Welfare state regimes and subjective well-being: a cross-national study.
Several studies have examined governmental responsiveness to the problems associated with unemployment as well as the impact of unemployment on individual subjective well-being. What has been missing, however, is a comparative analysis of public attitudes and behavior. We integrate research on political psychology and welfare state development of social provisions aimed at mollifying unemployment and other conditions of life. These two literatures may appear to be strange bedfellows but, as we will argue, connect together well. Utilizing the typology of Esping-Andersen (1990), while recognizing its limitations (see Bolderson and Mabbett 1995), we examine three types of welfare state regimes: Liberal; Conservative; and Social Democratic. The heart of the article compares the relationship between unemployment and subjective well-being across these regime types. The typology is particularly useful as it serves as an analytical framework to explore life satisfaction from a comparative perspective.
As economists have recently come to appreciate, employment status itself has been found to be a significant influence not only on an individual's economic condition but also on his or her social psychological outlook (see Clark and Oswald 1994). This has long been known to psychologists, sociologists, and others, at least in the USA and Britain. Specifically, being unemployed has been found to have a wide range of consequences, independent of income and other socioeconomic and demographic characteristics, affecting not only people's economic well-being, but also their personal well-being and happiness more generally (see especially Diener and Diener 1995, Myers and Diener 1995, Diener 1984, Kessler 1982, Campbell et at. 1976, Warr 1978, essays in Andrews 1986, Warr et al. 1988, Jackson et al. 1983, Veenhoven 1984, 1989, 1993). The effects of being unemployed are attributable not just to financial consequences, but also to additional influences such as a lack of productive activity (e.g., job satisfaction is related to subjective well-being; see Diener 1984, p. 555). Public opinion has been closely, though not perfectly, related to changes in the economy, particularly to the rate of unemployment itself (see Shapiro and Young 1989). For instance, Hibbs (1979), and Page and Shapiro (1992, ch. 4), have shown that Americans' concern for unemployment and support for government intervention to address the problems is highly responsive to real-world changes in unemployment. Similar patterns in public opinion have been found in European countries (Clarke and Dutt 1991, Maassen and de Goede 1991).
Previous research has focused on the attitudinal and behavioral consequences of unemployment within individual countries. The problem, though, is that we do not know fully and comprehensively how common their findings are to all industrial countries (e.g., see Clark and Oswald 1994).
While there has been some cross-national research on well-being and life satisfaction more generally, these studies have not often examined the particular effects of unemployment on individuals' happiness and overall satisfaction with life (cf. Inglehart 1990, Inglehart and Rabier 1986). Examining the recent literature on subjective well-being and happiness, we do, however, find trends and patterns of attitudes across nations that can be attributed to employment conditions and the economy as well as to country-specific factors. The lack of any increases in subjective well-being in Europe, for example, is quite understandable given the increases in structural unemployment rates. Similarly, the pattern of a slight gradual decline in happiness or personal satisfaction in the USA is consistent with the increased acceptance by experts and financial analysts of higher levels of unemployment than in the past (they accept a 5-6 percent level of unemployment and fear that further decreases in unemployment would spark inflation). Unemployment and the related sense of job insecurity is a cause of gradual decline in feelings of well-being (cf. Lane 1996).
Inglehart and Rabier's (1986) analysis illustrates the failure to consider unemployment and comparative differences. The authors show that personal satisfaction and happiness declined visibly in Europe, with the most striking drop occurring in Belgium, which experienced a 15 percent unemployment rate in the early 1980s (the European Community averaged 10 percent during the same period). This decline in well-being appeared to be less in countries with lower levels of subjective well-being to begin with and in the Netherlands (and to a lesser extent Denmark) where its welfare state could cushion the effects of unemployment. This, however, does not explain why Ireland and Britain paralleled or fared better than the Netherlands in this regard.
The election of so-called neoliberal parties in the USA and Western Europe created a new political context to study the dynamics of public opinion. These parties demanded reductions in welfare spending, particularly for the poor and unemployed, and made the elimination of inflation, rather than unemployment, their chief economic goal. Has this rise of neoliberal politics, then, been accompanied by significant changes in public attitudes? More specifically, have there been fundamental changes in the way the public perceives the problem of unemployment? Unfortunately, it is difficult to answer these and related questions about the nature of public opinion toward welfare state policies. Part of the reason why this is so difficult is that a well-defined research tradition has not yet been established within the study of comparative attitudes toward different welfare state policies (Shapiro and Young 1989, p. 144). Some recent work, however, shows promise in setting a research agenda here (Kaase and Newton 1995, van Deth and Scarbrough 1995, Borre and Scarbrough 1995). One recent study in particular makes good use of existing theories about attitudes toward the welfare state. Pettersen (1995) uses developmental theories (Rose and Peters 1978) to analyze attitudes toward social welfare policies.
Our study draws on the regime type literature because it provides a strong analytic basis for cross-national work. Several prominent works-including that of Esping-Andersen (1990), Palme (1990), and Castles and Mitchell (1902)--have used different methodologies to categorize countries according to types of social security programs. Others have argued that this analysis of regimes is problematic because of the difficulty of categorizing diverse programs covering unemployment, old age, etc. (Bolderson and Mabbett 1995). While these critiques are legitimate, we find regime theory a useful heuristic for identifying broad differences in western welfare states.
Esping-Andersen (1990) suggests the existence of three types of regimes: Liberal (e.g., USA, Australia); Conservative (e.g., France, Germany, Austria, UK); and Social Democratic (e.g., Norway, Netherlands, Denmark). This framework is helpful because it provides a set of expectations regarding variations in public opinion; the microfoundations of his aggregate analysis suggest a predictable pattern of attitudes. For example, Esping-Andersen's regime typology suggests that we should observe that citizen preferences in Social Democratic countries will be more favorable toward government intervention in securing a full employment guarantee. These countries should show higher support for active government policies regarding employment and social welfare. In comparison, Liberal regimes such as the USA and Australia should be the least disposed toward such activity, while Conservative regimes should fall in between. It is reasonable to assume that differences in well-being between the employed and unemployed will be the smallest in Social Democratic countries because of their comparatively expansive protections against the economic effects of unemployment. We borrow from Esping-Andersen not only to link the regime typology to subjective well-being in a more informal and exploratory manner, but more importantly, to develop three testable hypotheses concerning variations in welfare programs and public opinion. The first set of questions addresses the perceived severity of the problem of unemployment. Economic problems have historically been major public concerns in all developed and developing nations. Unemployment has been among the most, if not the most, important of these problems. There have been, however, some interesting trends over time as well as noteworthy differences across countries. How has the development of expansive welfare states affected these public attitudes? Have the more advanced safety-nets associated with the Social Democratic regimes dampened the fear of unemployment among individuals as they evaluate their personal circumstances? In other words, do we observe differences in attitudes among citizens of different types of welfare states? These general questions lead to hypothesis 1:
H1: citizens in countries that fall under the Social Democratic type, and to a lesser extent, the Conservative type, should perceive the threat of unemployment to be a lesser problem to them personally than in Liberal regimes. Conversely, Liberal regimes which offer less protection from employment swings should evoke greater sensitivity regarding unemployment.
We turn next to the role of government in providing policy solutions to unemployment. Esping-Andersen argues that established welfare states weaken the public's identification and dependence on private enterprise and promote strong government programs as an alternative source of goods and services. In short, we should see real differences in public attitudes across Liberal, Conservative and Social Democratic regimes concerning the role of government. Specifically, how do the public perceive the role of government in solving problems of unemployment? This can be expressed as hypothesis 2:
H2: because Social Democratic regimes put a higher value on government programs to promote full employment than their Liberal and Conservative counterparts, the public should be more supportive of government intervention against unemployment in Social Democratic regimes.
The third and most fundamental question concerns the impact of economic changes on public attitudes toward unemployment and people's perceived well-being. Of special interest and importance is the extent to which employment status--whether one is employed or unemployed-influences these attitudes. Have expansive welfare states been able to alleviate the psychological consequences of unemployment? This leads us to hypothesis 3:
H3: unemployment will have less impact on subjective well-being in Social Democratic countries than in Conservative and Liberal countries.
DATA AND METHODS
This study's analysis is based on data from the United States' National Opinion Research Center (NORC) General Social Surveys (GSS) 1972-1996; the International Social Survey Programme (ISSP) 1985-1993; and the Eurobarometer surveys (EB) 1973-1994. At some points, useful data and analyses that other researchers have provided are also cited.
We examine responses to survey questions that have been asked in a set of countries or over time in individual countries. Unfortunately, as is too often the case in public opinion research, the surveys do not provide complete time series for all of the questions of interest to us in the countries we are studying. The USA has the most extensive body of data, though it too is incomplete in important respects. Thus our comparisons over time and across countries are fragmentary, often relying on data from the 1970s and 1980s to reflect on contemporary conditions. Nonetheless, the data offer useful evidence to evaluate public opinion toward the problem of unemployment, government policy, and personal well-being in industrial nations.(1)
In addition, we investigate a wide range of differences and similarities in opinions among important subgroups of the populations. We compared attitudes and trends in attitudes according to sex, education, age, income, occupation, employment status, region and other variables, although all of this analysis could not be presented. Similar independent variables for the analysis across the GSS, EB and ISSP surveys were employed when possible.(2) In addition, multivariate analyses enabled us to identify differences in attitudes that were not related to or confounded by demographic characteristics. In particular, we used multiple regression analysis and ordered probit models for ordinal (ordered categorical) dependent variables.
The independent variables of greatest interest for our analysis are employment status (whether an individual is employed or unemployed), age, education, income, and sex. Comparing the attitudes of the employed versus unemployed is central to the question that this paper examines, including the importance of work and the consequences of employment for well-being and life satisfaction. Age is significant because life experiences such as retirement may affect attitudes regarding, for instance, the importance of government help to those still working. Age, in our view, is also especially important because it may preview future changes in attitudes as new generations of adults make up an increasing proportion of the population. The expectations of people with different levels of education may also differ. For example, the importance of work and the effect of being unemployed or dissatisfied with one's job may vary according to educational attainment. We might expect that being unemployed may have a more detrimental effect on the well-being of the better educated than on those with less education; education brings with it greater expectations for having a job and, in fact, a good job. Gender may be increasingly important as well, since women have become a larger part of the work force than in the past and have new expectations regarding work. There may also be sex differences in opinions toward government social welfare policies. In the USA, for example, women have tended to be somewhat more supportive than men of certain economic and other welfare policies (see Shapiro and Mahajan 1986).
PERCEIVED SEVERITY OF THE PROBLEM OF UNEMPLOYMENT
The available data over time show persistent support across western nations for fighting unemployment and providing unemployment benefits. Nonetheless, there are clear differences. The ISSP data provide relevant comparative evidence for the USA, Australia, Italy, the UK, and West Germany from 1985 to 1990, with information for Norway, Northern Ireland as well as East Germany, Hungary and Israel in 1990.
On the inflation versus unemployment tradeoff reported in Table 1, Americans and Australians (Liberal regimes) in 1985 were the most adverse toward inflation--near 50 percent or more, compared to 30-39 percent for the other countries. By 1990, however, the publics in all of the countries appeared to become marginally (though sometimes no more than expected from sampling error) more adverse toward inflation: a 5 percentage point increase in the USA, 4 percent in Australia, only 1 percent in Italy, 9 percent in (what was then) West Germany, and fully 27 percent in the UK. Italy remained the most supportive of government giving unemployment the highest priority. In Norway only 54 percent of the population ranked unemployment as being more of a problem than inflation. Of course, the previous finding could be the result of overall lower unemployment rates.
TABLE 1 Tradeoff between inflation and unemployment If the government had to choose between keeping down inflation or keeping down unemployment to which do you think it should give highest priority?
Country N Inflation Unemployment Percent Percent 1985 USA 564 49 51 Australia 1411 58 42 Austria 747 31 69 Italy 1444 31 69 UK 1390 30 70 West Germany 871 39 61 Grand Total 6427 40 60 1990 USA 952 54 46 Australia 2123 62 38 Norway 1218 46 54 Italy 944 32 68 UK 1094 57 43 Northern Ireland 719 52 48 West Germany 2322 48 52 East Germany 787 34 66 Israel 761 34 66 Hungary 928 74 26 Ireland 936 34 66 Grand Total 12784 49 51
Additional data come from the Eurobarometer surveys. According to the survey data in Table 2, which also includes some data for Spain, Portugal, and Greece, large majorities in 1989 (59-89 percent) responded that it was `very important' to fight against unemployment. The only countries for which this `very important' response was less than 70 percent were the Netherlands (59 percent), Luxembourg (67 percent), and the UK (65 percent). Previously, in the May 1976 Eurobarometer study of public opinion in France, Northern Ireland, Belgium, West Germany, Luxembourg, Ireland, Netherlands, Denmark and the UK (see Table 2), the public thought it was important to fight unemployment, with Luxembourg having a bit smaller percentage saying `very important' (53 percent compared to 70 percent or more for the other countries). For the inflation versus unemployment tradeoff, unemployment was thought to be of greater priority in each country except the UK and Luxembourg. In considering sources of unemployment, among the top two criticisms of multinational corporations that were mentioned, between 10 and 35 percent of respondents cited the use of cheap labor threatening employment in advanced European countries, with the Dutch expressing the most concern and citizens of Luxembourg the least (data not reported here). Subgroup differences in opinions were examined in simple bivariate tables and were also evaluated in multivariate analyses to estimate the independent effects of several demographic characteristics and to minimize any emphasis on relationships that might be spurious.(3) For our purposes, these multivariate results confirm or disconfirm that bivariate subgroup differences remain after controlling for other variables, so that we can reject the proposition that these relationships are spurious.
TABLE 2 Importance of fighting unemployment Country N Very Important important 1976 How important is this problem to you. Fighting unemployment France 1218 78 19 Belgium 916 77 19 Netherlands 894 76 22 Germany 995 73 21 Luxembourg 257 53 41 Denmark 956 80 17 Ireland 1005 87 11 UK 1008 68 29 Northern Ireland 311 85 14 Grand total 7560 77 20 1989 Importance of problem... The fight against unemployment. France 1031 79 19 Belgium 997 72 24 Netherlands 967 59 38 Germany 1191 70 26 Italy 1004 82 17 Luxembourg 294 67 25 Denmark 991 76 22 Ireland 1008 86 14 UK 948 65 31 Northern Ireland 289 83 Greece 968 79 18 Spain 983 82 18 Portugal 922 89 10 Grand total 11593 76 22 Little Not at all importance important Percent 1976 How important is this problem to you. Fighting unemployment France 2 1 Belgium 2 1 Netherlands 1 1 Germany 5 1 Luxembourg 5 1 Denmark 2 1 Ireland 1 1 UK 3 1 Northern Ireland 1 0 Grand total 2 1 1989 Importance of problem... The fight against unemployment. France 1 0 Belgium 4 0 Netherlands 3 0 Germany 3 1 Italy 1 0 Luxembourg 5 3 Denmark 2 0 Ireland 0 0 UK 3 1 Northern Ireland 17 10 Greece 2 0 Spain 0 0 Portugal 0 0 Grand total 2 0
In examining subgroup differences in the case of attitudes toward `fighting unemployment' in the Eurobarometer the most striking finding is the lack of variation across subgroups (tables not reported here, but available from the authors). Thus to the extent that unemployment is a very important problem, it is important across the population. With the exception of region, the multivariate coefficients are only rarely statistically significant. The variables that are important in some countries suggest that men and the better educated see the problem as slightly less important than do women and those who are less well educated. Only in the UK does union membership seem to make respondents more sensitive to the unemployment problem. Income and blue collar versus other occupations vary in their effect across countries with blue collar workers if anything also less concerned than others about unemployment and those with higher incomes less concerned about the problem. Being unemployed itself does not lead to systematic variation, indicating that the problem is noticed fully nationwide or collectively, and not just by those directly affected by it. The regional variations in opinion are evident but are not easily interpreted without detailed knowledge of the social, cultural, political, or other relevant influences specific to the geographies of each of the countries. It is very possible that the noticeable regional differences reflect variations in local unemployment rates or sensitivity to the problem related to other aspects of regional economies.
The bivariate findings confirm the multivariate findings. Over 90 percent of the unemployed and employed in all surveyed countries in both 1976 and 1989 believe that it is `important' or `very important' to fight against inflation. The bivariate results for education suggest that all levels of education see the fight against unemployment as being important, with the less educated slightly more likely to believe that it is `very important'.
The evidence supports two findings. First, it suggests lack of support for hypothesis 1. Public sensitivity to unemployment is strong across all western nations; where variation exists it does not correspond with the expected differences across welfare regimes. Table 1 indicates that individuals in Liberal regimes (the USA and Australia) are more concerned over inflation in the employment-prices tradeoff than their counterparts. There is no evidence for increased sensitivity to unemployment in these countries where social protections are less. Concern for unemployment versus inflation is likely related to changes in prices, unemployment rates, and regional variations, and not to welfare state regimes. Second, the fact that the multivariate analyses do not uncover systematic variations in opinion suggests that the problem is perceived as a national or collective issue and that attitudes are not largely motivated by individuals' personal characteristics and circumstances. The microfoundations of Esping-Andersen's expectations regarding individual motivation are not supported.
ROLE OF GOVERNMENT
We next turn to whether social Democratic regimes put a higher value on full employment than their counterparts, leading to more support for government intervention against unemployment. While data across regime types for an extended time period are not available, we have some pertinent data from the 1989 Eurobarometer study and the 1990 ISSP study. In addition to examining support across regime types, changes for individual countries over time are of interest. Support apparently varies over time depending on the severity of economic problems, with support increasing in bad times. In the case of social welfare policy and spending, Americans in particular distinguish between different kinds of social welfare policies (see Cook and Barrett 1992, Page and Shapiro 1992, ch. 4; for a review of literatures on other countries, see Coughlin 1980, Shapiro and Young 1989). Opinion toward `welfare' spending--with `welfare's' pejorative connotation--and other aspects of social welfare policy is affected by economic conditions, levels of welfare spending, and the party holding the presidency. As economic conditions worsen, support for social policies tends to increase; and support for social welfare spending tends to drop as this type of government domestic spending increases when a Democrat occupies the presidency (see Wlezien 1995, Durr 1993, and Weaver et al. 1995, Borre and Scarbrough 1995).
The 1989 Eurobarometer surveys in Table 3 indicate that in a number of European countries, 83 percent or more of the population in each, thought that it was `absolutely necessary . . . to be able to benefit from social welfare when needed.' In other words, social democratic welfare states did not, as expected, attract greater support. This lack of variation suggests somewhat surprisingly that regime type seems to matter very little in building support for government intervention to mitigate the effects of unemployment.
TABLE 3 Importance of social welfare programs Which things seem absolutely necessary? ... To be able to benefit from social welfare when needed, such as in the case of unemployment, sickness, handicap, old age.
Country N Not Mentioned Mentioned Percent 1989 France 1040 10 90 Belgium 1028 6 94 Netherlands 970 15 85 West Germany 1202 6 94 Italy 1011 5 95 Luxembourg 301 4 96 Denmark 1000 12 88 Ireland 1016 4 96 UK 957 17 83 Northern Ireland 291 9 91 Greece 1000 3 97 Spain 1003 3 97 Portugal 1000 3 97 Grand total 11819 7 93
The ISSP data for 1985 and particularly 1990, however, show variation between different welfare state regimes, though this is not as sharp a divide as Esping-Andersen assumes. Table 4 reveals the following: Americans and Australians are the least supportive--at under the 60 percent level--of government having a responsibility to provide a decent standard of living for the unemployed, with the UK, Italy (for 1985 and 1990), Norway, Ireland, Northern Ireland and East Germany (for 1990) the most supportive at the 80 percent level or better (combining the `definitely should' and `probably should' categories). There does seem to be a real difference in attitudes toward the role of government between Liberal and both Conservative and Social Democratic countries. We do not observe the difference between Conservative and Social Democratic regimes that Esping-Andersen anticipates.
TABLE 4 Provide for the unemployed Country N Definitely Probably should should 1985 On the whole, do you think it should be or should not be the government's responsibility to: provide a decent standard of living for the unemployed. USA 624 16 35 Australia 1448 15 44 Austria 895 18 51 Italy 1528 39 46 UK 1448 44 41 West Germany 1006 24 62 Grand total 6949 28 46 1990 Resp: Provide For Unemployed USA 1110 14 39 Australia 2292 6 50 Norway 1459 42 49 Italy 964 32 45 UK 1139 32 48 Northern Ireland 739 46 43 West Germany 2667 19 59 East Germany 972 52 42 Israel 955 32 31 Hungary 953 21 51 Ireland 992 48 42 Grand total 14242 28 48 Probably, Definitely not not Percent 1985 USA 33 17 Australia 32 9 Austria 21 1 Italy 11 4 UK 11 4 West Germany 12 3 Grand total 19 7 1990 Resp: Provide For Unemployed USA 34 13 Australia 36 8 Norway 8 1 Italy 16 6 UK 15 5 Northern Ireland 8 2 West Germany 17 5 East Germany 5 1 Israel 23 15 Hungary 21 7 Ireland 8 2 Grand total 19 6
During the 1985-1990 period, shown in Table 5, there was substantial support in all countries for having the government finance projects to create jobs, with a 70 percent figure even in the USA in 1990 (Table 5, combining favor and strongly favor). Strong support (strongly favor), however, fell off by about 10 percentage points in the UK, Australia, and Italy. Concerning the extent to which the publics in these countries think that it is the government's responsibility to provide a job for everyone who wants one (Table 6), Americans in 1985 showed the least support (35 percent, combining definitely should and probably should), followed by Australians (53 percent), then the UK (72 percent), West Germans (81 percent), Austrians (85 percent), and Italians (90 percent). But in 1990, Americans increased their support by 9 percentage points (to 44 percent), compared to drops in support in the other countries: by 12 percent in Australia, 5 percent in Italy, 8 percent in the UK, and 6 percent in West Germany. In the 1990 survey, Norway showed an 84 percent level of support--again not statistically different from several countries outside of their regime type (a higher percentage of `definitely should' responses did exist though). With the exception of Australia, support for maintaining or increasing government spending on unemployment benefits was high (70 percent or more) and stable (there were no data for Austria in 1990). In Australia, the 48 percent figure in 1985 dropped to 42 percent in 1990 (see Table 7).(4)
TABLE 5 Government financing to create new jobs Here are some things the government might do for the economy. Circle one number for each action to show whether you are in favor or against it. Government action for the economy: Government financing of projects to create new jobs.
Country N Strongly Favor Favor Percent 1985 USA 668 27 42 Australia 1459 25 52 Austria 957 28 44 UK 1503 38 51 Italy 1580 54 37 West Germany 1032 30 41 Grand total 7199 35 45 1990 USA 1185 25 45 Australia 2369 16 54 Norway 1440 28 56 Italy 975 44 47 UK 1178 28 56 Northern Ireland 743 31 56 West Germany 2752 31 42 East Germany 999 68 24 Israel 986 52 37 Hungary 944 47 31 Grand total 13571 34 46 Neither Against Strongly against Percent 1985 USA 16 3 12 Australia 11 3 9 Austria 22 1 4 UK 8 1 3 Italy 7 0 2 West Germany 15 5 9 Grand total 12 2 6 1990 USA 21 7 2 Australia 18 10 3 Norway 9 6 1 Italy 7 2 1 UK 11 1 1 Northern Ireland 8 4 1 West Germany 18 7 2 East Germany 6 2 1 Israel 8 3 1 Hungary 13 7 3 Grand total 13 6 2
TABLE 6 Provide jobs for all On the whole, do you think it should be or should not be the government's responsibility to: provide a job for everyone who wants one.
Country N Definitely Probably should should 1985 USA 637 14 21 Australia 1442 20 33 Austria 935 47 38 Italy 1553 51 39 UK 1441 38 34 West Germany 1020 35 46 Grand total 7028 36 36 1990 USA 1118 16 28 Australia 2334 11 30 Norway 1458 52 32 Italy 974 38 47 UK 1136 24 40 Northern Ireland 728 35 41 West Germany 2702 30 45 East Germany 978 63 32 Israel 982 56 32 Hungary 971 52 38 Ireland 987 38 34 Grand total 14368 34 36 Probably Definitely not not Percent 1985 USA 34 32 Australia 29 18 Austria 13 3 Italy 8 3 UK 17 12 West Germany 16 3 Grand total 18 10 1990 USA 34 23 Australia 46 12 Norway 13 4 Italy 10 6 UK 22 15 Northern Ireland 16 8 West Germany 21 4 East Germany 5 1 Israel 9 4 Hungary 9 2 Ireland 17 12 Grand total 21 8
Table 7 More spending on unemployment benefits Listed below are various areas of government spending. Please show whether you would like to see more or less government spending in each area. Remember that if you say `much more', it might require a tax increase to pay for it. More or less government spending for: Unemployment benefits.
Country N Much More Same Less Much more less Percent 1985 USA 641 8 18 49 17 7 Australia 1475 4 9 35 30 22 Austria 917 3 12 44 31 10 Italy 1511 15 41 27 11 7 UK 1465 12 29 40 15 4 West Germany 1006 8 27 52 11 3 Grand total 7015 9 24 39 19 9 1990 USA 1125 7 20 51 15 6 Australia 2375 2 8 32 39 19 Norway 1416 4 15 57 20 5 Italy 959 15 37 31 10 7 UK 1144 8 28 47 14 4 Northern Ireland 732 20 36 33 8 2 West Germany 2719 9 27 50 11 2 Israel 969 12 17 27 25 19 Hungary 943 10 38 27 16 9 Grand total 12382 8 23 41 19 8
The evidence is mixed and does not fully support hypothesis 2. Tables 5 to 8 indicate less support for government promotion of full employment and measures to offset unemployment in Liberal welfare state regimes as opposed to Conservative and Social Democratic regimes. On the other hand, differences did not-as Esping-Andersen predicted--emerge between the Conservative and Social Democratic regimes. Moreover, changes over time defied neat classification. For instance, the increase in American support for government provision of jobs and the drop in other countries contradicts Esping-Andersen's predictions.
WELL-BEING AND LIFE SATISFACTION: THE EMPLOYED VERSUS THE UNEMPLOYED
We now turn to the impact of unemployment on subjective well-being. The effect of unemployment across regime types should be noticeable. Within Social Democratic, and to a lesser extent, Conservative regimes, more extensive social welfare programs exist to mitigate against the destructive consequences of unemployment.
Figures 1 and 2 present the trends in the mean responses to the three category measure of self-reported happiness for the USA from 1972 to 1996 for both the employed and unemployed (GSS). What is striking in Figure 1 is the wide gap, as is expected between the employed and the unemployed (the greater volatility for the unemployed is largely due to greater sampling error because of fewer respondents).
[Figures 1-2 ILLUSTRATION OMITTED]
Figure 2 for the Eurobarometer data show the percentages of individuals `somewhat satisfied' or `very satisfied' with life by employment status in ten countries. The unemployed are significantly less satisfied with life than the employed. The gap was narrower in the mid to late 1970s and the 1990s for several countries. In the late 1980s this gap was as much as 50 percent in Luxembourg with approximately go percent of the employed compared to 40 percent of the unemployed being satisfied.
Figure 2 confirms hypothesis 3. Smaller differences between employed and unemployed individuals can be found in Social Democratic (the Netherlands and Denmark) than in Conservative and Liberal countries in the 1980s and 1990s. The type of social welfare programs appear to make a difference in subjective well-being for the unemployed.
Pinpointing the impact of being unemployed on well-being and overall life satisfaction is complicated, however, in two ways. First, it is possible that people who are less happy with their lives or otherwise psychologically troubled have difficulty in finding and holding on to jobs, so that the causal relationship is the reverse of that posed in the questions of interest here (cf. Verkley and Stolk 1989). Other studies, using alternative data, have provided some evidence that being without a job contributes to dissatisfaction and diminished well-being, not the reverse; that is, as people reenter the labor force after being unemployed, their well-being improves (see Jackson et al. 1983, Verkley and Stolk 1989). The relationship between employment and overall happiness may be a reciprocal one, but one in which being employed makes an important difference for one's well-being.
Second, the effect of being unemployed has to be examined multivariately. Our multivariate ordered probit and OLS models estimating subgroup differences in individuals' personal happiness or overall satisfaction based on the 1972-1996 NORC- GSS data (for the USA) and the 1976-1994 Eurobarometer surveys reveal some persistent differences (in France, Germany, Belgium, the Netherlands, Ireland, Italy, Great Britain, Luxembourg, Northern Ireland, Greece, Spain, Portugal, Denmark, Finland  and Norway ). After controlling for relevant influences, married people are happier or more satisfied than others; women are happier than men in the USA, but the results are more mixed for the Eurobarometer data on life satisfaction, with men tending to be less satisfied than women on average; people with more income and education are generally happier or more satisfied with their lives than those with less. Further, the coefficients for the age and age-squared terms reveal that happiness and life satisfaction in virtually all the countries we examined decline with age initially and then increase later in life. Most notably, after these income and life-cycle effects are taken into account, the unemployed are clearly and persistently less happy or less satisfied with their lives than those who have jobs. While the coefficients for this variable vary somewhat over time, the variations do not show a pattern of systematic increases or decreases in all countries. It should be noted that even in Denmark and the Netherlands when other variables are controlled, the unemployed consistently tend to be less happy than the employed. These results lead us to temper any strong conclusion that unemployed individuals in Social Democratic countries have felt better about their lives, relative to the employed, compared to Conservative and Liberal regimes (data of Profit and OLS models not shown, but available from the authors).
Even with the expansion of welfare states, employment is as important as ever for individuals' economic as well as social and psychological well-being. Our analysis indicates at best mixed support for the hypotheses. While differences did emerge between Liberal regimes as opposed to Conservative and Social Democratic regimes, we did not find the expected systematic variations between the latter two. No support emerged for hypothesis I and the expectation that Liberal regimes, which offer less government protection against the winds of economic change will be more sensitive regarding unemployment than their counterparts. The evidence was divided on whether individuals in Social Democratic regimes are more supportive--as suggested by hypothesis 2--of government intervention. Finally, the evidence provides mixed support for hypothesis 3. Subtle and important differences in trends (in the cases of Denmark and the Netherlands), observed through the gap between the employed and the unemployed, suggests that the unemployed may fare better in subjective well-being in Social Democratic countries.
Esping-Andersen and others have constructed elaborate theories of welfare state development. Our analysis raises questions about the microfoundations of these theories. What are needed are institutional and political analyses of comparative welfare state development that are anchored in accurate models of public perceptions, attitudes, and behavior.
An earlier version of this paper was presented at the August 29-September 1 1996 Annual Meeting of the American Political Science Association, San Francisco. The authors thank Maryanne Dawicki for research assistance; Greg Haley and Sue Zayac: of Columbia University's Electronic Data Service for help in acquiring data; Debi Gilchrest for assistance in administering the project; and Aida Llabally for her administrative support. They also thank Young-Sook Nam of the International Labour Office, Geneva, Switzerland for her helpful comments. The data were obtained (for Great Britain and other European countries) from the ESRC Data Archive at the University of Essex and (for the NORC General Social Surveys and some Eurobarometer surveys) from the U.S.'s Inter-University Consortium for Political and Social Research. The research reported here was conducted at Columbia University's Paul F. Lazarsfeld Center for the Social Sciences. The responsibility for the analysis and interpretation in this report is the authors'.
Andrews, Frank M. (ed.) (1986): Research on the Quality of Life. Ann Arbor, MI, Survey Research Center, Institute for Social Research, University of Michigan.
Bolderson, Helen and Mabbett, Deborah (1995): Mongrels or thoroughbreds: A crossnational look at social security systems', European Journal of Political Research, 28, 119-39.
Borre, Ole and Scarbrough, Elinor (eds.) (1995): The Scope of Government, Oxford, Oxford University Press.
Campbell, Angus, Converse, Philip E., and Rodgers, William L. (1976): The Quality of American Life, New York, Russell Sage Foundation.
Castles, F.G. and Mitchell, D. (1992): `Identifying welfare state regimes: The links between politics, instruments and outcomes', Governance, 50).
Clark, Andrew E., and Oswald, Andrew J. (1994): `Unhappiness and unemployment', The Economic Journal, 104, 648-59.
Clarke, Harold, D. and Dutt, Nitish (1991): `Measuring value change in Western industrialized societies: The impact of unemployment', American Political Science Review, 85, 905-20.
Cook, Fay Lomax, and Barrett, Edith J. (1992): Support for the American Welfare State: Views of Congress and the Public, New York, Columbia University Press.
Coughlin, Richard M. (1980): Ideology, Public Opinion, and Welfare Policy, Berkeley, CA, Institute of International Studies.
Diener, Ed (1984): `Subjective well-being', Psychological Bulletin, 95, 542-75.
Diener, Ed, Diener, Marissa, and Diener, Carol (1995): `Factors predicting the subjective well-being of nations', journal of Personality, and Social Psychology, 69, 851-64.
Durr, Robert H. (1993): `What moves policy sentiment,' American Political Science Review, 87, 158-70.
Esping-Andersen, Gosta (1990): The Three Worlds of Welfare Capitalism, Princeton, NJ, Princeton University Press.
Greene, William H. (1993): Econometric Analysis, 2nd edn. New York, Macmillan.
Hibbs, Douglas A., Jr. (1979): `The mass public and macroeconomic performance: Dynamic public opinion toward unemployment and inflation', American Journal of Political Science, 23, 705-31.
Inglehart, Ronald (1977): The Silent Revolution: Changing Values and Political Styles Among Western Republics, Princeton, Princeton University Press.
Inglehart, Ronald (1990): Culture Shift in Advanced Industrial Society, Princeton, Princeton University Press.
Inglehart, Ronald and Rabier, Jacques-Rene (1986): `Aspirations adapt to situations--but why are the Belgians so much happier than the French?'. In Andrews, Frank M. (cd.) Research on the Quality of Life, Ann Arbor, MI, Survey Research Center, Institute for Social Research, University of Michigan, 1-56.
Jackson, Paul R., Stafford, Eliabeth M., Banks, Michael H., and Warr, Peter B. (1983): `Unemployment and psychological distress in young people: The moderating role of employment commitment', Journal of Applied Psychology, 68 (August), 525-35.
Juster, F. Thomas, and Courant, Paul N. (1986): `Integrating stocks and flows in quality of life research'. In Andrews, Frank M. (ed.) Research on the Quality of Lift, Ann Arbor, MI, Survey Research Center, Institute for Social Research University of Michigan, 147-70.
Kaase, Max, and Newton, Kenneth (1995): Beliefs in Government, Oxford, Oxford University Press.
Kessler, Ronald C. (1982): `Socioeconomic status and psychological distress', American Sociological Review, 47, 752-64.
Lane, Robert E. (1996): `The joyless market economy', Paper delivered to the Conference on Economics, Values, and Organization. Yale University, New Haven, CN, April 19-21.
Liao, Tim Futing (1994): Interpreting Probability Models: Logit, Probit, and Other Generalized Models, Thousand Oaks, CA, Sage.
Maassen, Gerard, and Goede, Martijn de (1991): `Changes in public opinion on the unemployed: The case of the Netherlands', International Journal of Public Opinion Research, 3, 182-94.
Myers, David G. and Diener, Ed (1995): `Who is happy?' Psychological Science, 6, 10-19.
Page, Benjamin I. and Shapiro, Robert Y. (1992): The Rational Public: Fifty Years of Trends in Americans Policy Preferences, Chicago, University of Chicago Press.
Palme, J. (1990): `Models of old-age pensions'. In Ware, A. and Goodin, R. (eds.) Needs and Welfare, London, Sage.
Pettersen, Per Arnt (1995): `The Welfare State: The security dimension'. In Borre, Ole and Scarbrough, Elinor (ed.) The Scope of Government, Oxford, Oxford University Press, 198-233.
Rasinski, Kenneth A. (1989): `The effect of question wording on public support for government spending', Public Opinion Quarterly, 53, 388-94.
Rose, R. and Peters, B.G. (1978): Can Government Go Bankrupt? New York, Basic Books.
Shapiro, Robert Y., and Mahajan, Harpreet (1986): `Gender differences in policy preferences: A summary of trends from the 1960's to 1980's', Public Opinion Quarterly, 50, 42-61.
Shapiro, Robert Y. and Young, John T. (1989): `Public opinion and the Welfare State: The United States in comparative perspective', Political Science Quarterly, 104, 59-89.
van Deth, Jan W. and Scarbrough, Elinor (eds) (1995): The Impact of Values, Oxford, Oxford University Press.
Veenhoven, Ruut (1984): Conditions of Happiness: Summary Print, Dordrecht Holland, Reldel.
Veenhoven, Ruut (1989): How Harmful is Happiness? Consequences of Enjoying Life or Not, Rotterdam, Universitaire Pers Rotterdam.
Veenhoven, Ruut, (1993): Happiness in Nations: Subjective Appreciation of Life in 56 Nations 1946-1992, Rotterdam, Erasmus University, RISBO.
Verkley, Harry, and Stolk, Jan (1989): `Does happiness lead into idleness?' In Veenhoven, Ruut (ed.) How Harmful is Happiness? Consequences of Enjoying Life or Not, Rotterdam, Universitaire Pets Rotterdam, 79-93.
Warr, Peter (1978): `A study of psychological well-being', British Journal of Psychology, 69, 111-21.
Warr, Peter, Jackson, Paul, and Banks, Michael (1988): Unemployment and mental health: Some British studies', Journal of Social Issues, 44, 47-68.
Weaver, R. Kent, Shapiro, Robert Y, and Jacobs, Lawrence R. (1995): `Public opinion on welfare reform: A mandate for what?' In Dickens, William and Weaver, R. Kent (ed.) Looking Before We Leap: Social Science and Welfare Reform, Washington, DC, The Brookings Institution.
Wlezien, Christopher. (1995): `The public as thermostat: Dynamics of preferences for spending', American Journal of Political Science, 39, 981-1000.
John S. Lapinski, Charles R. Riemann and Matthew F. Stevens are doctoral students in political science at Columbia University. Robert Y. Shapiro is a professor of political science at Columbia University and associate director of the Paul F. Lazarsfeld Center for the Social Sciences. Lawrence R. Jacobs is an associate professor of political science at the University of Minnesota.
Correspondence should be addressed to J. S. Lapinski at Department of Political Science, Columbia University, 420 W. 118th Street, New York, NY 10027, USA.
(1) Our study uses single measures to tap `happiness' or life satisfaction. In contrast, other studies of the relationship between employment and personal well-being, sometimes use multiple measures to construct well established psychological scales that measure individuals' well-being or life satisfaction or `happiness' (see Clark and Oswald 1994, Juster and Courant 1986). The use of single measures may have produced results that are less valid and reliable than desired. However, since most of our findings are quite consistent and explicable in the context of past research that has made use of certain better measures, we are confident about the robustness of our results and analyses. Further, in the case of the NORC General Social Surveys for the United States, our results for the influences on general `happiness' as measured by responses to a single survey question were compared with those for a multi-item scale measuring overall fife `satisfaction'. The latter scale (not reported here, but available from the authors) is highly reliable (Cronbach's alpha reliability coefficient of .72) and is correlated with the individual level responses to the happiness item at .45, which is high correlation given the normal degree of measurement error found in these kinds of survey data). Our general as well as specific findings for both the happiness measure and the satisfaction scale are very similar, indicating that our findings and conclusion based on these data are robust and reliable.
(2) The independent variables used for the multivariate analyses were derived from questions posed in the EB, ISSP, and GSS. Variables were coded as similarly as possible across surveys: (age) actual age of respondent; (agesq) age of respondent squared; (bluecollar) EB and ISSP 1 = blue collar, 0 = not blue collar; (child) number of children living in household; (divorced) 1 =divorced and 0 =not divorced; (educ) ISSP and GSS respondents asked how many years of education were completed recoded 1 = 8 years or less; 2 = 9-11 Years; 3 = 12 Years; and 4 = 13 years or more; Eurobarometer surveys asked respondents how old they were when they finished their education. While this variable is not a perfect proxy for respondent education, it was the best variable available across EB surveys; (income) coded into quartiles 1 =lowest quartile and 4 = highest quartile; (married) dichotomous coding 1 = married and 0 = not married; region(s) based upon census division in the GSS; each regional dummy variable refers to a specific `regional administrative unit' defined by the administrators of the survey. Regional variables were exclusively for control purposes; (sex) dichotomous coding 1 = male, 0 = female; (separate) dichotomous coding 1 = separated from spouse and 0 = not separated; (unemp) dichotomous coding 1 =unemployed (looking for work) and 0 =employed, not looking for work; (union) dichotomous coding 1 = member of trade union and 0 = not member.
(3) The multivariate analyses, linear regression and ordered probit models, are quite complex to present and can at times be difficult to interpret fully and concisely. Interpreting the magnitude of the effect is more complicated for probit models than for linear regression models, for which the regression coefficient represents the effect on the estimated mean of the dependent of a one unit or category change in the independent variable. Probit coefficients are more appropriate when the dependent variable is dichotomous or has more than two ordered categories; in these cases it is technically not appropriate to estimate effects on the mean of the dependent variable but rather the probability of being in a particular category of the dependent variables. The coefficients themselves are not directly interpretable as changes in probabilities resulting from one unit or one category shifts in each independent variable (for a more extensive discussion of these nonlinear multivariate methods, see Greene 1993 and Liao 1994).
(4) Other data (not reported) show that the unemployed's support of the government's role in providing for the unemployed decreased. For example, support of the unemployed within the USA drops over 27 percent points between 1985 and 1990. Another interesting finding is that the gap between the employed and unemployed became smaller between 1985 and 1990. Also approval for the government's role in caring for the unemployed is strongest among those with lower incomes, with support of over 63 percent among the lowest income quartile in all countries. With regard to the role of government in providing a job to everyone who wants one, the unemployed and those with lower income show stronger support than the employed. In Great Britain and Austria, over 75 percent of the unemployed believe it is the government's responsibility to provide employment. Not surprisingly, much lower support exists in the USA and Australia. In 1900 approximately half of the unemployed favored such a position. All income groups in Great Britain showed high support for government financing projects to create jobs, with both the bottom and top quartiles in 1985 and 1990 showing support levels of at least 79 percent. In the USA and Australia, support was much lower among those in the highest income group. Support among the lowest income quartile was quite high with at least 72 percent believing that the government should finance new projects.
|Printer friendly Cite/link Email Feedback|
|Author:||Lapinski, John S.; Riemann, Charles R.; Shapiro, Robert Y.; Stevens, Matthew F.; Jacobs, Lawrence R.|
|Publication:||International Journal of Public Opinion Research|
|Date:||Mar 22, 1998|
|Previous Article:||Behind the Oval Office: Winning the Presidency in the Nineties.|
|Next Article:||Medium or message? Predicting dimensions of political sophistication.|