies or reforms that causes the greatest hardship for large numbers of individuals, rather than any specific, individual policy by itself. However, social welfare scholarship can do more to address the implications of cumulative impact and to incorporate its aggregative, interactive and longitudinal dimensions into the critique of individual policies and the policy-making process. These kinds of cumulative impact have been at the forefront of the environmental studies literature since the 1980s – we have some catching up to do.

Acknowledgement

Many thanks to Professor Charlotte O'Brien and the anonymous reviewer for their very helpful comments on this article.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. In addition to introducing Universal Credit, Personal Independence Payments, and the Benefit Cap, the Welfare Reform Act 2012 also included regulations to establish the new ‘claimant commitment’, the introduction of the so-called ‘bedroom tax’, additional penalties for fraud, the abolition of the ‘Social Fund’, reforms to Employment and Support Allowance, and changes to child support. Suffice to say it was a wide-ranging piece of legislation.

2. ‘Public Functions’ has the same meaning as functions of a public nature under s.150(5) Human Rights Act 1998.

3. Protected characteristics are specified under s.4–12 Equality Act 2010 and are Age, Disability, Gender reassignment, Marriage and civil partnership, Race, Religion or belief, Sex and Sexual orientation.