Worlds of Welfare and Attitudes to Redistribution: A Comparison of Eight Western Nations
Author(s): Stefan Svallfors
Source: European Sociological Review, Vol. 13, No. 3 (Dec,  1997), pp. 283-304
Published by: Oxford University Press
Stable URL: http://www.jstor.org/stable/522616
Accessed: 05/09/2008 13:05
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=oup.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
http://www.jstor.org
European Sociological Review, Vol. 13 No. 3, 283-304
283
Worlds of Welfare and Attitudes to Redistribution: A Comparison of Eight Western Nations
Stefan Svallfors
In this paper attitudes to redistribution in eight Western nations are analysed, using data from the International Social Survey Program (ISSP).The paper begins with a discussion of various 'regime types'as presented by Esping-Andersen and Castles and Mitchell, among others. Countries are then chosen to represent four 'twin pairs'of countries, approximating four 'worlds of welfare capitalism': the social democratic (Sweden/Norway), the conservative (Germany/Austria), the liberal (US/ Canada), and the radical (Australia/New Zealand). The empirical analysis assesses whether attitudes to redistribution and income differences are structured in the way suggested by the discussion of different cleavage structures in various regime types. It is concluded that while the level of attitudes regarding redistribution and income differences clearly is affected by regime type, group patterns are very similar between all the countries.
This article draws on two traditions of social research that have too often led separate lives. One is the comparative study of 'welfare state regimes': the attempts to distinguish and analyse distinct types of welfare states according to their institutional characteristics and/or distributive outcomes. The second is the comparative analysis of values, attitudes, and commitments among the populations of industrialized nations.
The studies of 'welfare state regimes' have to a large extent neglected regime characteristics in terms of the attitude structures and value commitments found among populations of different welfare regimes. Their sophisticated treatment of institutions and actual distributions of various goods has not been extended to any substantial analysis of the way in which different regimes further certain attitudes at the expense of others. The comparative studies of attitudes and values have, on the other hand, often neglected historical and institutional explanations and interpretations of attitudes or belief systems. This has left many such studies
marred by a lack of interpretation of the attitudinal patterns and developments they have registered.
This situation has been recognized by various people located in either tradition. In launching a wide-ranging research programme on social citizenship, Korpi (1980) emphasized that different types of social policy arrangements tended 'to generate very different patterns of coalition in the electorate' (Korpi, 1980: 305; see also Rosenberry, 1982). However, none of the subsequent analyses in the research programme has made any attempt to study this aspect of various social policy models (Korpi, 1989; Palme, 1990; Kangas, 1991; Wennemo, 1994; Korpi and Palme, forthcoming).
In a similar vein, Esping-Andersen (1980, 1985, 1990) has repeatedly stressed the repercussive effects of different welfare policy arrangements and regimes in constituting or diluting cleavage structures and conflict lines. Yet the empirical indicators to be found in Esping-Andersen's own work are restricted to the Scandinavian countries (Esping-Andersen, 1985: ch. 8), while comparisons of attitudes or values
© Oxford University Press 1997
284
STEFAN SVALLFORS
between different welfare regimes have been neglected.
Many comparative attitude and value researchers have, like Gallie, pointed to the importance of taking 'account of the profound institutional differences that can exist between capitalist societies at a broadly similar level of economic development' when trying to explain attitudinal patterns and differences between various nations (Gallie, 1983: 268). Yet most comparative attitude studies have been rather superficial in their treatment of institutional and historical conditions underlying the attitudinal patterns.
A thorough comparative assessment of attitudinal patterns and trends was for a long time blocked by a lack of adequate data. Researchers often had to rely on re-analysing existing national survey data, a practice which made comparisons difficult due to the different wording of survey questions (Coughlin, 1979,1980; Hibbs and Madsen, 1981). The alternative was usually to use highly specific samples of industrial workers, elite groups or other samples not representative of the population at large (Scase, 1977; Gallie, 1983; Verba era/., 1987).
The 1980s have witnessed the launching of several large-scale projects in which comparative data-sets on attitudes and values have been established. The European Values Study/World Values Study has been conducted twice, in 1981 and 1990, surveying attitudes and values across a broad range of issues and countries (Ester eta/., 1993). The International Social Survey Program (ISSP) was inaugurated in the mid-1980s. Within this research programme, annual surveys on various topics are conducted in what is now more than twenty industrialized nations (Davis and Jowell, 1989; Becker e/*/., 1990; Svallfors, 1996). The International Social Justice Project, probing attitudes to social justice in twelve Western and East European countries, was conducted in 1992 (Kluegel */*/., 1995).
Research programmes such as these have yielded a richness in comparative attitude data unknown to previous generations of scholars. The situation now may seem the complete reverse of what it was a decade ago: we are now rich on data, but qualified analyses and interpretations lag considerably behind. The more far-ranging analyses of these comparative data-sets have been more interested in establishing general trends and describing national
differences than in going into in-depth explanations and interpretations of national differences (Harding era/., 1986; Inglehart, 1990; Ester eta/., 1993; Jowell*/ a/., 1993; Kelley and Evans, 1993; Kluegel eta/., 1995; Svallfors, 1995a).
This paper attempts to link these two research fields in an analysis of how attitudes to redistribution are structured in various welfare state regimes. Data is taken from the 1992 ISSP survey on Social Inequality, the most encompassing data-set dealing with attitudes towards inequality and state redistribution to date. After presenting the major issues and arguments in the comparative study of welfare state regimes, eight countries, representing four different regime types, are selected. Attitudes among the populations of these countries are analysed in order to detect (a) the overall support for state intervention and redistribution in the various countries, (b) the range of income differences that are considered legitimate in various countries, and (c) what social cleavages are dominant in structuring attitudes to (redistribution.
Welfare State Regimes and Social Cleavages
The debate on welfare state regimes to a large extent revolves around the work of Esping-Andersen (1990). His typology is an attempt to classify contemporary Western welfare states as belonging to one of three 'worlds of welfare capitalism'. He argues that welfare states should not be classified according to any continuous measure such as social spending; they are qualitatively different configurations of institutions and distributive outcomes. Different historical actors have been instrumental in creating different welfare states.
Esping-Andersen discerns one 'liberal' welfare state regime, in which the market is the primary arena in the distribution of resources and in providing protection for the employed. State provisions are comparatively low and are provided as flat-rate benefits or in response to proven need. The second regime type is the 'conservative'one, in which state provisions are income-related and encompassing, but distributed strictly according to previous earnings. This regime type strongly emphasizes the family in various ways. Women are encouraged to
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
285
stay at home while their children are small and tax deductions and payment bonuses are used to supplement the bread-winner's income. The last regime type is the 'social democratic', in which most of the population is included on a citizenship basis. Provisions are income-related as in the conservative regime, but the floor level under which no-one is allowed to fall is set at a higher level. In this regime type, focus is on the individual rather than on the family as the basic unit of society.
As suggested by the labelling of the regime types, Esping-Andersen sees the three regime types as tightly wedded to specific political actors. He also suggests that the various regime types tend to create specific social cleavages and conflicts in the transition from industrial to post-industrial societies, which is of crucial importance for this paper (Esping-Andersen, 1990: ch 9). In the liberal regime type, class cleavages tend to dominate other possible sources of inequality and conflict. Even the impact of race and ethnicity in a country such as the United States tends to recede as class differentiation within ethnic groups increases over time. The conservative regime type creates strong conflicts between 'insiders' and 'outsiders', between groups with a good labour-market position and ensuing adequate protection from social insurance, and a growing surplus population of unemployed and others with a weak labour-market attachment or performance. The social democratic regime type tends to create a strong gender and sectoral conflict between a public sector mainly populated by women and a private sector dominated by men. On issues of redistribution, taxation, and public services, women in the public sector and men in the private are increasingly likely to take different sides.
Comments on and criticisms of Esping-Ander-sen's theses have been many and varied. Some argue that further regime types need to be added, either to include countries such as those of the 'Mediterranean Rim' (Leibfried, 1992), to distinguish between different types of 'liberal' welfare states (Castles and Mitchell, 1992,1993), or to add a distinction between 'basic security' and 'targeted' welfare states (Palme, 1990; Korpi, 1994; Korpi and Palme, 1995). Still others argue that the predominant class perspective in Esping-Andersen's work should be complemented by a gender perspective, in which issues of welfare state services and the role of the
family should be the focus (Orloff, 1993; Sainsbury, 1994)'.
The reworked typology offered by Castles and Mitchell (1992, 1993) is of particular interest from the perspective of this paper. They argue that the label 'liberal' in fact covers two sets of welfare states that are fundamentally different both in terms of the political actors behind their development and in their institutional characteristics and distributive outcomes. One is the truly liberal world, which combines low social expenditure, low benefit equality, and low levels of taxation with a weak position for labour parties and trade unions. The prime example is, of course, the United States.
The other they denote is the 'radical' world, made up of countries like Australia and New Zealand. In these countries, low taxation and low social expenditures are combined with high benefit equality. Incomes and capital taxes are rather high while consumption taxes and social security contributions have been low or non-existent. These countries also have fairly strong labour movements that have been countered by a strong and united right wing which has made it difficult to translate left electoral strength into government incumbency. In these latter countries, there has been a strong emphasis on 'primary welfare', that is, on regulating wages and other work conditions, and less on welfare state redistribution (Castles, 1985, 1988). State intervention has not been regarded as inherently suspect, as in the liberal world, but as preferably targeted towards the least well off in society.
What makes Castles and Mitchell's typology perhaps more interesting than some of the other critics' attempts is that they link the institutional characteristics of regime types to the historical and contemporary strength of various political actors. In the same vein as Esping-Andersen, they see the regime types as configurations of institutions, actors, and distributive outcomes, rather than just combinations of variable values.
It should be pointed out, however, that the whole debate on regime types suffers from some degree of confusion regarding the theoretical status of the regime concept. Are they ideal types in the Weberian sense or are they empirical categories? While the whole idea of a neat package of institutions, actors, and distributive outcomes surely looks more like an ideal type than like a  description of a complex
286
STEFAN SVALLFORS
reality, both Esping-Andersen and most of his later critics treat them rather as empirical categorizations, where the 'lack of fit' between regime type characteristics and existing welfare states becomes a troublesome issue.
None of these critics has made any serious attempt to question or to test the propositions of emerging social conflicts found in the last part of Esping-Andersen's (1990) work. The aim of the following analysis is to do exactly this, but following the reworked typology offered by Castles and Mitchell. The purpose is, first, to assess whether different welfare state regimes actually tend to promote different ways of valuing market distributions and the redistributive responsibilities of government. Should the state intervene to redress inequalities created by market processes? What income differences are considered legitimate in various nations? Second, the analysis aims to study the impact of various structural cleavages on patterns of attitudes in different national contexts.
In order to achieve this, careful thought has to be given to the selection of countries for the analysis. Simply patterning differences between four countries, one from each regime type, is less satisfactory, since differences will always occur between different countries. Instead, a plausible case must be made that countries approximating the same regime type are more similar to each other than they are to countries approximating other regime types.
Countries in the following analysis are, then, chosen to represent four 'twin pairs' of regime type countries. The quotation marks should indeed be taken seriously, since no two countries, even within the relatively narrow frame o f affluent Western countries, are in any way duplicates, but it can still be argued that some countries are certainly more similar than others in regime type characteristics.
The eight countries chosen are Sweden/Norway (social democratic/encompassing), Germany^Aus-tria (conservative/corporatist), Australia/New Zealand (radical/targeted) and Canada/United States (liberal).2 The choice of countries was guided by the availability of relevant attitudinal data, which disqualified countries like France and Switzerland that did not participate in the ISSP at this time and the Netherlands that did not conduct the 1992 survey. It was also important to avoid borderline cases
like Britain, where little agreement exists on classification.
In appendix table Al, selected indicators regarding the income distribution and redistribution, the scope and social ambitions of state intervention, and the characteristics of labour markets are displayed. The intention is not to make any comprehensive or in-depth comparisons, of which there already are several.3 It is instead to give a rough overview of some of the differences between these countries.
The indicators of income distribution show that the two social democratic countries clearly have the most equal distribution of both market and disposable incomes, followed by Germany. The outstanding position of Sweden is primarily an effect of a far-ranging redistribution via the welfare state, but it is also due to a relatively egalitarian distribution of market incomes among the active population. Australia, Canada, and the United States display a much more unequal income distribution. It is interesting to note, however, that Australia seems to share a more equal wage distribution with Germany and Sweden (Bradbury, 1993). Its unequal income distribution occurs later in the redistributive chain, affected by factors such as employment and unemployment, and redistribution by the welfare state.
The figures on government spending and social security transfers clearly separate the social democratic from the conservative and other countries. In the former, high government spending and high social security transfers are combined. The conservative welfare states spend just as much on social security as the social democratic states, but their total government outlays are lower, revealing their underdeveloped social service sector (Esping-Andersen, 1990,1993). The radical and liberal countries share low overall government spending and low social security transfers.
The labour-market regimes also display substantial differences in labour-market participation and unemployment. The encompassing welfare states combine low unemployment and high labour-market participation both among men and women.The corporatist countries have higher unemployment and a much lower rate of labour-market participation, especially among women. The Anglo-Saxon countries have unemployment levels in the same range as Germany, and occupy a middle-range position regarding labour market participation.4
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
287
The strength of political actors is indicated by the strength of unions and the political left in elections, parliaments, and governments. These indicators show the outstanding position of the political left in Sweden and Norway, but also, and perhaps less expected in terms of the regime type discussion, the historically strong position of the Austrian social democrats. The problems of the Antipodean left in translating electoral and union strength into government incumbency are clearly displayed, along with the fundamental marginalization of the political left in North America.
All these factors will serve as a background when analysing and interpreting results from the empirical analysis. Will the varying strength of the political left correspond to different attitudes to state redistribution in various countries? Will more encompassing welfare states be accompanied by more interventionist attitudes in the population? What range of income differences are considered legitimate in countries with a more equal real income distribution compared to countries with a more unequal distribution?
Data and Methods
The data on which the analysis builds comes from the International Social Survey Program (ISSP). Within this programme, an attempt has been made to create a truly comparative data-set with which to analyse attitudes and values among the populations of industrialized nations (Davis and Jowell, 1989; Becker era/., 1990).The programme was inaugurated in the mid-1980s, and now involves more than twenty industrialized countries from four continents. A wide variety of topics have been surveyed, and from 1990 onwards previous modules have been replicated, allowing comparison both across nations and through time (Svallfors, 1996). Data for the analysis in this paper come from the 1992 replication of the module on Social Inequality, which was fielded in 17 countries including the eight discussed above.5 This module includes both questions about the responsibility of government in redistributing resources and life chances, and questions about legitimate income differences between various occupations, which are displayed in the next section.
While comparative attitude research is potentially very fruitful, it is also fraught with difficulties which may make results and interpretations fragile. The most important problem is probably how to establish cross-national validity of indicators. Attitudes are by their very nature context-dependent, which is why we want to compare them across nations in the first place, and this is of course not a problem in itself. A problem arises if we find that it is not values and attitudes that vary across nations, but the meaning and connotations of various concepts.6 There is an immanent danger of creating research design artefacts instead of comparing and explaining substantial findings.
This problem has been dealt with as far as possible within the ISSP. The questionnaire design is truly a cross-national exercise, involving drafting groups comprised of people from several countries and requiring approval from the whole 25-nation group. The chances of finding and eliminating problems of cross-national validity have thus been uniquely great within the ISSP. However, in order to make sense of the survey findings, researchers would still be advised to use a few strategically chosen countries rather than including as many countries as possible in the analysis. It should be emphasized that the countries chosen in this paper all come from the relatively narrow range of Western, multi-party, welfare capitalist nations. Problems of establishing cross-national validity in these cases do not seem to be insurmountable.
A second problem concerns sampling and non-responses. Samples within the ISSP should all be of probability type, but the 1992 module nevertheless contains two data-sets (those from Australia and the US), in which the questionnaire was administered to a panel of respondents who had previously answered a national survey. This should be a minor problem, however, since the original sample was a random one and because sample characteristics do not differ substantially from the strictly random ISSP samples (ISSP, 1992).
Most non-response rates in the other countries vary in the range from 25 to 35 per cent (details can be found in ISSP, 1992). The German case, however, is problematic in that its non-response rate is 49 per cent. The 1992 survey in Austria reports no response rates, but the survey of which the Austrian ISSP92 is a part reports a response rate of 75 per cent. The
288
STEFAN SVALLFORS
Austrian data also lack the questions on legitimate income differences, so data on these variables from the 1987 ISSP survey has been used for Austria instead. The deficiencies of the German and Austrian data should be kept in mind when assessing the findings for the 'corporatist/conservative' welfare regime.
Readers should therefore be warned not to take absolute numbers literally, especially gross percentages on single items. It appears, however, as if the structure of non-responses looks fairly similar in all countries, with a slight over-representation of higher occupational strata (ISSP, 1992), so the problems of sampling and non-responses should not create any insurmountable problems for comparison. Any bias due to non-responses should work in a fairly similar way across nations, probably, foreshadowing conclusions from later sections, leading to a slight bias in an anti-egalitarian direction.
Caution is also motivated by the difficulties in creating truly identical class categorizations. In the ISSP, a variety of different occupational codings are used, which have been transformed into a common class scheme described in Appendix 2 and in the text below. Obviously, there will always be minor discrepancies between class codings in different countries.
What the added data difficulties point to is not that comparisons are impossible, but that caution is required in both survey design and in interpretation of the results. Interest should be focused not on small differences in gross percentages on single items, but on larger patterns, on how various groups
differ in different countries, using compounded indices rather than single items. It is also important to establish carefully whether different items really correlate in the same way across national contexts (Küchler, 1987; Scheuch, 1989).
The analysis that follows starts by displaying distributions on the various items and indicators used, and proceeds via the construction of attitudinal indices, to the use of multiple regression analysis to studygroup differences in attitudes. Appendices display additional information on indices and variables.
Attitudes to Redistribution: An Overview
Let us first take a look at the views on government responsibility for correcting market outcomes. In Table 1 the three items on this topic from the 1992 ISSP survey are displayed along with the percentage of those answering that they agreed with the propositions in each country.7
As is clear from the table, there is neither a perfect fit of the welfare regime model nor a completely random distribution. The first item, on government responsibility for income redistribution, shows populations in the conservative regime being most in favour of such a responsibility, followed by the social democratic regimes, with the citizens of the radical and liberal world mixed in the bottom positions. For the second item, responsibilities for full employment, the fault line runs between the social democratic and conservative countries on one hand,
Table 1. A ttitudes to redistribution in eight nations. Percentage agreeing with certain propositions
S we          Nor          Ger          Aut          Aus
It is the responsibility of the government to reduce the differences between people with high incomes and those with low incomes
The government should provide a job for everyone who wants one
The government should provide everyone with a guaranteed basic income
53.7
74.1
45.5
60.0          65.5          69.5
42.6
78.3          66.3          72.1
78.4          58.1          51.2
39.4
50.9
NZ
53.1
49.1
60.5
Can          USA
47.9          38.3
40.1          47.1
48.6          34.2
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
289
Table 2. Index distributions
Swe          Nor          Ger
Government index
Mean                                     4.00          4.71           4.21
Standard deviation                  1.83          1.66          1.96
Cronbach's alpha                     0.63          0.64          0.70
Source: ISSP92
and the radical and the liberal countries on the other. The third item, a guaranteed minimum income, shows the most surprising pattern, with Norway in a division of its own, the United States in a secure bottom position, and all the others mixed up in the middle. The low endorsement of the proposition in Sweden is perhaps the most surprising.8
We should let interpretation of the figures wait until after the multivariate analysis. For this purpose, an additive index was constructed from the three items, dividing between 'strongly agree' and 'agree' (2); 'neither agree nor disagree' (1); and 'disagree' and 'strongly disagree' (0). These items were summed, creating an index which may vary between 0 and 6, where 0 means disagreeing with all three propositions and thereby endorsing a clear-cut anti-interventionist stand and 6 means a strong interventionist standpoint. Index distributions and measures of reliability (Cronbach's alpha) are displayed in Table 2, and show the surprisingly low index values for Sweden. In other respects there is a good fit between welfare regimes and redistributive
Table 3.  Legitimate income differences in eight countries
	Swe	Nor
Unskilled factory worker	100	100
Farm worker	114	124
Skilled factory worker	121	123
Owner of a small shop	129	147
Doctor (GP)	195	207
Cabinet minister	226	199
Chairman of a large national company	239	228
Income distribution index	1.96	1.82
(stddev.)	(0.80)	(0.57)
N	742	1320
Aut          Aus          NZ          Can          USA
4.29          3.26          3.73          3.24          2.86
1.71           2.10          2.01           2.18          2.29
0.52          0.70          0.63          0.73          0.79
attitudes, running from the social democratic regime, through the conservative and radical worlds, to the liberal world.
Measures of reliability for the index are for the most part acceptable, with the exception of Austria, where the Cronbach's alpha measure is lower than would be preferred.9 The low reliability of the index is yet another reason to regard the results for Austria with some caution in the following analysis.
Moving from attitudes to government intervention to legitimate income differences, the results in Table 3 show a slightly different picture from those found in Tables 1 and 2. The table shows the legitimate incomes ('how  much  do you  think a-------
should earn') for a number of occupations in relation to the legitimate income for an unskilled factory worker.10
The social democratic countries now stand out as by far the most egalitarian. Australia and New Zealand also display a certain level of egalitarianism, while the conservative and liberal countries, with the partial exception of Canada, display a far more
Ger	Auta	Aus	NZ	Can	USA
100	100	100	100	100	100
117	116	113	116	113	135
146	155	129	141	154	186
241	205	194	199	221	291
384	438	326	351	420	614
446	604	317	352	312	500
711	615	480	419	512	1114
4.32	4.53	3.27	3.13	3.39	5.46
(3.16)	(2.57)	(1.57)	(1.38)	(1.62)	(5.25)
1897	840	1915	1058	831	1080
Noře: unskilled workcr=100. Sources: ISSP87; ISSP92; a1987.
290
STEFAN SVALLFORS
inegalitarian pattern. The differences between various populations in the Western world are staggering. While Swedes and Norwegians think that a chairman should earn less than 2.5 times as much as an unskilled worker, US citizens think that s/he should earn more than 11 times as much.
The support for welfare-state intervention in the conservative welfare states, as displayed in Tables 1 and 2, does not extend to egalitarianism in views about income distribution. Germans and Austrians believe in an encompassing but highly stratified redistributive order. In this respect, their attitudes are the reverse of the Antipodean inhabitants of the radical world of welfare. They have a much more circumscribed view of government redistribution, but clearly more egalitarian views on income distribution than those found in the conservative world of welfare.
As an indication of what constitutes legitimate income differences, we use the ratio of the three top and the three bottom occupations in every country. This measure is used in the next section in order to compare various groups in each country and produces more reliable estimates than simply using the ratio of the top and bottom occupations. In other respects, results when doing the latter are mainly the same as those reported in the following tables.11
Class, Gender, and Sector: Comparing Groups
Why and how should we expect social cleavages to vary across welfare regimes in their impact on attitudes to redistribution? This question may be separated into two: why should we expect different groups to differ in their views on redistribution, and why should we expect such differences to vary between welfare regimes?
Answers to the first question have often focused on the highly unequal distribution of risks and resources in capitalist societies, specifically embodied in class differences. Resources, such as money or qualifications and credentials, as well as risks of unemployment, sickness, and poverty are systematically tied to positions in the labour market, and thus constitute links between class positions and welfare policies. Weaker positions in labour markets would,   ceteris paribus,   imply  greater   reliance   on
welfare policies. As Matheson points out, contemporary class conflicts are most often fought on the terrain of the welfare state, involving 'attitudes and behaviour towards state intervention in economic and social life with the effect of overriding the distribution of resources via market mechanisms, especially the labour market' (Matheson, 1993: 57).
Many authors have argued that the privileged role of class and 'class-related' cleavages around welfare policies are complemented or even superseded by other structural cleavages in contemporary capitalism. Some have argued that gender is coming to the fore as an important cleavage when it comes to support for the welfare state. This is due, first, to the fact that women are more dependent on the welfare state, both as employees, as family members relieved of heavy and unrewarded care work, and as recipients of benefits from the state. Women often have a more precarious market position than men, leaving them either dependent on a male breadwinner or as more dependent on the state than men are. In many ways, the latter dependency may be regarded as more desirable (Hemes, 1987#, 1987/?; Borchorst and Sum, 1987). Second, the specific experiences of women may make them more inclined to embrace a 'rationality of caring' in which concern, consideration, and devotion to others are more prominent (Prokop, 1976; Waerness, 1987). The institutionalization of caring services bring into the public realm what was previously a private matter, thus transforming a 'moral economy of domesticity' into support for state welfare (Piven, 1985).
Issues of gender are tightly wedded to the question of public and private sector location. Divisions between public and private sector employees can, just like gender divisions, be separated into those emanating from the self-interest of those employed by the public sector and those emerging from specific socialization experiences in the public sector compared with the private one. Those employed by the public sector have an obvious interest in guarding their employment, wages, and working conditions (Dunleavy, 1980; Zetterberg, 1985), at the same time as their working conditions may create bonds of sympathy and solidarity with fellow public-sector employees and their clients, patients, and other 'welfare dependants' (Lafferty, 1988). According to many commentators, this suggests   possibilities   for   'paternalistic' alliances
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
291
between welfare clients and higher-level administrators in the public sector (Kolberg and Pettersen, 1981; Zetterberg, 1985; Dunleavy, 1986; Cloward and Piven, 1986; Joppke, 1987).
There are of course other social cleavages than those discussed above that may have an impact on welfare-state attitudes. What is specific about gender and sector divisions is that they are to a large extent institutionalized in the welfare state and that conflicts are fought above all on the terrain of the welfare state. Specific traits of female care responsibilities and the gendered labour-market segmentation, and the self-interest and special socialization among public-sector employees, and the livelihood of those excluded from the labour market, are all tightly wedded to the scope and organization of welfare policies. Any changes in the distribution of resources between private and public sector, or re-organization of welfare policies, are therefore likely to have important effects not only for class cleavages, but also for relations between men and women, between those employed by the private sector and public-sector employees, and between the employed and those outside the paid labour force.
What has most often been conspicuously lacking in the debate on 'new'social cleavages is a comparative perspective. Specific national experiences have often been taken as general tendencies (Svallfors, \99Sk 70). Why, then, should we expect group differences such as these to vary across welfare regimes? Although this is the least developed part of the discussion on welfare regimes, some hints can be found in the literature.
The importance of applying a comparative perspective lies in the opportunity it offers to map the influence of national institutions and historical traditions in constituting or diluting various conflict lines. The identities and interests of social actors are not pre-determined from their structural positions. They are created in a process where the institutional framework within which people act, and the historical traditions through which events and processes are interpreted, have a decisive impact. The weight of exposure to different institutional regimes creates diverging world-views even between people in similar structural locations, something that is implied but rarely empirically illuminated in the debate on welfare-state regimes.
As noted above, Esping-Andersen (1990) maintains that in the liberal regime, class will emerge as the most important cleavage, the conservative regime will nurture an 'insider—outsider' cleavage, and the social democratic regime will evolve towards a sector or gender conflict. According to Castles and Mitchell (1992, 1993), we should also expect their fourth regime, the 'radical', to display mainly a class pattern. Taylor-Gooby (1991) suggests that in the social democratic regime, gender conflicts will be subsumed under private vs. public sector conflicts, while the 'insider-outsider' conflict in the conservative regime will emerge mainly as a gender conflict. In the liberal regime, both gender and sector conflicts will be subsumed under class conflicts in the labour market.
Some previous attempts to test the salience of various structural cleavages in different regimes have both offered support for these assumptions, and also questioned them in crucial respects. Coughlin (1979, 1980), using data from the 1960s and early 1970s, found substantial class differences in Sweden, the United Kingdom, and the US, but hardly any group differences at all in Germany in attitudes towards welfare policies. Even if his was not an attempt to test propositions from the literature on welfare regimes, it appears to support suggestions about minor class differences in attitudes in the conservative welfare regime compared to the liberal and social democratic ones.
These findings were supported by Svallfors (1993) in an explicit attempt to test the impact of various social cleavages on attitudes to redistribution in liberal, conservative, and social democratic regimes. Using data from the 1987 ISSP on Social Inequality, Svallfors finds that Sweden is clearly dominated by class cleavages, even if gender differences also have some impact. Britain displays the most fragmented conflict pattern, where both class, sector, and gender cleavages are clear, while West Germany displays very minor attitudinal differences between various groups.
Further support for this suggestion is found in Matheson (1993). Using data from the ISSP 1990 module on The Role of Government, Matheson finds substantial differences in attitudes regarding welfare-state intervention in Germany between 'insiders' and 'outsiders' such as part-time workers, and  unemployed and  retired people, while class
292
STEFAN SVALLFORS
differences are clearly minor. In the United States and Norway the pattern is reversed, so that class, and a 'class-related' factor such as education, are clearly dominant in structuring attitudes, while differences along non-class cleavages are small. Australia, representing a second 'liberal' nation in Mathesons analysis, displays both class and non-class attitudinal differences to a considerable extent (Matheson, 1993: 1605). While the findings on both Germany and United States give considerable support to the theses presented by Esping-Andersen, the findings for Norway seem to point in the same direction as Svallfors (1993) in emphasizing the dominant role played by class differences in the social democratic regime.
Obviously, both class and non-class variables need a thorough coding in order to be cross-nationally comparable. What is perhaps the most crucial variable in comparing social cleavages, class, is at the same time the most problematic to make strictly comparable across nations. In the selected countries, five different occupational codings are used.12 The strategy used for making these classifications comparable is to recode them into a six-class version of the class schema devised by Goldthorpe and colleagues and used in a multitude of empirical studies of social mobility and political sociology. The logic behind this schema is to distinguish classes according to the work and market situation that various occupations entail (Erikson and Goldthorpe, 1992: 28-47). It has been proved to have high internal validity in terms of actually discriminating between occupations with different work and market situations (Evans, 1992). It has also been useful as an empirical predictor of voting and various other attitudes (Heath era/., 1991; Marshall era/., 1988; Baxter */*/., 1991).
The class schema exists in versions of different detail. The version used here distinguishes between unskilled workers, skilled workers, routine non-manual employees, service class II (lower-level controllers and administrators), service class I (higher-level controllers and administrators), and the self-employed. This is the most detailed version that can be created for all eight countries using the information in the ISSP92 datafile. When reclassifying the various occupational codings into the common schema, the reclassifications made by Ganzeboom and   his   colleagues   proved   immensely   helpful
(Ganzeboom and Treiman, 1994). The recodings used in this paper are provided in Appendix 2.13
Apart from class codings, a few variables are used in order to capture the differences across various non-class cleavages. The analysis compares men with women and private-sector employees with public-sector employees.14 Additionally, four categories of labour-market status are constructed in order to compare attitudes across the insider—outsider cleavage: those currently in the labour market are compared with the unemployed, with the retired, and with other groups currently outside the labour force (such as housewives and students).15
In appendix Table A2, group means are displayed for all countries, both indices and all available variables. A number of findings may be obtained from these data before proceeding to multivariate analysis. The first is that patterns between groups are strikingly similar across all nations. In spite of the problems of creating a truly comparable class categorization, differences in attitudes between classes are quite similar across all nations and indices. Higher-level non-manuals are substantively more sceptical towards government redistribution and more inclined to accept large differences in income than workers are. Class differences are statistically significant (at the 0.05 level) in all the countries studied.
The second is that public- versus private-sector employment does not seem to constitute a particularly important fault line. Differences are seldom statistically significant, and attitudinal differences point in different directions on different indices, sometimes showing public-sector employees to be more in favour of redistribution and small income differences than private-sector employees, sometimes the opposite. However, some support is found for the hypothesis about the sector cleavage being most important in the social democratic regime. It is only in Norway and Sweden that sector appears to have any importance at all when explaining (re)distributive attitudes, but even here differences are not very substantial.16
Instead, gender clearly appears to be a more relevant factor. Women within all the countries are more in favour of government redistribution and less in favour of large differences in income than men are, and gender differences are, with few exceptions, statistically significant.17
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
293
Table 4.  Structural determinants and government index: multiple regression (OLS) unstandardi^ed regression coefficients
Norway
Germany
Australia
USA
Gender (men—0)
0.61*
0.54*
0.24*
0.61***
Skilled worker Routine non-manual Service class II Service class I Self-employed
Unskilled worker (Reference category)
0.11	0.36**	0.11	-0.17
0.21	-0.17	-0.41**	-0.78***
0.68***	_0 44***	-0.53**	— 1.27***
1 29***	_ | |3***	—1.04***	-1.66***
0.25	-0.19	-0.78***	-0.92***
0.75***	0.71**	0.37	0.35
0.27*	0.01	0.09	-0.33
0.06	0.20	0.10	0.09
Unemployed
Retired
Others not in labour force
Employed (Reference category)
Constant R2
12.4%
3.96***
5.7%
3.8%
3.28*** 9.3%
Levels of significance: ***=T-value significant at 0.001 level; **=0.01 level; *=0.0^ level.
Various categories in regard to labour-market status also differ in their views about (redistribution. The unemployed are more inclined to favour government redistribution than other groups are, and the retired accept larger income differences than the other groups.18
Since the relationship between various structural cleavages and attitudes is so similar across countries, and since relevant data are missing for some of them, four countries were selected for multivariate analysis.19 Thus, Norway, Germany, Australia, and the United States represent the four regime types in the regressions presented in Tables 4 and 5.20 Since sector did not have any substantive impact on attitudes, and since this variable was not included in the US and was put only to those currently in the labour force in the other three countries (see n. 14), it is not included in the regressions.21
As shown in Table 4, both class and gender variables have clear effects in all four countries. Men and higher-level non-manuals are clearly less supportive of government redistribution than women and workers are. One difference between countries regarding the dummy variables for class is that the self-employed in Australia and the US are clearly less supportive of redistribution than workers are, while in Norway and Germany such differences are insignificant.
In Norway and Germany, differences between the unemployed and the employed are statistically significant. Other labour-market statuses seem to have less effect. The patterns, for both class and non-class variables, are very similar across nations, although the effects are larger and the amount of variance explained is higher in Norway than in the other countries, especially Australia.
In Table 5, there are also great similarities across nations in the way attitudes are structured. Men, higher-level non-manuals, and the retired are more prone to support large income differences than women, workers, and those currently employed are. The amount of explained variance is roughly the same in all four nations. The Australian views are the most determined by the variables, in contrast to the government index.
Since group patterns in attitudes are so similar between countries, it makes sense to include all four countries in a common regression, where countries appear as separate dummy variables. Table 6 displays results of such regressions for both indices. The table iterates the findings of substantial class and gender differences, differences between the unemployed and other groups regarding government redistribution, and between the retired and the rest regarding income differences.
294
STEFAN SVALLFORS
Table 5. Structural determinants and legitimate income differences multiple regression (OLS) unstandardi^ed regression coefficients
Norway            Germany           Australia               USA
Gender (men-0)
-0.16*
- Q 94***
-0.37*
-0.59
Skilled worker Routine non-manual Service class II Service class I Self-employed
Unskilled worker (Reference category)
0.06	-0.48*	-0.14	-0.40
0.06	0.08	0.33**	0.53
0.15**	0.51*	0.39***	0.75
0.27***	0.93**	0.95***	2.23***
0.06	0.05	0.49**	0.83
0.11	0.12	-0.20	-0.61
0.14**	1 21***	0.60***	1.90***
0.13**	0.24	0.07	0.40
Unemployed
Retired
Others not in labour force
Employed (Reference category)
Constant R2
1.78*** 5.6%
4.40*** 5.3%
3.07*** 7.3%
4.80***
4.2%
Levels of significance: ***—T-value significant at 0.001 level; **=0.01 level; *=0.0s level.
Table 6. Structural determinants, welfare regimes, and attitudes to distribution: multiple regression (OLS) unstandardi^ed regression coefficients
	Government	Income
	index	difference index
Gender (men=0)	0.46***	— 0 54***
Skilled worker	0.16	-0.29*
Routine non-	-0.34***	0.24*
manual		
Service class II	-0.64***	0.42***
Service class I	-1.22***	1.04***
Self-employed	-0.54***	0.32*
Unskilled worker	(Reference category)	
Unemployed	0.64***	-0.19
Retired	0.04	0.96***
Others not in	0.07	0.22*
labour force		
Employed (Reference category)		
Norway	1.87***	-3.62***
Germany	1.20***	-1.08***
Australia	q 44***	-2.17***
US (Reference category)		
Constant	2 93***	5.26***
R2	16.2%	17.3%
Levels of significance: ***=T-vaIue significant at 0.001 level; **=0.01 level; *=0.05 level.
The table also shows the large differences between various countries, holding other factors constant. Although welfare regimes appear to have small effects in structuring the effects of other variables, they have strong effects on the level around which other group differences occur.
In summary, then, the results in this section give support to some of the earlier research on cleavages in different welfare regimes, but run counter to other previous findings. The most striking result is the great similarity we find in cleavage structures. Research which suggested that the conservative regime would be characterized by small or non-existent class differences in attitudes does not receive much support from the findings presented here.
Suggestions about private-public sector employment constituting a new important fault line also receives precious little support. In most countries attitudinal differences between private- and public-sector employees are simply non-existent. Some support is found, however, for the notion that such differences should be more prevalent in the social democratic regime type, but even in Sweden and Norway we find rather small differences between private- and public-sector employees in this respect.
The suggestions that men and women differ over (re)distributive matters clearly receive support. Such differences occur in every country, and appear to be
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
295
little affected by regime type. It should be emphasized that significant gender differences remain even after taking into account that women have different class positions and labour-market statuses compared with those of men.
Perhaps the most important finding is that class differences are clear-cut and persistent across all the countries included in the analysis. Research which found class to be the dominant cleavage in the social democratic regime receives clear support from the analysis, as do theses on the influence of class in the liberal and radical regimes. Even in the conservative regime we find clear class differences, something which will be discussed further in the concluding section.
Conclusion
How should these findings be interpreted in relation to the discussion about various regime types with which this article started? The four regime types appear as four rather clear-cut configurations regarding the aggregated levels of attitudes. The social democratic countries combine strong support for welfare-state intervention with egalitarianism regarding income differences. The conservative countries combine strong support for welfare-state intervention with inegalitarian views on income distribution. The radical countries combine low support for welfare-state intervention with fairly egalitarian views on income distribution. Lastly, the liberal world combines low support for government redistribution with inegalitarian views on income distribution. Two deviations from this general pattern are the less than wholehearted support for government redistribution in Sweden, and the relatively egalitarian views on income distribution in Canada. To sum up, however, there is indeed reason to speak of four distinct regime types in attitudes to (re)distribution.
The differences that occur even between nations at the same level of economic development and within the same framework of Western welfare capitalism are worth emphasizing. Between the citizens of the United States and those of Norway there is a considerable difference in the way inequality and redistribution are viewed and valued. Suggestions that all Western nations share a common political
culture that includes views on how distributive matters should be arranged must be discarded. The anti-redistributive values found in the United States are by now well known (McClosky and Zaller, 1984; Verba and Orren, 1985; Kluegel and Smith, 1986), but the distinctiveness of Norwegian egalitarianism is no less clear (Martinussen, 1988; Jenssen and Martinussen, 1994). It should be pointed out that Norwegians display the most egalitarian and pro-welfare state attitudes, even more so than their Nordic neighbours in Sweden, in spite of the fact that Sweden has both a historically stronger labour movement and a more egalitarian actual income distribution (Table Al).
One thing that could explain this is that Norwegian employer's seem to have been much less politically articulate than their Swedish counterparts and the Norwegian Conservative Party much less right-wing than the Swedish 'Moderate Party' (Hatland, 1992: ch 6; Svallfors, 1989: chs. 8-9). On a more general level, this points to the importance of taking into account the political articulation by various organized interests - such as political parties, trade unions, and employer federations - when trying to explain why attitudes vary across national contexts. This topic has only been touched upon in this article, due to lack of space, but it would seem worthwhile to conduct comparative studies, not only of mass attitudes, but also of strategies and articulations among organized interests.
When it comes to differences between groups, it is rather the similarities between regime types than the differences that are noteworthy. Both class and gender differences are clear-cut and persistent across various national contexts, and there is very little difference in the way in which attitudes are structured by various background factors. Some support for stronger sectoral conflicts in the social democratic regime could be found, but even here sectoral differences in attitudes were minor. Thus the various arguments suggesting that different regime types tend to create different cleavage structures in the population receives very little support. It should be emphasized that this finding is not only supported by the multivariate analyses in Tables 4-6, where only four nations could be included, but also by the bivariate analysis including all eight nations displayed in appendix Table A2.
296
STEFAN SVALLFORS
As pointed out above, this is a result that flies in the face of much existing research. Why has previous research, indicating small or even non-existent class differences in the conservative regime type, come to such different conclusions? An elaborated answer to this question will have to await further research, spanning time series and using other data-sets and issues. One suggestion, however, is that the weakness of the class variables in virtually all the previous research might explain the failure to find any substantial class differences in the conservative regime type. Perhaps the use of status scales or rough class proxies distorted the German results even more than was the case for other countries.
In any case, the results in this paper indicate that not even the German system, geared as it is towards social harmony and bridging class conflicts (Hancock, 1989; Stjerno, 1995), precludes clear class distinctions in views about redistribution. The often-alleged Nordic consensus (Andersen, 1984; Heclo and Madsen, 1987), and the assumed relative classlessness of the 'new worlds' in North America and Australia22 also receive little support from the analysis.
The arguments within feminist theory about specific attitudes being prevalent among women compared to men receives considerable support from the analysis. Even after taking the different class positions of men and women into account, gender differences are clear-cut in all regime types. Suggestions about the varying political potential of gender between various welfare regimes were not empirically substantiated. Men and women disagree everywhere and to roughly the same extent, at least on questions about redistribution.
It should be emphasized that the present study is limited to one point in time, and that it covers only a small subset of various attitudes towards (redistribution. Future studies will reveal whether, for example, the unexpectedly low endorsement of redistributive attitudes in Sweden is due to some peculiarity of the Swedish public in the spring of 1991. Other studies using ISSP data have, however, revealed a large stability in attitudes between 1985 and 1990 (Papadakis and Bean, 1994; Pettersen, 1995), so the likelihood of any large attitudinal swings in the short term must be regarded as small.
It would also have been useful to have had access to comparative data on  more detailed questions
about various aspects of welfare policies. Previous analyses using ISSP data have indicated interesting differences between general and specific questions about welfare policies (Huseby, 1995). They have also pointed to differences in attitudes to selective and universal welfare policies, and differences between attitudes towards the security and equality dimensions of the welfare state (Pettersen, 1995; Roller, 1995). The ISSP surveys that these analyses were built on did not, however, contain enough of the relevant countries in order to permit any full-scale test of Esping-Andersen's or Castles and Mitchell's theses. Any ultimate conclusions regarding the arguments about different cleavage structures in different welfare regimes will have to await further research.23
In the meanwhile, the results presented here deserve serious attention. What might be regarded as 'the feedback effect' of policies upon politics, or the sense in which 'policies produce politics' (Pierson, 1993: 595), might then be regarded as one in which the baseline from which attitudes are formed is clearly affected by regime type, when it comes to attitudes towards redistribution among the populations of Western nations. However, when it comes to structuring various groups' attitudes, it seems as if the dominant role played by market and production relations throughout the Western world, as well as the gendered division of labour also prevalent across regime types, is creating a fairly similar pattern in attitudes to (redistribution. Rather than different welfare states creating different social cleavages, it would seem that general mechanisms linked to class and gender are creating fairly similar attitudinal cleavages across welfare regimes.
Notes
1. Here, as elsewhere in the article,'Germany' refers to former West Germany and after reunification to the Länder which made up former West Germany. The citizens of former East Germany were exposed to a fundamentally different welfare regime, which is also reflected in the substantial differences between former East Germans and former West Germans in their attitudes to welfare policies and redistribution (Braun and Kolosi, 1994; Roller, 1994). The ISSP data used come only from what used to be West Germany.
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
297
2.     The alternative labels depend on whether regimes should be classified according to their institutional characteristics or by their dominant political traditions. The term 'liberal' could in this respect be treated either as a political label or as an institutional characteristic in which reliance is mainly on the market.
3.     For some of the more far-ranging studies, see Flora (1986), Esping-Andersen (1990), Korpi and Palme (forthcoming).
4.   During   the   1990s   labour-market   conditions   have
severely deteriorated in the social democratic regime countries, particularly in Sweden. These developments have occurred after the attitudinal surveys analysed here were conducted, so they have no bearing on the findings of this paper.
5.     The Swedish survey was conducted in 1991 as a one-off replication of the 1987 ISSP survey on Social Inequality. It does not contain all the standard ISSP variables. The other surveys were conducted between February 1992 and July 1993. Time lags as small as these do not seem to pose any large problems for comparisons, since attitudes tend to change rather slowly (cf. Papadakis and Bean, 1994, who compared the 1985 and 1990 ISSP surveys on The Role of Government).
6.     For example, anyone interested in comparing attitudes to welfare policies quickly realizes that asking about välfdrd in Sweden and 'welfare' in the United States is to pose two entirely different questions. The respondents in Sweden will tell you what they think about pensions, public health care, and the social security system. The respondents in the United States will tell you what they think about means-tested programmes targeted towards poor people, and they will in most cases also come up with a vague picture of undeserving unwed ghetto mothers or some other highly negative image (Smith, 1987).
7.     The Swedish and the Austrian data also include items about 'provide more chances for children from poor families to go to university','spend less on benefits for the poor'and 'provide a decent standard of living for the unemployed'. They do not, however, appear to have any substantial implications for the answers on the other items, since answers to the three items that were posed both 1987 and 1992 are generally very stable from the 1987 survey, where all countries posed all six terns.
8.     The large differences between Sweden and Norway on this specific item could lead one to suspect that there are problems in the translation from the ISSP master questionnaire. The Swedish expression 'Regering/ riksdag bör hálla alia med en garanterad minimiin-
komsť, points to specific institutions, the government (in the narrow sense) and parliament, as having a responsibility, while the Norwegian translation 'Myndighetena bor sorge for att alle har en garantert minstelonn' is more vague and does not point to specific 'public authorities' as having a responsibility. This is, however, also the case for the first two items, where no such large attitudinal differences occur. The Norwegian translation also asks about 'wages' (lenri) while the Swedish one asks about'income'.The combined effect of the two differences in translation might explain at least some of the surprisingly large differences between Sweden and Norway. 9. The low reliability is not caused by any single item, but correlations are lower between all three items in Austria.
10.     The Austrian data are from the 1987 ISSP module, since these items were not posed in the 1992 survey. The Swedish survey contained the same list of occupations as the 1987 survey (see n. 5), a list that was changed somewhat in 1992. The 1987 list of occupations contains'bricklayer','bank clerk','secretary', and 'bus driver' in addition to those found in the table. The 1992 list contains instead 'shop assistant in a department store', 'solicitor', 'own er-man age r of a large factory', and 'appeal court judge' (ISSP, 1987; ISSP, 1992).
11.     The only substantial difference is that when using the ratio of a chairman and an unskilled worker, no significant class differences are found in Norway. Estimates are generally less reliable, as indicated by larger standard errors in the equations, than those reported in the following tables.
12.     Norway, Germany, Canada, and Austria use the 1968 version of International Standard Classification of Occupations (ISC068), at a three-digit level. The US uses the same classification, but on a four-digit level. Australia and New Zealand use the new ISC088 classification. Sweden uses the Nordic classification of occupations, which is an adapted version of ISC068, and the Socio-Economic Classification (SEI), which is very similar to the EGP class schema used in this paper. Since the question about self-employment was not asked of spouses in all countries, the classification builds on the occupation of the respondent.
13.     The reclassifications from ISC088 and ISC068 into Erikson—Goldthorpe classes may be obtained as data-files from Harry Ganzeboom (Utrecht University), e-mail: ganzeboo@cc.ruu.nl. Slight deviations from Ganzeboom's reclassifications occur in the class schema used here, since his recodings from ISC068
298
STEFAN SVALLFORS
are from the four-digit version. The deviations seem to be very minor at the level of detail aimed at here. The US classification has been transformed from ISC068 to ISC088 before recoding it into classes, on recommendation from Ganzeboom. The reclassification from the Swedish occupational classifications were made by Robert Erikson, Jan O Jonsson, and Michael Táhlin at the Swedish Institute for Social Research. They may be obtained from Robert@sofi.su.se.The classification used here is not as detailed as theirs, since some of the relevant information is missing from the Swedish ISSP data. A file copy of the recodings in Appendix 2 may be obtained from the author, Stefan.svallfors @soc.umu.se.
14.     The public- vs private-sector variables only include those currently employed, that is, the self-employed and those currently outside the labour force are excluded. In most countries, the question was only posed to those currently in the labour force. The exception is Sweden, where those currently outside the labour force are included (but not the self-employed). Since the question on labour-force participation was not posed in Sweden (see nn. 5 and 15), it is not possible to use the same demarcation line here as in the other countries. For the US the data are missing for this variable. It would have been interesting to also analyse differences between various sections of those employed in the public sector, along the lines of the analysis by Wright and Cho (1992). The available data, unfortunately, do not allow such comparisons.
15.     The relevant data for this classification is missing from the Swedish data, and is less detailed in the Austrian 1987 ISSP data-set.
16.     Sectoral differences are statistically significant only for 4 of the possible 16 indices: the government index and the income difference index in Norway, the income difference index in Sweden, and the government index in Canada.
17.      It is only on the government index in Austria and on the income difference index in the US that gender differences are not statistically significant.
18.     With the sole exception of government index in Canada, differences between categories are statistically significant.
19.     In addition to the missing and faulty data reported in nn. 5,7,9,10,14, and 15, occupation was asked for only from those currently in the labour force in Austria, Canada, and New Zealand, thereby excluding the retired, homeworkers, etc. In the other countries, last occupation (if any) was asked for from those currently not in the labour force.
20.     Regression analysis (OLS) assumes that the numeric dependent variable is normally distributed. The technique is, however, fairly robust as long as the violations are not extreme, such as is the case when the dependent variable is dichotomous. The two dependent variables used in the analyses below do not pose any big problems in that respect. OLS also assumes an addititive model, that is, that the effect of a certain independent variable is more or less equal regardless of the value on other independent variables. This assumption can be tested by creating new 'merged' variables that are included instead of the 'main' variables, for example, by creating a new 'class-gender' variable that is entered instead of separate class and gender variables. Various tests including the entering of such merged variables resulted in minuscule improvements compared to results presented in Tables 4-6, which means that the assumption of addititivity is reasonable (the SPSS runs are available on request from author).
21.     As the sector cleavage had some impact in Norway, separate regressions were run for Norway where the sector variable was included. The effects were small and not significant (government index) or barely significant (income difference index) at the 0.05 levd. Other estimates were virtually identical to the ones reported in Tables 4—5.
22.    See the work cited in Vanneman (1980: 769-770). Vanneman's own conclusion, like that of this paper, refutes such arguments, as he finds virtually no difference in class identification between Britain and the US.
23.     The 1996 ISSP module Role of Government III will offer ample opportunities for comparing welfare regimes using a broader set of countries and re-distributive attitudes.
Acknowledgements
The research on which the paper is based has been financially supported by the Bank of Sweden Tercentenary Foundation, the Swedish Council for Social Research, and the Swedish Council for Research in the Social Sciences and Humanities. A previous version of the article was presented at the Second European Conference for Sociology, Budapest, 1 September 1995, at the department of political science in Goteborg, 11 September 1995, and at the workshop on welfare state research and political sociology, Sánga Säby, 4 October 1995. The author gratefully acknowledges the data and information provided by Eero Carroll, Robert Erikson, Harry Ganzeboom, and Rolf Uher. Thanks to Clive Bean, Eero Carroll, Jonas Edlund,
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
299
Phil Gendall, Knud Knudsen, Gordon Marshall, George Matheson, Lennart Nilsson, Tapio Salonen, Tom W Smith, Katarina Stenberg, Steinar Stjerno, Eva Sundström, Rune Áberg, Torsten Österman, and anonymous referees for helpful comments, and to Charlott Nyman for her excellent editing.
References
Andersen, B. R. (1984) Rationality and irrationality of the
Nordic welfare state. Daedalus 113, 109-139. Baxter, J., Emmison, M. Western, J. (1991) (eds.) Class
Analysis   and   Contemporary  Australia.   Macmillan,
Melbourne. Becker, J. W., Davies, P. E., and Mohler, P. P. (1990)
(eds.) A ttitudes to Inequality and the Role of Government.
Sociaal en Cultureel Planbureau, Rijswiik. Borchorst,   A.   and   Sum,   B.   (1987)   Women   and   the
advanced welfare state: a new kind of patriarchal
power? In Showstack Sassoon, A. (ed.) Women and the
State. Hutchinson, London. Bradbury, B. (1993) Male wage inequality before and after
tax: a six country comparison. Discussion Paper no.
42, Social Policy Research Centre, Sydney. Braun,   M.   and   Kolosi,   T.   (1994)   Wandel   der   Einstellungen zu sozialer Ungleichheit in Deutschlanda
und Ungarn. In Braun, M. and Mohler, P. P. (eds.)
Blickpunkt Gesellschaft, iii. Einstellungen und Verhalten
der Bundesbürger. Westdeutscher Verlag, Opladen. Castles, F. G. (1985) The Working C lass and Weif are. Allen
& Unwin, Sydney. Castles, F. G. (1988) Australian Public Policy and Economic
Vulnerability. Allen & Unwin, Sydney. Castles, F. G. and Mitchell, D. (1992) Identifying welfare
state regimes: the links between politics, instruments
and outcomes. Governance•, 5, 1-26. Castles, F. G. and Mitchell, D. (1993) Worlds of welfare
and families of nations. In Castles, F. G. (ed.) Families
of  Nations:   Patterns   of  Public   Policy   in   Western
Democracies. Dartmouth, Aldershot. Clo ward, R. A. and Fox Pi ven, F. (1986) The welfare state
in an age of industrial decline. Smith College Studies in
Social Work, 56, 132-155. Coughlin, R. M. (1979) Social policy and ideology: public
opinion   in   eight   rich  nations.   Comparative Social
Research, 2, 3-40. Coughlin, R. M. (1980) Ideology, Public Opinion and Welfare
Policy. Institute of International Studies, Berkeley,
Calif. Davis, J. A. and Jowell, R. (1989) Measuring national differences: an introduction to the International Social
Survey Programme *(ISSP). In Jowell, R., Wither-
spoon, S., and Brook, L. (eds.) British Social A ttitudes: SpeciallnternationalReport. Gower, Aldershot.
Dunleavy, P. (1980) The political implications of sectional cleavages and the growth of state employment. Political Studies, 28, 364^383, 527-549.
Dunleavy, P. (1986) The growth of sectional cleavages and stabilization of state expenditures. Society and Space, 4, 129-144.
Erikson, R. and Goldthorpe, J. H. (1992) The Constant Flux: A Study of Class Mobility in Industrial Societies. Clarendon Press, Oxford.
Esping-Andersen, G. (1980) Social Class, Social Democracy and State Policy: Party Policy and party Decomposition in DenmarkandSweden. New Social Science Monographs, Copenhagen.
Esping-Andersen, G. (1985) Politics against Markets: The Social Democratic Road to Power. Princeton University Press, Princeton, NJ.
Esping-Andersen, G. (1990) Three Worlds of Welfare Capitalism. Polity Press, Cambridge.
Esping-Andersen, G. (1993) (ed.) Changing Classes: Stratification and Mobility in Post-Industrial Societies. Sage, London.
Ester, P., Halman, L. and de Moor, R. (1993) (eds.) The
Individualizing Society:   Value  Change in Europe and
North America. Tilburg University Press, Tilburg.
" Evans, G. (1992) Testing the validity of the Goldthorpe
class schema. European Sociological Review, 8, 211-232.
Flora, P. (1986) (ed.) Growth to Limits: The Western European Welfare States since World War IL De Gruyter, Berlin.
Gallie, D. (1983) Social Inequality and Class Radicalism in France and Britain. Cambridge University Press, Cambridge.
Ganzeboom, H. B G and Treiman, D. J. (1994) Internationally comparable measures of occupational status for the 1988 International Standard Classification of Occupations. Department of Sociology, Utrecht University (mimeo).
Hancock, M. D. (1989) West Germany. The Politics of Democratic Corporatism. Chatham House Publishers, Chatham, NJ.
Harding, S. and Phillips, D. with Fogarty, M. (1986) Contrasting Values in Western Europe: Unity, Diversity and Change. Macmillan, London. Hatland, A. (1992) Til den som trenger det mest? 0konomisk behovspreving i norsk sosialpolitik. (To the One Who Needs It Most? Economic Means-testing in Norwegian Social Policy). Universitetsforlaget, Oslo. Heath, A., Jowell, R., Curtice, J. Evans, G. Field, j. and Witherspoon, S. (1991) Understanding Political Change: The British Voter 1964-87. Pergamon, Oxford.
300
STEFAN SVALLFORS
Heclo, H. and Madsen, H. (1987) Policy and Politics it-Sweden:  Principled Pragmatism.   Temple. University
Press, Philadelphia. Hemes, H. M. (1987*) Welfare State and Woman Power,
Universitetsforlaget, Oslo. Hemes, H. M. (1987b) Women and the welfare state: the
transition  from  private  to  public  dependence.  In
Showstack  Sassoon, A.  (ed.)   Women and the State.
Hutchinson, London. Hibbs, D. A. and Madsen, H. J. (1981) Public reactions to
the growth of taxation and government expenditure.
World Politics, 33, 413-435. Huseby, B. (1995) Attitudes towards the size of government. In Borre, O. and Scarbrough, E. (eds.) The
Scope   of   Government.    Oxford    University    Press,
Oxford. Inglehart, R. (1990) Cultural Shift in Advanced Industrial
Society. Princeton University Press, Princeton, NJ. ISSP (1987) International Social Survey Program  1987-
Social Inequality. Datafile compiled at Zentralarchiv,
Universität zu Köln. ISSP (1992) International Social Survey Program  1992 -
Social Inequality II. Datafile compiled at Zentralarchiv,
Universität zu Köln. Jenssen, A. T. and Martinussen, W. (1994) (eds.) Velferds-
stateniväre hjerter. (The Weifare State in Our Hearts).
Ad Notam Gyldendal, Oslo. Joppke, C. (1987) The crisis of the welfare state: collective
consumption and the rise of new social actors. Berkeley
journal of Sociology, 32, 237-260. Jowell, R., Brook, L. and Dowds, L. (1993) International
Social Attitudes:   The   10th   BS A   Report.   Go wer,
Aldershot. Kangas, O. (1991) The Politics of Social Rights: Studies on the
Dimensions of Sickness Insurance in OECD countries.
Almqvist & Wiksell, Stockholm. Kelley, J. and Evans, M. D. (1993) The legitimation of
inequality:   occupational  earnings  in  nine nations —►
American journal of Sociology, 99, 75-125. Kluegel, J. R and Smith, E. R. (1986) Beliefs A bout Inequality: A mericani View of What Is and What Ought to Be.
Aldine de Gruyter, New York. Kluegel, J. R., Mason, D. S. Wegener, B. (1995) (eds.)
Social justice and Political Change. Aldine de Gruyter,
New York. Kolberg, J. E. and Pettersen, P. A. (1981) Om velferdssta-
tens politiske basis. (The political basis of the welfare
state). Tidsskriftfor samfunnsforskning, 22, 193-222. Korpi, W. (1980) Social policy and distributional conflict
in     the     capitalist     democracies:     a     preliminary
comparative framework.   West European Politics, 3,
296-316.
: Korpi, W. (1989) Power, politics and state autonomy in the development of social citizenship: social rights during sickness in eighteen OECD countries since 1930. American Sociological Review, 54, 309-328.
Korpi, W. (1994) Class, gender and power in the development of social citizenship. Swedish Institute for Social Research, Stockholm (mimeo).
Korpi, W. and Palme, J. (1995) The strategy of equality and the paradox of redistribution. Swedish Institute for Social Research, Stockholm (mimeo).
Korpi, W. and Palme, J. (forthcoming) (eds.) Contested Citizenship: A Century of Social Policy in the Western World.
Küchler, M. (1987) The utility of surveys for cross-national research. Social Science Research, 16, 229-244.
Lafferty, W. (1988) Offentlig-sektorklassen. (The public-sector class). In Bogen, H. and Lageland, O. (eds.) Offentlig eller privat f'FAFO, Oslo.
Leibfried, S. (1992) Towards a European welfare state: on integrating poverty regimes into the European Community. In Ferge, S. and Kolberg, J. E. (eds.) Social Policy in a Changing Europe. Campus/Westview, Frankfurt/Boulder, Colo.
LIS. Luxembourg Income Study. Database on Income Distribution. LIS, Luxembourg.
Marshall, G., Rose, D., Newby, H., and Vogler, C (1988) Social Class in Modern Britain. Unwin Hyman, London.
Martinussen, W. (1988) Solidaritetensgrenser. (The limits of solidarity). Universitetsforlaget, Oslo.
Matheson, G. (1993) The decommodified in a commo-dified world. Unpublished PhD thesis, University of New England, Armidale.
McClosky, H. and Zaller, J. (1984) The American Ethos: Public Attitudes Towards Capitalism and Democracy. Harvard University Press, Cambridge, Mass.
OECD (1992) Labour Force Statistics 1970-90. OECD, Paris.
Orloff, A. S. (1993) Gender and the social rights of citizenship. American Sociological Review, 58, 303-328.
Oxley, H. Maher, M., Martin, J. P., and Nicoletti, G. (1990) The Public Sector: Issues for the 1990s. Working Paper no. 90, OECD Department of Economics and Statistics, Paris.
Palme, J. (1990) Pension Rights in Welfare Capitalism: The Development of O Id-Age Pensions in 18 OECD countries 1930 to 1985. Almqvist & Wiksell, Stockholm.
Papadakis, E. and Bean, C. (1994) Mass attitudes and the welfare state: comparisons between institutional regimes 198590. Paper presented at the 6th Conference on Socio-Economics, Paris 15-17 July.
Pettersen, P. A. (1995) The welfare state: the security dimension. In Borre, O. and Scarbrough, E. (eds.)
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
301
The Scope of Government. Oxford University Press, Oxford. ^ Pierson, P. (1993) When effect becomes cause: policy feedback and political change. World Politics, 45, 595-628.
Piven, F. F. (1985) Women and the state: ideology, power, and the welfare state. In Rossi, A. S. (ed.) Gender and the Life Course. Aldine, New York.
Prokop, U. (1976) Weibliche Lebens^usammenhang. Suhr-kamp, Frankfurt.
Roller, E. (1994) Ideological basis of the market economy: attitudes toward distribution principles and the role of government in Western and Eastern Germany. European Sociological'Review, 10, 105-117.
Roller, E. (1995) The welfare state: the equality dimension. In Borre, O. and Scarbrough, E. (eds.) The Scope of Government. Oxford University Press, Oxford.
Rosenberry, S. A (1982) Social insurance, distributive criteria and the welfare backlash: a comparative analysis. British journal of Political Science, 12, 421-447.
Sainsbury, D. (1994) (ed.) Gendering Welfare States. Sage," London.
Saunders, P., Stott, H., and Hobbes, G. (1991) Income inequality in Australia and New Zealand: international comparisons and recent trends. Review of Income and Wealth, 37/1, 63-79.
Scase, R. (1977) Social Democracy in Capitalist Society. Croom Helm, London.
Scheuch, E. K. (1989) Theoretical implications of comparative survey research: why the wheel of cross-cultural methodology keeps on being reinvented. International Sociology, 4, 147-167.
SCIP Social Citizenship Indicator Program. Database on the development of social security in 18 OECD countries 1930-1990. Swedish Institute for Social Research, Stockholm.
Smith, T. W. (1987) That which we call welfare by any other name would smell sweeter. Public Opinion Quarterly, 51, 75^83.
Stjerno, S. (1995) Mellom kirke og kapital. Scandinavian University Press, Oslo.
Svallfors, S. (1989) Vem älskar välfärdsstaten? Attityder, organiser ade intr essen och svensk välfärdspolitik. Arkiv, Lund.
Svallfors, S. (1993) Policy regimes and attitudes to inequality: a comparison of three F^uropean nations. In Boje, T. P. and Olsson Hort, S. E. (ed.) Scandinavia in a New Europe. Scandinavian University Press, Oslo.
Svallfors, S. (1995^) (ed.) In the Eye of the Beholder: Opinions on Welfare and Justice in Comparative Perspective. Bank of Sweden Tercentenary Foundation, Stockholm.
Svallfors, S. (1995/7) The end of class politics? structural cleavages and attitudes to Swedish welfare policies. ActaSociologica, 38, 53-74.
Svallfors, S. (1996) National differences in national identities: an introduction to the International Social Survey Program. New Community, 22, 127-134.
Taylor-Gooby, P. (1991) Welfare state regimes and welfare citizenship. Journal of European Social Policy, 1, 93-105.
Vanneman, R. D. (1980) US and British perceptions of class. American JournalofSociology, 85, 769-790.
Verba, S. and Orren, G. R. (1985) Equality in America. The View from the Top. Harvard University Press, Cambridge, Mass.
Verba, S., Kelman, S., Orren, G. R., Miyake, I., Watanuki, J., Kabashima, I., and Ferree, G. D., Jr (1987) Elites and the Idea of Equality. Harvard University Press, Cambridge, Mass.
Waerness,K. (1987) On the rationality of caring. In Show-stack Sassoon, A. (ed.) Women and the State. Hutchinson, London.
Wennemo, 1.(1994) Sharing the Costs of Children: Studies on the Development of Family Support in the OECD countries. Almqvist & Wiksell, Stockholm.
Wright, E. O. and Cho, D. (1992) State employment, class location and ideological orientation: a comparative analysis of the United States and Sweden. Politics and Society, 20, 167-196.
Zetterberg, H. (1985) An Electorate in the Grips of the Welfare State. Swedish Institute for Opinion Polls, Stockholm.
302
STEFAN SVALLFORS
Appendix 1
Table Al. Income distributions, labour markets, and welfare policies in eight nations: Selected indicators"
Swe         Nor         Ger          Aut          Aus          NZ
Can
USA
1. Distribution of factor income c. 1985	0.42	0.34	0.43	-	0.42	-	0.39	0.44
(Gini)b	(6)	(2)	Ü)		(5)		(3)	(8)
2. Disposable income	0.20	0.22	0.24	0.20	0.28	(0.30)	0.28	0.33
(Gini)	0)	(4)	(5)	(2)	(8)		(7)	(12)
3. Factor income 25-59 years	0.30	0.25	0.32	R	0.36	—	0.35	0.39
(Gini)	(3)	0)	(5)		Ü)		(6)	(9)
4. Disposable income 25-59 years	0.18	0.19	0.23	0.18	0.26	-	0.27	0.32
(Gini)	(i)	(4)	(5)	(3)	(7)		(8)	(12)
5. Government outlays	60.6	52.9	45.1	49.4	34.3	30.9	44.2	36.5
(of GDP 1989)	0)	(4)	(10)	(6)	(15)	(17)	(H)	(14)
6. Social security transfers	21.3	19.8	18.0	20.1	9.7	12.0	12.2	11.0
(of GDP 1989)	(3)	(J)	(10)	(6)	(18)	(15)	(14)	(17)
7. Labour-market participation0	86.1	79.8	68.8	67.7	73.9	74.0	77.6	78.5
(1990)	0)	(3)	(12)	(13)	(H)	(10)	(6)	(5)
8. Female 1-m participation	83.5	72.6	54.5	55.4	61.9	63.8	69.0	69.6
(1990)	0)	(4)	(14)	(13)	(9)	(8)	(6)	(5)
9. Unemployment	2.4	3.0	6.7	3.3	7.2	4.8	9.2	7.0
(mean 1980-90)	(2)	(4)	(8)	(5)	(10)	(7)	(14)	(9)
lOtf. Union density'1	78	55	32	56	46	42	28	17
(1985)	0)	(6)	(11=)	(5)	(7)	(9)	(14=)	(17)
b. Union density	71	55	31	59	51	43	27	24
(mean 1950-85)	0)	(4)	(13)	(2)	(7)	(9)	(15)	(17)
11. Lefť share of. . . .								
a. votes	50.5	49.2	38.8	47.7	45.9	44.1	15.7	0
	0)	(2)	(11)	(3)	(5)	(8)	(16)	(18)
b. seats in parliament	51.1	50.3	39.7	47.6	44.2	47.6	8.2	0
	0)	(2)	(10)	(3)	(9)	(4)	(17)	(18)
c. seats in government	81.4	68.0	26.8	62.8	26.2	32.0	0	0
(means 1950-90)	0)	(2)	(8)	(3)	(9)	(6)	(16=)	(16=)
lRank orders are among the 12 countries in the Luxembourg Income Study (rows 1  4) or among 18 major OĽCD countries (rows 5-11). The income measure is family income corrected for family size according to the OKCD standard equivalence scale; the Gini coefficient is a measure-income inequality, with complete equality=0; factor incomc=incomc from work and capital; disposable income = factor incomc+transfcrs —taxes. 'Labour-market participation=people 16+ in the work force/total population aged 16  64. 1 Union density = union members as a share of the labour force outside farming. ■"Political left=social democratic, socialist, and communist parties. Sources: LIS (1  4); Oxlcy era/., 1990 (5,6); OKCD, 1992 (7, 8, 9); SCI P (10,11); Saunders eta/, 1991 (2: NZ).
WORLDS OF WELFARE AND ATTITUDES TO REDISTRIBUTION
303
Table A2. Index values in various groups
Swe
Nor
Ger
Aut
Aus
NZ
aData for Austria arc for 1987. Including everyone outside the labour force. =missing data. Sources: ISSP, 1987; ISSP, 1992.
Can
USA
G overwent index								
Men	3.71	4.42	3.92	4.19	3.11	3.48	2.93	2.49
Women	4.33	5.02	4.49	4.37	3.43	3.97	3.57	3.13
Unskilled worker	4.78	5.17	4.75	4.75	3.75	4.43	3.35	3.61
Skilled worker	4.43	5.03	4.50	4.44	3.63	4.01	3.62	3.29
Routine non-manual	4.15	4.98	4.25	3.93	3.26	3.70	3.62	2.94
Service class II	3.50	4.42	3.85	3.65	3.08	3.54	3.19	2.40
Service class I	2.81	3.60	3.04	3.32	2.51	2.55	1.97	1.84
Self-employed	3.62	4.68	3.99	4.36	2.77	2.94	3.22	2.57
Private sector	3.83	4.48	4.13	4.04	3.06	3.70	3.48	-
Public sector	4.10	4.73	3.95	4.11	3.25	3.68	2.99	-
Employed	-	4.52	4.04	4.04	3.08	3.51	3.17	2.76
Unemployed	-	5.44	4.93	4.80	3.85	4.83	3.23	3.47
Housewife/student	-	4.83	4.54	4.55	3.54	4.04	3.42	3.25
Retired	-	4.99	4.15	4.43	3.39	3.98	3.59	2.69
Income difference index*								
Men	2.09	1.89	4.77	4.77	3.47	3.30	3.52	5.78
Women	1.81	1.74	3.88	4.34	3.02	2.96	3.25	5.20
Unskilled worker	1.70	1.70	3.95	4.28	2.83	2.64	2.88	4.74
Skilled worker	1.86	1.72	4.12	4.71	3.01	2.82	3.30	4.57
Routine non-manual	1.81	1.77	4.13	4.36	3.21	3.05	3.41	5.31
Service class II	2.01	1.86	4.70	4.77	3.35	3.16	3.26	5.45
Service class I	2.61	2.02	5.36	6.27	4.00	3.45	3.89	7.05
Self-employed	2.11	1.83	4.37	4.14	3.43	3.30	3.40	5.46
Private sector	2.06	1.85	3.98	4.47	3.16	3.02	3.45	-
Public sector	1.86	1.74	4.13	4.56	3.23	3.04	3.17	-
Employed	-	1.80	4.14	4.32	3.23	3.11	3.34	5.25
Unemployed	-	1.65	4.01	4.43	2.75	2.78	3.15	4.14
Housewife/student	-	1.86	3.90	-	3.00	2.86	3.25	5.18
Retired	-	1.90	5.24	4.76b	3.78	3.54	4.03	6.94
Appendix 2: The Class Variable
All commands are in SPSS for Windows. Variable numbers come from ISSP 1992.
♦Germany Austria Norway Canada*
COMPUTE egpclass=vl06.
RECODE egpclass(001=3)(002=l)(011 thru
029=1)(031 thru 039=2)(040 thru 053=1)(054=2)(060
061=1)(062=2)(063=1)(064=2)(065=1)(066=2) (067=1)(068 069 071=2)(072=3)(073 thru 079=2)(081 082=1)(083 084=2)(090 thru 131=1) (132 thru 191=2)(192=1)(193 thru 199=2)(201 202 203 211=1)(210=2) (212 thru 310=2)(321 thru 342=3)(351 352=2)(359 360=3)(370=5)(380 thru 399=3) (400=2)(410=6)(421 422 431 =2)(432=3)(441 442 443=2)(451 452 490=3)(500=2)(510=6) (520=3)(531=4)(532=3)(540 thru 560=5)(570 580 581=4) (582=2)(589=5)(591 592=3)(599=5)(600 610
304
STEFAN SVALLFORS
611 612=6)(621 thru 649=5)(700=3)(711=4)(712=5) (713 thru 728=4)(729=5)(731 732=4)(733 thru 752=5) (753 thru 756=4)(759=5)(761 762=4)(770 771=5)(772 773=4)(774 775=5)(776 777 778=4)(779 thru 789=5)(791 thru 794=4)(795 796 799=5) (801=4)(802 803=5)(811=4)(812=5)(819 thru 833=4)(834=5)(835 thru 844=4)(849=5)(851 852=4) (853=5)(854 thru 861=4)(862=3)(871 thru 892=4) (893 thru 910=5)(921 thru 931=4)(939=5)(941=4) (942 943=5)(949 951 952=4)(953=5)(954 955=4) (956 957=5)(959=4)(961 969 971=5)(972 973 974=4) (979 thru 999=5).
IF (vllO eq 1 and egpclass ge 2) egpclass=6. VALUE LABELS egpclass 1 'Service ľ 2 'Service Iľ 3 'Routine n-m' 4 'skilled man' 5 'unskilled man' 6 'selfemp'.
♦Australia New Zealand USA*
*USA *. [ISC068 is first transformed to ISC088J
COMPUTE isco88=v106.
RECODE isco88 [this recoding may be obtained from the
author]
COMPUTE egpclass=ausl06 [nzl06; isco88]. RECODE egpclass (110=1)(1000 thru 1120=1)(1130 thru 1143=2)(1200 1210 1220=1)(1221=6)(1222 thru 1239=1)(1240=2)(1250 1251=1)(1252 1300 1310=2) (1311=6)(1312 thru 1319=2)(2000 thru 2131 = 1)(2132 2139=2)(2140 thru 2147=1)(2148=2)(2149 thru 2229=1)(2230 2300=2)(2310=1)(2320 thru 2340=2)(2350 2351 2352=1)(2359=2)(2400=1) (2410=2)(2411=1)(2412 2419=2) (2420 thru 2429=1) (2430 2431 2432=2)(2440 thru 2443=1)(2444=2) (2445=1)(2446 thru 3142=2)(31433144=1)(3145 thru 3229=2)(3230 3231 3232=3)(3240 3241 3242=2) (3300 thru 3340=3)(3400 thru 3429=2)(3430=3) (3431 3432=2) (3433=3)(3434=2)(3439=3)(3440 thru 3451=2)(3452 3460=3)(3470 thru 3475=2) (3480 thru 4141=3)(4142=5)(4143 thru 5120=3)(5121 =2)(5122=4)(5123 5130=5)(5131 =3) (5132=5)(5133=3)(5139=5)(5140 5141 =4) (5142=5)(5143=4)(5149=5) (5150 5151
5152=2)(5160=5)(51615162=4)(5163=5)(5164=4) (5169=5)(5200 thru 5230=3)(6000 thru 6130=5) (6131 6132 6133=6)(6134 thru 6154=5)(6200 6210=6) (7000=4)(7100 thru 7113=5)(7120=4)(7121 7122 7123=5)(7124 7129 7130=4)(7131=5)(7132 7133 7134=4)(7135=5)(7136 thru 7141=4)(7142 7143=5) (7200 thru 7233=4)(7234=5)(7240 thru 7313=4) (7320 7321 7322=5)(7323 7324=4)(7330 7331 7332=5) (7340 thru 7420=4)(7421=5)(7422 7423=4)(7424=5) (7430=4)(7431 7432=5)(7433 thru 7500=4)(7510=3) (7520=4)(7530 thru 8143=5)(8150 thru 8172=4) (8200 thru 8310=5)(8311=4)(8312 thru 8331=5) (8332 8333=4)(8334 thru 9000=5)(9100 thru 9113=3)(9120 thru 9333=5). IF (vllO eq 1 and egpclass ge 2) egpclass=6. VALUE LABELS egpclass 1 'Service ľ 2 'Service Iľ 3 'Routine n-m'4 'skilled man' 5 'unskilled man' 6 'selfemp'.
* Sweden*
COMPUTE egpclass = sl73.
RECODE egpclass (56,57,60=1)(46=2)(33,36=3)
(79=6)(89=6)(21,22=4) (11,12=5).
IF (vllO eq 1 and egpclass ge 2) egpclass=6.
IF (sl73 eq 33 and (sl06 ge 400 and sl06 le 499))
egpclass=5.
IF (sl73 eq 22 and (sl06 eq 105 or sl06 eq 106 or sl06
eq 107 or sl06 eq 111 or sl06 eq 112 or sl06 eq 123 or
sl06 eq 131 or sl06 eq 912 or sl06 eq 913)) egpclass=3.
IF (sl73 eq 33 and (sl06 ge 931 and sl06 le 949)) egp-
class=5.
VALUE LABELS egpclass 1 'Service ľ 2 'Service IF
3 'Routine n-m'4 'skilled man' 5 'unskilled man' 6
'selfemp'
Author's Address
Department of Sociology, Umeá University, 901 87 Umeá, Sweden, e-mail: Stefan.Svallfors@soc.umu.se.
Manuscript received: September 1996