Introduction
The UK government's current strategy combines radical spending cuts with a restructuring of most areas of public provision. The immediate objective, as stated on the opening page of the 2010 Emergency Budget and repeated in 2011, is to reduce costs and hence the budgetary deficit (HM Treasury, Reference Treasury2010: 1, Reference Treasury2011: 1). The new policies go far beyond this. As well as cutting state spending, they will expand the role of non-state, especially for-profit, providers, make provision more diverse and tighten work incentives. Many commentators argue that a generous and inclusive state welfare bolsters the legitimacy of government in unequal, but democratic, capitalist societies. The implication is that cuts and restructuring on the scale currently being pursued will tell in the opposite direction. Legitimacy is about the extent to which citizens accept the government's authority. Most empirical work on legitimacy uses indirect measures to do with attitudes to and trust in government. This article considers a more direct approach using overt manifestations of dissent and rejection of authority such as riots, demonstrations and political strikes.
The article falls into four sections which discuss the relationship between state welfare and political legitimacy, describe current policy changes in the UK and their impact, explore the relationship between reform and dissent across a number of countries using data from OECD Social Expenditure and other databases and the Comparative Democracies Time-Series Dataset, and draw conclusions about the likely effect of current policy changes. The study uses cross-national data to examine the relationship between the policies now being pursued by the current UK government and disorder. Since the Economic Adjustment Programmes now being imposed on Ireland, Portugal and Greece by the IMF, EU and ECB share many features in common with UK reforms, the findings may have a broader relevance. The EU ‘roadmap’ for economic recovery currently being finalised implies much closer economic surveillance for all member states (European Commission, 2011). This suggests that it will become even more difficult for governments to increase spending in response to popular pressures. The political legitimacy of welfare states across Europe may come under even more intense pressure.
Many factors other than welfare policy may lead to disorder. These include foreign, nuclear or environmental policies, the impact of international market changes in a globalised world, multi-cultural and ethnic minority policies and police tactics. Any empirical analysis that seeks to link overt dissent to policy changes is likely to encounter considerable ‘noise’ in the data. The relationship between disorder and welfare state reform is of interest to readers in this country because we have experienced major riots and political demonstrations in the context of reforms (Guardian/LSE, 2011) and because the full effect of the Coalition programme has not yet impacted on public provision and the living standards of those on low incomes. This article seeks to demonstrate that it is possible to show that welfare state restructuring may contribute to disorder, despite these problems. The article also contributes by exploring an empirical approach to welfare state legitimacy that goes beyond that of studies based on indirect measures such as attitudes.
Legitimacy and welfare
Many scholars have argued that state welfare contributes to political legitimacy as well as to economic progress and the satisfaction of citizen needs in democratic capitalist societies. Legitimacy is to do with acceptance of the authority of government, and is central to the orderly operation of democracies (Weber, Reference Weber and Parsons1964: 382). The Marxist tradition has identified two main roles for welfare states in capitalist societies. These are to do with reconciling ‘two basic and often mutually contradictory functions: accumulation and legitimation’ (O'Connor, Reference O'Connor1973: 9). The provision of welfare directly helps capital to expand because it ensures that a skilled, healthy and adequately housed work force is available. It also makes an indirect contribution by helping to secure acceptance of the existing social order as fair, because it helps address needs that are not met by the operation of the market. Gough analyses the legitimation role of social welfare as ‘indirectly productive’ (Reference Gough1979).
Offe's work starts out from a more sociological perspective, analysing legitimation in relation to social order rather than economic production. He identifies a continuing tension in modern western societies between capitalism (based on inequality and competition) and welfare (based on respect for individuals and meeting need), and argues that neither side ‘would be prepared to abandon the welfare state’ because it ‘performs essential and indispensable functions for the accumulation of capital as well as for the economic and social well-being of the working class’ (Offe, Reference Offe, Pierson and Castles2006: 75). Other writers make similar points: Williams sees the roles of the state as ‘maintaining political harmony, stability and social control’ (Reference Williams1989: 35, compare Lister, Reference Lister2010: 61). Daly argues that ‘not only does the welfare state shore up capitalism, but it legitimates it as well’ (Reference Daly2011: 69).
Legitimacy is theoretically important because it links government activity in a democracy to social order and contributes indirectly to economic success. It is difficult to assess the relationship between particular policy stances and political legitimacy empirically. The theme of welfare state contributions to legitimacy has been explored quantitatively chiefly through work that relies on the indirect measures provided by attitude surveys (see van Oorschott, Reference van Oorschott2000; Rothstein, Reference Rothstein2005; Mau and Veghte, Reference Mau and Veghte2007; Svallfors and Taylor-Gooby, Reference Svallfors and Taylor-Gooby2007; Svallfors, Reference Svallfors2011; Taylor-Gooby and Martin, Reference Taylor-Gooby and Martin2010). The main findings are that state welfare is in general popular, but that there is a persistent bias across all countries towards stronger support for welfare that meets needs, and particularly those of deserving groups (van Oorschott, Reference van Oorschott2006). These arguments suggest that policies which conflict with these assumptions by cutting back on social spending or by harming the poor and vulnerable, run the risk of damaging the capacity of state welfare to contribute to legitimacy. However, policies which favour deserving groups, such as those that reinforce the work ethic, may not necessarily have this effect. An important theme in recent analyses of welfare policy deals with the question of how governments can retrench on social spending and retain popularity. This concern underlies much of the debate about the emergence of a ‘new politics of welfare’ (see Pierson, Reference Pierson2001). As legitimacy and social stability come under increasing pressure, these issues are of even greater importance.
Another strand in empirical work relates welfare policies more directly to the need to secure legitimacy. Piven and Cloward's pioneering study Regulating the Poor (Reference Piven and Cloward1993) analyzes poor relief as a strategy to maintain social order. The authors argue, on the basis of analyses of the US Aid to Families with Dependent Children (now Temporary Assistance for Needy Families) and related programmes, such as food stamps, that ‘expansive relief policies are designed to mute civil disorder and restrictive ones to reinforce work norms’ (1993: xv). This argument has been extremely influential, but there appears to be relatively little UK work that relates benefit spending directly to disorder. This may well result from the fact that disorder tends to take place in specific localities. UK benefit rates are determined nationally and cannot respond to local pressures. Targeted inner city, urban renewal, social exclusion and neighbourhood programmes in the UK are often understood as emerging in response to threats of social unrest (see Power, Reference Power2012: 20 for a review).
More broadly, social scientists have sought to relate overall spending by national governments and by trans-national agencies to social order, mainly outside the world of developed welfare states. IMF interventions have been shown to impact on social stability in African (Morrison et al., Reference Morrison, Lafay, Dessus, Cornia and Helleiner1994) and in developing countries more generally (Haggard et al., Reference Haggard, Lafay and Morrisson1995). Ponticelli and Voth examine the effect of austerity programmes over the last ninety years across Europe and show a strong and consistent link between a broad measure of social unrest (including assassinations and attempted revolutions) and budget cuts (2011). In this article, we seek to focus more precisely on changes in welfare state policy and examine their relationship with disorder of the kind that might be experienced within a modern welfare state.
Welfare state reform
The data under discussion here refer to the period from 1980 to 2005, before the 2007 crisis and subsequent recession. In this section, we discuss current policy developments to identify the changes whose likely effect on legitimacy is of most interest.
The policies of the 2010 Coalition have been extensively discussed (Institute for Fiscal Studies, 2011; Yeates et al., Reference Yeates, Haux, Jawad and Kilkey2011; Taylor-Gooby and Stoker, Reference Taylor-Gooby and Stoker2011; Taylor-Gooby, Reference Taylor-Gooby2012). We can identify two main policy directions: cut-backs, to reduce spending and help the government achieve its primary budgetary objective of deficit reduction, and restructuring, with objectives that are less evident. The cuts include reductions in public spending for all cash benefits, apart from first tier pensions, local government services and all capital programmes (including social housing, the NHS and education). Current spending on the NHS and the schooling component of education is maintained in cash terms but affected by the continuing programmes of cuts and efficiency savings agreed previously. These cuts are the largest since the Geddes Axe in 1921–2 and are to be achieved over the relatively short period of five years (for details see Institute for Fiscal Studies, 2011; Yeates, et al., 2011).
Cuts in taxes and cash benefits affect the top and bottom two deciles of the income distribution most severely. The impact of the (larger) cuts in spending on services in kind is controversial, but appears likely to impact most harshly on low income groups (Brewer et al., Reference Brewer, Browne and Joyce2011). Further work shows that the cut-backs will disproportionately affect families with children, particularly those on low incomes, single parents and women (Cooper, Reference Cooper2010; Crawford, Reference Crawford2010; O'Dea, Reference O'Dea2010; Fawcett Society/Women's Budget Group, 2011).
The cut-backs are accompanied by a programme of restructuring affecting virtually every area of government. This includes the NHS (NHS Bill, 2011), local government (Localism Act, 2011), education (Education Bill, 2011), child care, universities, cash benefits (Welfare Reform Bill, 2011), the employment service and employment protection, and the open public services programme (Cabinet Office, 2011). The reforms contained within this programme are not yet fully implemented but appear to have three common features:
They expand the opportunities for non-state (mainly private commercial) agencies to take a greater role in provision, in the NHS, schooling, universities, pensions and employment and local government services. This effect is likely to be amplified by greater voluntary use of the private sector, as pressures on the NHS and on care services increase and as the automatic enrolment second-tier pension is implemented;
They ensure that provision is even more work-centred, most obviously in the universal credit programme, reform of the employment services and treatment of disabled people, but also in the impact of higher education fees on choice of courses, the proposed tightening of employment protection rules and the effect of lower benefits on incentives; and
They will lead to greater diversity, both on an area basis and between client groups and providers, in health and social care, in local government standards and perhaps in benefit rates and state sector pay (Price Waterhouse Coopers, 2010; Local Government Association, 2011; House of Commons Library, 2011).
There has been some discussion of a fourth common feature, a stronger role for the third sector as part of a ‘Big Society’ programme. The cuts in local and central funding for this sector, the reluctance of government to include it in current contracting out of services, the difficulties it experiences in offering common national services and the mismatch between the areas in which third sector activity is concentrated and the areas of highest welfare need suggest this is unlikely to progress far (Community Development Foundation, 2009; Lyon and Sepulaveda, Reference Lyon and Sepulveda2009; Breeze, Reference Breeze2010; Heims et al., Reference Heims, Price, Pollock, Miller, Mohan and Shaoul2010; Mohan, Reference Mohan2011).
The programme of combining restructuring with cuts has been discussed in terms of the politics of coalition building (Gamble, Reference Gamble2011), the project of embedding the cuts so that (unlike previous episodes of cut-back in the UK) they become permanent (Hood et al., Reference Hood, Emerson and Dixon2010; Taylor-Gooby, Reference Taylor-Gooby2012) and the object of constructing a new political economy of welfare as part of a new UK growth model (Fairbairn and Irving, Reference Fairbairn and Irving2011; Gough, Reference Gough, Fairbairn and Irving2011; Hay and Wincott, Reference Hay and Wincott2012). In this article, we are more concerned with attempting to chart their likely impact on welfare state legitimacy, as assessed through social disorder.
The outcomes of the Coalition reforms are likely to be as follows:
• First, from the cut-backs: greater poverty and inequality (as cut-backs impact disproportionately on groups that are already disadvantaged);
• Secondly, from the restructuring: a stronger role for the private sector and a smaller role for the state sector (as private business enters health and social care, higher education and local government services more vigorously); stronger work incentives (from tighter benefit rules, the payment-by-results Employment Service and a harsher employment regime); and greater variation in provision by area, social group and provider (as local state and private providers exert greater autonomy in health care, education and other services).
Disorder and the impact of cuts and restructuring
In this section, the impact of the kinds of policy directions that are being pursued by the UK Coalition government on political legitimacy, understood in the most basic terms as social order and disorder, is examined. The research question considers the relationship between the kinds of welfare state reforms pursued as part of austerity packages across Europe, of which current UK policies provide a clear and well-developed example, and social disorder. The above discussion generates the hypotheses that the following five factors tend to generate social disorder and damage the capacity of the welfare state to contribute to political legitimacy:
• spending cuts;
• privatisation;
• greater poverty;
• a more work-centred benefit and labour market regime; and
• greater variation in provision by place, social group and provider.
Evidence that these factors damage social stability will suggest that the Coalition policies are very likely to do so.
There are a number of practical problems in obtaining good quality data, in conceptualising and operationalising legitimacy in terms of disorder and in measuring the five policy outcomes of interest in current policies identified above. It has not been possible to gather the data in a suitable form to include the last-mentioned in the analysis reported here.
Data sources
The dependent variable and contextual data are derived from material gathered for the Cross-National Time-Series Dataset (Banks, Reference Banks2011). This dataset has been updated and is made available by Professor Norris as part of the Democracy Time-Series Dataset. The data cover 191 nations between 1975 and 2008, and for these purposes include material on major riots, political strikes and political demonstrations gathered from the pages of the New York Times, checked to eliminate duplications and corrected from other sources. The data were assembled by one of the leading scholars in the field. They are widely regarded as providing a good quality, consistent source of cross-national information and were used as the basis for the book Driving Democracy (Norris, Reference Norris2008). The analysis reported here covers the nations for which material is available that can conveniently be used to represent the policies relevant to legitimacy and disorder identified above. This restricts the analysis to twenty-six countries with established welfare states for the period 1980–2005.Footnote 1
The dependent variable: conceptualising legitimacy as disorder
Overt manifestations of collective dissent are relatively rare in the civic life of the established democratic welfare states under consideration. The data base includes large political demonstrations and strikes directed against government policies, and major riots,Footnote 2 as well as assassinations, revolutions, purges and political coups. There are too few of the last four in the countries and time-period under consideration to permit successful analysis. In order to maximise the use of the data, a combined variable summing riots, demonstrations and strikes was generated. There are 141 instances of these events at the six time-points covered by the data on welfare state policies (1980, 1985, 1990, 1995, 2000 and 2005).
One issue is that while these events may reasonably be said to indicate strong dissatisfaction and the strikes and demonstrations are limited to large events in opposition to government policies, they do not necessarily reflect concern about social policies and welfare state restructuring. Political strikes may be directed at industrial policy and demonstrations at a range of issues (green and immigration policies for example). Riots may reflect policing practices. The coding excludes foreign-policy-related demonstrations, so that the major anti-war rallies or the demonstrations against closed borders in Eastern Europe of recent years are omitted. Other imperfections remain. The material available can only offer an exploratory broad-brush analysis. It should be noted that many of the issues which are included and are not directly related to welfare are likely to be part of the market-centred approach to public policy with which we are concerned. Examples would be public disorder or a political demonstration or strike in response to a refusal by government to support particular industries or against restrictive trade union legislation or rejection of fair trade or public health legislation.
The independent variables
Public and private social expenditure variables are derived from the OECD Social Expenditure database (OECD, 2011a). This covers thirty-four countries, reporting at five-year intervals. The twenty-six on which there was reasonably good data coverage from 1980 to 2005 and which have established welfare states are included in this analysis. This material covers public and private social spending as a percentage of GDP. It can be used to represent changes in welfare state spending and privatisation. Other OECD data sources provide material on the percentage of the population living below the 60 per cent of median income poverty line (OECD, 2011c) and employment protection (OECD, 2011b). This is used as a measure of the work-centredness of public policy. No convenient measure was identified for the greater fragmentation of provision within nations by provider, social group and area. This factor was omitted from the analysis.
The preceding section identified processes of change: spending cut-backs, and increases in poverty, privatisation and work-centredness. The independent variables used in the analysis are the five-year changes in the relevant measures, for example public or private social spending changes between 1980 and 1985, 1985 and 1990, 1990 and 1995, 1995 and 2000 and 2000 and 2005. The fact that the model relates manifestations of disorder to policy changes during the previous five years helps to address the problem of taking account of the implementation of major reforms over time.
The privatisation variable used includes total mandatory and voluntary private spending. The reform programme is likely to have effects in both areas. It will expand the role of private providers for NHS and local government services and in education. It will also impose auto-enrolment in second-tier private pensions and damage standards in a number of state services through harsh spending cuts, so that private alternatives become more attractive to those who can afford them. Analysis using a variable restricted to changes in mandatory private spending produces similar results.
The public spending variable is the OECD measure of total public social spending across all benefits and services for all population groups (old age, survivors, incapacity, health, family, active labour market programmes, unemployment, housing and other social programmes). A separate analysis was carried out using a variable adjusted to omit spending on pensions, incapacity, survivors and health care, since these areas are to some extent protected from cuts in the UK. The adjusted variable focuses attention on the groups identified in the Institute for Fiscal Studies and Women's Budget Group analyses cited earlier as most severely affected by the reforms, those of working age and particularly women and children in families. It produced similar results. This suggests that it is the scale and not the detail of public social spending that affects collective social behaviour in a broad-brush analysis of the kind reported here.
Poverty is measured as the relative poverty rate before taxes and transfers at the 60 per cent of median income poverty line. The work-centredness of national policies is more difficult to measure since it includes both quantitative aspects, such as the level and duration of entitlement to out-of-work benefits that forms the basis of Esping-Andersen and Scruggs’ decommodification measures (see Scruggs, Reference Scruggs2006) and also qualitative aspects, to do with conditionality, labour market interventions and the management of unemployed people. Decommodification measures are only available for eighteen of the twenty-six countries for the period from 1971 to 2002.
A convenient and consistent comparative measure of employment protection is available from OECD (2011b). The Employment Protection Index draws together information on the ease with which individual workers can be dismissed, the costs to the employer of collective dismissals and the level of regulation of temporary contracts. It reflects the extent to which legislation and government interventions impose flexibility on the labour market. While this does not equate to the work-centredness of benefit rules, it goes hand in hand with it, since the object of the new benefit rules is to increase labour market flexibility by ‘helping people to move into and progress in work’ (Department for Work and Pensions, 2011). The current UK reforms include a re-categorisation of contracts that have lasted between one and two years from permanent to temporary which will reduce the country's score on the index.
The analysis: stage I
The analysis proceeded in two stages, first examining the dependent variable and the way in which it related to the independent variables specified in the discussion above and then constructing a regression model to explore the relationship between the various factors in combination and social disorder.
Figure 1 shows the distribution of scores on the disorder variable by year. There is clearly considerable variation over time, with a peak in the late 1980s and early 1990s. The correlation between disorder and year remains significant at the 5 per cent level when controlled by nation, indicating that it is not the result of activity in a particular country but more generalised. Further exploration shows that riots and political strikes and demonstrations were most noticeable in East and Central European countries, associated with the end of the Cold War, in the UK, associated with the imposition of Poll Tax, and in France, Germany and elsewhere, associated with responses to the 2001–2 Eurozone and US recessions. This development is reflected in Ponticelli and Voth's analysis of the impact of austerity on a more general measure of social disorder (Ponticelli and Voth, Reference Ponticelli and Voth2011: Figure 1).
Disorder index by year
More generally, disorder may reflect cultural shifts, issues of communication between nations and overall trends not captured in welfare state reforms. The European riots and demonstrations of 1968 or the Arab spring on 2011 were plausibly influenced by events elsewhere at the time, as well as by particular developments within societies. For this reason, we include in the model the year to which the data refer. It also seems likely that the kind of overall differences in the general direction and social role of welfare summed up in the notion of welfare state regime may relate to legitimacy and dissent. All things being equal, the ‘people's home’ ideology of the social democratic ideal type may command stronger commitment and generate higher levels of legitimacy than corporatist or liberal welfare settlements. Regime type is also included in the model with nations grouped together following Esping-Andersen's framework (Reference Esping-Andersen1990), modified to take into account the identification of Mediterranean welfare state regime (Ferrera, Reference Ferrera2005). The Nordic countries (Denmark, Finland, Norway and Sweden) are categorised as Social Democratic, the European social insurance welfare states (Austria, Belgium, France, Germany and Switzerland) as corporatist, the Anglo-Saxon nations (Australia, Canada, New Zealand, United Kingdom and the United States) as Liberal and the Southern European regimes (Greece, Italy, Portugal and Spain) as Mediterranean. The former East European and the East Asian countries were not allocated to categories since the consensus among scholars on the classification of welfare systems in of these countries is weaker.
The first stage of the analysis examined the relationship between the combined measure of disorder and the variables measuring changes in private and state spending, employment protection and poverty (see Table 1). Changes in private spending and increases in poverty are both related positively, strongly and highly significantly to the kind of disorder that indicates declining legitimacy, as hypothesised. However, the correlations between employment protection and state spending and disorder are low and not significant. This suggests that in itself changes in state spending and in the level of employment protection do not contribute to disorder of the kind under discussion. This may in part be due to the weakness of the employment protection variable in measuring the relevant aspect of restructuring. However, it is suggestive evidence that greater strictness in this area may not be damaging to welfare state legitimacy. This interpretation is supported by the cross-national evidence that work-ethic is widely valued and contributes to perceptions of deservingness identified by van Oorschott (Reference van Oorschott2006) and discussed earlier.
Disorder, cuts, restructuring, regime and year (Pearson correlations)
Note: ** = significant at 1 per cent level.
Social democratic regime type relates strongly, significantly and negatively to disorder, probably reflecting widespread political commitment to social provision in these, the most generous and inclusive group of welfare states. The association between year (1990) and disorder noted in the discussion of Figure 1 is also confirmed.
The analysis: stage II
In the second phase of the analysis, a multi-variate model was constructed to examine the effect of cuts and restructuring on political legitimacy, understood as a low level of disorder. The model included the four independent variables (changes in state spending, in private spending, in employment protection index and in poverty over each of the five five-year periods), dummy variables for the years for which the dependent variable is measured and regime type according to Esping-Andersen's categorisation. The year 2000 is omitted to avoid the dummy variable trap of creating an exact linear relationship between the variables. The modelling faces data limitations since the variables are only available for a limited number of observations for relatively few states. For this reason, the data are pooled from 1980 to 2005, and the manifestations of disorder are treated as separate events in the model rather than as a time-series. The model chosen uses ordinary least squares regressions to maximise the use of the data measured at interval level.
Table 2 shows the regression model. The tolerance statistics are all above 0.52, indicating limited multi-collinearity. The R squared statistic of 0.256 falls to 0.180 when adjusted to take account of the relatively small sample size. This shows reasonable explanatory power for a model of this type. The coefficients in the model generally follow the pattern of the zero-order correlations in Table 1, confirming the assumptions about multi-collinearity. Four points should be noted: public social spending is more strongly (and negatively) correlated with disorder when the other factors are taken into account, and becomes significant at the 6 per cent level. The negative relationship between social democratic regime type and disorder persists, but becomes rather less significant. The year 1990, which is significant in correlations in Table 1, ceases to be significant in the regression model, since other factors which coincide in that year have greater explanatory power. The employment protection variable, used to measure work-centredness in welfare reform, does not relate to disorder either in uni- or multi-variate analysis. It is unclear whether this reflects the relationship between work-centred policies and political legitimacy or the relative weakness of the conceptual link between reduced employment protection and work incentive reforms. The tension between adequate benefits and commitment to the work ethic is entrenched in public attitudes (van Oorschott, Reference van Oorschott2006; Park et al., Reference Park, Clery, Curtice, Phillips and Utting2012: Figure 2.2). Individual responsibility to pursue paid work has been a strong theme in recent statements by leading politicians (Miliband, Reference Miliband2012).
Disorder, cuts, restructuring, regime and year (OLS regressions)
Notes: ** = significant at 1 per cent level.
* = significant at 5 per cent level.
+ = Significant at 6 per cent level.
The findings appear plausible and fit with Ponticelli and Voth's finding that austerity programmes are associated with greater social instability. Increased poverty is the most important factor in national policy which tends to increase social disorder. Privatisation, which reduces access to services for most people, comes next. Cuts in public spending are also associated with disorder. Social democratic welfare states are much less likely to experience overt dissent even when the variables discussed here are taken into account.
4 Discussion and conclusions
The analysis reported in this article is based on data from a period stretching back more than two decades before the current recession and round of welfare state cuts and restructuring. It suggests that, in the context of the other factors, cuts of the kind the UK government is pursuing play a real but not overwhelming role. It is the detail of the cuts and restructuring package that is of greater importance. The fact that the Coalition has chosen to implement cuts in a way that bears most heavily on the poorest groups plays the strongest role in relation to legitimacy. The aspect of restructuring that increases the role of the private sector is also significant. It is unclear whether the work-centredness of reforms has an effect, apart from its likely indirect impact in increasing poverty by denying vulnerable groups adequate benefits at a time of high unemployment. It is noteworthy that all main parties now endorse a work-centred agenda, with an attendant rhetoric that values ‘hard-working families’ over unemployed people (for example, Miliband, Reference Miliband2012; Duncan-Smith, Reference Duncan-Smith2012). A majority (54 per cent) of the mass public continues to believe that ‘unemployment benefits are too high and discourage people from working’ although the lack of jobs is demonstrated by rising unemployment, a doubling of involuntary part-time working and greater job insecurity (Office for National Statistics, 2008 and 2012, Table 3; British Social Attitudes, 2012; Taylor-Gooby, Reference Taylor-Gooby2012: 10). The dominance of a work-first ideology indicates the size of the challenge that faces progressives in building a truly inclusive politics.
This analysis is exploratory and is affected by limitations to do with the data and the operationalisation of the variables. The main weaknesses are three. First, the relatively small number of countries covered and relatively short time period for which comparable data are available give a limited number of data points for multi-variate analysis. In view of these limitations, the research is pursued using a pooled rather than multi-level time-series regression. Secondly, the dependent variable may be influenced by factors other than welfare state restructuring. This gives rise to considerable ‘noise’ in the data. Thirdly, work-centredness is multi-faceted, and the measure used reflects only one aspect of a complex process, so that the implications for this policy area are unclear. Addressing these limitations requires substantial further work, particularly in constructing more detailed and more wide-ranging historical data-sets and in undertaking cross-national qualitative work to chart the importance of cultural and other factors in social disorder. The fact that the analysis generates strong, significant and plausible results indicates that the effects being measured are sufficiently powerful to emerge, despite these imperfections.
The research extends analysis of the impact of policies such as those pursued by the current government to include the effects on more abstract issues such as the contribution of social welfare to political legitimacy, as well as immediate questions of poverty, inequality and equal access to good quality services. It experiments with a way of conceptualising and measuring legitimacy that has been little used in quantitative work on state welfare, and may supplement the use of attitudinal measures in this field. It shows that increased poverty and privatisation are likely to have real effects in damaging social order and that spending cuts probably have an influence in the context of the other changes. The policies and traditions of social democratic welfare states provide some protection against these effects.
In relation to current political issues in the UK, the most important points to emerge are two:
• The increase in poverty, resulting from the combination of benefit cuts disproportionately affecting those on the lowest incomes and the public sector cut-backs which increase unemployment, is likely to have a real effect in generating social disorder;
• The policies which expand the role of the private sector are also likely to contribute in undermining legitimacy and social stability. The cut-backs themselves have a weaker, but still real, influence. It is the way they are channelled to bear on those already on low incomes that is of more importance.
Considerable controversy surrounds the questions of whether cut-backs are necessary or desirable from an economic perspective. This analysis points out that, from the viewpoint of social order, it is not so much the fact of cut-backs in social spending as the groups affected and the detail of the restructuring of the welfare state that matter. This suggests that the Coalition might well be able to pursue its primary stated objective of eliminating the deficit without imposing cuts and privatisations in a way that is so injurious to the legitimacy of government. From the viewpoint of opposition parties, this finding strengthens the arguments for more egalitarian and inclusive social measures. From the viewpoint of abstract analysis, the implication is that Coalition is prepared to take risks in relation to legitimacy in order to secure additional objectives of weakening the bargaining power of employees and damaging the welfare state as one of the main working class gains of the past century. However, the research also indicates that work-centredness in general does not necessarily damage the contribution of social policy to political legitimacy.
Acknowledgements
I wish to thank Professor Pippa Norris at Harvard for making available the Comparative Democracies Time-Series Dataset (Reference Norris2011), and OECD for the Social Expenditure, Poverty and Employment Protection datasets.
