The Journal of Socio-Economics 37 (2008) 19061920

Contents lists available at ScienceDirect

The Journal of Socio-Economics

journal homepage: www.elsevier.com/locate/soceco

Jobs as Lancaster goods: Facets of job satisfaction and overall job satisfaction
Ali Skalli a , Ioannis Theodossiou b, , E Vasileiou a
a b

LEM, Universit e Panth eon-Assas, Paris 2, France Centre for European Labour Market Research, Aberdeen University Business School, Economics, University of Aberdeen, Aberdeen, UK

a r t i c l e

i n f o

a b s t r a c t
Overall job satisfaction is likely to reect the combination of partial satisfactions related to various features of ones job, such as pay, security, the work itself, working conditions, working hours, and the like. The level of overall job satisfaction emerges as the weighted outcome of the individuals job satisfaction with each of these facets. The purpose of this study is to determine the extent and importance of partial satisfactions in affecting and explaining overall job satisfaction. Using the European Community Household Panel (ECHP) a two layer model is estimated which proposes that job satisfaction with different facets of jobs are interrelated and the individuals reported overall job satisfaction depends on the weight that the individual allocates to each of these facets. For each of the 10 countries examined, satisfaction with the type of the job is the main criterion by which workers evaluate their job for both the short and the long term. 2008 Elsevier Inc. All rights reserved.

Article history: Received 5 January 2007 Received in revised form 25 February 2008 Accepted 6 April 2008 JEL classication: C23 C25 I31 Keywords: Overall job satisfaction Earnings Working conditions Working time Job security Type of work

1. Introduction Job satisfaction is an important, readily available measure of the workers utility derived from the job. It allows the identication of those characteristics which have a differential impact on the workers utility. This is important since higher job satisfaction is likely to result in higher performance at work, decreased absenteeism and tardiness (Lawler and Porter, 1967; Locke, 1969; Hamermesh, 1977; Freeman, 1978; Borjas, 1979). There are a number of empirical studies which investigate the effect of individual and job characteristics on job satisfaction1 although the effect of any individual or job characteristic on workers job satisfaction is also dependent on a number of features like institutions or social norms (Sousa-Poza and Hennerger, 2000).2 Fischer and Sousa-Poza (2006)

Corresponding author. E-mail address: theod@abdn.ac.uk (I. Theodossiou). 1 Recent studies include Sousa-Poza and Hennerger (2000), Sousa-Poza and Sousa-Poza (2000a,b), Fischer and Sousa-Poza (2006), Blanchower et al. (2001), Huang and Vliert (2003a,b, 2004), Medgyesi and Robert (2003), Stier and Lewin-Epstein (2003), Ahn and Garcia (2004), Bockerman (2004), Deloffre and Rioux (2004), De Witte et al. (2004), Hui et al. (2004), Sweeney and McFarlin (2004), Clark (2005) Clark and Senik (2006), Clark and Postel-Vinay (2008), Green and Tsitsianis (2005), Kaiser (2007), Pouliakas and Theodossiou (2005), Bender et al. (2006) Diaz-Serrano and Vieiri (2005) and Davoine and Erhel (2006). 2 Sousa-Poza and Sousa-Poza (2000a,b) analysed job satisfaction on the assumption that it depends on the balance between work-role inputs (education, working time, effort) and work-role outputs (wages, fringe benets, status, working conditions, intrinsic aspects). Thus, if work work-role outputs (pleasures) increase relative to work-role inputs (pains), then job satisfaction will increase. 1053-5357/$ see front matter 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.socec.2008.04.003

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

1907

nd a positive relation between self-reported health measures and job satisfaction. Clark and Oswald (1996) and Groot and van den Brink (1999) establish that there is a U-shaped relationship between age and job satisfaction. Studies show that job satisfaction increases with wage (Lydon and Chevalier, 2002). The effect of gender on job satisfaction is not unambiguous. Clark (1997) reports that British females are more satised than their male counterparts and Bender et al. (2006) show that this is due to the fact that female workers are able to be in jobs that offer exibility. Yet, Kaiser (2002) and Moguerou (2002) report that females appear to be more satised than their male counterparts in a number of continental European countries and in the U.S. Clark and Oswald (1996) and Drakopoulos and Theodossiou (1997) show that the level of job satisfaction diminishes as the number of working hours increases. Drakopoulos and Theodossiou (1997) and Sloane and Williams (2000) show that workers who work in small rms report higher job satisfaction levels compared to those working in large rms. This implies that working in a small unit offers more job control and less repetitive tasks. However, satisfaction with job security increases with rm size (Idson, 1996; Lang and Johnson, 1994). Kaiser (2002) shows that the effect of occupational hierarchy on job satisfaction is country-specic. Davoine (2005) nds that the effect of education on job satisfaction is also country specic. The fundamental assumption of the literature reviewed above is that individuals make a judgement about their job as a whole. Yet, research has shown that job satisfaction also depends on the aspects involved in the job tasks (Warr, 1999; Frey and Stutzer, 2002). Thus, like other emotional judgments, job satisfaction arises from a variety of feelings related to the characteristics of the job, like feelings about the working conditions, about the level of earnings, about the risk of losing the job, about the opportunity for personal control and so on. Satisfaction with the level of earnings is not the same as satisfaction with job security, which is not the same as satisfaction with working conditions, but all are components of overall job satisfaction that occupy different points on the scale of satisfaction. Hence, overall job satisfaction can be viewed as a weighted outcome of the individuals satisfaction with each one of the aspects or facets of the job. This is similar to Lancasters theory of consumption behaviour (Lancaster, 1966, 1971) where the utility that is derived from consuming a given good depends on the utilities that are associated with its characteristics. This view postulates that individuals do not simply make a judgement about their job as a whole but rather that their level of overall job satisfaction is a combination of different levels of satisfaction with the different characteristics of the job. This implies that there is the possibility that individuals may value differently the different aspects of the job, intrinsic of extrinsic, so an individual may remain equally satised with her or his job when certain aspects of job satisfaction change, provided that there is an accompanied compensating change of other aspect or aspects of the job. Hence, the same stated job satisfaction level can be obtained through different combinations of job facets reecting satisfaction with intrinsic and extrinsic features of the job. In this view, each job is thought to have a number of aspects or facets from which utility is derived. The different mix of such aspects leads to a differentiated job. Thus, two different mixes of characteristics for the same job may be viewed by the worker as equally attractive, provided that a low content in one desirable aspect is compensated by an increase in another. By considering the mix of properties, the various qualities of individual jobs can be incorporated into the analysis. The same stated job satisfaction level can be obtained through different combinations of job facets reecting various features of the job. This approach to investigating job satisfaction is particularly well suited to address a number of important issues such as the effects of the major changes of work organisation that rms have experienced during the last three decades on job satisfaction. These changes have had an impact on pay practices, job contents, working conditions and environment and job security (Lindbeck and Snower, 1996; Blanchower and Oswald, 1999; Aaronson and Sullivan, 1998; Nickell et al., 2002). This paper explores the relationship between overall job satisfaction and satisfaction with important aspects of the work environment which are linked to organisational changes in the workplace. Hence, it does not focus on evaluating the effect of different individual and job characteristics on the individuals job satisfaction with the different aspects of the job or identifying national differences. This is dealt with by the plethora of studies some of which are reviewed above. This study focuses on evaluating the effect of satisfaction of various facets such as satisfaction with the type of work, satisfaction with working conditions, with working time, with job insecurity and with earnings on the overall job satisfaction. It develops a two-layer model3 where job satisfaction is viewed as an aggregate concept consisting of different components or facets. The results show that satisfaction with different facets of jobs are interrelated and the reported overall job satisfaction depends on the weight which is attached to each of these facets by the individual worker. The econometric methodology accounts for unobserved sources of individual heterogeneity and for time-specic effects. The remainder of the paper is organised as follows: Section 2 discusses the estimation methodology; Section 3 discusses the data used in this study; Section 4 presents the results and their interpretation; and Section 5 concludes. 2. The empirical model Assume that the number, L, of job facets signicant to overall job satisfaction S is nite. Let sl , l = 1, . . ., L, denote satisfaction with respect to each job facet. It is assumed that S can be explained by the satisfaction levels, sl , with respect to all job facets, l =1, . . ., L. Suppose that the researcher observes a set of K individual and job characteristics, xk , k = 1, . . ., K that are potential determinants of the levels of the job satisfaction facets, sl , l =1, . . ., L.

This is similar to Van Praag et al. (2002) and Van Praag and Ferrer-i-Carbonell (2004).

1908

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

The modelling strategy should account for this interdependence since an individuals level of overall job satisfaction depends on her/his level of satisfaction with each of the job facets which in turn depends on a number of exogenous explanatory variables. This can be done within a two-layer model framework of the type: S = f (s1 , s2 , . . . , sL ), s = (s , s , . . . , s
l 1 2 l1

(1a) ,s
l+1

, . . . , s , x1 , x2 , . . . , xK ),

l = 1, . . . , L. sl

(1b) in the set of (2a)

A reduced-form from which such interactions have been eliminated could be obtained by replacing each partial satisfaction Eq. (1b) by its corresponding expression as a function of xk , k = 1, . . ., K. This yields: S = f (s1 , s2 , . . . , sL ), s = (x1 , x2 , . . . , xK ),
l

l = 1, . . . , L.

(2b)

Which is an L + 1-equation model that describes the way satisfaction levels with respect to different facets of a job interact. For instance, the satisfaction level that an individual might report with respect to working conditions is likely to be conditional on how satised she/he is with respect to earnings and vice versa. However, estimation of Eqs. (2a, 2b) requires taking into account of a number of specication issues. Although there are a large number, L, of job satisfaction facets likely to inuence overall job satisfaction, only a few, say J < L, of them are observed in the data. This implies that (i) not all partial satisfaction equations in (2b) can be estimated and (ii) the reduced form of the model should also account for the fact that for the L J unmeasured partial satisfaction levels, only their determinants can be taken into account in Eq. (2a). The system should therefore be written as: S = g (s1 , s2 , . . . , sJ , x1 , x2 , . . . , xK ), s = (x1 , x2 , . . . , xK ),
j

(3a) (3b)

j = 1, . . . , J < L.

However, the estimation of system (3) is likely to result in biased estimates as there are two important sources of unobserved individual heterogeneity. The rst relates to satisfaction with respect to L J job facets which are indirectly measured via their observed determinants. The second potential source of bias is of a general nature. Even if satisfaction with all job facets could be measured, a large number of unobserved determinants of partial job satisfaction levels would remain unaccounted for. Regardless how large is the set of observable determinants, xk , k = 1, . . ., K, included in the model, the model cannot account for all the determinants of ones level of satisfaction. Let yh , h = 1, . . ., H denote the set of unobserved individual and job characteristics which have an inuence on either the measured or the unmeasured partial satisfaction levels. Some, if not all, of these determinants would also explain the level of overall job satisfaction. Given the above, Eqs. (3a, 3b) can be written as: S = g (s1 , s2 , . . . , sJ , x1 , x2 , . . . , xK , y1 , y2 , . . . , yH ), s = (x1 , x2 , . . . , xK , y1 , y2 , . . . , yH ),
j

(4a) (4b)

j = 1, . . . , J < L.

The simultaneous nature of Eqs. (4a, 4b) is a source of bias. The partial satisfaction Eq. (4b) imply that the partial satisfaction levels, sj , j = 1, . . ., J, are correlated with the unobserved variables yh, h = 1, . . ., H. Hence, if the yh variables are left in the error term of the overall satisfaction Eq. (4a), any estimate of the effect of partial satisfaction levels on the overall job satisfaction would suffer from endogeneity bias. To overcome this problem the methodology of Van Praag et al. (2002) is used. First, the J partial satisfaction equations in (4b) are estimated and the corresponding residuals are calculated. These are then used to derive a surrogate variable reecting the part of individual unobserved heterogeneity attributable to the variables, yh, h = 1, . . ., H; that is, the part common to all the residuals. The surrogate variable is derived as the rst principal component of the J J error covariance matrix of the residuals retrieved after estimating the auxiliary partial satisfaction Eq. (4b). After the inclusion of the surrogate variable in the overall satisfaction Eq. (4a), one can reasonably assume that the remaining error in that equation is no longer correlated with the partial satisfaction variables sj , j = 1, . . ., J. Furthermore, the rst principal component of the J J error covariance matrix of the residuals from the J Eq. (4b) is expected to include a large part of the L L error covariance matrix of the unobserved heterogeneity due to satisfaction with the unobserved job facets. The latter assertion stems from the fact that satisfaction with earnings is included in the set of the observable job satisfaction facets. The theory of compensating differentials (Rosen (1986)) predicts that earnings act as a utility equalising mechanism so that satisfaction with earnings is likely to be highly correlated with a number of job attributes. Hence, the residuals from the satisfaction-with-earnings equation include information about satisfaction with other job attributes which are unobservable. The categorical nature of the dependent variables and the simultaneous nature of the system (3) complicate the estimation procedure. To deal with this, the approach chosen in this study is as follows: The overall satisfaction equation has ordered qualitative variables on both sides of the equality. Although this poses no particular econometric problem, it implies that there should be a large number of dummy variables on the right-hand side.4

If the number of satisfaction levels is, say, m, then the number of dummy variables that should be included is J(m 1).

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

1909

This makes the overall satisfaction equation computationally unattractive and yields results that are not easily interpretable. Freeman (1978) shows that a categorical variable (in this case job satisfaction) can be linearised by transforming it to z-scores. This transformation has computational advantages for the approach used in this study. Hence, the overall job satisfaction S and the partial satisfaction variables, sj , j = 1, . . ., J are rescaled according to the standard normal distribution. With this unit transformation, the z-scores overall job satisfaction and partial job satisfactions measure the number of standard deviations between a given response and the mean. The longitudinal dimension of the data is exploited in two ways. First, Eqs. (4a), (4b) are estimated by individual random effects by controlling for time xed effects through a set of year dummy variables. However, this may be problematic since the individual random effect model may overlook the potential correlation between individual random effects, such as innate ability, with some explanatory variables, such as wage or household income. To circumvent this problem, the Mundlak (1978) methodology is applied. This approach amounts to decomposing each of the disturbance terms into a sum of a zero-mean individual random effects term and a zero-mean pure error term. To illustrate the procedure, consider that the following random effect model: zit = at +
H h=1

bh vhit + i + it ,

i = 1, 2, . . . , N,

t = 1, 2, . . . , T

(5)

where the individual random effect, i is likely to be correlated with some, if not all, the explanatory variables vh . Suppose that the correlation takes place only through the long-term components of the vh variables and that these can be captured hi variables over time. In this case instead of estimating Eq. (5), one could estimate the following by the average of the v specication: zit = at +
H h=1 H

bh vhit +

H h=1

hi + i + it , ch v

i = 1, 2, . . . , N,

t = 1, 2, . . . , T

(6)

is such that E(i vhi ) = 0, h = 1, . . ., H. This is applicable to the explanatory variables which show c v where i = i h=1 h hi signicant variability over time. The above procedure introduces some dynamics in the model by distinguishing between permanent and transitory effects for important explanatory variables xk , k = 1, . . ., K, in each of the auxiliary Eq. (4b), and for the partial satisfaction variables sj , j = 1, . . ., J, in Eq. (4a). Following Van Praag et al. (2002), it can be shown that for any exogenous variable, xk , to include in the ki ) + (bk + ck ) ki is equivalent to including bk (xkit x right-hand side of the estimated equation the expression bk xkit + ck x xki . This allows explicit decomposition of the effect of a variable xk into two distinct effects. Differences across individuals in ki , measure ki measure between effects and the individual deviations from the averages per individual xkit x the averages x within effects. The coefcients bk reect short term or transitory effects and bk + ck measure long term or permanent effects. 3. The data This study uses the eight waves (19942001) of the European Community Household Panel (ECHP) which has a unique structure since it offers data for the European Union countries derived from identical questionnaires across counties. The ECHP provides information on a variety of socio-demographic and job characteristics of the respondents. Crucially for the purpose of this study it offers information on the individuals evaluation about his or her work environment based on job satisfaction ratings. Indicators of the overall job satisfaction and of the satisfaction with ve job facets are derived from the following questions: (1) (2) (3) (4) (5) (6) How satised are you with your work or main activity? How satised are you with your present job in terms of earnings? How satised are you with your present job in terms of job security? How satised are you with your present job in terms of type of work? How satised are you with you present job in terms of working times (day time, night time, shifts etc)? How satised are you with your present job in terms of working conditions/environment?

The answers are ranked in 6-level scale from 1 (completely dissatised) to 6 (completely satised). Due to data limitations the analysis is restricted to 10 countries (Austria, Belgium, Denmark, Finland, France, Greece, Italy, The Netherlands, Portugal and Spain).5 The self-employed are excluded and the sample is restricted to employees of working age (1765). The 10 European countries are studied separately in order to identify potential national similarities and differences for the effect of each facet of job satisfaction on the overall job satisfaction. However, rst, the satisfaction with the job facets (in Eq. (4b)) need to be estimated. Hence, a number of key variables identied in the literature are used to explain partial job satisfactions. In particular, along with a set of personal characteristics (such as gender, marital status, experience, education,

5 The ve remaining countries included in the ECHP dataset (UK, Ireland, Sweden, Luxembourg, Germany) cannot be used in the analysis, due to the data limitations on the satisfactions questions. The questions on the ve facets for these countries are not available.

1910

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

health and past unemployment experience), the establishment size, private/public sector, industrial sectors, occupation, the type of contract (permanent and temporary), the personal labour income and the equivalised household income6 are used as regressors. Table A1 details the denitions of these variables. Table A2 reports the means of the Overall Job Satisfaction and the Facet Satisfactions of the eight waves of the ECHP (19942001) for the 10 European countries.7 Comparison of the unweighted averages across countries, show that workers in northern Europe report both higher overall job satisfaction and for each single facet of job satisfaction compare to those living in the Mediterranean countries (Greece, Portugal, Italy and Spain). Workers in countries such as Denmark and Austria report a noticeably higher average on almost every partial job satisfaction compared to the remainder. Interestingly, for every European country except in the Netherlands, the lowest scores are associated with satisfaction with earnings. These differences among European countries may reect differences in the general macroeconomic environment or, as Deloffre and Rioux (2004) argue, such cross-national differences may reect differences in cultural background. The econometric evidence in this paper shows that an important part of these differences may be attributed to the different way that workers value the different facets of their jobs across countries.

4. The relationship of overall job satisfaction with the satisfactions derived from the job facets The estimation of the ve auxiliary Eq. (4b) corresponding to the ve facets of job satisfaction (satisfaction with earnings, with job security, with the type of work, with working conditions and with working times) are reported in Tables A3A7, but they are not discussed as they are auxiliary to the focus of this paper. Their specication is based on the literature reviewed in section I and the results are in line with this literature. Furthermore; although there are important similarities as to the role of some of the determinants of partial facets of job satisfaction across countries, there are also notable differences.8 The Mundlak approach is applied only on variables of interest, namely on log gross hourly wages and log gross equivalised household income, as not all of the explanatory variables show signicant variability over time. Table 1 reports the results from the estimation of the relationship between overall job satisfaction and satisfaction with important aspects of the work environment which are linked to organisational changes in the workplace. Overall, the ndings suggest that jobs do not appear to be one-dimensional. Job satisfaction is a multi-dimensional which is compatible with the view that there are partial satisfactions arising from different aspects of the job that occupy different points on the scale of satisfaction. The overall job satisfaction or utility derived from a job is an aggregation of these partial satisfactions. Different mix of facets of the job may generate the same overall level of job satisfaction. The results also show that endogeneity is indeed a crucial issue. This is evidenced by the fact that the surrogate instrumental variable (IV)9 used to account for endogeneity bias is statistically highly signicant for all countries with the exception of the Netherlands and Spain. Satisfaction with each of the ve job facets investigated in this study is a highly signicant determinant of the overall satisfaction for all ten countries.10 The Mundlak transformation captures the transitory and permanent effects. The transitory or short-term effects appear to be different than the long term or permanent effects. In the simple application of this study, the transitory effect is approximated by the coefcients on each of the ve included partial satisfaction measures whereas the permanent effect is approximated by the coefcients on the mean satisfaction of the respective facet. The interpretation of the coefcients is straightforward. For example, for Austria, if transitory satisfaction with earnings increase by one percentage point there is a 37% increase in individual overall satisfaction. In addition, a one-percentage point increase in permanent satisfaction with earnings causes an additional 7% increase in individual overall satisfaction. However, if transitory satisfaction with job security increase by one percentage point there is a 25% increase in individual overall satisfaction whereas a one-percentage point increase in the permanent satisfaction with job security causes a 0.8% decrease in individual overall satisfaction. The results show that the transitory effect is always an important contributor to overall job satisfaction. However, the permanent effect does not turn out to be always statistically signicant. Thus, in the case of satisfaction with earnings and satisfaction with the type of work, both transitory and permanent satisfaction effects are signicant for all countries with the exception of Denmark and the Netherlands. This highlights the importance of these two aspects of the job in shaping the individuals overall job satisfaction. In contrast, the permanent effect of the remaining job facets on overall satisfaction does not appear to be as important for all the countries studied. In particular, the coefcients on mean satisfaction with job

6 The equivalised household income takes into account the family structure. It is used by several authors on the job satisfaction literature (for example Van de Stadt et al., 1985). 7 Note that Austria and Finland have entered the ECHP survey in 1995 and 1996, respectively. 8 Yet, the interpretation of these differences is difcult in the absence of specic knowledge on the potential causes. 9 Recall that this surrogate instrumental variable is derived as the rst principal component of the residual vectors from the partial satisfaction Eq. (4b). The rst component explains 4252% of the total variance depending on the country examined. 10 Since individuals approaching retirement might be more inclined to retire if they are unsatised with their jobs, the estimates may be sensitive to the inclusion or exclusion in the sample of the older workers. To investigate this issue the model is also estimated for the 1755 age group. The results and the ranking of the job satisfaction facets turn out to be remarkably robust to this. The results are available from the authors on request.

Table 1 Overall job satisfaction (ordered probit technique with individual random effects and xed time effects) Coefcient Austria Satisfaction with earnings Satisfaction with job security Satisfaction with type of work Satisfaction with working conditions Satisfaction with working times Mean satisfaction with earnings Mean satisfaction with job security Mean satisfaction with type of work Mean satisfaction with working times Mean satisfaction with working conditions Year dummies IVa log likelihood Observations 0.371*** 0.228*** 0.536*** 0.384*** 0.210*** 0.070*** 0.008 0.258*** 0.003 0.025 Yes 0.180*** 17,346 18,154 Belgium 0.252*** 0.100*** 0.570*** 0.362*** 0.155*** 0.134*** 0.043 0.362*** 0.002 0.005 Yes 0.154*** 15,358 13,350 Denmark 0.191*** 0.079*** 0.573*** 0.150*** 0.103*** 0.008 0.017 0.306*** 0.026 0.028 Yes 0.097*** 19,840 20,338 Finland 0.304*** 0.082*** 0.509*** 0.222*** 0.096*** 0.058** 0.003 0.369*** 0.010 0.083*** Yes 0.080*** 17,404 17,504 France 0.285*** 0.222*** 0.480*** 0.324*** 0.438*** 0.048*** 0.026 0.331*** 0.040** 0.021 Yes 0.186*** 27,680 27,021 Greece 0.430*** 0.456*** 0.596*** 0.255*** 0.173*** 0.040* 0.034 0.169*** 0.064*** 0.060*** Yes 0.104*** 22,258 20,040 Italy 0.386*** 0.270*** 0.801*** 0.213*** 0.127*** 0.040** 0.011 0.344*** 0.056*** 0.022 Yes 0.147*** 42,721 36,750 The Netherlands 0.130*** 0.076*** 0.351*** 0.168*** 0.134*** 0.023 0.039** 0.178*** 0.043** 0.045** Yes 0.040 26,381 25890 Portugal 0.314*** 0.363*** 0.732*** 0.288*** 0.089*** 0.126*** 0.013 0.272*** 0.009 0.038* Yes 0.240*** 29,931 33,419 Spain 0.273*** 0.201*** 0.658*** 0.137** 0.158*** 0.050*** 0.037** 0.238*** 0.052*** 0.021 Yes 0.010 40,996 34,438

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920 1911

Note: *, **, *** indicate signicant improvement at 10, 5, 1% levels, respectively. a Recall that this surrogate instrumental variable is derived as the rst principal component of the residual vectors from the partial satisfaction Eq. (4b). The rst component explains 4252% of the total variance depending on the country examined.

1912

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

Table 2 Level Effects of Satisfaction with Job Facets on Overall Job Satisfaction Country Satisfaction with Earnings Austria Belgium Denmark Finland France Greece Italy Netherlands Portugal Spain 0.441 0.386 0.199 0.362 0.333 0.470 0.426 0.153 0.44 0.323 (Rank) (2) (2) (2) (2) (4) (3) (2) (4) (2) (2) Job security 0.22 0.143 0.096 0.079 0.196 0.490 0.281 0.115 0.376 0.164 (Rank) (4) (5) (4) (5) (5) (2) (3) (5) (3) (3) Type of work 0.794 0.932 0.879 0.878 0.811 0.765 1.145 0.529 1.004 0.896 (Rank) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) Working conditions 0.409 0.357 0.178 0.139 0.345 0.195 0.191 0.213 0.326 0.158 (Rank) (3) (3) (3) (3) (3) (4) (4) (2) (4) (4) Working times 0.213 0.157 0.077 0.106 0.398 0.109 0.071 0.177 0.098 0.106 (Rank) (5) (4) (5) (4) (2) (5) (5) (3) (5) (5)

Note: ranks in parentheses.

security are signicant only in France, the Netherlands and Spain, on mean satisfaction with working times only in Greece, Italy and Spain and on mean satisfaction with working conditions only in Finland and Greece. In general, transitory effects are much larger than the permanent effects. This result indicates that in most cases the transitory effects have a sizeable and important impact on the overall job satisfaction. The transitory effects of satisfaction with earnings and with the type of work are systematically higher compared to the corresponding permanent effects. In addition, the sign of the coefcients show that higher satisfaction with earnings and/or higher satisfaction with the type of work yields higher overall job satisfaction both in the long and in the short term. This is not the case of satisfaction with the other job characteristics. The transitory effects on overall job satisfaction of satisfaction with job security, working conditions and working times are systematically positive whereas their respective permanent effects are negative (when they are statistically signicant). This implies that although the transitory effect of satisfaction with the above facets contributes positively to overall job satisfaction in some countries, any lasting impact of these aspects on overall job satisfaction slowly diminishes with time as workers adapt to the workplace or as these aspects of work deteriorate with time. The results show that the permanent effect of satisfaction with working times declines in Greece, Italy and Spain and the permanent effect of satisfaction with working conditions declines in Greece and Finland. Table 2 reports the cumulative transitory and permanent effects of satisfaction with each facet on the overall job satisfaction. These cumulative effects are obtained by summing up the transitory and permanent components for each satisfaction facet. In Table 2, the numbers in parentheses report the satisfaction ranking of the specic job facet according to the importance attached by individuals of each country to this job facet in terms of the cumulative satisfaction effect. It is shown that regarding the weight that workers attach to each of the ve job facets, there are cross-country differences. This suggests that cross-country differences in the overall job satisfaction levels are formed by the importance that workers attach to each of the characteristics of their job which, in turn, depends on the countrys economic, institutional and cultural environment. A striking feature of the reported results is that satisfaction with the type of work is revealed to be the most important determinant of overall job satisfaction for all ten countries. This result seems to emphasise the value which workers attach to some intrinsic attributes of the job since, clearly, the pecuniary aspects of the job do not rank as having the highest importance in forming individuals job satisfaction in any of the ten countries. These ndings are in line with a number of studies (Frey and Stutzer, 2002; Ryan et al., 1996; Bender et al., 2006) which show that workers are primarily concerned with the type of work which they are contracted to perform. Occupational psychology literature shows that workers are interested in the non-monetary features of the job such as autonomy, the degree of skill utilisation, the challenge in performing job tasks and the like. Lawler and Porter (1967) argue that satisfaction with some characteristics, such as extrinsic aspects, satisfy mainly lower level needs whereas satisfaction with other aspects, such as intrinsic characteristics, satises higher order needs. According to Frey and Stutzer (2002), the fact that workers value satisfaction with the work itself so highly can also explain why many people undertake unpaid work (volunteer work and charity). However, Table 2 shows that although earnings are not ranked rst in any of the countries, they are ranked second in most, except in Greece where they are ranked third and in the Netherlands and France where they are ranked fourth. In the southern European countries (Spain, Portugal, Italy, and Greece) satisfaction with job security in terms of low risk of job loss is ranked higher compared to the Northern European countries reecting either the higher likelihood of job loss in the former countries or the nature of the welfare institutions in the latter countries which provide a comprehensive safety net for the unemployed workers. Thus, in Greece, satisfaction with earnings ranks third after satisfaction with job security, a result that mirrors the importance which individuals attach to having a job in a country with relatively high unemployment rates and a limited safety net. In contrast, in France, satisfaction with working conditions and with working schedules rank higher compared to satisfaction with earnings, but job security appears to be the least important job facet. These results reect the current labour market conditions in France; a country with relatively high unemployment rates, where employment

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920 Table A1 Denitions of the variables used Variables Job satisfaction Satisfaction with earnings Satisfaction with job security Satisfaction with working conditions/environment Satisfaction with type of work Satisfaction with working times Male Married Lower secondary Upper secondary Tertiary Experience Supervisor Intermediate Non-supervisory position Manager Professional Technicians Clerks Sales worker Skilled agriculture worker Craft worker Machine operators Elementary occupations Private Firm size 20 Firm size 100 Firm size 500 Firm size 500+ Agriculture Manufacturing Services Permanent contract Fixed term contract Good health Unemployment spells lnwage lnhousinc Mean wage Mean income Note: the frequencies of the above variables appear in Table A8. Denitions

1913

Standardised score of satisfaction with the job or main activity Standardised score of satisfaction with earnings Standardised score satisfaction with job security Standardised score of satisfaction with working conditions or environment Standardised score of satisfaction with the type of work Standardised score of satisfaction with working times Dummy variable with value 1 for male workers Dummy variable with value 1 for married workers Dummy variable with value 1 for workers whose highest qualication is from primary or lower secondary education Dummy variable with value 1 for workers whose highest qualication is from upper secondary education Dummy variable with value 1 for workers whose highest qualication is from tertiary education Potential labour market experience Dummy variable with value 1 if the respondent has a supervisory position Dummy variable with value 1 if the respondent has an intermediate position Dummy variable with value 1 if the respondent has a non-supervisory position Dummy variable with value 1 for managers, legislators and senior ofcials Dummy variable with value 1 for professionals Dummy variable with value 1 for technicians and associate professionals Dummy variable with value 1 for clerks Dummy variable with value 1 for service, shop and market sales workers Dummy variable with value 1 for skilled agricultural and shery workers Dummy variable with value 1 for craft and related trades workers Dummy variable with value 1 for plant and machine operators Dummy variable with value 1 for elementary occupations Dummy variable with value 1 for private sector employees Dummy variable with value 1 if employer size is less than 20 regular paid employees Dummy variable with value 1 if employer size is greater than 20 and less than 100 regular paid employees Dummy variable with value 1 employer size is greater than 100 and less than 500 regular paid employees Dummy variable with value 1 employer size is greater than 500 regular paid employees Dummy variable with value 1 for workers in the agricultural sector Dummy variable with value 1 for workers in the industrial sector Dummy variable with value 1 for workers in the sector of services Dummy variable with value 1 for workers with a permanent contract Dummy variable with value 1 for workers with a xed-term contract Dummy variable with value 1 if the respondent has reported a good or a very good health status Number of unemployment spells during the ve years before the individual joined the survey log of CPI-deated wage log of CPI-deated equivalised household income Mean wage over the eight survey years Mean household income over the eight survey years

security has decreased between 1985 and 1995 (OECD, 1997), but in which there are comprehensive employment protection practices that largely mitigate the workers feeling of job insecurity (Deloffre and Rioux, 2004). An additional feature of Table 2 is that it allows the calculation of the marginal rates of substitution or trade off ratios between the levels of satisfaction associated with different job facets. The trade off ratios show by how much the satisfaction with a job aspect has to improve when the satisfaction with some other job facet declines, in order for the worker
Table A2 Means of overall and partial job satisfaction in ECHP (19942001) Mean France Overall Job satisfaction Satisfaction with earnings Satisfaction with security Satisfaction with type of work Satisfaction with working times Satisfaction with working conditions 4.40 3.54 4.17 4.58 4.32 4.23

Greece 3.88 3.26 4.03 4.03 4.03 3.92

Netherlands 4.75 4.39 4.65 4.84 4.84 4.33

Spain 4.26 3.25 4.14 4.34 4.13 4.23

Denmark 4.96 4.33 4.79 4.9 5.00 4.81

Belgium 4.49 3.94 4.42 4.65 4.61 4.43

Italy 4.05 3.31 4.08 4.24 3.99 4.02

Portugal 3.95 3.23 3.99 4.17 4.05 4.18

Austria 4.48 4.10 4.94 5.09 4.99 5.05

Finland 4.58 3.92 4.40 4.49 4.61 4.47

1914

Table A3 Satisfaction with earnings (GLS z-score with individual random effects and xed time effects) Coefcient Austria Male Lower secondary Upper secondary Private Firm size 100 Firm size 500 Firm size 500+ lnwage Mean wage lnhousinc Mean income Intercept R2 : within R2 : between R2 : overall No.: obs. 0.166*** 0.150*** 0.077* 0.078*** 0.037** 0.020 0.012 0.440*** 0.142*** 0.184*** 0.181*** 6.1748*** 0.030 0.107 0.098 18,154 Belgium 0.164*** 0.074** 0.059** 0.028 0.014 0.027 0.011 0.568*** 0.066 0.146*** 0.064 5.914*** 0.024 0.087 0.068 13,487 Denmark 0.010 0.201*** 0.106*** 0.120*** 0.072*** 0.001 0.004 0.648*** 0.346*** 0.093*** 0.102** 4.7051*** 0.0337 0.102 0.086 20,484 Finland 0.070*** 0.142*** 0.073*** 0.132*** 0.012 0.013 0.043 0.535*** 0.004 0.076*** 0.078* 6.050*** 0.033 0.112 0.103 17,504 France 0.134*** 0.030 0.003 0.106*** 0.064** 0.020 0.010 0.289*** 0.072* 0.196*** 0.030 4.6119*** 0.023 0.128 0.116 27,033 Greece 0.133*** 0.000 0.019 0.108*** 0.063*** 0.016 0.128*** 0.910*** 0.049 0.235*** 0.049 13.666*** 0.110 0.342 0.259 20,048 Italy 0.162*** 0.254*** 0.216*** 0.070*** 0.029** 0.060*** 0.021 0.957*** 0.055 0.112*** 0.175*** 14.87*** 0.059 0.227 0.177 36,790 Netherlands 0.182*** 0.019 0.011 0.008 0.013 0.003 0.040* 0.526*** 0.058 0.106*** 0.082** 4.738*** 0.023 0.100 0.078 25,896 Portugal 0.024 0.126*** 0.066** 0.078*** 0.032** 0.065*** 0.089*** 0.740*** 0.182*** 0.118*** 0.042 7.009*** 0.062 0.176 0.135 33,443 Spain 0.149*** 0.148*** 0.045** 0.065*** 0.021 0.022 0.033* 0.669*** 0.058* 0.124*** 0.062** 9.173*** 0.049 0.182 0.140 34,540

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

Note: *, **, *** indicate signicant improvement at 10, 5, 1% levels, respectively. Additional variables included in the regression are: experience, experience square, number of unemployment spells, two dummy variables for occupational status, seven year dummy variables, eight occupational dummy variables, two industry dummy variables, one marital status dummy variable, one dummy variables for being in good health, and one dummy variable for having a xed term contract.

Table A4 Satisfaction with job security (GLS z-score with individual random effects and xed time effects) Coefcient Austria Male Lower secondary Upper secondary Private Firm size 100 Firm size 500 Firm size 500+ lnwage Mean wage lnhousinc Mean income Intercept R2 : within R2 : between R2 : overall No. obs. 0.009 0.093** 0.103** 0.303*** 0.065*** 0.052** 0.023 0.066** 0.013 0.035 0.101** 1.8715*** 0.016 0.171 0.112 18,154 Belgium 0.069 0.057 0.028 0.088** 0.051 0.061 0.053 0.086 0.163* 0.017 0.110 1.1867 0.0259 0.111 0.084 13,496 Denmark 0.007 0.035 0.002 0.054** 0.067** 0.055** 0.024 0.043 0.017 0.006 0.101** 0.39 0.063 0.192 0.141 20,487 Finland 0.004 0.005 0.009 0.024 0.022 0.024 0.075*** 0.025 0.135*** 0.070*** 0.042 1.709*** 0.096 0.261 0.208 17,504 France 0.014 0.010 0.055** 0.416*** 0.042* 0.002 0.001 0.059** 0.138*** 0.042** 0.052 2.188*** 0.055 0.318 0.229 27,050 Greece 0.046** 0.009 0.013 0.383*** 0.044*** 0.022 0.077*** 0.237*** 0.240*** 0.115*** 0.029 7.439*** 0.081 0.530 0.430 20,049 Italy 0.072*** 0.129*** 0.120*** 0.283*** 0.055*** 0.086*** 0.075*** 0.341*** 0.046 0.091*** 0.215*** 7.82*** 0.034 0.344 0.246 36,804 Netherlands 0.05** 0.059 0.047 0.087*** 0.066*** 0.053** 0.088*** 0.155*** 0.102** 0.030 0.092** 1.105*** 0.049 0.142 0.089 25,952 Portugal 0.017 0.059* 0.063* 0.120*** 0.029** 0.022 0.089*** 0.199*** 0.016 0.081*** 0.006 3.257*** 0.064 0.234 0.169 33,440 Spain 0.030* 0.110*** 0.036** 0.112*** 0.048*** 0.052*** 0.030* 0.198*** 0.016 0.051*** 0.048* 3.231*** 0.087 0.359 0.290 34,567

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

Note: see the legend of Table A3.

1915

1916

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

to remain equally satised with the job. They are calculated by using the ratio of the coefcients associated with two different job facets within the same country. For instance, if Greek workers are to remain equally satised after a decrease in their satisfaction with job security, satisfaction with earnings should increase by a factor 0.490/0.470 = 1.04. If they are to remain equally satised after a decrease in their satisfaction with the type of work, satisfaction with earnings should increase by a factor 0.765/0.470 = 1.62. This can be compared to the case of Denmark, for instance where if Danish workers are to remain equally satised after a decrease in their satisfaction with job security, satisfaction with earnings should increase by a factor 0.096/0.199 = 0.48 and if they are to remain equally satised after a decrease in their satisfaction with the type of work, satisfaction with earnings should increase by a factor 0.879/0.199 = 4.41. Using this approach one is able to evaluate whether a change in the satisfaction of a given job facet requires more or less than proportional compensation in the satisfaction of other job facet so as to keep the overall workers satisfaction unaltered. Overall, the above ndings imply that, in line with Lancasters model, jobs are not one-dimensional. A worker may be willing to accept a job involving less of a given desirable facet if he or she is compensated with more of another desirable attribute. The implied utility model is compatible with a multi-characteristic space accommodating the existence of a number of sub-utilities associated with different characteristics. Thus, the overall utility of a job is the result of an aggregation of all sub-utilities related to a specic mix of the job characteristics. Hence, a different mix of these job characteristics may produce a different level of overall utility.

5. Concluding remarks This paper differs from the conventional approach in investigating the effect of job characteristics on job satisfaction in that it assumes that jobs are evaluated by workers through a vector of sub-utilities derived from separate job characteristics. Therefore, it is the judgment that individuals make about each of these characteristics which determines their level of overall job satisfaction. Overall job satisfaction is assumed to be the aggregate outcome of partial satisfactions with different job facets. This approach requires the estimation of a two-layer model comprising a set of partial satisfaction equations and an overall job satisfaction equation. Scitovsky (1976) argued that the most cherished values are priceless and are not for sale and that intrinsic work enjoyment yields greater satisfaction than pay. He proposed that the difference between liking and disliking ones work may well be more important than the differences in economic satisfaction that the disparities in our income lead to (p. 103). This study supports this view. For each of the ten countries examined, satisfaction with the type of work is the main criterion which workers use to evaluate their job. This is true for both the short and the long term. In addition, the results clearly show that satisfaction with each of the ve job facets is a highly signicant contributor to the overall job satisfaction although a different mix of these job characteristics may produce a different level of overall job satisfaction. The results of this study are important in terms of human resources management. Workers value highly some characteristics offered by the job task which they perform. Thus, much attention should be given to the design of the job tasks which lead to the fulllment of some important needs of the workers. In most jobs, the employers can only monitor their employees very partially (Holmstrom and Milgrom, 1991) and hence it is essential for the employers to nd ways to enhance the employees job satisfaction. The organisational environment which offers high job satisfaction is more likely to be conducive to a situation where employees are successfully engaged in their job tasks and hence make up a productive workforce.

Table A5 Satisfaction with type of work (GLS z-score with individual random effects and xed time effects) Coefcient Austria Male Lower secondary Upper secondary Private Firm size 100 Firm size 500 Firm size 500+ lnwage Mean wage lnhousinc Mean income Intercept R2 : within R2 : between R2 : overall No. obs. 0.024 0.130** 0.115** 0.117*** 0.105*** 0.100*** 0.132*** 0.042 0.010 0.044 0.093* 2.019*** 0.0104 0.095 0.061 18,154 Belgium 0.128*** 0.017 0.046 0.057* 0.029 0.050 0.093*** 0.121** 0.075 0.041 0.017 4.719*** 0.012 0.065 0.049 13,496 Denmark 0.010 0.052* 0.003 0.057** 0.073*** 0.086*** 0.196*** 0.118*** 0.149*** 0.014 0.017 0.224 0.016 0.078 0.044 20,487 Finland 0.069** 0.097*** 0.002 0.051** 0.091*** 0.102*** 0.057* 0.040 0.173*** 0.035 0.020 2.424*** 0.013 0.100 0.071 17,504 France 0.031 0.071** 0.031 0.066** 0.099*** 0.116*** 0.146*** 0.086*** 0.015 0.018 0.169 0.840*** 0.031 0.092 0.068 27,050 Greece 0.051** 0.132*** 0.059*** 0.165*** 0.039** 0.073*** 0.110*** 0.154*** 0.117*** 0.183*** 0.214 5.5967*** 0.025 0.344 0.249 20,049 Italy 0.013 0.017 0.004 0.142*** 0.012 0.013 0.032 0.223*** 0.020 0.041** 0.160*** 4.980*** 0.013 0.200 0.133 36,804 Netherlands 0.099*** 0.196*** 0.156*** 0.049** 0.076*** 0.108*** 0.106*** 0.141*** 0.000 0.065*** 0.121*** 0.856*** 0.017 0.054 0.035 25,952 Portugal 0.015 0.009 0.038 0.098*** 0.051*** 0.027 0.076*** 0.239*** 0.015 0.062*** 0.031 3.173*** 0.016 0.155 0.095 33,440 Spain 0.029 0.119*** 0.034* 0.133*** 0.063*** 0.051*** 0.059*** 0.185*** 0.011 0.048** 0.813*** 1.712*** 0.015 0.162 0.103 34,567

Note: see legend of Table A3.

Table A6 Satisfaction with working conditions (GLS z-score with individual random effects and xed time effects) Coefcient Austria Male Lower secondary Upper secondary Private Firm size 100 Firm size 500 Firm size 500+ lnwage Mean wage lnhousinc Mean income Intercept R2 : within R2 : between R2 : overall No. obs. 0.001 0.032 0.052 0.022 0.087*** 0.141*** 0.148*** 0.017 0.084* 0.032 0.123** 0.668 0.004 0.055 0.038 18,154 Belgium 0.096*** 0.009 0.019 0.068* 0.050 0.025 0.075* 0.074 0.055 0.000 0.076 5.418*** 0.008 0.048 0.036 13,496 Denmark 0.020 0.056* 0.011 0.007 0.048* 0.012 0.003 0.087** 0.091** 0.038 0.039 0.018 0.008 0.054 0.038 20,487 Finland 0.046* 0.057* 0.006 0.006 0.027 0.003 0.029 0.046 0.045 0.039 0.046 0.641* 0.004 0.068 0.050 17,504 France 0.050** 0.006 0.025 0.007 0.136*** 0.101*** 0.082** 0.053* 0.043 0.001 0.029 0.101 0.012 0.062 0.046 27,050 Greece 0.187*** 0.107*** 0.009 0.044** 0.017 0.000 0.073** 0.041 0.006 0.121*** 0.092*** 2.8788*** 0.014 0.252 0.168 20,049 Italy 0.074*** 0.107*** 0.054* 0.078*** 0.019 0.006 0.043* 0.072** 0.132*** 0.029 0.074*** 0.552 0.008 0.114 0.072 36,804 Netherlands 0.005 0.022 0.025 0.022 0.075*** 0.067*** 0.099*** 0.012 0.016 0.042** 0.068* 0.376 0.010 0.040 0.023 25,952 Portugal 0.007 0.029 0.032 0.050** 0.007 0.012 0.036 0.143*** 0.117*** 0.056*** 0.030 1.253*** 0.006 0.042 0.0255 33,440 Spain 0.088*** 0.046** 0.008 0.026 0.061*** 0.076*** 0.077*** 0.056** 0.118*** 0.027 0.012 0.556* 0.005 0.078 0.047 34,567

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

Note: see legend of Table A3.

1917

1918

Table A7 Satisfaction with working times (GLS z score with individual random effects and xed time effects) Coefcient Austria Male Lower secondary Upper secondary Private Firm size 100 Firm size 500 Firm size 500+ lnwage Mean wage lnhousinc Mean income Intercept R2 : within R2 : between R2 : overall No. obs. 0.014 0.014 0.057 0.118*** 0.033 0.037 0.043* 0.108*** 0.067 0.017 0.139*** 1.217*** 0.005 0.054 0.036 18,154 Belgium 0.075** 0.152*** 0.118*** 0.115*** 0.015 0.020 0.018 0.076 0.094 0.025 0.058 0.265 0.005 0.046 0.034 13,496 Denmark 0.031 0.048* 0.033 0.164*** 0.028 0.021 0.078** 0.001 0.028 0.014 0.084* 0.980*** 0.008 0.084 0.045 20,487 Finland 0.062** 0.007 0.041* 0.013 0.011 0.009 0.004 0.107*** 0.079* 0.041 0.027 0.223 0.007 0.067 0.056 17,504 France 0.020 0.001 0.003 0.183*** 0.009 0.073** 0.048 0.011 0.005 0.013 0.043 0.2332 0.034 0.059 0.054 27,050 Greece 0.048** 0.126*** 0.066*** 0.275*** 0.003 0.083*** 0.032 0.037 0.116** 0.106*** 0.068* 1.080*** 0.013 0.186 0.136 20,049 Italy 0.035* 0.003 0.017 0.215*** 0.017 0.041** 0.023 0.053* 0.133*** 0.006 0.228*** 0.530 0.004 0.107 0.070 36,804 Netherlands 0.036 0.050 0.060** 0.128*** 0.017 0.013 0.017 0.160*** 0.104** 0.029 0.015 0.625** 0.008 0.033 0.025 25,952 Portugal 0.045** 0.015 0.047 0.167*** 0.028* 0.005 0.021 0.121*** 0.051 0.070*** 0.040 1.030*** 0.006 0.050 0.030 33,440 Spain 0.034* 0.036* 0.008 0.256*** 0.004 0.024 0.011 0.042* 0.038 0.026 0.042 0.629** 0.003 0.063 0.044 34,567

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

Note: see legend of Table A3.

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920 Table A8 Frequencies of variables (%) France Male Married Lower secondary Upper secondary Tertiary Supervisor Intermediate Non supervisory position Private Manager Professionals Technicians Clerks Sales workers Skilled Agriculture worker Craft worker Machine operators Elementary occupations Firm size 20 Firm size 100 Firm size 500 Firm size 500+ Agriculture Manufacture Services Permanent contract Fixed contract Good health: Experience (mean) Wage (in national currency) Income (in national. currency) 56.31 59.86 41.94 32.45 25.62 12.59 20.22 67.19 69.58 4.79 8.30 19.22 83.43 12.74 1.35 13.48 12.91 8.5 24.12 18.07 12.84 10.21 1.39 29.25 69.36 74.72 9.85 69.90 19.2 years 10,136 8328 Greece 62.01 65.56 32.81 36.83 30.37 6.14 7.26 85.74 61.52 2.27 16.42 8.08 17.73 13.67 1.11 18.20 9.66 9.39 53.66 23.36 5.87 4.75 1.36 29.34 69.30 65.26 19.45 93.49 16.5 years 176,831 165,169 Netherlands 61.14 67.68 18.26 57.55 24.18 12.93 16.06 68.18 70.71 11.12 18.35 21.45 13.70 9.41 1.38 9.94 6.43 4.44 16.91 21.32 22.64 28.99 1.35 21.72 67.75 77.07 6.47 84.49 20.4 years 2670 2436 Spain 65.51 64.33 52.25 20.07 27.68 5.81 13.60 57.78 80.49 7.97 11.68 9.84 8.82 14.39 5.30 18.24 9.00 13.24 53.28 19.44 9.96 10.57 7.45 30.62 61.93 42.55 23.64 81.93 20.1years 132,743 124,381 Denmark 53.21 57.75 19.05 46.09 34.86 13.18 12.92 62.46 61.05 6.29 15.18 18.44 11.92 11.72 2.01 10.39 7.03 7.71 31.47 20.95 13.33 10.77 3.17 22.12 61.65 66.38 8.68 87.89 21.8 years 16,972 11,646 Belgium 55.87 67.65 21.21 34.40 44.40 10.06 17.62 56.25 66.55 4.87 17.37 12.11 16.39 7.49 0.99 7.33 4.54 7.84 25.41 16.56 12.21 15.53 1.48 20.48 56.67 63.58 7.73 84.86 18.3 years 39,543 45,716 Italy 63.51 69.37 46.47 42.34 11.19 5.88 11.45 54.15 71.93 3.21 9.36 10.91 17.81 13.04 3.96 19.47 6.89 10.66 54.27 17.81 9.58 7.08 6.53 31.35 62.13 53.70 8.05 72.00 18.5 years 1090 1510 Portugal 58.04 68.96 79.64 12.29 8.08 2.85 4.19 64.41 80.40 5.78 5.70 7.96 8.69 13.93 12.51 20.54 8.26 14.60 60.69 18.72 9.08 4.45 15.05 31.28 53.66 50.36 12.37 62.68 21.5 years 66,654 88,251 Austria 57.96 61.03 20.04 71.82 8.13 8.55 20.14 52.63 75.39 6.48 4.60 15.71 14.22 15.39 10.74 18.44 6.69 7.73 46.02 24.89 15.51 13.58 11.20 31.46 57.34 75.20 6.29 85.27 20.7 years 13,358 17,163

1919

Finland 49.88 66.04 20.75 42.49 36.76 12.12 13.61 53.42 65.53 9.19 17.85 15.46 8.72 11.05 9.67 10.43 6.56 5.13 50.96 24.91 15.56 8.55 9.05 19.09 48.58 66.76 12.39 74.04 22.1 years 5562 7270

Acknowledgements The authors would like to thank Keith Bender, Franc ois Gardes, Nikos Georgantzis, Joseph Lanfranchi and participants of the departmental seminar at the Department of Economics, University of Wales, Swansea, UK and participants at the international conference Capabilities and Happiness at the University of Milano-Bicocca, Italy, the 23d Journees de Microeconomie at the Universite de Nantes, France, the 8th IZA European summer school in labour economics, Germany and the Appliquee 23d conference of the International Society of Quality of Life Studies, Rhodes University, Grahamstown, South Africa and a referee of this journal, for helpful comments. The nancial support of the European Commission under the Fifth Framework Programme Improving Human Potential (contract number: HPSE-CT-2002-00143) is gratefully acknowledged. Appendix A See Tables A1A8. References
Aaronson, D., Sullivan, D.G., 1998. The decline of job security in the 1990s: displacement, anxiety, and their effect on wage growth. Federal Reserve Bank of Chicago. Economic Perspectives 22 (First Quarter), 1743. Ahn, N., Garcia, J.R., 2004. Job satisfaction in Europe. Document de travail, 16 FEDEA. Bender, K., Donohue, S.M., Heywood, J.S., 2006. Job satisfaction and gender segregation. Oxford Economic Papers 57, 479496. Blanchower, D., Oswald, A., Stutzer, A., 2001. Latent entrepreneurship across nations. European Economic Review 45, 680691. Blanchower, D., Oswald, A., 1999. Well-Being, Insecurity and the Decline of American Job Satisfaction. Working paper, University of Warwick. Bockerman, P., 2004. Perceptions of job instability in Europe. Social Indicators Research 67, 283314. Borjas, G.J., 1979. Job satisfaction, Wages and Unions. Journal of Human Resources 14, 2140. Clark, A.E., Senik, C., 2006. The (unexpected) structure of rents on the French and British labour markets. Journal of Socio-Economics 35, 180196. Clark, A.E., 2005. What makes a good job? Evidence from OECD countries. In: Bazen, S. Lucifora, C., Salverda, W. (Eds.), Job Quality and Employer Behaviour, Palgrave, pp. 1130. Clark, A.E., Postel-Vinay, F., 2008. Job security and job protection. Oxford Economic Papers, in press. Clark, A.E., 1997. Job satisfaction and gender: why are women so happy at work? Labour Economics 4, 341372. Clark, A.E., Oswald, A.J., 1996. Satisfaction and comparison income. Journal of Public Economics 61, 359381. de lEmploi: une perspective europeenne. de lemploi, La decouverte, Davoine, L., Erhel, C., 2006. La Qualite La qualite Paris. Davoine, L. 2005. Les determinants de la satisfaction au travail en Europe:Limportance du contexte. Working paper. De Witte, H., Halman, L., Gelissen, J. 2004. European work orientations at the end of the twentieth century. In: Arts, W., Halman, L. (Eds.), European Values at the Turn of the Millennium. Vol. 7., Brill, Leiden, The Netherlands. Deloffre, A., Rioux, L., 2004. Do Workers Correctly Evaluate their Job Security? An European Comparison. Working paper.

1920

A. Skalli et al. / The Journal of Socio-Economics 37 (2008) 19061920

Diaz-Serrano, L., Vieiri, J.A.C., 2005. Low pay, higher pay and job satisfaction within the European Union: empirical evidence from fourteen countries. IZA DP, 1558. Drakopoulos, S.A., Theodossiou, I., 1997. Job satisfaction and target earning. Journal of Economic Psychology 18, 693704. Fischer, J., Sousa-Poza, A. 2006. Does Job Satisfaction Improve Health? New Evidence using Panel Data and Objective Measures of Health. University of St. Gallen FAA Discussion Paper No. DP-110. Available at SSRN: http://ssrn.com/abstract=955734. Freeman, R.B., 1978. Job satisfaction as an economic variable. American Economic Review 68, 135141. Frey, B., Stutzer, A., 2002. Happiness & Economics Princeton. Princeton University Press, New Jersey, USA. Green, F., Tsitsianis, N., 2005. An investigation of national trends in job satisfaction in Britain and Germany. British Journal of Industrial Relations 43, 401430. Groot, W., van den Brink, H.N., 1999. Job satisfaction of older workers. International Journal of Manpower 20, 343360. Hamermesh, D.S., 1977. Economic aspects of job satisfaction. In: Ashenfelter, O., Oates, W.E. (Eds.), Essays in Labour Market Analysis. Wiley, New York, pp. 5372. Holmstrom, B., Milgrom, P., 1991. Multi-task principal agent analysis: incentive contracts, asset ownership and job design. Journal of Law, Economics and Organization 7, 2452. Huang, X., Vliert, E.V.D., 2003a. Comparing work behaviors across cultures, a cross-level approach using multilevel modeling. International Journal of Cross Cultural Management 3, 167182. Huang, X., Vliert, E.V.D., 2003b. Where intrinsic job satisfaction fail to work: national moderators of intrinsic motivation. Journal of Organizational Behavior 24, 159179. Huang, X., Vliert, E.V.D., 2004. Job level and national culture as joint roots of job satisfaction. Applied Psychology: An International Review 63, 329348. Hui, M.K., Au, K., Fock, H., 2004. Empowerment effects across cultures. Journal of International Business Studies 35, 4660. Idson, T.L., 1996. Employer size and labor turnover. In: Solomon, W., Polachek (Eds.), Research in Labor Economics. JAI Press, Greenwich Conn. and London, pp. 273304. Kaiser, L.C., 2007. Gender-Job satisfaction differences across Europe: an indicator for labour market modernization. International Journal of Manpower 28, 7594. Kaiser, L.C., 2002. Job satisfaction: A Comparison of Standard, Non Standard, and Self-Employment Patterns across Europe with a Special note to the Gender/Job satisfaction paradox. EPAG Working Paper 27. Lancaster, K., 1971. Consumer Demand: A New Approach. Columbia University Press, New York. Lancaster, K., 1966. A new approach to consumer theory. Journal of Political Economy 74, 132157. Lang, J.R., Johnson, N.B., 1994. Job satisfaction and rm size: an interactionist perspective. The Journal of Socio-Economics 23, 405423. Lawler III, E.E., Porter, L.W., 1967. The effect of performance on job satisfaction. Industrial Relations 7, 2028. Lindbeck, A., Snower, D.J., 1996. Reorganization of rms and labor market inequality. American Economic Review 86, 515521. Locke, L.A., 1969. What is job satisfaction? Organizational Behavior and Human Performance 4, 309336. Lydon, R., Chevalier, A., 2002. Estimates of the Effects of Wages on Job Satisfaction. Discussion Paper, Centre for Economic Performance, p. 531. Medgyesi, M., Robert, P., 2003. Satisfaction with work in a European perspective: center and periphery, old and new marker economies compared. Review of Sociology 9, 4368. Moguerou, P., 2002. Job satisfaction among US PhD graduates: the effects of gender and employment sector. IREDU Working Paper. Mundlak, Y., 1978. On the Pooling of Time Series and Cross Section Data. Econometrica 46, 6985. Nickell, S., Jones, T., Quintini, G., 2002. A picture of the job insecurity facing British men. The Economic Journal 112, 127. OECD, 1997. Employment Outlook 1997. Pouliakas, K., Theodossiou, I., 2005. Confronting Objections to Performance Pay: An Analysis of the Incentives-Job Satisfaction Relationship after Controlling for Endogeneity. University of Aberdeen, Mimeo. Rosen, S., 1986. The theory of equalizing differences. In: Ashenfelter, O., Layard, R., (Ed.), Handbook of Labor Economics. Volume 1. Handbooks in Economics Series, no. 5, North Holland, pp. 641692. Ryan, A., Schmitt, M., Johnson, R., 1996. Attitudes and effectiveness: examining relations at an organizational level. Personnel Psychology 49, 853882. Scitovsky, T., 1976. The Joyless Economy: An Inquiry into Human Satisfaction and Dissatisfaction. Oxford University Press, Oxford. Sloane, P.J., Williams, H., 2000. Job satisfaction, comparison earning and gender. Labour 14, 473502. Sousa-Poza, A., Hennerger, F., 2000. Work attitudes, work conditions and hours constraints: an explorative, cross-national analysis. Labour 14, 351372. Sousa-Poza, A., Sousa-Poza, A.A., 2000a. Taking another look at the gender/job-satisfaction paradox. Kyklos 53, 135151. Sousa-Poza, A., Sousa-Poza, A.A., 2000b. Well-being at work: a cross-national analysis of the levels and determinants of job satisfaction. Journal of SocioEconomics 29, 517538. Stier, H., Lewin-Epstein, N., 2003. Time to work: a comparative analysis of preferences for working hours. Work and Occupations 30, 302326. Sweeney, P.D., McFarlin, D.B., 2004. Social comparisons and income satisfaction: a cross-national exanimation. Journal of Occupational and Organizational Psychology 77, 149154. Van de Stadt, H., Kapteyn, A., Van de Geer, S., 1985. The relativity of utility: evidence from panel data. The Review of Economics and Statistics 67, 179187. Van Praag, B.M.S., Ferrer-i-Carbonell, A., 2004. Quantied Happiness: A Satisfaction Calculus Approach. Oxford University Press, Oxford, UK. Van Praag, B.M.S., Frijters, P., Ferrer-i-Carbonell, A., 2002. The anatomy of subjective well-being. Journal of Economics Behavior and Organization 51, 2949. Warr, P.B., 1999. Well-being and the workplace. In: Kahneman, D., Diener, E., Schwarz, N. (Eds.), Well-being: The Foundations of Hedonic Psychology. Russell Sage Foundation, New York.