Abstract
The Conservative-Liberal Democrat coalition government abolished the Discretionary Social Fund and Council Tax Benefit in the Welfare Reform Act 2012 as part of their programme of austerity, with powers to design replacement schemes devolved to local authorities in England. Discretionary Housing Payments, which had long been the responsibility of local authorities, were given an expanded role – to soften the edges of welfare reform being pursued by central government. This article presents analysis of a new quantitative dataset constructed by the author detailing variations in these three payments across local authorities in England. This analysis explores the variation in provision that now exists across England and examines the extent to which the political makeup of elected councils, as well as economic and demographic differences, can explain the variations in provision that now exist. We find that there has been substantial retrenchment in the local social security schemes in the period since their localisation, indicating that the devolution of powers alongside budget cuts has proved a successful mechanism for implementing austerity. We also find that the political makeup of elected councils is associated with the degree of cutbacks in these schemes, with Labour-led councils less likely to retrench across all three payments when compared with councils led by the Conservative party, suggesting that politics remains possible even in a harsh financial climate such as that faced by local authorities in England.
Introduction
The imposition of austerity in the UK since 2010 by Conservative-led governments has resulted in spatially unequal outcomes and has hit those in poverty hard. Local government has borne the brunt of particularly severe cuts (Gray and Barford, 2018). To date, the literature on the spatial impact of cuts to social security during this period of austerity has largely focussed on cuts to national payments (e.g. Beatty and Fothergill, 2016, 2018), the impact of which has been felt unequally across the UK. But the Conservative-Liberal Democrat coalition government also created a new spatial dimension in relation to social security provision through changes to three schemes, which we label collectively as ‘local social security schemes’ in this article.
The Welfare Reform Act of 2012 abolished two schemes – Council Tax Benefit and the Discretionary Social Fund – and tasked local authorities in England 1 with designing local replacements. Council Tax Benefit was a centrally designed but locally administered scheme, while the Discretionary Social Fund, as its name implies, was based on localised discretion, but the administration of the scheme had been ‘centralised to 20 Benefit Delivery Centres (BDCs) under previous Labour governments’, which the Conservatives viewed as impeding the discretionary nature of the scheme (Grover, 2012: 355). A third scheme, Discretionary Housing Payments, which are top-up payments made by local authorities to help residents meet rent payments, were given an expanded remit to include the mitigation of (some of) the effects of welfare reform.
These local welfare schemes were argued by government to soften the edges of cuts made to national social security payments. In a debate in the House of Commons in March 2013, the Labour MP Stephen Timms asked about a disabled woman whose council-owned home had been adapted at public expense but who, with the introduction bedroom tax, was faced with the prospect of being forced to move. In response, the then Secretary of State for Work and Pensions, Iain Duncan Smith, replied, The right hon. Gentleman knows very well that that is the point of discretionary housing payments . . . we have put more money into discretionary payments to sort these things out than they ever did when they brought these in. The reality is that there is money for them to do just that.
2
When pressed by Labour MP Kate Green about pressures on the Local Welfare Assistance schemes due to cuts in local authority budgets, the Parliamentary Under-Secretary for Communities and Local Government (and Conservative MP) Brandon Lewis spoke to the supposed efficiency of these schemes, suggesting that The hon. Lady is right; local authorities are running a range of really good schemes. That is why they have been passed to local authorities. What has been highlighted is that many authorities are running good, efficient schemes and spending way below the amount of money originally put forward.
3
These are far-from-unique exchanges. When faced with criticism about cuts to national schemes made in the period since 2010, government figures have on many occasions pointed to the new emphasis on local authority provision. The localisation of these schemes opened up the prospect of a new spatial terrain of social security support across England. And yet, we know relatively little about the extent to which a variable patchwork of provision has in fact emerged, due in part to the limited reporting requirements placed on local authorities in relation to the design of, and the financial allocations made to, these schemes. This matters because these schemes are significant for people living on the lowest incomes and because the devolution of powers in the context of fiscal consolidation is a neglected aspect of the governance of austerity.
The aims of this article are twofold: to examine variations in these three payments since their ‘localisation’ in April 2013 and to explore whether the political makeup of elected councils can help to explain this variation. To do this, we present analysis from a new dataset compiled by the author, consisting of publicly available information as well as data collated from two Freedom of Information (FOI) requests by Church Action on Poverty and the New Policy Institute, which they have very kindly shared. This new dataset enables examination of variation in relation to these three payments jointly for the first time.
In the next section, we provide a discussion of relevant literature on welfare state change that has informed the present study and outline our hypotheses. In the third section, we describe the three local schemes in question and discuss what we know about what has happened to these schemes since their localisation in April 2013. We then outline, in the fourth section, the data and methods, in which we describe the dataset that has been constructed and the variables that it contains. The fifth section contains the empirical analysis and this is presented for the three payments in turn, while the final section concludes.
Austerity, localism, and welfare politics
This study draws on the literature on welfare state change in seeking to understand our empirical cases of the localisation of three payments following the Welfare Reform Act 2012. In this tradition, there is debate about whether in the supposed ‘era of austerity’ we observe large-scale retrenchment or whether policies prove resilient to attempts to dismantle them (e.g. Hemerijck, 2013; Pierson, 1994). These debates are located primarily at national rather than local level. The contribution here is in terms of providing an examination of partisanship in the context of new (constrained) cases, as we explain in the following section. In doing so, we aim not only to shed light on these important cases but also on one of the neglected aspects of the programme of welfare reform that has been implemented in the period since 2010.
This debate acts as a starting point for thinking about the three local welfare schemes considered here, but the context in which these schemes have emerged differs in notable respect from that of the wider debate. First, the passage of the Welfare Reform Act is treated not as a variable but as a known fact – the puzzle is what the localisation provisions within the Act have meant for the policies enacted at local level. Second, the resilience perspective faces significant challenge in this context since local government finances are heavily dependent on grants received from central government. With these grants constrained significantly in the period in question one could argue that the localisation of these payments is an example of ‘responsibility without (effective) power’ (Brien, 2018: 6). An NAO report from 2018 finds that ‘spending power’ (central government funding plus Council Tax receipts) fell by 28.7% in real terms between 2010/2011 and 2017/2018, with rising Council Tax receipts partially offsetting deep cuts to central government grants (NAO, 2018: 15).
Third, even within the new institutionalist school of thought, which has tended to emphasise the political difficulty of imposing austerity, the path-dependent nature of institutions and the resilience this implies, there is a recognition that there are conditions that make the successful implementation of austerity more likely. Each of Pierson’s (1994: 19–24) three strategies for minimising resistance to retrenchment can be seen in relation to the Welfare Reform Act 2012. The first, obfuscation, can be observed in terms of the minimal reporting requirements placed on local authorities following the localisation of two payments (CTB and LWA), which has ensured that the consequences of localisation have minimal visibility. The second, division, might be seen in terms of the demand that pensioners be protected in the local CTB schemes that were designed, thereby concentrating losses on working-age residents. The third, compensation, applies to the Welfare Reform Act more broadly, where the increase in funding for Discretionary Housing Payments was used to offset losses for some claimants, providing both a real (yet limited) and rhetorical response to the critics of austerity. This context, then, suggests that resistence to retrenchment is likely to have been more difficult than where the successful deployment of these strategies could not be achieved.
There is then the question of how variations in provision can be explained. This matters because a key concern of any localisation process is that it creates policy variations between areas which raise questions of justice between citizens (Chaney, 2013). To consider this, we draw on work suggesting the importance of partisan politics for explaining variations in social provision. In The Possibility of Politics, Stein Ringen (1987: 205; see also Korpi and Palme, 2003) argued that ‘it is possible to effect change via social reform, and second, that the harder we try the more impact we may expect to have’. However, this work, and the literature on partisanship more broadly, is almost exclusively based at the level of the nation-state, where welfare politics are primarily located. One reason why we might not expect this dynamic to be similarly apparent in relation to these localised schemes is that local government finance is determined to a considerable extent by central government, which limits the extent to which partisan ideology can translate into policy variation. Local authorities of all hues are operating within a restrictive financial climate, forcing them to make difficult choices in terms of which services to prioritise and which to cut. Given this, two other types of variation we might expect to observe are in relation to the needs and wealth of the local authorities in question: that is, we expect that local authorities with greater demographic challenges and with greater competition for resources will be more likely to have responded to this new localism by imposing cuts to these three schemes.
These theoretical traditions inform our three hypotheses:
Political: Faced with equivalent problems in terms of the need to square service and scheme provision within a given, and constrained, resource envelope, authorities of different political hues will make distinct – and predictable – choices. In particular, this hypothesis predicts that Conservative-led councils will be more likely to retrench local welfare schemes than Labour-led administrations.
Demographic: LAs with higher age-dependency ratios face greater financial challenges and will cut back on local social security provision for working-age residents to a greater extent than authorities whose residents are predominantly younger. This is in part because, faced with greater demand in areas of statutory responsibility (e.g. adult and child social care), local authorities will have little choice but to make cuts on discretionary spending items such as the local social security schemes.
Economic: Resource competition is greater in more deprived authorities and therefore increases the pressure to retrench local social security provision in these authorities. This can be because of lower capacity to raise revenue (e.g. via Council Tax or Business Rates, half of which are retained by local authorities) as well as greater demand on locally provided services and schemes, including the local social security schemes.
It is to the three payments that make up this local social security provision that we now turn.
The three local welfare schemes and what became of them
The first of the three policies we examine are Local Welfare Assistance schemes, which replaced some of the discretionary elements of the Social Fund following the Welfare Reform Act of 2012. The Social Fund was made up of two components – a regulated scheme and a discretionary scheme. The regulated scheme provided ‘maternity, funeral, cold weather and winter fuel payments for people who satisfy certain qualifying conditions’ (Department for Work and Pensions, 2018: 6). The discretionary scheme, as its name implied, provided no right to an award, even where qualifying conditions were met, and budgets were cash limited. Grover (2012: 354) notes that ‘the discretionary elements of the Social Fund were introduced as part of the Social Security Act 1986 as a means of relieving the “exceptional expenses” of the poorest social assistance recipients’. These payments acted as a ‘scheme of last resort’ for those unable to meet unexpected short-term expenditures – in Royston’s (2017: 240) words ‘to give aid in a crisis’.
From April 2013, upper tier local authorities were given the responsibility of designing and administering replacement schemes. Central government would provide local authorities with funding for these new schemes, but at national level there were major cuts in the aggregate budget from about £330m in 2010/2011 to £178m in 2013/2014 (Royston, 2017: 242), the point of their localisation. From 2015/2016, the funding for Local Welfare Assistance schemes was amalgamated into the Revenue Support Grant received by councils (Gibbons, 2017: 5), rather than remaining a distinct budget item.
The Coalition government sought to emphasise the link between the localisation of this support and an enhanced ability to respond to local need, arguing ‘that local knowledge is best suited for the relief of “exceptional needs”’. Grover (2012: 356) continues, ‘in this sense, LWA is framed by a concern with the imperfect knowledge of central government that, as we have seen, is crucial to the localism critique of central government functions’. Inevitably, the capacity of them to respond to local need is bound up with the financial allocations to them. Even if locally designed schemes were more efficient in terms of targeting a given set of resources at areas of greater local need, significant cuts to the budgets of such schemes cannot but reduce their potential to do so (see also Hick and Lanau, 2019).
Importantly, there were no requirements placed on upper tier authorities in terms of the design of their Local Welfare Assistance schemes, even in the most minimal of senses in terms of requiring that such schemes must remain open. Moreover, there were no requirements placed on authorities to report to central government on the design of, or the financial allocations to, the schemes that they implemented (Ayrton et al., 2019: 32). Drakeford and Davidson (2013) argue the devolution of LWA to local authorities in England displayed ‘an overall drive to avoid any central responsibility for future policy decision-making’, resulting in high levels of local authority discretion in terms of the policies that they would pursue.
The second schemes are Council Tax Support schemes, which replaced the previous Council Tax Benefit in 2013, again as a result of the Welfare Reform Act 2012. Under the Council Tax Benefit system, means-tested social security claimants would automatically be entitled to full support for Council Tax via ‘passporting’, with claimants of other payments applying on a means-tested basis (Royston, 2017: 88). Council Tax Benefit/Support is not a cash benefit but is rather a waiver on otherwise liable Council Tax payments and is thus significant from the point of view of the council as it impacts on the revenue that councils receive.
Following the Welfare Reform Act, Council Tax Support schemes were, similarly, devolved to LAs in England with a 10% cut in their budget. The devolution of this benefit along with a cut left, in the words of Hastings and colleagues (2017: 2021) ‘“few alternatives” but to download austerity to the poor’. Unlike the Local Welfare Assistance schemes, central government mandated that pensioners be protected from any cuts within in these locally designed Council Tax Support schemes. The devolution of the scheme in conjunction with a budget cut, when combined with a requirement to protect pensioners, meant that local authorities would be pushed towards making savings from working-age residents, especially in authorities with higher proportions of pensioner households (Royston, 2017: 243).
The third schemes are Discretionary Housing Payments. These are local authority-provided payments made to recipients of Housing Benefit, or of the housing element of Universal Credit, to enable them to make rent payments. They are regular, rather than one-off, payments, though they are typically time-limited. They are of an older provenance than the 2012 Welfare Reform Act, but they have taken on a new importance in softening the edges of welfare reform (Meers, 2019), with central government allocations to local authorities determined in part by estimates of the impact of the Benefit Cap, the Bedroom Tax, and changes to Local Housing Allowance at LA level. Reflecting this newfound role, the total allocation to DHPs increased from £20m in 2010/2011 to £130m in 2013/2014 (Royston, 2017: 236–237) as local authorities were provided with funding to offset some losses in relation to welfare reform. The total central government allocation in 2018/2019 was £153m.
While DHP funding is provided by central government, local authorities are permitted to top-up their allocation and to spend a maximum of two and a half times the central government allocation. Ayrton et al. (2019: 31) suggest that ‘significant over-spending relative to the funds provided by central government is taken as evidence that a borough is forming its own view of local need’. However, authorities are not required to spend the allocations they receive. Given the importance of DHPs in preventing families from becoming homeless, Meers (2019) argues that ‘it is perhaps surprising, therefore, that not all local authorities spend the entirety of their DHP budgets; 33% of authorities spend less than 95% of their DWP allocation (Department for Work and Pensions, 2018)’.
Much of what we know about what has happened these schemes since their localisation has come from a number of think-tanks and third-sector organisations (e.g. Aitchison, 2018; Ayrton et al., 2019), who have sought to investigate the extent of provision some years after the devolution of powers to local authorities. These existing works paint a picture of substantial retrenchment in local social security provision in the period since localisation in April 2013. Adam et al. (2019) find that, by 2018–2019, 90% of English councils had made changes to their CTS scheme, which they note is up from 82% in 2013/2014. Ayrton (2016) shows that the most common change made by LAs to their Council Tax Support schemes was to introduce ‘minimum payments’ – that is, a minimum amount of Council Tax that a household would have to pay irrespective of their income. But other approaches were adopted too, including amending taper rates that govern the extent to which eligibility is withdrawn as income increases, restricting eligibility via reducing the threshold of savings people are allowed before eligibility is withdrawn, and by introducing ‘band caps’, which prevents eligibility for residents in higher Council Tax-band properties (Ayrton et al., 2019; Royston, 2017).
We also know that Local Welfare Assistance schemes have, in many places, been cut back or even closed. Gibbons (2017: 18) finds that ‘Reviewing the websites of all English upper tier local authorities as at the end of April 2017, we find that twenty six (16 percent of the total) of these have now closed their local welfare schemes’, and that ‘a further 40 local authorities have set a budget for local welfare provision which is at least 60 percent lower than their initial 2013/14 allocations from DWP, excluding administration costs’. Research by Church Action on Poverty finds that the aggregate amount spent on Local Welfare Assistance schemes has declined from £172m, provided by central government to English local authorities in 2013/2014, to an estimated £46.4m provided for by English local authorities in 2017/2018 (Aitchison, 2018: 4). Variations in Discretionary Housing Payments have not attracted the same degree of attention but are of interest for the reasons discussed above.
While we have some understanding of how these schemes have changed over time, we have less information about which kinds of authorities have cut their schemes back. An exception to this is by Adam et al. (2019), who find that, after adjusting for local authority differences, Labour-led councils were less likely to have imposed minimum Council Tax Payments than those led by the Conservative party. These findings, however, related to only one of the three schemes considered here. This article is, to the best of the author’s knowledge, the first attempt to bring together empirical data on the three schemes jointly in the period since localisation.
Data and methods
The analysis presented in this article has been facilitated by the compilation of a new dataset. It is the first time, to the author’s knowledge, that these data on the three local welfare schemes have been analysed jointly. Data in relation to the Local Welfare Assistance schemes have kindly been shared by Church Action on Poverty. In 2018, Church Action on Poverty issued FOI requests to 163 local authorities to provide information about their Local Welfare Assistance schemes (Aitchison, 2018). There are 151 upper tier and unitary authorities in England, but in a small number of cases upper tier authorities have themselves devolved responsibility and budget for LWA schemes to district councils so there are a slightly larger number of relevant authorities. Our data on Local Welfare Assistance schemes include a binary measure of whether an LWA scheme remained open in 2018/2019 for 151 of the 163 authorities who were asked to provide information on this, and a measure capturing the percentage change in the LWA budget – i.e. between the amount council’s received in 2013/2014 from central government to provide initial funding for these schemes to the financial allocations councils had set for the financial year 2018/2019 – for 129 authorities.
In relation to Council Tax Support, we draw on data which have been collected through FOI requests by the New Policy Institute and kindly shared by them. There are a number of ways that councils can change their CTS schemes. As Ayrton (2016) notes, by far the most common amendment made has been to introduce minimum payments for working-age claimants, and we include this variable here. Another way that CTS schemes can be amended is by introducing band caps, as discussed above, and we include this as a secondary measure in our sensitivity analysis. Our analysis for this scheme is based on data for 322 of the 326 lower tier or unitary councils.
In relation to Discretionary Housing Payments, data on financial returns at local authority level are provided by central government on a bi-annual basis (UK Government, n.d.). This provides us with data for six financial years between 2013/2014 and 2018/2019. In 2018/2019 £153m was allocated to local authorities for the purpose of Discretionary Housing Payments. Two-thirds of DHP allocations are itemised as partially offsetting specific aspects of welfare reform (in descending order of size, on the Bedroom Tax/Spare Room Subsidy, the Benefit Cap, LHA allowance reforms and a combination of changes in these payments). The remaining one-third is not identified as relating to welfare reform per se. Central government provide local authority-level data about the amounts spent on DHP during the financial year and the allocations from central government made to local authorities for this purpose. From this we construct a measure of the proportion of their central government DHP allocation that authorities spend, averaged over the 6 years for which we have data. We have data for 313 of the 326 lower tier or unitary councils for this measure.
In terms of independent variables, the political makeup of elected councils across England is drawn from the database compiled by Keith Edkins on this topic. 4 Our primary political variable is the party of control of elected councils in England. This is based on composition data following the local elections held on 2 May 2018, a period which thus corresponds to the most recent period of interest in terms of the policy variables.
Our demographic variable, the age dependency ratio, captures the proportion of residents of 60 years or above as a proportion of those aged between 16 and 59 years in each local authority. We rely on two economic measures. The first is a measure of Gross Value Added by local authority, data for which is available from the Office for National Statistics. GVA captures ‘value added by production activity in a region to the resident population of that region’ (ONS, n.d.). The most recent year for which we have data are 2015, which is the year analysed here. There are a small number of extreme values in these data, so we top-code the original variable at £50,000 (which affects seven cases) and express values in thousands. Our second measure is quantiles of the 2015 Index of Multiple Deprivation, which have been aggregated up from Lower Super Output Area (LSOA) to local authority (LA) level. The IMD is a multidimensional composite indicator based on the dimensions of income, employment, education, health, crime, housing, and living environment (Department for Communities and Local Government, 2015). The dataset contains both the composite measure and its component dimensions. Our primary measure of deprivation captures the average of multidimensional deprivation scores of LSOAs within a local authority. Our analysis is based on IMD deprivation deciles for the analysis of Council Tax Support and Discretionary Housing Payments; the analysis of Local Welfare Assistance is based on a smaller sample of unitary and upper tier councils; we therefore rely on IMD deprivation quintiles. In the sensitivity analysis, we also focus on the income and employment subcomponents. Finally, in one model, we include a dummy for councils in London and the South East, where cost pressures are especially acute, and for another model, we include a control for Housing Benefit receipt (in 2015), data for which have been taken from the DWP’s StatsXplore site.
The analytic strategy is to present descriptive information about what has happened to each scheme since the period of localisation and then to move towards regression models in which we examine the effect of partisan makeup of elected councils, while controlling for economic and demographic differences. The analysis relies on Ordinary Least Squares regression which is appropriate since the dependent variables are continuous and do not depart excessively from normality. We include stars indicating statistical significance in the regression tables, though of course we are dealing with a near-census of local authorities and thus the logic underpinning statistical significance does not apply. For this reason, we concentrate our attention on effect sizes in discussing the results. In the penultimate section, we summarise the key points from the sensitivity analysis that has been conducted, which includes models with additional controls, alternative operationalisation of key variables, and non-linear model specification.
Analysis
Local Welfare Assistance schemes
We begin our analysis with the Local Welfare Assistance schemes, the data for which have been collated and kindly shared by Church Action on Poverty. There are two variables of interest in relation to these schemes – a binary measure indicating whether these schemes remained open and, for authorities where the scheme remains open, how the budget has changed in the period 2013/2014 to 2018/2019. Table 1 demonstrates that, by 2018/2019, 83% of local authorities had kept their LWA schemes open, with 17% having closed their scheme.
Does the Local Welfare Scheme remain open?
Turning then to how the scheme budgets have changed over time, 5 in Table 2, we observe that the median change to the budget was a cut of more than three-quarters, and more than 8 in 10 have cut their schemes by one-half. Thus, while most schemes remain open, the budgets for these schemes have typically been severely curtailed.
Change in LWA budget between 2013/2014 and 2018/2019.
In Table 3, we explore the association between party of control and whether the Local Welfare Assistance scheme remains open. Table 3 is consistent with the partisanship hypothesis, with Labour-led councils more likely than Conservative administrations to have kept their LWA schemes open. Three-quarters of Conservative-led councils have kept their schemes open, compared with 90% of councils that are led by majority Labour-led administrations and four in five councils which were under No Overall Control or were controlled by either the Liberal Democrats or by Independents.
Does LWA scheme remain open by party of government.
When we focus on changes to the budget by party of control (see Figure 1), we observe that Conservative-led councils are somewhat more likely to have made larger cuts to their schemes. Of the five councils that have cut their LWA budgets by less than a quarter, three are led by Labour administrations (Slough, Islington, North Tyneside – the latter two increasing their budgets in this area) while two are Conservative administrations (Peterborough, Derbyshire).
Reduction in LWA budget between 2013/2014 and 2018/2018.
In Table 4, we begin to examine the effect of party of control, while adjusting for differences in the characteristics of the authorities in question – namely, in terms of the age dependency rate, the deprivation rate, and the geography of the LAs in question. This matters because the party-of-government effect is in part obscured by the other variables of interest. For example, Labour-led councils are substantially more likely to be more deprived than Conservative-led councils. They also have substantially lower age dependency rates (mean values 0.32 vs 0.47).
OLS regression on changes to LWA scheme budget.
t statistics in parentheses.
p < 0.05; **p < 0.01; ***p < 0.001.
In Table 4, we model changes to LWA budgets between 2013/2014 and 2018/2019. In Model 1, using all available cases for which we have scheme budget information, we observe that the average cut to LWA scheme budget was 4.5 percentage points (ppts) lower for Labour-led councils than those administered by the Conservatives. When we turn to the completed case analysis in Model 2, we find that the partisan effects elevate to −7 ppts in the case of Labour-led administrations and −11 ppts in the case of councils under No Overall Control, or led by the Liberal Democrats or by Independents. The 24 cases lost in moving between Models 1 and 2 are all county councils, for whom we do not have demographic or economic information. In Model 3, we see that age dependency is associated with lower LWA budget cuts, contrary to what we expect; our economic variables behave inconsistently – higher GVA is associated with lower LWA cutbacks (which we expect) but higher rates of deprivation as measured by the IMD are also associated with lower cutbacks (contrary to what we expect). The inclusion of these variables reduces the partisan effect to 5.1 ppts. In Model 4, we include a dummy for councils in London and the South-East, where LWA budget cuts are more common than elsewhere. The inclusion of a London and South East dummy lowers the partisan effect to 2.8 ppts; the age dependency rate now falls in the expected direction, with councils with higher rates of age dependency more likely to impose LWA cuts. In Model 5, we exclude five influential observations where values of Cook’s D exceed the traditional cut-off of 4/n. This improves the model fit and changes the direction of the deprivation coefficient (with more deprived authorities more likely to cut back their LWA schemes, as expected). The exclusion of these influential observations leaves us with a model that estimates that the average cut for LWA budgets in Labour-led administrations was 3.9 ppts lower than those controlled by the Conservatives.
Council Tax Support schemes
The second type of scheme we consider here are Council Tax Support (CTS) schemes, data for which have been collated and kindly shared by the New Policy Institute. We know from the existing literature that the primary way that these schemes have been amended has been via the introduction of a minimum payment (as a percentage of a resident’s council tax liability). Figure 2 displays the distribution of minimum payments by local authority. This demonstrates that by 2018/2019 fewer than 20% of councils did not impose a minimum payment. The most common option was to impose a minimum payment of 20% of the eligible payment, with a maximum payment of 50%. In Figure 5 in the Appendix we present how the cumulative distribution of minimum payments has changed over time. From this we observe that most councils imposed minimum payments at the earliest opportunity in 2013/2014 and that these have progressively increased over time.
Distribution of minimum payments (%) by local authority, 2018/2019.
We find that 81% of councils imposed minimum payments in 2018/2019, up from 70% in 2013/2014. When we look at the party of control effect (Table 5), we observe that there is a differential, with Conservative-led councils more likely than those led by Labour, the Liberal Democrats or NOC/Independents to have imposed minimum payments.
Whether councils imposed minimum payments in 2018/2019, by party of government.
If we look at the distribution of minimum payments between Labour - and Conservative-led councils (see Figure 3), we observe that this is reasonably similar: the median value for Conservative-led councils is somewhat higher than for those administered by Labour, but the 75th percentile value is lower. It is two Conservative-led councils who impose the highest minimum payments (North Lincolnshire and Kettering, at 50% and 45%, respectively).
Distribution of minimum payments for councils, by party of government.
In Table 6, we present results from a set of linear regression models where the dependent variable is the value of the minimum payment imposed on residents in relation to their Council Tax (if any), as presented in Figure 2. We observe that, relative to the reference category of Conservative-led councils, Labour councils imposed minimum payments that were 0.9 ppts lower, while the equivalent figures for those under NOC/Independent and Lib Dem councils were −2 ppts and −8 ppts, respectively.
Regression model (OLS) predicting minimum Council Tax Payment (2018/2019).
t statistics in parentheses.
p < 0.05; **p < 0.01; ***p < 0.001.
There is a negative association, contrary to our hypothesis, between age dependency rates and the minimum payment value imposed; controlling for age dependency elevates the coefficient for Labour-led administrations somewhat. In Model 3, we observe that, in line with our hypothesis, higher levels of Gross Value-Added and lower rates of multiple deprivation are associated with lower minimum payments being imposed on residents. Including the economic variables into the model substantially increases the effect size for Labour-led councils – rising to −7.4 ppts. In Model 4, we exclude 15 influential observations. This sharpens the effect the partisan effect for Labour-led councils to -8.9 ppts. We find, then, most councils have made changes to their CTS schemes by requiring minimum Council Tax payments even for low-income residents of working age. We observe a very modest partisan difference in terms of the minimum payments imposed by councils in our descriptive analysis but this partisan effect is considerably stronger when we control for characteristics of the councils themselves.
Discretionary Housing Payments
The third payments of interest are Discretionary Housing Payments, which have become a central mechanism for councils to partially offset the effects of welfare reform. Our measure here is the proportion of central government DHP allocations spent by local authorities, averaged over 6 years. The total (nominal) value of this central government allocation has fluctuated significantly between 2013/2014 and 2018/2019, from £180m in 2013/2014, falling to £125m in 2015/2016, before rising to £153m in 2018/2019, our primary year of interest.
Over the 6 years, the median proportion of DHP allocation spent was 96%. A quarter of authorities spent less than 85% of their allocation, while 1 in 20 spent less than 70% of their allocation. In contrast 1 in 10 councils spent more than 5% more than the allocation they received. Thus, while there is a significant degree of clustering around the 100% mark (see Figure 4), there is also variation around it.
Proportion of DHP allocation spent, by party of government.
In Table 7, we identify the local authorities spending the lowest and highest proportions of the DHP allocations on this purpose, alongside the party of control of these authorities. This demonstrates, at a descriptive level and for these most extreme cases only, that three of the five councils spending the lowest proportion of their DHPs were Conservative councils, while three of the five authorities spending the highest proportions of their allocations were Labour-led councils.
Councils with the five lowest and highest spending on DHPs as proportion of allocation (2013/2014–2018/2019) and party of government.
More generally, the average Conservative-led council spends 91% of the contribution it receives, whereas Labour-led councils spend 100% of this. Moving beyond averages, we display the distribution in the boxplot shown in Figure 4. What we find is that there is somewhat less variation in the behaviour of Labour-led councils than Conservative-led ones. For Labour-led councils, the interquartile range is very small indeed – just 8 ppts, with three-quarters of Labour-led councils spending in excess of 95% of their allocation.
In Table 8, we see that in Model 1, Labour-led councils spend 8.5 ppts more of their allocations than Conservative-led administrations. Adding in the demographic variable (in model 2), we find that higher age dependency ratios are associated with lower proportionate spending on DHP, as we predict. When we include the economic variables (Model 3), we find that GVA per person is associated with higher proportions of DHP allocations being spent (as we expect) while multiple deprivation exerts a very modest negative effect on the proportion of DHP allocation spent. In Model 4, we include the number of housing benefit claims per working-age adult in the local authority. This is positively associated with spending more on DHP, as we might anticipate. This leaves us with a partisan effect of 7.3 ppts for Labour-led councils when compared to those led by the Conservative Party. Finally, in Model 5, we exclude 18 influential observations. This serves to weaken the partisan effect to 4.5 ppts.
Regression (OLS) estimating the effects of party-of-government, demographics, and economics on the proportion of DHP allocation spent.
t statistics in parentheses.
p < 0.05; **p < 0.01; ***p < 0.001.
Unlike the two previous payments considered, where there have been substantial cutbacks to local-level schemes, the most common response in relation to Discretionary Housing Payments has been to allocate all, or almost all, of the awards received from central government for this purpose. As we have argued earlier, this is perhaps understandable as DHPs are often used to ensure that low-income families can meet housing-related payments, and which may therefore prevent homelessness, which may in turn result in a charge on council resources. Despite this overall tendency to make awards broadly in line with the amounts allocated, we observe that partisanship does appear to have a modest effect on the propensity to spend DHP allocations received from central government in full, even adjusting for differences in the characteristics of councils.
Sensitivity analysis
A series of tests of sensitivity have been explored. These include (1) estimating ordinal logit models so as to relax assumptions regarding the normal distribution of the main dependent variables, (2) using quantiles based on the income and employment components of the Index of Multiple Deprivation instead of the composite index, (3) replacing party of control variable with a measure reflecting Labour’s seat share in each council in 2018, (4) using party of control data for 2013, and (5) controlling for claimant counts and regional differences. In each case, we compare results to the full models presented above before influential observations have been removed and we discuss the primary differences here.
For Local Welfare Assistance schemes, we find that relying on the 2013 electoral makeup of councils lowers partisan effect for Labour-led councils to −2.3 ppts (in comparison to −2.8 ppts in the main model), while reliance on income deprivation quintiles sharpens this to −3.7 ppts. One notable departure occurs if we replace LA party-of-control with Labour’s seat share. While in the descriptive analysis, councils with less than one-quarter Labour representation are associated with the highest LWA cutbacks (this category captures most of the Conservative-led councils), once controls were added in the regression models, councils with 50.1%–75% Labour representation were associated with the greatest reductions in LWA budgets. Thus, a party-of-control variable provides partially inconsistent results to a focus on Labour’s seat share (for this scheme only), perhaps suggesting that there are important differences within Labour-led councils that are not captured by the party-of-control variable. Finally, the effect of age deprivation is negligible if a full range of regional controls are added.
In relation to Council Tax Support schemes, the partisan effects are found to be highly consistent across the sensitivity analysis we conduct. In addition, we re-run the analysis using a binary measure of whether a council has imposed a council tax band cap on eligibility. We continue to observe a partisan effect in relation to this, especially in relation to the comparison between Labour- and Conservative-led administrations.
In relation to Discretionary Housing Payments, we find that the use of 2013 party data moderates the Labour effect to 6 ppts and the income decile model to 6.6 ppts (from 7.3 ppts in the main model). Overall, then, the sensitivity analysis largely confirms our preceding findings, and, in doing so, provides confidence in the findings that we have reached.
Discussion and conclusion
The Welfare Reform Act of 2012 created a new localised system of social security payments across England. The Act abolished Council Tax Benefit and the Discretionary Social Fund, and tasked local authorities with designing local replacements to these schemes, with limited restrictions placed on them in terms of scheme design and with minimal requirements on them to report to central government about the replacement schemes that eventuated (if any). Allocations made to local authorities for Discretionary Housing Payments were significantly increased, with these payments given a new remit to mitigate the sharpest impacts of welfare reform. These new localised powers opened up the prospect of a new, localised social security terrain, but it is one that remains under-analysed empirically. In this article, we have presented analysis of a newly-constructed dataset examining provision in relation to these three payments jointly for the first time, providing empirical evidence on these important cases.
We find that there have been substantial cuts in local social security provision in the period since the devolution of payments to LAs in England, leaving a patchwork of local welfare support. In relation to Local Welfare Assistance, we find that while most schemes remained open in 2018/2019, there have been very substantial cuts made to the budgets to most schemes, with the overwhelming majority (80%) experiencing a budget cut of more than one-half. In relation to Council Tax Support schemes, we find that more than 80% of councils had imposed a minimum payment on working-age residents and that these minimum payments have ratcheted upwards over time. The proportion of DHP allocations that councils spend for this purpose have been more stable, which is perhaps understandable given their role in seeking to mitigate housing need, but the absolute amounts allocated by central government has declined. Thus, while schemes have typically not been abandoned altogether, there have been substantial cutbacks to their budgets, reflecting the cuts imposed by central government. Considerable variation now exists in terms of provision across England. The localisation of these schemes has proved to be a rather successful mechanism for implementing austerity.
We have also sought to explore whether differences in the political makeup of elected councils helps to explain variation in local welfare provision. Our analysis has drawn on political, demographic and economic hypotheses. In relation to political partisanship, we find that Conservative-led councils have been more likely to cut back on all three payments than those led by the Labour Party, even after we control for local authority differences in terms of demographic and economic profiles. These partisan effects are evident for two of the three schemes in the univariate analysis (i.e. Model 1 in Tables 4, 6, and 8; the effect in relation to Council Tax Support is very modest), and they are evident for all three schemes after we control for local authority characteristics and adjust for influential observations. Thus, within what is a very restricted climate, the evidence suggests that partisan politics at a local level continued to play a role in terms of the subsequent policies that emerged.
The demographic variable was in line with our hypothesis for two of the three payments. Higher age dependency ratios were associated with councils making larger cuts to their Local Welfare Assistance scheme budgets and spending lower proportions of their DHP awards for that purpose. Local authorities with higher age dependency rates were associated with the imposition of lower minimum payments as part of their Council Tax Support schemes, which is contrary to our hypothesis. These outcomes provide a reasonable level of support for our demographic hypothesis.
The economic hypothesis received greater support. More deprived local authorities were associated with larger cuts to the budgets of their Local Welfare Assistance schemes, imposed greater minimum payments in respect of their Council Tax Support schemes (consistent with our hypothesis) and spent a lower proportion of their DHP awards (again, consistent with our hypothesis, but observed for one of our two measures only).
There has been a substantial degree of retrenchment in the three schemes examined here since the Welfare Reform Act of 2012 and the policy landscape in relation to this provision across England is now highly variable. The cuts made to these schemes, which are often relied upon by citizens on the lowest of incomes, have been harsh but are consistent with what we might expect given the emphasis on fiscal consolidation as an over-arching policy priority of central government, and with the substantial degree of cuts made to local government funding. But despite this harsh financial climate, it has remained possible to make different choices, and we have observed how local social security provision varies in important ways by the political makeup of local government and by the demographic and economic challenges faced by local government administrations. Politics remains possible even in the harshest of policy environments.
Footnotes
Appendix
Acknowledgements
I am particularly grateful to Gavin Aitchison of Church Action on Poverty and to Peter Kenway of the New Policy Institute, who kindly shared the data they had collected on Local Welfare Assistance and Council Tax Support schemes, respectively. For comments and suggestions, I also thank participants at the survey data users’ brown bag (SURDUBB) seminar group at Cardiff University, at the Social Policy Association conference in Durham and at the People, Place and Policy conference in Sheffield, at which drafts of this paper were presented in 2019. I am also grateful for the thoughtful and constructive comments received from two reviewers and the Editor.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
